Prepared for the IPCS by the International Commission
    for Protection Against Environmental Mutagens and Carcinogens

    This report contains the collective views of an international group of
    experts and does not necessarily represent the decisions or the stated
    policy of the United Nations Environment Programme, the International
    Labour Organisation, or the World Health Organization.

    Published under the joint sponsorship of
    the United Nations Environment Programme,
    the International Labour Organisation,
    and the World Health Organization

    World Health Orgnization
    Geneva, 1985

         The International Programme on Chemical Safety (IPCS) is a
    joint venture of the United Nations Environment Programme, the
    International Labour Organisation, and the World Health
    Organization. The main objective of the IPCS is to carry out and
    disseminate evaluations of the effects of chemicals on human health
    and the quality of the environment. Supporting activities include
    the development of epidemiological, experimental laboratory, and
    risk-assessment methods that could produce internationally
    comparable results, and the development of manpower in the field of
    toxicology. Other activities carried out by the IPCS include the
    development of know-how for coping with chemical accidents,
    coordination of laboratory testing and epidemiological studies, and
    promotion of research on the mechanisms of the biological action of

        ISBN 92 4 154191 1 

         The World Health Organization welcomes requests for permission
    to reproduce or translate its publications, in part or in full.
    Applications and enquiries should be addressed to the Office of
    Publications, World Health Organization, Geneva, Switzerland, which
    will be glad to provide the latest information on any changes made
    to the text, plans for new editions, and reprints and translations
    already available.

    (c) World Health Organization 1985

         Publications of the World Health Organization enjoy copyright
    protection in accordance with the provisions of Protocol 2 of the
    Universal Copyright Convention. All rights reserved.

         The designations employed and the presentation of the material
    in this publication do not imply the expression of any opinion
    whatsoever on the part of the Secretariat of the World Health
    Organization concerning the legal status of any country, territory,
    city or area or of its authorities, or concerning the delimitation
    of its frontiers or boundaries.

         The mention of specific companies or of certain manufacturers'
    products does not imply that they are endorsed or recommended by the
    World Health Organization in preference to others of a similar
    nature that are not mentioned. Errors and omissions excepted, the
    names of proprietary products are distinguished by initial capital






    2.1. Bacterial mutation assays
         2.1.1. Principles and scientific basis of the assay
         2.1.2. Relevance and limitations of the assay
         2.1.3. The procedure
        Outline of the basic procedure
        Critical factors in the procedure
         2.1.4. Presentation and interpretation of data
        Data-processing and presentation
        Interpretation of data in terms of 
                         positive and negative
        Dealing with ambiguous results
         2.1.5. Discussion
        How the most critical factors identified 
                         above can influence the validity of the 
        Interpretation of the results in terms of 
                         the intrinsic mutagenic activity of the 
                         test material
    2.2. Genotoxicity studies using yeast cultures
         2.2.1. Introduction
         2.2.2. Genetic end-points
        Point mutation
         2.2.3. Information required
         2.2.4. Interpretation
        Significance of positive results in yeast 
        Negative results in yeast assays
    2.3. Unscheduled DNA synthesis in cultured mammalian cells
         2.3.1. Introduction
         2.3.2. Chemical exposure and UDS
         2.3.3. Procedure
        The choice of a suitable cell line
         2.3.4. The elimination of semiconservative replication
         2.3.5. Chemical exposure
         2.3.6. Radiolabelling procedures
         2.3.7. Detection of UDS
         2.3.8. Data processing and presentation
         2.3.9. Discussion
        Choice of cell line
        Choice of protocol
        Method of activating proximate carcinogens

    2.4.  In vitro cytogenetics and sister-chromatid exchange
         2.4.1. Introduction
         2.4.2. Procedure: chromosomal aberrations
        Cell types
        Culture methods
        Chromosome assay
         2.4.3. Procedure: sister chromatid exchange
         2.4.4. Procedure: scoring
         2.4.5. Extent of testing
         2.4.6. Data processing and presentation
         2.4.7. Discussion
        Critical factors
        Experimental design and analysis
         2.4.8. Conclusions
    2.5.  In vitro cell-mutation assays
         2.5.1. Principles and scientific basis of the assay
         2.5.2. Relevance and limitations
         2.5.3. Procedure
        Outline of the basic technique
        Cell types and selective systems
        Culture conditions
        Expression time
        Choice and concentration of selective 
        Stability of the spontaneous mutant 
        Provision for metabolic conversion
        Controls and internal monitoring
       Population size, replicates, and
         2.5.4. Data processing and presentation
        Treatment of results
        Evaluation of results
        Ambiguous results
         2.5.5. Discussion
        Mutant selection
        Expression time
        Cell numbers
        Metabolic conversion
         2.5.6. Conclusions
    2.6. The use of higher plants to detect mutagenic chemicals
         2.6.1. Introduction
         2.6.2. Test systems
        Detection of mitotic chromosome damage
        Detection of aberrations in meiotic
        Detection of gene mutations at specific or 
                         multiple loci
         2.6.3. Discussion
    2.7. The  Drosophila sex-linked recessive lethal assay (SLRL)
         2.7.1. Introduction

         2.7.2. Procedure
        Test organism life cycle
        Stock cultures
        List of nomenclature
        Equipment and laboratory techniques
         2.7.3. Principle of the recessive lethal assay
         2.7.4. Metabolic activation
         2.7.5. Test performance
        Treatment procedures
        Toxicity tests
        Control and replicate experiments: sample 
         2.7.6. Data processing and presentation
         2.7.7. Discussion
        Disadvantages of the recessive lethal test
        Weak mutagens and non-mutagens
        Data base
        Correlation with mammalian carcinogenicity 
        Recent developments
    2.8.  In vivo cytogenetics: bone marrow metaphase analysis and 
         micronucleus test
         2.8.1. Introduction
        Current understanding of the formation of 
                         chromosomal aberrations
        Classification of chromosomal aberrations
        The basis for micronucleus formation
         2.8.2. Procedure
        Experimental animals
        Treatment and sampling
        Dose levels
        Number of cells scored per animal
        Positive and negative controls
        Preparation procedure for bone-marrow 
        Preparation procedure for micronuclei
        Microscopic analysis
         2.8.3. Data processing and presentation
        Chromosomal aberrations
        Statistical evaluation
         2.8.4. Discussion
        Possible errors in microscopic evaluation
        Comparison of test sensitivity
        Application of the method to other tissues
         2.8.5. Conclusions
    2.9. Dominant lethal assay
         2.9.1. Introduction
         2.9.2. Procedure for male mice
        Standard and test conditions
        Test conditions to be established by each 
         2.9.3. Dominant lethals in female germ cells

         2.9.4. Data processing and presentation
         2.9.5. Discussion


    3.1. Introduction
    3.2. Laboratory facilities and equipment
         3.2.1. Microbial laboratories
         3.2.2. Tissue culture laboratories
         3.2.3. Facilities for other procedures
    3.3. Good laboratory practice in genetic toxicology
         3.3.1. Origins and nature of GLP
         3.3.2. GLP requirements
         3.3.3. Summary of resources and records needed


    4.1. Introduction
    4.2. Selection of assays
         4.2.1. Detection of the major types of genetic damage
        Gene mutations
        Chromosomal damage
        DNA damage
         4.2.2. Scientific validity
        Genetic basis
        Metabolic capability
         4.2.3. Predictive value
        Mutagenic activity
        Carcinogenic activity
        Relevance to chemical class
         4.2.4. Available expertise and facilities
    4.3. Application of assays
         4.3.1. The phased approach
        Phase 1 - the basic screen
        Supplementary tests
         4.3.2. Nature and extent of potential human exposure
        Limited or negligible distribution
        Medium distribution, limited exposure 
        Extensive distribution, intentional or 
                         unavoidable exposure
         4.3.3. Regulatory requirements
    4.4. Acceptablility and reliability of data
    4.5. Interpretation of results and significance for human 
         hazard assessment
         4.5.1. General principles
        Results of individual assays
        Results from combinations of assays
         4.5.2. Influence of the extent of exposure and
        Pharmaceutical compounds
        Chemical compounds in food
        Domestic chemical compounds
        Chemical compounds used in industry




    Every effort has been made to present information in the 
criteria documents as accurately as possible without unduly 
delaying their publication.  In the interest of all users of the 
environmental health criteria documents, readers are kindly 
requested to communicate any errors that may have occurred to the 
Manager of the International Programme on Chemical Safety, World 
Health Organization, Geneva, Switzerland, in order that they may be 
included in corrigenda, which will appear in subsequent volumes. 

                       *    *    *

    A detailed data profile and a legal file can be obtained from 
the International Register of Potentially Toxic Chemicals, Palais 
des Nations, 1211 Geneva 10, Switzerland (Telephone no. 988400 - 


The following experts participated in the preparation of this

Dr I.-D. Adler, Institute for Genetics, GSF, Neuherberg, Federal 
   Republic of Germany

Dr J. Cole, MRC Cell Mutation Unit, University of Sussex, Brighton, 
   Sussex, United Kingdom 

Dr N. Danford, University College of Swansea, Swansea, United

Dr U. Ehling, Institute for Genetics, GSF, Neuherberg, Federal
   Republic of Germany

Dr M. Parry, Genetics Department, University College of Swansea, 
   Swansea, United Kingdom 

Dr R. Roderick, Sittingbourne Research Centre, Sittingbourne, Kent, 
   United Kingdom

Dr S. Venitt, Chester Beatty Research Institute, Institute for 
   Cancer Research, Royal Cancer Hospital, Pollards Wood Research 
   Station, Chalfont St. Giles, Buckinghamshire, United Kingdom

Dr W. Vogel, Department of Radiation Genetics and Chemical 
   Mutagenesis, Sylvius Laboratories, Leiden, The Netherlands

Dr R. Waters, Genetics Department, University College of Swansea, 
   Swansea, United Kingdom


    During the past few years, great advances have been made in 
understanding the processes leading to malignant disease.  It is 
clear that alterations in genetic material are involved in these 
processes and that a great many carcinogens are capable of inducing 
such alterations under appropriate conditions.  Heritable 
alterations in germ cells may also be induced by certain chemicals 
and may constitute a genetic risk.  Numerous short-term tests have 
been developed to detect the ability of chemicals to cause such 
changes and are being used routinely and successfully, on a large 
scale.  There is a widespread desire to evaluate the data obtained 
from short-term tests and to generate such data in areas of the 
world where the necessary combinations of expertise may not yet be 

    The International Commission for Protection Against 
Environmental Mutagens and Carcinogens (ICPEMC), an assembly of 
scientists with expertise in the fields of environmental 
mutagenesis, carcinogenesis, genetic toxicology, and epidemiology, 
was therefore pleased to respond to the request of the 
International Programme on Chemical Safety of the World Health 
Organization, to prepare a document containing guidance in the 
field of short-term testing for mutagens and carcinogens with 
genetic activity.  This document represents the views of ICPEMC and 
is published by IPCS in an attempt to stimulate scientific 
discussion, as well as to provide guidance on the use of 
genotoxicity tests in chemical safety programmes.  Although short-
term tests to screen for mutagenicity and carcinogenicity are 
useful, they have their difficulties and limitations.  For example, 
while the majority of chemical initiators of carcinogenesis give 
positive results in tests for genetic change, it is not necessarily 
true that all chemicals with genetic activity are carcinogens. 
Moreover, there are carcinogens and cocarcinogens that are not 
readily detected by mutagenicity tests and that may act by 
mechanisms of quite a different nature.  Such substances are 
necessarily excluded from consideration here, but that does not 
mean that they may not be of the same importance.  There are 
differences of approach in genetic toxicology, as in most other 
branches of science.  The present document reflects a widely-used 
approach that may be regarded as good contemporary practice.  It 
should be emphasized that it does not claim to be definitive or to 
contain recommendations for regulatory action either in connection 
with the kind or number of tests that should be carried out, or the 
regulatory decisions that may be taken on the basis of the results 
of such tests.  It is designed to explain the types of test that 
are commonly employed and the meaning that the results of such 
tests may have in the assessment of possible human hazard, as far 
as is possible with current scientific knowledge. 

    It is obvious that any assessment of test results in terms of 
mutagenic or genotoxic hazard can be properly made only in the 
context of the whole toxicological profile of a substance and its 
use.  ICPEMC is currently working towards a position with regard to 
the selection of short-term tests and its recommendations should be 
available in the near future. 

    Current practice is still rapidly evolving and should not be 
considered as fixed.  Moreover, what might be considered feasible 
in scientifically advanced countries with large resources of 
expertise might be quite inappropriate in developing countries.  
The latter, however, provide the raison d'être of the present 
document, which is offered in a spirit of helpfulness in the hope 
that it may enable short-term genotoxicity tests to be used in a 
reasonable manner. 


    It has been known for many years that some chemicals can cause 
cancer in man.  More recently, there has been a growing awareness 
of the possibility that chemicals may also produce mutations in 
human germ cells thus influencing the frequency of genetic or 
heritable diseases.  Many thousands of chemicals, including 
pharmaceutical products, domestic and food chemicals, pesticides, 
and petroleum products are present in the environment and new 
chemicals are being introduced each year.  In addition, there are 
many compounds that occur naturally, which are known to be 
mutagenic and/or carcinogenic (e.g., mycotoxins in foods).  It is 
important, therefore, that chemicals to which people are exposed, 
either intentionally (e.g., therapeutically), in the course of 
their daily life (e.g., in domestic products, cosmetics, etc.), or 
inadvertently (e.g., in pesticides) are tested for their potential 
to produce cancer and genetic damage (mutations). 

    A few chemicals have been identified as carcinogenic because of 
their known association with cancer in man.  However, carcinogenic 
activity is usually determined by the ability of a chemical to 
produce tumours during the life-time exposure of laboratory 
animals.  Studies of this kind may last for two or three years and 
require the use of scarce resources and expertise.  This has led to 
the search for alternative ways of detecting chemicals with cancer-
producing potential and a number of relatively inexpensive assays 
have been developed, many using biological systems rather than 
whole mammals.  Because such assay systems need far less time to 
complete than classical long-term studies in rodents they are 
referred to as "short-term tests". 

    Although the results of epidemiological studies have confirmed 
that exposure to a number of chemicals, such as vinyl chloride and 
beta-naphthylamine, can cause cancer in man, convincing 
epidemiological evidence that chemicals constitute a mutagenic 
hazard for man is not yet available.  It is known that genetic 
defects cause a significant proportion of human diseases, but the 
contribution of environmental chemicals to genetic disease is 
unknown.  This is not surprising as the possibility of such a 
danger to health has only been apparent for about one generation. 

    The information that determines the characteristics of a cell 
or organism is contained in the genetic material of the cell, which 
is composed of deoxyribonucleic acid (DNA).  DNA is composed of 
sub-units of deoxyribonucleotides, which themselves consist of a 
pentose sugar (2-deoxyribose), a phosphate ester, and a purine or 
pyrimidine nitrogenous base.  These sub-units form a 3-dimensional 
helical double-stranded structure (Watson & Crick, 1953).  Each of 
the two strands consists of a linear array of the deoxyribose sugar 
molecules linked together in a chain by phosphate molecules.  The 
strands are joined side-by-side by hydrogen bonds between 
complementary pairs of the purine and pyrimidine bases.  The 
complementary pairs of bases are guanine (a purine) paired with 
cytosine (a pyrimidine) and adenine (a purine) paired with thymine 
(a pyrimidine).  The unique sequence of bases, taken in groups of 

three, or triplets, forms the genetic code, each triplet coding for 
a particular amino acid.  Sequences of triplets provide uniquely 
the information necessary for the synthesis of a functional protein 
or enzyme.  Such a functional sequence of bases is known as a gene.  
Genetic information is passed from one cell generation to the next 
by precise duplication of the strands and equal distribution of the 
DNA, prior to cell division (i.e., the mitosis of somatic cells or 
the meiosis of germ cells) and is responsible for the faithful 
handing on of all the characteristics of one generation to the next 
generation.  This fundamental genetic process is common to all 
organisms ranging from a simple bacterial cell to a complex mammal 
or plant.  In higher organisms, the long strands of DNA are bound 
to proteins (histones) and are organised into a number of complex 
structures called chromosomes, located in the nucleus of the cell. 

    With the exception of the germ cells, which carry a single set 
of chromosomes and are termed haploid, the cells of higher 
organisms contain duplicate sets of chromosomes, one set derived 
from each parent, i.e., they are diploid. 

    Diploid cells, therefore, carry a pair of each of the 
functional genes, which occupy a precise position or locus along 
the length of the DNA of a particular chromosome.  The paired genes 
may be identical (homozygous) or functionally different 
(heterozygous); heterozygous forms of the same gene are called 
alleles.  In many cases, one of the two alleles has a dominant 
function over its partner.  Such partners are called dominant (or 
partially dominant) and recessive (or partially recessive).  
Occasionally, a pair of alleles are expressed independently of each 
other and are regarded as co-dominant. 

    Genes carried by the X-chromosome behave differently.  As males 
have only one X-chromosome, the genes are not carried in pairs and 
a gene on this chromosome is expressed as a dominant or recessive 
X-linked characteristic.  Although females have two X-chromosomes, 
a similar, though more complex situation exists, as only one of the 
two X-chromosomes expresses its genes in a particular cell type.  
The fact that the basic helical structure of DNA and the genetic 
coding is common to all living organisms, whether they are 
bacteria, plants, or mammals, means that data obtained from studies 
of the effects of a chemical on one species can be used to predict 
the possible genetic response of another species to the same 

    Alterations to the information carried by the DNA occur as a 
result of small changes in the structure of the DNA molecule, 
whereby the base sequence transmitted to the next generation is 
changed, and this may result in descendants with different 
characteristics to the parent.  The alterations are, in effect, 
mutations and, though many of them are detrimental, some mutations 
are compatible with a normal, healthy existence, being responsible 
for the subtle differences between members of a species and 
constituting the driving force of evolution. 

    Mutagenic chemicals interact with DNA causing changes in its 
structure.  This may result in the loss, addition, or replacement 
of bases, thus altering their sequence in the DNA and affecting the 
fidelity of the genetic message. 

    The effect of these mutational changes may be to prevent the 
synthesis of functional proteins completely (i.e., inhibition of 
gene expression) or may lead to the synthesis of proteins with 
modified structure and enzymes with altered activity and 
specificity.  Where mutations lead to changes in the genetic 
information carried by male or female germ cells, the progeny of 
such affected parents may express the mutation as some form of 
heritable abnormality or disease.  When mutations occur in the 
somatic cells of a complex organism, they may produce irreversible 
changes in the cell that may ultimately be involved in producing a 
cancerous growth. 

    Mutation of one of a pair of genes may lead to a change in gene 
expression that can override the function of the normal partner and 
is called a dominant mutation.  Recessive mutations are expressed 
only when both genes carry the same recessive mutation.  For 
example, a recessive mutation inherited from the male parent will 
only be expressed in the progeny if the same recessive gene is also 
inherited from the female parent. 

    Cells can survive potentially lethal damage to the DNA, because 
of the activity of a series of enzyme-mediated processes that are 
generally termed DNA repair.  The simplest form of DNA repair 
involves removal (excision) of the chemically-altered portion and 
the repair of the gap left in the DNA strand by the synthesis of 
new DNA using the undamaged sister strand as a template.  Damage to 
DNA that is not repaired by this mechanism interferes with normal 
DNA synthesis (i.e., DNA replication).  This stimulates another 
kind of DNA repair (post-replication repair) which, because it is 
not always accurate (i.e., it is error-prone), may lead to 
mutagenic changes in the DNA.  The detection of unusual excision 
DNA repair activity, so called unscheduled DNA synthesis, in 
mammalian cells (as a response to damage to the DNA) forms the 
basis of an important assay for identifying chemicals that cause 
damage to DNA (section 2.3). 

    Damage to DNA may be expressed as mutations at the chromosome 
level (i.e., chromosomal aberrations or chromosomal mutations) or 
at the level of the gene (i.e., gene or point mutations).  
Chromosomal mutations may be observed as changes in the structure 
of the chromosome (structural aberrations) or in the number of 
chromosomes in a cell (numerical aberrations or aneuploidy).  
Structural aberrations are a consequence of DNA damage while 
numerical changes are usually caused by defects in the accurate 
distribution of chromosomes during cell division, i.e., DNA damage 
may not be involved. 

    Mutations contribute significantly to human diseases and 
congenital malformations, though the extent of this contribution is 
not precisely known.  Some diseases such as Down's syndrome 

(Trisomy 21) are associated with structural or numerical 
chromosomal abnormalities.  Others are a result of mutations of 
single genes and there are other diseases and congenital 
abnormalities for which a genetic alteration may be partly 
responsible.  Sickle cell anaemia is a disease caused by the 
inheritance of a single mutant gene that is responsible for the 
synthesis of an abnormal haemoglobin molecule.  In the homozygous 
state, i.e., where the mutant gene is inherited from both parents, 
the resulting disease is severe.  Heterozygous individuals carry 
only a single mutant gene and suffer from the relatively mild 
sickle cell trait.  Because of the resistance of sickle cell red 
blood cells to the malarial parasite, carriers of the trait have a 
selective advantage over unaffected individuals and the mutant gene 
is maintained at a high frequency in certain populations. 

    Although there is no definitive evidence that exposure to 
chemicals is responsible for any of the known human genetic 
disorders, experimental evidence from other mammals have shown that 
chemicals can produce both chromosomal and gene mutations of the 
type that are associated with human genetic diseases.  There is 
little direct evidence to suggest that man is any less susceptible 
than other mammals to the effects of exposure to mutagenic 

    Alterations in the structure and function of DNA are believed 
to play a crucial role in the production of cancer by chemicals.  
Carcinogenesis is a multistage process that may take years to 
evolve and a number of different factors influence the progression 
from a normally functioning cell to an invasive neoplastic tumour.  
(A carcinogen is defined as an agent that significantly increases 
the frequency of malignant neoplasms in a population; carcinogens 
may be physical, chemical, or biological agents).  The complex 
mechanisms by which chemicals induce malignancy are not fully 
understood, but there is evidence that suggests the occurrence of 
four major stages following an adequate exposure of a mammal 
(including man) to a chemical carcinogen: 

    (a) transport of the chemical from the site of entry into 
        the body and, in many cases, metabolic modification
        of the chemical (principally in the liver) to a more
        reactive form;

    (b) interaction of the molecule or its reactive metabolite 
        with the molecular target in the cell (the most 
        important of which is DNA);

    (c) expression of the DNA damage as a potentially
        carcinogenic lesion; and

    (d) progression, influenced by modifying factor(s), and
        proliferation to form a malignant tumour.

Some carcinogenic chemicals appear to be responsible for only one 
part of the process and are not regarded as complete carcinogens.  
For example, many chemicals that interact with DNA and are thus 

mutagenic appear to initiate the process by inducing the primary 
DNA lesion.  These are called initiators and the damage they cause 
is generally irreversible.  Other compounds have been shown to 
influence the expression and progression of the initial DNA change 
and are called tumour enhancers.  Some of these do not interact 
with DNA, they are not mutagenic and include the so-called tumour 
promoters.  A third group includes chemicals known as complete 
carcinogens in that they are probably capable of both initiating 
and promoting activity.  All chemicals that produce DNA damage 
leading to mutations or cancer, including initiators and complete 
carcinogens, are described as genotoxic. 

    The animal bioassay for detecting carcinogenic chemicals is a 
large, complex and very expensive scientific study using some 
hundreds of rodents to which the suspect chemical is administered 
for most of their life span.  Similarly, the specific locus test in 
mice (Searle, 1975), which is one of the few currently available 
assays for detecting heritable gene mutations in mammals, requires 
the examination of many thousands of offspring and is equally 
expensive and time-consuming.  Thus, of the multitude of chemicals 
introduced into the environment this century, and the hundreds of 
new compounds being synthesized each year, only a small fraction 
can be tested in conventional animal studies.  For this reason, the 
last decade has seen the introduction of a number of relatively 
rapid tests for detecting mutagenic and carcinogenic chemicals.  
Such tests are economical in resources and produce results in a 
matter of weeks.  Almost all of these short-term procedures are 
based on the demonstration of chromosomal damage, gene mutations, 
or DNA damage, and many of them are  in vitro assays (i.e., 
conducted in experimental biological systems without the use of 
live animals).  As will be described in section 2, the test 
organisms range from bacteria and yeasts to insects, plants, and 
cultured animal cells and there are also short-term tests in which 
laboratory animals are exposed to test chemicals for periods of a 
few hours to, at most, a few weeks.  In practice, a suspect 
chemical is first tested using  in vitro procedures, to study its 
ability to react with DNA and thus induce mutations.  It may then 
be necessary to determine its genotoxicity for intact animals by 
testing in short-term mammalian ( in vivo) assays. 

    The concept that carcinogenic chemicals cause cancer by a 
mutagenic mechanism is the basis of the somatic mutation theory of 
cancer induction.  Between 1955 and 1970, there were many attempts 
to demonstrate the mutagenicity of carcinogens using simple 
bacterial assays but these experiments failed to show a direct 
relationship between mutation and cancer.  Following the pioneering 
studies of Miller & Miller (1966), it was realized that the 
majority of chemical carcinogens were biologically inactive until 
they were enzymatically converted into reactive molecules.  Such 
chemicals are referred to as pro-carcinogens.  Intermediate 
metabolites that are the precursors of the ultimate reactive 
molecule (i.e., the molecule capable of reacting with DNA) are 
known as proximate carcinogens.  Cancer-inducing chemicals are 
often poorly soluble in water and, like most foreign compounds 
entering an organism, undergo a sequence of metabolic reactions 

intended to detoxify them and if necessary convert them into more 
water-soluble products, which can then be excreted by the kidneys.  
In some cases, these metabolic reactions produce carcinogenic 
chemicals, converting them into proximate and ultimate carcinogenic 
metabolites.  Most ultimate carcinogenic molecules are 
electrophilic reactants capable of binding with nucleophilic sites 
on DNA and other macromolecules (Miller & Miller, 1971). 

    The major classes of chemical carcinogens are activated by an 
oxidative reaction catalysed by microsomal mixed-function oxidases, 
though other enzyme systems influence this activation.  Appropriate 
enzymes are present in most mammalian tissues, the highest activity 
occurring in the liver.  In bacteria and most cultures of mammalian 
cells, mixed-function oxidase activity is either absent or very low 
and they are therefore not capable of activating the majority of 
carcinogens at a significant rate.  Insects have a complex 
metabolic capability and appear to be capable of activating a wide 
range of pro-carcinogens.  Yeast cells are also capable of limited 
foreign-compound metabolism.  In the early 1970s, Garner et al. 
(1971) and Malling (1971), recognizing the significance of the 
metabolic activation of chemicals, devised experiments in which 
mammalian enzymes and bacterial cells were combined in a single 
assay.  This led to the introduction of the first useful screening 
assay for carcinogenic chemicals, i.e., the  Salmonella typhimurium  
reversion test described by Professor Bruce Ames and his colleagues 
(Ames et al., 1973).  Essential aspects of mammalian metabolism are 
now introduced into many short-term  in vitro assays by the 
incorporation of an enzyme-rich, cell-free extract of mammalian 
tissues.  The most commonly used preparation is the post-
mitochondrial supernatent (referred to as the "S9" fraction), 
obtained after high speed centrifugation of a homogenate of rat 

    Although more than a hundred "test systems" for investigating 
genotoxicity have been described in the literature, ranging through 
the biological phyla from bacteriophage to mammals, less than 20 
are in regular use and some of these are only available in 
specialised laboratories.  The most widely-accepted systems are 
summarized below and described in detail in section 2. 

    Assays that involve the use of bacteria for detecting mutagenic 
chemicals are the most-extensively used and, in general, the most 
thoroughly validated.  Unlike higher organisms, in which the DNA is 
organized into complex chromosomal structures, bacteria contain a 
single circular molecule of DNA that is readily accessible to 
chemicals that can penetrate the cell wall.  Bacterial tests also 
have the advantage that a population of many millions of cells with 
a relatively short generation time can be tested in a single assay.  
In the classical techniques, strains of bacteria are used that 
already contain mutations of specific genes.  For example, a 
mutation at the histidine locus in  S. typhimurium removes the 
ability of the bacterium to synthesize histidine and such mutants 
cannot survive in culture medium lacking this nutrient.  Reversion 
at this locus enables the cell to synthesize histidine again and 
thus proliferate in medium lacking the amino acid.  Mutations 

induced by test-chemicals, i.e., "reverse" mutations, are detected 
by the growth of the "revertant" bacteria to form colonies in 
appropriate selective culture media.  Reverse mutation refers to 
reversion of an existing mutation, while forward mutation refers to 
the formation of a new mutation (section 2.1).  Bacterial assays 
can be adapted for detecting mutagenic metabolites in body fluids 
(e.g., urine, blood, plasma) from exposed animals or human beings. 

    Yeast and fungi occupy a position between bacteria and animal 
cells in terms of genetic complexity.  The structure of fungal DNA 
and its organisation into chromosomes is similar in many ways to 
that of mammals.  Both haploid and diploid forms can be used in 
genetic assays.  Tests using yeasts, such as  Saccharomyces 
 cerevisiae, are available for detecting both forward and reverse 
mutations and a variety of other genetic changes (section 2.2).  
Certain strains of yeast can be used to detect chemicals that 
induce aneuploidy (i.e., unequal distribution of chromosomes during 
cell division) and there is some evidence that non-genotoxic 
carcinogens can be identified using these strains. 

    The demonstration of DNA repair activity in mammalian cells is 
indirect evidence of DNA damage.  DNA repair can be detected in a 
simple mammalian cell culture assay that involves the measurement 
of "repair" or "unscheduled" DNA synthesis (UDS) (section 2.3).  
The assay is based on the fact that thymidine is incorporated into 
DNA during both normal and repair synthesis.  Cells treated with 
the suspect chemical are exposed to radioactive isotope-labelled 
(i.e., tritiated) thymidine at a stage when normal DNA synthesis is 
dormant or suppressed.  The amount of radiolabelled thymidine 
detected in the DNA is a measure of DNA repair synthesis and, thus, 
an indication of primary DNA damage. 

    Chemicals can be tested for their ability to induce chromosomal 
damage either in mammals, insects, cultured mammalian cells, or in 
plants.  Mammalian cell cultures provide a convenient test system 
and either established cell lines or human blood lymphocytes can be 
used (section 2.4).  Analysis of metaphase chromosomes in cells 
from the bone marrow of rats, mice, or hamsters is a well-
established technique for studying chromosome damage  in vivo  
(section 2.8).  Alternatively, chromosome fragments can be 
identified as micronuclei in certain bone marrow cells and in other 
tissues and the "micronucleus test" has proved to be a relatively 
simple assay for detecting chemicals capable of damaging 

    Cultures of mammalian cells can also be used to investigate the 
induction of gene mutations by chemicals (section 2.5).  The 
principles involved are similar to those of microbial assays, i.e., 
cells are cultured in medium containing the suspect chemical and 
are then sub-cultured into a selective medium in which only mutant 
cells can survive.  The number of cells that proliferate to form 
colonies is a measure of the number of cells that have undergone a 
forward mutation at the specific gene locus. 

    As described earlier in this section, bacteria, yeasts, and 
cultured mammalian cells may lack the enzymes necessary for the 
conversion of many carcinogens and mutagens to a molecular form 
that will react with DNA.  Assays using these systems must, 
therefore, be supplemented with a suitable mammalian microsomal 
enzyme preparation.  Many insects are able to activate a wide range 
of genotoxic chemicals and the demonstration of genetic changes in 
 Drosophila melanogaster (the fruit fly) forms a useful assay for 
investigating carcinogenic and mutagenic chemicals (section 2.7). 

    Because of the tremendous advances made in the use of microbial 
and mammalian cell procedures in genetic toxicology, plant material 
is less often used for studying mutagenic chemicals than 
previously.  However, the use of plants such as the bean  (Vicia 
 faba), the onion  (Allium cepa), the spiderwort  (Tradescantia 
 paludosa), maize  (Zea mays), barley  (Hordeum vulgare), and the 
soybean  (Glycine max) may have significant advantages over other 
systems and their value in screening chemicals for mutagenic 
activity has still to be fully explored (section 2.6).  
Investigation of genetic changes at both the gene and chromosomal 
level can be conducted in plants without the complicated laboratory 
facilities required for other types of assay and this may be a 
great advantage under certain circumstances.  A possible 
disadvantage is that the metabolic pathways in plants differ in 
many respects from those in mammals.  Thus, meaningful 
extrapolation to man of data obtained in plant studies is uncertain 
at present. 

    Data obtained from non-mammalian organisms and cultured 
mammalian cells determine whether a chemical or its metabolite(s) 
is capable of interacting with DNA and producing genetic damage.  
Two procedures are described in section 2 that are used to 
investigate the mutagenic activity of chemicals in the intact 
animal.  The first (section 2.8) is designed to detect chromosome 
damage in the somatic cells of rodents.  The second (section 2.9) 
is the dominant lethal assay that can identify chemicals capable of 
inducing genetic damage in the reproductive or germ cells of 
animals.  In this test, male rats or mice are dosed with the 
suspect chemical and mated with untreated females.  Certain types 
of chromosomal damage induced in the male germ cells are lethal to 
the fertilised ova and this can be detected by examination of the 
uterine contents. 

    Before short-term assays can be used to screen chemicals for 
potential carcinogenic activity with any degree of confidence, 
their sensitivity and accuracy for this purpose must be thoroughly 
validated.  The first comprehensive validation of bacterial tests 
for detecting carcinogens was conducted by Ames and his colleagues 
(McCann et al., 1975) using a combination of bacteria  (Salmonella)  
and mammalian microsomal enzymes.  In this study, in which 300 
chemicals were tested, approximately 90% of carcinogens were 
bacterial mutagens and 90% of non-carcinogens failed to show 
mutagenic activity.  Following this, Purchase et al. (1978) 
investigated 120 carefully-selected chemicals in a series of six 
short-term tests.  Again, the Ames bacterial mutation test gave a 

predictive value for carcinogenic potential of about 90%.  Analyses 
of these and later studies showed that the success rate of 
mutagenicity tests for detecting carcinogens was influenced by the 
type or class of chemical selected for testing and the criteria on 
which the carcinogenic activity in animals was judged.  Rinkus & 
Legator (1979) reviewed data from bacterial tests on 465 known or 
suspected carcinogens.  The compounds were divided into a number of 
separate categories, depending on their chemical structure.  The 
chemicals that showed the best correlation (94%) between mutagenic 
and carcinogenic activity were those that either could react 
directly with DNA (i.e., ultimate electrophiles) or could be 
activated by metabolic enzymes to DNA reactants.  Chemicals that, 
from their structure, appeared unlikely to react with DNA, showed a 
very poor correlation between mutagenic and carcinogenic activity.  
These chemicals appear to cause cancer by a different, possibly 
non-genotoxic, mechanism. 

    The most ambitious validation exercise to date was the 
International Program for the Evaluation of Short-term Tests for 
Carcinogenicity (de Serres & Ashby, 1981) in which some thirty  in 
 vitro and  in vivo assays involving more than fifty laboratories 
were evaluated for their ability to discriminate between 
carcinogenic and non-carcinogenic compounds.  Twenty-five 
carcinogens and 17 non-carcinogens, including 14 pairs of 
carcinogenic/non-carcinogenic analogues, were tested in most of the 
assays.  Animal cancer bioassay data from the 42 chemicals were 
critically evaluated by an expert committee.  Bacterial mutation 
assays gave the best overall performance producing reliable results 
in a large number of laboratories, and were confirmed as the first 
choice as an initial screening test.  However, it was also noted 
that some known rodent carcinogens were not detected or were only 
detected with difficulty by the standard Ames procedure.  Other 
assays that discriminated well between carcinogens and non-
carcinogens included  in vitro tests for chromosome damage and 
unscheduled DNA synthesis and assays using yeasts, and, although 
the database was smaller, results from  Drosophila tests and  in 
 vitro gene mutation assays suggested that they were useful 
components of a testing battery.   In vivo tests for chromosome 
damage showed their ability to differentiate between a number of 
carcinogen/non-carcinogen pairs and thus confirmed their value for 
investigating the  in vivo behaviour of chemicals found to be 
mutagenic in  in vitro tests. 

    A further major collaborative study, which was designed to 
complement the International Program referred to above, was 
conducted under the auspices of the International Programme on 
Chemical Safety.  The object of this project was to identify the 
best  in vitro test or tests to complement the bacterial reversion 
assay (Ashby et al., 1985). 

    Although bacterial mutation assays have a high predictive value 
for carcinogenicity, in most validation studies at least 10% of 
compounds give results that conflict with the animal cancer data.  
For this reason, it is generally accepted that bacterial assays 

should not be used in isolation for the testing of chemicals and it 
is common practice to use a "battery" or "package" of short-term 
assays as a preliminary screen. 

    Carcinogenic and/or mutagenic chemicals may induce one or more 
of a number of genetic changes and an assessment of the possible 
genotoxic hazard posed by a chemical should normally contain assays 
capable of detecting changes at both the gene and chromosomal 
level, and in some cases, tests for DNA damage.  Some authorities 
such as the Organisation for Economic Cooperation and Development 
(OECD), the European Economic Community (EEC) and the US 
Environmental Protection Agency (US EPA), require specific tests to 
be carried out on certain types of chemicals.  The application and 
interpretation of short-term tests are discussed in detail in 
section 4 and additional information is presented by Dean et al. 

    Section 2 contains descriptions in phylogenetic order of the 
most commonly used and most widely-accepted assays, some of which, 
however, are used more often than others.  In each procedure, 
specific minimal scientific and technical criteria can be 
identified that are critical factors in obtaining data of 
acceptable quality and reliability.  These factors are emphasized 
in section 2 and should be carefully considered by scientists 
contemplating setting up a testing facility and by those who are 
responsible for judging the validity of submitted data and 
assessing the genotoxic hazard associated with the use of 
chemicals.  Additional criteria that relate to good laboratory 
practice in genetic toxicology and to the type and quality of 
laboratory facilities are discussed in section 3. 


2.1.  Bacterial Mutation Assays

2.1.1.  Principles and scientific basis of the assay

    Bacteria have proved to be most suitable for the study of 
mutations, which are rare events, occurring naturally at a specific 
locus in less than one in a million bacteria at each cell division.  
However, as bacteria are single-celled organisms that divide 
rapidly and can be grown in large numbers in a few hours, it has 
proved to be relatively easy to grow tens of millions of organisms 
under circumstances where the one in a million event can be 
detected.  Furthermore, a great deal is known about the 
biochemistry and genetics of bacteria, and it has, therefore, been 
possible to develop special strains that are sensitive to a wide 
range of mutagens. 

    Bacterial test systems fall into 3 main classes, namely, those 
that detect backward mutations, those that detect forward 
mutations, and those that rely on a DNA repair deficiency.  By far 
the most widely-exploited method is the induction of backward or 
reverse mutations in  Salmonella typhimurium or, less frequently, 
 Escherichia coli.  The important point of this type of test is 
that, from the very beginning, each strain of bacterium already 
possesses  a selected mutation that prevents it performing an 
essential biochemical function, such as the synthesis of one out of 
the twenty or so amino acids necessary for the synthesis of 
proteins, unlike a non-mutant, "wild-type", strain of the same 
species.  A wild-type (prototrophic) strain can synthesize all its 
amino acids from inorganic nitrogen (e.g., ammonium phosphate) when 
provided with a suitable source of carbon (e.g., glucose).   The 
strains of  S. typhimurium used in reverse mutation tests cannot 
synthesize the amino acid histidine and are therefore designated as 
 "his-".  Similarly,  E. coli strains that cannot synthesize the 
amino acid tryptophan have been used and are referred to as "trp-".  
Such strains are said to be auxotrophic.  The reverse mutation test 
is so named because it can show whether a test material can reverse 
the effect of the pre-existing mutation (e.g.,  his-) by causing a 
second mutation which allows the bacterium to synthesize histidine 
from inorganic nitrogen.  This process is often abbreviated to 
 "his-" to  "his+" and is referred to a reversion from auxotrophy 
to prototrophy.  The resulting mutants are also called revertants. 

    In order to make the test more sensitive, the tester strains 
have been made more susceptible to mutagens by genetically changing 
the structure of their cell walls so that they become more 
permeable to large fat-soluble molecules.  Other genetic 
manipulations have reduced their ability to repair regions of DNA 
that have been damaged by chemicals and various types of radiation.  
Further sensitivity in the tester strains has been attained by 
introducing plasmids (small DNA molecules) that carry genes that 
interfere further with DNA repair, making the host bacteria even 
more vulnerable to the mutagenic effects of chemicals. 

    Because there are many different types of DNA damage, and 
because the use of the reverse mutation test requires the test 
chemical to hit a very small target in order to overcome the effect 
of the pre-existing mutation, several different targets are 
presented to the test chemical.  This is achieved by using several 
strains of the same species of bacterium, each one carrying a 
different pre-existing mutation in the same amino-acid gene.  Two 
types of mutation are employed:  base-pair substitution and 
frameshift.  For example, there are several different mutations in 
the histidine gene of  S. typhimurium, each different mutation 
being carried in a different strain, but all strains sharing the 
other traits (e.g., DNA-repair defects and cell-wall defects that 
make them very sensitive). 

    As mentioned in section 1,  E. coli and  S. typhimurium lack 
most of the enzymes that can perform the type of metabolic 
activation characteristic of mammalian biotransformation.  The 
enzymes are therefore added in the form of a liver extract prepared 
from laboratory animals, usually rats.  The rats are given 
chemicals that increase the amount of metabolic-activation enzymes 
in the liver and are then left for a few days before they are 
killed, to allow the enzymes to build up.  This is known as 
induction, and the chemicals that are used are called "inducers".  
The most widely used inducer is Aroclor(R) 1254, a mixture of 
polychlorinated biphenyls.  Phenobarbital plus 5,6-benzoflavone is 
also used for induction.  The liver is ground up and centrifuged at 
high speed; the supernatant liquid contains the metabolic enzymes 
(some of which are bound to membranes (microsomes)) and is called 
S9 (short for "9000  g supernatant"). 

    In a bacterial mutation assay,  his- bacteria are mixed with S9 
and several doses of the test chemical and are allowed to divide by 
providing a small amount of histidine.  If the test chemical is 
itself mutagenic, or if the enzymes in the S9 act on the test 
chemical to produce substances (metabolites) that are mutagenic, 
this will be shown by the appearance of a small proportion of 
bacteria, which will continue to grow and divide, even when the 
supply of histidine has been used up.  These revertants can be 
detected easily, since their DNA has been permanently changed so 
that they can make histidine from inorganic nitrogen, and can grow 
indefinitely without added histidine.  Thus, each mutant eventually 
grows into a microscopic colony and it is the count of these that 
is the end-point of the assay. 

2.1.2.  Relevance and limitations of the assay

    Bacterial mutation assays are used in a large number of 
laboratories throughout the world.  Several large-scale trials, 
carried out to test the usefulness of these assays in detecting 
potential carcinogens and mutagens (Purchase et al., 1978; McMahon 
et al., 1979; Bartsch et al., 1980; de Serres & Ashby, 1981), have 
shown that bacterial mutation assays are very good at picking out 
chemicals known to cause cancer.  Moreover, relatively few 
chemicals that do not cause cancer have given positive results in 
these tests.  In general, therefore, chemicals that are mutagenic 

in bacteria are more likely to cause cancer than chemicals that are 
not mutagenic, i.e., mutagenicity is a characteristic property of a 
large number of carcinogens.  It is important to understand, 
however, that there does not seem to be any useful quantitative 
relationship between the ability of a chemical to cause mutations 
in bacteria and cancer in animals or people.  In other words, a 
chemical that is a strong mutagen in bacteria is not necessarily a 
strong carcinogen in animals, nor is it always the case that a weak 
bacterial mutagen will be a weak carcinogen. 

    A second limitation of bacterial mutation tests is their 
inability to detect chemicals that are thought to induce cancer, 
not by causing DNA damage, but by other means, as yet poorly 
understood.  Such substances include asbestos, nickel, arsenic, and 
hormone-like chemicals such as diethylstilboestrol.  There are 
other substances, e.g., phorbol esters, which are extracts of 
certain species of the plant genus  Euphorbia, and certain 
secondary bile acids, which usually do not cause cancer when given 
alone to animals, but which increase the effects of other cancer-
causing chemicals.  This so-called tumour promotion does not come 
about because of the production of DNA damage.  Thus, it will not 
be detected by mutagenicity assays which detect only substances 
that can initiate cancer.  Nevertheless, promoters may well play a 
significant part in human cancer and it is important to recognise 
that bacterial mutagenicity tests cannot detect promoting activity. 

    Substances are known that cause genetic defects (and possibly 
cancer) in higher organisms by interfering with the machinery that 
controls the exact distribution of chromosome sets from one 
generation of cells to the next, and from parents to children, 
causing mistakes in the number and structure of chromosomes 
delivered to cells during cell division.  Such substances do not 
always cause DNA damage, and will not be detected in bacterial 
mutagenicity tests, since bacteria do not have chromosomes. 

    The use of cell-free extracts (S9) from rats to represent the 
metabolism of chemicals in mammals is another limitation that must 
be borne in mind, when interpreting the results of bacterial 
mutation tests.  Studies have shown that breaking up liver cells 
can, in some cases, distort the pattern of metabolism, resulting in 
levels and proportions of metabolites that would not be produced in 
the intact liver.  Moreover, because the test is carried out in a 
test-tube rather than in an animal, it is impossible to allow for 
several other factors, which can in some cases give a misleading 
impression of the mutagenic or carcinogenic effects of a chemical.  
The way the chemical enters the body and is distributed to the 
various organs, how each organ metabolises it, and how the chemical 
or its metabolites leave the body can all play a part in 
determining if, and to what extent, the chemical is mutagenic or 
carcinogenic for the animal.  None of these factors can be 
reproduced in a test-tube containing bacteria, S9, and a chemical. 

    Despite these limitations, bacterial mutation tests have been 
found by trial and error to be extremely valuable as the first in a 
series of tests for screening chemicals for potential mutagenic and 
carcinogenic activity.  Moreover, bacterial tests have been 
validated in far greater detail than any other tests currently used 
in genetic toxicology. 

2.1.3.  The procedure

    The bacterial mutation test that forms the basis of all 
screening programmes was devised by B.N. Ames and his co-workers 
and is usually referred to as the " Salmonella/microsome test".  It 
is essential that workers who intend to use this test, and those 
who review the results of such tests read the following papers: 
Ames et al. (1975), McCann et al. (1975), McCann & Ames (1976), 
Maron et al. (1981), Levin et al. (1982), and Maron & Ames (1984).  
The following technical details are not intended as a defined 
recommended protocol, but represent good current practice and good 
criteria for successful bacterial tests.  Outline of the basic procedure

    In the  Salmonella/microsome test, several  his- strains of
 S. typhimurium are used in order to detect several different types 
of DNA damage.  A set of sterile test-tubes is held at 45 °C.  
Molten soft agar ("top agar") (2 ml) containing a low concentration 
of histidine is added to each tube followed by 0.1 ml of a culture 
of the required bacterial strain, which has been grown over the 
previous night in a very rich nutrient liquid ("nutrient broth").  
This "overnight culture" contains about 1 x 109 bacteria per ml, so 
that each tube contains about 1 x 108 bacteria.  A range of doses 
of the test chemical (dissolved in a suitable solvent) is then 
added, each dose to a separate tube.  Dimethylsulfoxide (DMSO) is 
the most widely used solvent.  It dissolves numerous different 
kinds of chemicals, is miscible with water and, at the amount used 
in the test (0.1 ml or less), is not toxic to bacteria. 

    Several tubes are set aside to act as "controls", i.e., tubes 
that will receive the solvent but not the test chemical and will 
therefore indicate the background (spontaneous) level of mutation.  
It is essential to know the level of background mutation for each 
bacterial strain in each experiment in order to tell whether the 
test chemical has had any mutagenic effect.  Finally, 0.5 ml of 
S9-mix is added to each tube and the contents mixed thoroughly by 
rapid shaking.  S9-mix consists of S9 (usually between 4 and 30% by 
volume) to which has been added nicotinamide-adenine dinucleotide 
phosphate (NADP) and glucose-6-phosphate (which together provide 
energy for metabolism), phosphate buffer to maintain pH, and salts 
of magnesium and potassium.  A set of tubes is also prepared 
without S9.  This is to check whether the test chemical can cause 
mutation without the need for metabolic activation.  Chemicals of 
this type are directly-acting mutagens:  certain directly-acting 
mutagens can be made non-mutagenic by S9; thus, it is important to 
include this check. 

    The additions of bacteria, test chemical, and S9-mix are made 
in rapid succession, in order to avoid the potentially harmful 
effects of the rather high temperature (45 °C) necessary to keep 
the soft agar molten.  As soon as possible after mixing, the 
contents of each tube are poured on to the surface of 30 ml of 
solid 1.5% agar ("bottom agar") which contains glucose, ammonium 
and other salts, and phosphate buffer in a 9-cm petri dish 
("plate").  The plate is shaken to distribute the top agar in a 
thin, even layer over the bottom agar.  The lid of the plate is 
replaced and each plate is placed on a level surface:  the top agar 
then cools and solidifies.  When all the tubes have been poured, 
and the plates have cooled, they are inverted and placed in an 
incubator at 37 °C for 48 h. 

    This is called the plate incorporation technique, since all the 
ingredients of the test are incorporated into a thin layer of soft 
agar on the surface of harder agar in a plate.  During the first 
few hours of incubation, all the  his- bacteria will grow, since 
there is a trace of histidine present.  At the same time that the 
bacteria are dividing, the enzymes in the S9-mix, supported by the 
energy provided by the NADP and G-6-P, may act on the test chemical 
to form metabolites that can enter the rapidly dividing bacteria. 
Some of these metabolites, or the test chemical itself, may react 
with the bacterial DNA, causing DNA damage, some of which will lead 
to mutation in a very small fraction of the progeny of the 100 
million bacteria present at the start of incubation.  When all the 
histidine has been used up, the bulk of bacteria will stop 
dividing, and a thin, visible confluent lawn of bacteria will have 
formed in the soft agar.  However, bacteria that have sustained DNA 
damage leading to a mutation with the effect of reverting the  his- 
gene to  his+ will continue to divide, since they can now 
synthesize their own histidine from the ammonium salts in the 
bottom agar.  Each single revertant  his+ (mutant) bacterium can 
produce enough daughter bacteria in 48 h to form a single colony of 
bacteria, easily visible to the naked eye.  Therefore, the number 
of such colonies on the plate is an accurate reflection of the 
number of  his+ revertants that have arisen spontaneously or by the 
action of the test chemical.  If there are significantly more 
revertant colonies on treated plates than on control (solvent only) 
plates, and if the numbers of revertants rise with increasing dose, 
the result of the test is positive, and the chemical is a bacterial 
mutagen.  Critical factors in the procedure

    There are several conditions that must be met in order to 
ensure an adequate test:  these are briefly discussed below.  More 
extensive discussion can be found in IARC (1980a) and Venitt et al. 

    Base-line protocol

    It is essential that a base-line protocol should be written 
before starting a screening programme.  Methods for the preparation 
and storage of S9 and bacterial strains, and other procedures 

should be thoroughly checked by performing assays with reference 
mutagens and authenticated bacterial strains, under conditions 
prescribed by the chosen protocol.  Advice should be sought from 
experienced investigators. 

    Choice, checking, storage, and culture of bacterial strains

    The following strains of  S. typhimurium are most-commonly used 
for routine screening (Ames et al., 1975):  TA 1535, TA 1538, TA 98, 
and TA 100.  Strains TA 97 and TA 102 are also considered useful, 
under some circumstances (Maron & Ames, 1983).  In addition, 
 E. coli WP2uvrApKM101 is often included.  This  trp- strain is very 
sensitive to a wide range of mutagens (McMahon et al., 1979; de 
Serres & Ashby, 1981; Matsushima et al., 1981; Venitt & Crofton-
Sleigh, 1981). 

    Bacterial strains should be regularly checked for their 
characteristic genetic traits, including:  amino acid requirement; 
background mutation; induced mutation with reference mutagens; 
presence of plasmids where appropriate; presence of cell-wall and 
DNA-repair mutations. 

    Authenticated "master cultures" should be stored at a 
temperature below -70 °C.  Overnight cultures for routine assays 
should be prepared by inoculation from master cultures or from 
plates made from a master culture - never from a previously-used 
overnight culture.  The overnight culture should contain at least 
109 viable bacteria per ml, and should be freshly prepared for each 

    Negative and positive controls

    Each assay should include negative controls (addition of the 
solvent but no test chemical) in order to check the background 
mutation and positive controls (addition of reference mutagens to 
check that the assay is performing correctly).  A list of 
appropriate positive control mutagens is given by Maron & Ames 
(1984).  Where possible, the compounds selected as positive 
controls should be structurally related to the compound under test. 

    Test material and solvents

    All data available on the substance to be tested should be 
provided and recorded, including its lot or batch number, physical 
appearance, chemical structure, purity, solubility, reactivity in 
aqueous and non-aqueous solvents, temperature- and pH-stability, 
and sensitivity to light.  A sample of each substance to be assayed 
for mutagenicity should be retained for reference purposes. 

    Solutions of test substances should be freshly prepared for 
each experiment, and unused portions should be discarded.  The 
nature and percentage of impurities should be given:  if a known 
impurity is present in the test substance, it too should be assayed 
for mutagenicity at doses equivalent to those that would be present 
in the chosen doses of the major constituent.  If a mixture is to 
be tested, this should be stated. 

    The proposed uses of the test substance should be known, since 
antibiotics, surfactants, preservatives, and biocides pose special 
problems in bacterial mutation assays. 

    It is essential to devise operating procedures that minimize 
the hazards from storage, handling, weighing, pipetting, and 
disposing of mutagens and carcinogens, and that deal with 
accidental contamination (Montesano et al., 1979; IARC, 1980b; 
University of Birmingham, 1980; MRC, 1981).  Laboratories should 
follow the guidelines laid down for Good Laboratory Practice (GLP) 
(PMAA, 1976; Federal Register, 1978).  These important matters are 
discussed further in section 3. 

    In most cases, DMSO is the best solvent, but, in cases where it 
is unsuitable, other solvents may be used (Maron et al., 1981). 

    Preparation and use of S9

    The animals should be free of disease and infection, kept clean 
and at a reasonable temperature and should not be stressed by 
careless handling.  Dosing with inducing agents should be 
consistent from one batch of animals to the next.  Animals should 
be killed humanely and the livers removed and chilled as soon as 
possible.  S9 should be stored at, or below, -70 °C. 

    Optimum mutagenesis with a particular test compound depends on 
the amount of S9 added per plate.  Too much as well as too little 
S9 can drastically lower the sensitivity of the test.  The optimum 
S9 level for a given compound should therefore be checked.  The 
amount of S9 per plate is best expressed as mg liver protein per 
plate calculated from the protein concentration of the S9. 

    There are two widely-accepted methods of using S9-mix: 

    (a) S9-mix is mixed with the top agar, bacteria, and test
        substance, and the whole mixture is immediately
        poured on to the surface of the bottom agar (Ames et
        al., 1975); and

    (b) in the pre-incubation method, the test substance,
        bacteria, and S9-mix are mixed and incubated for 30
        min; top agar is then added, and the mixture is
        poured on to the bottom agar.  This modification is
        often more efficient in detecting certain classes of
        mutagens, for example, aliphatic  N-nitroso compounds
        (Bartsch et al., 1976; Yahagi et al., 1977).

    Design of experiments

    A minimum of three plates per dose should be used in all 
experiments.  Doses of test chemicals should be spaced at intervals 
differing by factors of less than 5.  Narrow spacing of doses 
avoids missing mutagens that are very toxic and that produce very 

steep dose-response curves with sharp cut-offs.  Combining the 
requirement for narrow spacing of doses with the need to encompass 
a very wide range of doses, two strategies emerge: 

    (a) a large experiment with closely-spaced doses ranging
        from sub-microgram to milligram levels, using 7 or 8
        doses together with positive and solvent controls; and

    (b) two experiments, the first using sub-microgram to tens
        or hundreds of micrograms.  If the results are
        positive, this should be confirmed in the strains and
        the dose range in which the positive effect was
        observed.  If negative results are obtained, the
        second experiment should be carried out at a higher
        dose range, using the highest dose from the first
        experiment as the lowest dose in the second
        experiment, and extending the dose-range well into
        the milligram range.

    All experiments should be repeated at least once.  If the first 
experiment produces a weak or equivocal result (e.g., a dose-
related but less than 2-fold increase in revertants per plate), the 
experiment should be repeated until a consistent picture emerges. 

    Incubation and examination of plates

    Plates should be incubated at 37 °C for at least 48 h before 
being scored.  It is important to ensure that volatile test 
compounds and gases are incubated in closed systems.  After 
incubation, it is essential to inspect the background lawn of both 
treated and control plates with a dissecting microscope in order to 
check for toxic effects (thinning of the lawn) or excess growth, 
which may indicate the presence of amino acids in the test 

2.1.4.  Presentation and interpretation of data  Data-processing and presentation

    The description of the protocol should be detailed enough to 
allow independent replication of the assay.  If a published 
protocol has been used, this should be referred to, and any 
deviations from it should be indicated.  The following information 
should be included in reports:  source of the S9 (strain and 
species of animal); details of inducers; percentage of S9 in the 
S9-mix; mg liver protein per plate; concentration of buffer and 
cofactors; items bought in from proprietary sources (e.g., S9, 
ready-poured plates) should be noted.  "Raw" data should be 
provided:  individual values of numbers of mutant colonies per 
plate should be tabulated in ascending order of dose, starting 
with the solvent controls.  Data from positive controls should be 
clearly identified and separated from the results obtained for the 
test substance.  The doses of test compound should be expressed by 
weight per plate and not by volume.  If the test substance is a 
formulation or mixture, results should also be expressed per weight 

of active ingredient(s).  Providing that a complete raw data set is 
provided, it is also useful to present graphs showing dose-response 
curves.  Interpretation of data in terms of positive and negative

    For a substance to be considered positive in a plate-
incorporation test it should have induced a dose-related and 
statistically significant increase in mutations compared with 
appropriate concurrent controls, in one or more strains of 
bacteria, in the presence and/or absence of S9, in at least two 
separate experiments.  Experience has shown that a doubling or more 
of the background mutation, combined with a dose-response curve, 
indicates a positive response. 

    A test substance is considered negative if it does not produce 
any increase in mutation at any dose, in at least 2 separate 
experiments that complied with the base-line protocol submitted 
with the test report.  This protocol should include the following 
requirements:  the strains used, e.g.,  S. typhimurium strains TA 
1535, TA 1538, TA 98, TA 100;  E. coli uvrA(pKM101); testing at 
doses spaced at 4-fold intervals or less and extending to the 
limits imposed by toxicity or solubility, or, where the substance 
is very soluble, into the milligram range; adequate concurrent 
negative and positive controls, including positive controls to test 
the efficiency of the S9-mix; tests in the presence and absence of 
S9-mix; and finally, evidence of the identity of the bacterial 
strains used in each experiment.  Dealing with ambiguous results

    An ambiguous result arises when, at one or more doses, there 
are more revertants per plate than are seen on concurrent control 
plates, but there is not a clear dose-response relationship.  This 
increase may be consistent in two or more experiments.  The effect 
might occur in just one tester strain and at one particular level 
of S9 in the S9-mix.  Such a result cannot be classified as 
negative, neither is it positive.  The use of historical control 
values to interpret ambiguous results is not recommended. 

    Ambiguous results may be caused by a technical problem, such as 
the presence of nutrients in the test substance or the 
bacteriostatic effect of the test substance; or it might be an 
indication that a change in experimental procedure is required.  In 
addition, in the course of several replicate experiments, one or 
two assays might be positive, and some might be negative.  Results 
of this type may be classified as "irreproducible".  Under these 
circumstances, the use of alternative protocols may resolve the 
problem.  See Venitt et al. (1983) for further discussion. 

2.1.5.  Discussion  How the most critical factors identified above can influence 
the validity of the data 

    The conduct of bacterial mutation tests requires close 
attention to every aspect of the experimental procedure.  Success 
in running large numbers of such tests in routine screening 
programmes depends on the establishment of consistent methods for 
every phase of the experiment.  A deficiency in just one area will 
jeopardize the whole enterprise.  Interpretation of the results in terms of the intrinsic 
mutagenic activity of the test material

    A bacterial mutagenicity assay simply determines whether the 
substance under investigation is or is not a bacterial mutagen in 
the presence and/or absence of an exogenous metabolizing system 
derived from a mammal (S9).  Such a test cannot determine whether 
the test substance is mutagenic and/or carcinogenic in any other 
species.  However, it may be concluded that a substance found to be 
mutagenic in properly-conducted bacterial mutation assays should be 
regarded as potentially mutagenic or carcinogenic for mammals 
(including man) until further evidence indicates otherwise. 

2.2.  Genotoxicity Studies Using Yeast Cultures

2.2.1.  Introduction

    The budding and fission yeasts  Saccharomyces cerevisiae and 
 Schizosaccharomyces pombe, respectively, are among the most 
extensively studied of the eukaryotes and provide convenient tools 
for use in genetic toxicology studies of environmental chemicals.  
The internal structure of the yeast cells shows strong similarities 
to that of the cells of higher organisms, in that they possess a 
differentiated nucleus containing a nucleolus.  The accurate 
functioning of cell division depends on the synthesis of a spindle 
apparatus; however, unlike mammalian cells, yeasts and other fungi 
maintain their nuclear membrane during cell division. 

    The budding and fission yeasts are distantly related and differ 
significantly in the persistance of the diploid phase of the life 
cycle.   S. cerevisiae haploid strains of the  a and mating type, and 
diploid cultures heterozygous and homozygous for mating-type may be 
cultivated in the vegetative phase.  In contrast, in  S. pombe, the 
vegetative haploid cells of mating type  h+ and  h- fuse to produce 
a zygote, which undergoes immediate reduction division (meiosis) to 
produce 4-spored haploid asci.  Thus,  S. cerevisiae strains are 
suitable for routine use as both vegetative haploids and diploids, 
whereas  S. pombe strains are suitable for use only for the 
measurement of genetic end-points detectable in haploids.  Diploid 
vegetative cultures of  S. pombe have been produced by special 
treatments but have not been used in genotoxicity studies so far. 

    Yeasts are physiologically-robust organisms, tolerating pH 
values between 3 and 9; they survive at temperatures from freezing 
to above 40 °C, and growth can occur over a range of approximately 
18 °C - 40 °C.  Growth is optimal at 28 °C and 32 °C for 
 S. cerevisiae and  S. pombe, respectively, using a carbon source 
such as glucose. 

    Diploid cultures of  S. cerevisiae undergo meiosis under a 
variety of conditions, such as those found in exhausted medium and 
in the presence of 1% potassium acetate.  Thus, by varying the 
medium, it is possible to study  S. cerevisiae during both mitotic 
and meiotic cell division.  The uncontrolled induction of meiosis 
and spore formation in exhausted vegetative growth medium can lead 
to problems during long periods of treatment. Such problems are 
eliminated by the use of diploid strains such as  JDI (described 
later), which are unable to undergo meiosis and spore formation. 

    For both fission and budding yeast, there is an extensive data
base of experiments involving their use in studies on the 
genotoxicity of chemicals.  This data base has been reviewed by the 
US Environmental Protection Agency Gene-Tox Program (Loprieno et 
al., 1983; Zimmermann et al., 1984).  However, readers should be 
aware that significant numbers of chemicals in the yeast data base 
were screened prior to the introduction of techniques involving  in 
 vitro mammalian activation mixes.  Thus, many of the apparently 
negative results in the literature may stem from the use of 
unsuitable protocols. 

    The primary advantages of yeasts in genotoxicity studies can be 
summarized as follows: 

    (a) eukaryotic chromosome organization;

    (b) variety of genetic end-points can be assayed;

    (c) cost-effective assays requiring limited technical and
        laboratory facilities using a "robust" organism.

2.2.2.  Genetic end-points

    When used in genotoxicity studies, yeast cultures have 
significant advantages over other test systems in terms of the 
comprehensive range of genetic end-points that can be assayed.  
These end-points include: 

    (a) point mutation in chromosomal and mitochondrial genes;

    (b) recombination, both between and within genes; and

    (c) chromosome aneuploidy during both mitosis and meiosis.

    A number of other variables including membrane damage, 
differential killing in repair-deficient strains, and selective 
effects have also been studied.  However, the data base for such 
events is still limited, and this section will be confined to the 
events classified into groups (a), (b), and (c).  Point mutation

    A variety of forward mutation systems has been used with yeast 
cultures.  However, the most extensively studied system has been 
the one based on the induction of defective alleles of the genes 

for adenine synthesis.  The system involves the use of cultures 
carrying defective mutations of the genes Adenine-1 and Adenine-2 
of  S. cerevisiae and Adenine-6 and Adenine-7 of  S. pombe, the 
presence of which results in the production of red-pigmented 
colonies in  S. cerevisiae and red/purple colonies in  S. pombe,  
owing to the presence of an intracellular pigment (aminoimidazole 
carboxylic acid ribonucleotide in the case of adenine-2 mutations). 

    Forward mutations are detected in such strains by the induction 
of further mutations at 5 genes that precede the production of the 
red/purple pigment in the adenine synthetic pathway.  Such 
mutations result in the production of doubly-defective colonies, 
which can be visually observed as white colonies or sectors.  The 
system can be illustrated as shown below: 

Strain P1 of  S. pombe genotype: ade6-60, rad10-198, h-

ade6-60     ---------------->   ade6-60 ade x <- new mutation
red/purple  forward mutation    white
colonies    produced by both    colonies
            and frameshift

    The assay involves the treatment of red/purple cultures with 
the test agent and the visual screening of colonies produced on 
low-adenine medium for the production of whole white or sectored 
colonies (Loprieno, 1981). 

    The most widely-used yeast strain for the detection of reverse 
mutation is the haploid strain of  S. cerevisiae XV185-14C, 
developed by von Borstel and his colleagues (Mehta & von Borstel, 
1981), which has the following genotype: 

     a ade2-1, arg4-17, lys1-1, trp5-48, his1-7, hom3-10

    The markers  ade2-1, arg4-17, lys1-1, try5-48 are ochre 
"nonsense" mutations, which are revertible by base-substitution 
mutagens that induce site-specific mutations or ochre-suppressor 
mutations in t-RNA loci.  The marker  his1-7 is a missense mutation 
that is reverted mainly by second site mutations;  hom3-10 is 
believed to be a frameshift defect because of its response to a 
range of diagnostic mutagens.  As with most, if not all, of the 
frameshift mutations identified in yeast, the  hom3-10 allele 
reverts at a relatively low frequency and its use in testing 
protocols requires the screening of large populations of cells. 

    In diploid strains of  S. cerevisiae, the only mutation marker 
that has been extensively used is the ilvl-92 mutation that is 
present in the homoallelic condition in strain  D7.  Unfortunately, 
the marker responds to only a limited range of mutagens and it 
would be inappropriate to regard it as a comprehensive point 
mutation screening system for environmental chemicals. 

    The induction of mutations that lead to defects in 
mitochondrial function in yeast may be detected by the assay of the 
frequency of respiration-deficient "petite" colonies, which are 
incapable of aerobic respiration.  Such colonies are characterized 
by their small size and their inability to grow on non-fermentable 
carbon sources such as glycerol.  Petite colonies may be produced 
by the induction of both chromosomal and extrachromosomal events 
but, in diploid cells, those detected are predominantly of 
extrachromosomal origin.  In yeasts, extrachromosomal mutations are 
induced  at high levels by a wide range of chemical mutagens.  
However, at present, the significance of such events is far from 
clear (Wilkie & Gooneskera, 1980).  Recombinationa

    In eukaryotic cells, genetic exchanges between homologous 
chromosomes are generally confined to a specialized stage of 
meiotic cell division which, in yeasts, occurs during the process 
of sporulation.  Recombinational events in yeasts may also be 
detected during mitotic or vegetative division, though the 
spontaneous frequency is generally at least 1000 times less than 
that observed during meiosis.  Sporulating yeast cultures can be 
used to study the rates of both spontaneous and induced meiotic 
recombination, but it is generally the mitotic events that are of 
practical value in genotoxicity studies. 

    Mitotic recombination can be detected in yeasts both between 
genes and within genes.  The former event is called mitotic 
crossing-over and generates reciprocal products whereas the latter 
is most frequently non-reciprocal and is called gene conversion.  
Crossing-over is generally assayed by the production of recessive 
homozygous colonies or sectors produced in a heterogenous strain.  
Gene conversion is assayed by the production of prototrophic 
revertants produced in a heteroallelic strain carrying two 
different defective alleles of the same gene.  Mitotic gene 
conversion can be distinguished functionally from point mutation by 
the elevated levels of prototrophy produced in heteroallelic 
strains compared with levels in homoallelic strains (carrying two 
copies of the same mutation). 
a  Nomenclature
    Genetic loci in this paper are labelled as follows: 

        abbreviation for gene
          ade 6     -    60
        specific       mutant
          gene         number

Capital letters indicate the wild-type form of the gene and lower 
case the mutant form.  The suffixes r and s indicate resistance and 
sensitivity to antimicrobial agents, respectively.  rad loci - 
indicate genes involved in DNA repair.  ‚------- represents a 
chromosome with its centromere. 

    The value of assaying mitotic recombination in yeast in 
genotoxicity studies stems from the observation that both events 
are elevated by exposure to genotoxic chemicals.  These increases 
are produced in a non-specific manner, i.e., levels are increased 
by all types of mutagens, irrespective of their mode of action.  
Thus, the primary advantage of assaying for the induction of 
mitotic recombination is that the events involved are reflective of 
the cell's response to a wide spectrum of genetic damage.  A number 
of suitable strains of  S. cerevisiae have been constructed for use 
in genotoxicity testing.  However, for the purposes of this 
document it will be confined to those most frequently used and 
convenient for use. 

    Mitotic gene conversion can be assayed using selective medium 
in the diploid  S. cerevisiae strain D4 (Zimmermann, 1975).  The 
genotype of D4 is as follows: 

                       a                             gal2
Chromosome III ‚--------------- Chromosome XII ‚---------------
               ‚---------------                ‚---------------
                    alpha                            GAL2

                    ade2-2                        leu1  trp5-12
Chromosome XV  ‚--------------- Chromosome VII ‚---------------
               ‚---------------                ‚---------------
                    ade2-1                        LEU1  trp5-27

     ade2-2, ade2-1, trp5-12, and  trp5-27 are heteroalleles at
the  ADE2 and  TRP5 loci, respectively.  These alleles undergo
mitotic gene conversion to produce prototrophic colonies
carrying one wild-type allele which makes growth possible on
selective medium lacking either tryptophan or adenine, e.g.,

                                       mitotic gene

                     cell division   ‚-----------------
                    with chromosome  ‚-----------------
      ade2-2          replication   /        ade2-2
‚-----------------  --------------->
‚-----------------                  \         or
      ade2-1                         \
                                      \      ade2-1

    The  D4 strain has been extensively used in the study of 
genotoxic chemicals (Zimmermann et al., 1984) and has proved to be 
a valuable tool.  However, the use of the strain is limited by the 
relatively high spontaneous reversion frequencies of the  ADE-2  
marker which means that, if this loci is to be used, cultures with 
low spontaneous prototroph frequency must be selected prior to 
chemical treatment. 

    Mitotic gene conversion can also be assayed in the strain  JDI  
(Sharp & Parry, 1981), which is capable of simultaneously assaying 
mitotic crossing-over on chromosome XV.  The genotype of JDI is as 

                          a           his4c
    Chromosome III  --------------‚-------------
                        alpha        his4ABC

                        ade2        ser1      his8
    Chromosome XV   ‚-------------------------------------
                        ADE2        SER1      HIS8

    Chromosome VII  ‚-------------------

     his4C, his4ABC, trp-U9, and  trp5-U6 are heteroalleles at the 
 HIS4 and  TRP5 loci, respectively.  These alleles undergo mitotic 
gene conversion to produce proto colonies carrying one wild-type 
allele which makes growth possible on selective medium lacking 
either tryptophan or histidine.  Mitotic crossing-over can be 
assayed by the production of red colonies or sectors homozygous for 
 ade-2 and the markers distal on chromosome XV, e.g., 

                           ade2  ser1  his8
                           ade2  ser1  his8
    ade2  ser1  his8     \  /                       ade2  ser1  his8
‚-------------------      \/                      ‚------------------
O------------------- -->  /\               ---->  O------------------
    ade2  ser1  his8     /  \                       ade2  ser1  his8
                           ADE2  SER1  HIS8
                           ADE2  SER1  HIS8

    white colonies       mitotic crossing-over     red colonies or
                                                   sectors auxotropic
                                                   for adenine serine
                                                   and histidine

    Thus, using the strain  JDI, it is possible to assay mitotic 
gene conversion at two separate loci and also to detect one of the 
possible homozygous chromosome combinations produced by mitotic 
crossing-over (but not both reciprocal products).  The strain has 
been selected for its inability to undergo sporulation and is thus 
suitable for long periods of treatment.  Protocols are available 
for the use of this strain under conditions of optimal cytochrome 
P-450 concentrations (Kelly & Parry, 1983a). 

    A particularly convenient multipurpose strain of S.  cerevisiae  
is D7 (Zimmermann, 1975) which carries a set of genetic markers 
that allow the simultaneous assay of mitotic crossing-over, gene 
conversion, and point mutation. 

    The genotype of D7 is as follows:

                    a                         trp5-12    cyhr 2
Chromosome III  ‚-------    Chromosome VII  ‚----------------------
                ‚-------                    ‚----------------------
                  alpha                       trp5-27    CYHS 2

                  ade2-40                       ilvl-92
Chromosome XV   ‚--------   Chromosome V    ‚----------------
                ‚--------                   ‚----------------
                 ade2-119                       ilvl-92

    The heteroalleles of the  TRP-5 locus,  tryp5-12 and  trp5-27,  
undergo mitotic gene conversion to produce prototrophic colonies 
carrying one wild-type allele which makes growth possible on 
selective medium lacking tryptophan.  While  ade2-40 is a 
completely inactive allele of  ADE-2 that produces deep red 
colonies,  ade2-119 is a leaky allele (only partially defective) 
causing accumulation of only a small amount of pigment and thus 
producing pink colonies.  In heteroallelic diploids, the  ade2-40  
and  ade-2-119 alleles complement to give rise to white adenine-
independent colonies.  Mitotic crossing-over in  D7 may give rise 
to the production of cells homoallelic for the  ade2 mutations and 
thus lead to the observation of both red and pink reciprocal 
products, e.g., 

                   ade2-40            ade2-40
               ‚--------------     ‚---------------  red colonies
 ade2-40           ade2-40         O---------------
‚---------     ‚--------------    /   ade2-40
O--------- --> \/                /               
 ade2-119      /\ ade2-119     / 
               O--------------  \
               O--------------   \   ade2-119
                   ade2-119       ‚---------------    pink colonies

    The frequency of induced reciprocal mitotic crossing-over can 
be unambiguously confirmed in  D7 by the visual observation of 
treated colonies.  Mitotic crossing-over can also be assayed in  D7  
by the use of the recessive cycloheximide resistant cyhr 2 allele 
on chromosome VII.  Crossing over between CYH2 and the centromere 

of chromosome VII results in the production of colonies that are 
capable of growth on medium containing cycloheximide (Kunz et al., 
1980), e.g., 

                           cyhr 2
     cyhr 2                cyhr 2                cyhr 2
  ‚----------         ‚-----------------      ‚-------------
  O----------  --->                   --->    O-------------
     CYHS 2                                      cyhr 2
                           CYHS 2
                           CRHS 2

  cycloheximide-      mitotic crossing-over   cycloheximide-
  sensitive colonies                          resistant

    The final genetic event that can be assayed in strain D7 is the 
induction of base-substitution mutation at the homallelic- ilvl-92  
markers by the production of prototrophs that grow on selective 
minimal media that lack isoleucine.  Aneuploidy

    Abnormal chromosome segregations leading to the production of 
numerical chromosome aberrations can be detected in yeasts by 
genetic means using appropriate yeast strains.  Suitable strains of 
 S. cerevisiae are available that are capable of detecting and 
quantifying the reduction of monosomy (chromosome loss) during 
mitotic cell division from the 2n to the 2n-1 condition and the 
production of disomy (chromosome gain) and diploidisation in spores 
produced during meiotic cell division (sporulation) (Fig. 1). 

    Although a number of strains have been developed for the 
detection by genetic means of chromosome aneuploidy, only two have 
been extensively used in the screening of environmental chemicals.  
These are  D6 described by Parry & Zimmermann (1976) and  DIS13  
described by Sora et al. (1982), which have been developed for the 
assay of induced aneuploidy during mitotic and meiotic cell 
division, respectively. 

    The genotype of  S. cerevisiae diploid strain D6 is as follows: 

        ade3       leu1    trp5    cyhr 2    met13
        ADE3       LEU1    TRP5    CYHS 2    MET13

                     ade2                              a
Chromosome XV   ‚--------------   Chromosome III  ‚------------
                ‚--------------                   ‚------------
                     ade2                            alpha


    This strain forms red colonies because of the presence of  ade2 
in a homozygous condition and is sensitive to the presence of 
cycloheximide in the medium because the  cyhr 2 resistance allele is 
recessive.  The loss of the chromosome VII homologue carrying the 
dominant wild-type allele of this group of linked markers results 
in cells that form white (due to the expression of  ade3) and 
cycloheximide-resistant (due to the expression of the selective 
 cyhr 2 marker) colonies that also express the markers  leu1,  trp5 
and  met13 (defined as Go).  Treatment protocols may involve the 
treatment of stationary phase cells in an appropriate buffer, 
exponential phase cells in buffer for a short period before they 
enter Go, or growing cells in nutrient medium or on overlay plates. 

    The vast majority of the experimental studies on chemical 
mutagens using yeasts have involved liquid-suspension assays.  
Plate assays are nevertheless also possible and have been used 
for the assay of mitotic crossing-over, point mutation, and gene 
conversion (Fink & Lowenstein, 1971; Parry et al., 1976; Kunz et 
al., 1980).  The advantages of liquid-suspension assays with regard 
to their ability to quantify cellular toxicity has, however, led to 
a preponderance of the studies in the literature.  Cells of both 
yeast species have also been used extensively in host-mediated 
assays (Fahrig 1975; Loprieno et al., 1976) where they appear to 
tolerate incubation in mammals for long periods without eliciting 
host reactions. 

    Liquid-suspension assays involve treatment of cells with test 
chemicals for periods of preferably less than 24 h, removal of the 
test chemical, followed by plating on nutrient and selective medium 
for quantitation of both cell viability and the genetic end-point.  
Appropriate treatment media for the strains described here can be 
found in the publications of Loprieno, (1981), Mehta & von Borstel 
(1981), Parry & Sharp (1981), Sharp & Parry (1981), and Zimmermann 
& Scheel (1981) for both  S. pombe and  S. cerevisiae.  Specific 
stages of mitotic cell division such as G1, S, and G2, can be 
investigated using synchronized cultures or, more conveniently, the 
separation of exponential phase cells by means of a zonal rotor 
(Davies et al., l978). 

    Exposure of yeast cells to test chemicals is generally 
performed at the optimal growth temperature for the two species, 
i.e., 28 °C and 30 °C for  S. cerevisiae and  S. pombe,
respectively.  When mammalian metabolic activation preparations are 
used (see later) it may be appropriate to incubate cultures at 
37 °C for a proportion of the total treatment time.  In all such 
treatments, it is essential that media are adequately buffered at 
pH 7.0, as yeast cultures rapidly acidify their media.  However, 
advantage can be taken of the pH tolerance of the organisms for the 
testing of chemicals that are biologically active at acid pHs.  
When direct comparisons have been made between liquid yeast 
suspension assays and bacterial plate assays, there has been a 
close similarity in the sensitivity of the two assays (Parry & 
Wilcox, 1982). 

    The assessment of the genotoxicity of chemicals in yeasts 
during meiosis involves the treatment of cells during the process 
of sporulation.  Sporulation can be induced by the transfer of 
vegetative cultures to a medium containing only potassium acetate, 
but maximum levels of sporulation are obtained if the culture is 
pre-grown in a pre-sporulation medium containing both acetate and 
nutrients.  Chemicals can be assayed by exposure of cells 
throughout the sporulation period or by treatment at specific 
stages during meiosis.  Suitable protocols for the assay of the 
effects of chemicals, during meiosis, on mutation and chromosome 
aneuploidy have been described in detail by Kelly & Parry (1983b) 
and Parry & Parry (1983), respectively.  During sporulation, the 
treatment medium undergoes an alkaline pH change which may result 
in the detoxification of some test compounds. 

    Mammalian metabolic activation preparations have been employed 
in the assay of genotoxic chemicals using both fission and budding 
yeasts, and suitable formulae for such mixes have been described by 
Loprieno (1981) and Sharp & Parry (1981), for use with 
 Schizosaccharomyces and  Saccharomyces, respectively.  Most 
preparations are based on those used in bacterial assays, which 
have been described earlier in this report.  Relatively few studies 
have been performed with yeasts using various enzyme-inducing 
agents, mammalian species, and liver fractions, though there is 
considerable scope for such studies (Wilcox et al., 1982).  There 
is now evidence that, unlike  Salmonella, yeast cells have a 
significant endogenous metabolic capacity of their own (Callen & 

Philpot, 1977) and protocols have been developed that produce 
relatively high levels of cytochrome P-450 for periods of up to 
18 h of chemical treatment (Kelly & Parry, 1983a). 

    In yeast liquid-suspension assays, the time of exposure to the 
test chemical depends on the nature of the protocols used, the 
specific chemical being tested, and the yeast strain and genetic 
end-point being studied.  Thus, no specific recommendations can be 
made with regard to the optimal time of exposure required to 
adequately test chemicals.  There are a number of factors that 
should nevertheless be borne in mind when designing an experiment: 

    (a) In studies with vegetative cells, care must be taken
        that, when diploid cells are used, the exposure times
        are not such as to lead to the induction of
        sporulation.  If long exposure times are necessary,
        cells should be checked for spore formation or use
        made of non-sporulating strains such as  JDI.

    (b) Exposure times should be sufficient to allow for
        entry of the test chemical into the cell and the
        production of the damage that provides the substrate
        for the induction of the specific end-point being

    (c) Provision must be made to allow for the expression of
        the end-point, e.g., in the case of the assay of
        induced chromosome aneuploidy a period of
        post-treatment cell division must take place before
        exposure to a selective agent.

    Dose selection is another parameter that is highly dependent on 
a number of experimental variables such as:  the culture used, the 
end-point measured and the nature of the chemical being tested.  In 
general, dose ranges should be selected on the basis of 
cytotoxicity and solubility to include concentrations that range 
from the "no-effect" dose level up to 90% cell lethality with 
approximately 1/3 log spacing.  Dose selection is more difficult 
with assays such as that for chromosome loss where "humped" dose-
response curves are a common feature (Parry et al., l980) and 
maximum induction of the end-point may occur at non-toxic doses. 
Similar problems of dose selection have also been encountered with 
specific chemicals such as the dinitropyrenes, whereas in yeasts, 
the induction of mitotic gene conversion is detectable only at non-
toxic doses and is reduced at higher concentrations (Wilcox & 
Parry, 1981).  In such cases, there is probably no alternative but 
to test a chemical down to arbitrary concentrations of at least 
0.1 µg/ml or to relate the minimum test concentrations to potential 
exposure levels. 

    After exposure to the test chemical, yeast cells are washed and 
plated after dilution on nutrient medium and the appropriate 
screening medium for the end-point under test.  In the case of cell 
viability and genetic end-points that provide a large number of 
scorable events per plate, 3 replicate plates are appropriate.  

However, in the case of relatively rare events, such as mutation in 
frameshift marker strains, and non-selectable events, such as 
forward mutation at the adenine loci, the plate numbers must be 
increased to ensure the statistical significance of the data, as 
with small numbers the standard deviation of the plate counts will 
be large.  There is no generally agreed method of analysing the 
data generated by yeast genotoxicity tests and a number of suitable 
methods have been described, e.g., Loprieno et al. (1976), Sharp & 
Parry, (1981), and Kelly & Parry (1983a). 

    It is essential that all experiments using yeast cells should 
be independently repeated and ambiguous results may require further 
experimentation with careful selection of sample size, treatment 
concentrations, culture stage, or metabolic activation system.  The 
aims of such repeated experiments should be to increase the 
statistical validity of the results. 

2.2.3.  Information required

    Data are best presented in tabular form supplemented with the 
appropriate graphical treatment.  A test report should include the 
following information: 

    (a) strain of yeast used and genotype;

    (b) description of the test conditions, including growth
        phase of cells used, whether growing or non-growing;
        details should be provided of length of treatment,
        dose levels, toxicity, medium, and treatment
        procedure; the negative and positive controls used
        should be clearly specified;

    (c) raw data should be provided to include plate counts
        of viability and colony type selected, calculations
        of survival and frequency of genetic end-point under
        study, and dose-response relationships if applicable;

    (d) the results should be evaluated using an appropriate
        statistical procedure and interpretation provided.

2.2.4.  Interpretation  Significance of positive results in yeast assays

    1.  A positive response in mutation assays is indicative
        of the ability of a chemical to induce point
        mutations in eukaryotic DNA.

    2.  A positive response in assays for mitotic
        recombination indicates the potential of a chemical
        to produce DNA interactions in a eukaryotic cell.
        The majority of such chemicals will be capable of
        producing either point mutations or chromosome
        aberrations in mammalian cells.

    3.  A positive response in assays for chromosome
        aneuploidy indicates the potential of a chemical to
        produce changes in chromosome number in eukaryotic
        cells.  However, in at least a proportion of such
        chemicals, the effect may be specific for yeasts and
        requires confirmation in a mammalian system.

    Such a response should be reproducible in independent 
experiments and should be significant when evaluated by an 
appropriate statistical test.  However, care should be taken to 
ensure that the cultures used showed the "normal" levels of 
spontaneous frequency for the event scored and that treatment 
conditions were not such as to induce sporulation in vegetative 
cultures.  Considerably more weight can be placed on results if an 
unambiguous dose response is observed, though deviations from 
linearity are common for many of the genetic end-points of yeast.  
There are no published data suggesting that yeast assays produce 
false-positive results for any consistent reason.  Negative results in yeast assays

    Negative results may be obtained in yeast assays of chemicals 
for genetic activity for a number of reasons: 

    (a) the test compound is inactive in eukaryotic cells;

    (b) the compound has not been exposed to the appropriate
        metabolic activation system;

    (c) the relevant genetic end-point is not detectable in
        yeast cultures, e.g., the induction of chromosome
        aberrations; or

    (d) the compound has not been tested over an appropriate
        dose range, e.g., a fungicide may lead to cellular
        toxicity before the genetically active cellular
        concentrations are achieved.  Another such example
        may be found in assays such as those for chromosome
        aneuploidy which frequently generate "humped"
        dose-response curves where the genetically active
        range requires the use of extensive concentration

    The validity of a negative result with a test chemical will be 
of more general relevance if data are accompanied by appropriate 
responses from positive control chemicals. 

2.3.  Unscheduled DNA Synthesis in Cultured Mammalian Cells

2.3.1.  Introduction

    The ability of living cells to remove damage induced in DNA was 
first reported in 1964 (Boyce & Howard-Flanders, 1964; Setlow & 
Carrier, 1964).  It is now clear that cells can cut out portions of 
DNA damage in one strand of the double helix, replace the excised 

portion with undamaged DNA nucleotides by using the opposite strand 
as a template and rejoin the newly synthesized section to the pre-
existing DNA strand (Hanawalt et al., 1979) (Fig. 2).  This process 
is called excision repair and restores the original integrity of 
the DNA molecules. 


    Unscheduled DNA synthesis (UDS) is the term used to describe 
the synthesis of DNA during the excision repair of DNA damage and 
as such is distinct from the semiconservative replication that is 
confined to the "S" phase of the eukaryotic cell cycle.  Rasmussen 
& Painter (1964) first reported the incorporation of 3H thymidine 
into the DNA of cultured mammalian cells during the repair of 
damage induced by ultraviolet irradiation.  These authors used 
autoradiography to detect UDS.  This method involves culturing 
cells on glass slides, exposing them to a DNA-damaging agent in the 
presence of a medium containing high specific activity 3H 
thymidine, and observing the radiolabel incorporated during UDS 
into cells that are not semiconservatively replicating DNA.  This 
is done by way of an emulsion or film that detects the beta 
emission from the tritium.  The ability of substances to induce UDS 
in cultured cells is now widely used routinely to assess the 
genotoxic activity of compounds in mammalian systems.  The assay is 
therefore a measure of the amount of repair produced and monitors 

neither the original lesion nor the consequences of repair.  The 
amount of DNA replication associated with UDS is relatively low 
compared with the amount associated with semiconservative 
replication.  If autoradiography is used to monitor this process, 
"S" phase cells that are undergoing semiconservative replication 
are readily eliminated from the analysis because of their heavy 
labelling indices.  In this section, the measurement of 
radiolabelled thymidine incorporated during UDS by either 
autoradiography (Cleaver & Thomas, 1981) or liquid scintillation 
counting (LSC) will be considered (San & Stich, 1975; Martin et 
al., 1978).  Unfortunately, the second method cannot distinguish 
between semiconservative and repair replication.  It measures the 
total amount of DNA replication by monitoring the incorporation of 
3H thymidine into the total DNA, including cells that are actively 
replicating DNA semiconservatively.  The elimination of the 
semiconservative replicative process is therefore an essential 
prerequisite for this approach and can be achieved by various 
methods to be discussed later. 

    UDS has been detected in cells cultured from many mammalian 
species, in various cell types, and with different inducing agents.  
Human fibroblasts (San & Stich 1975) or human transformed cell 
lines such as HeLa (Martin et al., 1978) are often used.  One 
disadvantage of these cell lines is that they do not possess the 
ability to activate proximate carcinogens as does the liver  in 
 vivo, and thus additional metabolic activation in the form of a 
liver microsomal extract is required during chemical exposure.  
Alternative approaches to metabolic activation have been developed 
using epithelial cells derived from liver.  These have been shown 
to retain some of the ability to activate proximate carcinogens 
(Williams, 1977; Dean & Hodson-Walker, 1979). 

2.3.2.  Chemical exposure and UDS

    A large number of mutagens/carcinogens capable of inducing 
many types of DNA damage are known to induce UDS.  The exact 
amount of this repair synthesis depends on (a) the particular 
mutagen/carcinogen in question, (b) the type of DNA repair process 
that operates on the damage induced, and (c) the size of the repair 
patch that is cut out to remove the damage prior to subsequent 
resynthesis.  For example, it is known that some types of DNA 
damage, such as those induced by gamma or X irradiation, are 
repaired relatively quickly in mammalian cells and involve the 
excision of only one to three bases per lesion (Regan & Setlow, 
1974).  Other types of damage such as UV radiation-induced 
pyrimidine dimers, are repaired more slowly and involve the 
replacement of from twenty to seventy bases per lesion (Regan & 
Setlow, 1974).  The ability to detect UDS is further influenced by 
more obvious factors such as whether the cells used take up and 
incorporate (3H)thymidine readily, the concentration and specific 
activity of the (3H)thymidine, and the efficiency of the 
scintillation counter or film used to detect the radiolabel. 

2.3.3.  Procedure  The choice of a suitable cell line

    Transformed cell lines, i.e., those possessing an infinite life 
("immortal" cells) are preferred by some workers, but others have 
used primary cell lines with a finite life of some 30 - 40 
generations.  Generally, transformed cells (e.g., Hela) are easier 
to culture and grow more rapidly than primary cells.  Rapid growth 
is an advantage for general cell culture, but it also means that 
the number of cells actively undertaking semiconservative DNA 
replication in the population at a given time is higher, and the 
sensitivity of the assay is thus slightly reduced.  This fact is 
less significant if autoradiography is used as the means of 
detecting UDS, since replicating "S" phase cells are readily 
excluded from cells at other stages of the cell cycle.  However, if 
scintillation counting is used as the means of detecting UDS, 
precautions must be taken to greatly reduce the amount of 
semiconservative replication in the cell population.  So far, it 
has not been possible to eliminate background levels of residual 
semiconservative replication entirely.  In this respect, 
untransformed cell lines have a distinct advantage in that they 
stop dividing when they reach confluence because of contact 
inhibition when the amount of residual semiconservative DNA 
replication is considerably less than that seen in an actively 
dividing population. 

    A second important factor influencing the choice of cell line 
involves the activation of proximate carcinogens.  Activation can 
be undertaken by the cell itself or by a microsomal liver extract 
added to the cell culture.  The advantages of adding microsomal 
extract are that the same extract can be used in a range of other 
tests on different organisms, hence facilitating legitimate 
comparisons between various end-points.  Furthermore, the source of 
extract can be easily varied if there is concern about the effects 
of a compound in a particular species, organ, or tissue.  The major 
disadvantage is that certain concentrations of the extract iself 
can be toxic to some cell lines.  The use of hepatocytes, because 
of their ability to activate proximate carcinogens without 
microsomal extract, would seem to offer a considerable advantage.  
However, a careful analysis of the hepatocyte line selected is 
essential prior to its use for the routine monitoring of DNA 
damaging agents, because it is also known that that such cells can 
exhibit a reduction in activation ability after a number of 
passages.  Consequently, many research workers prefer to use 
freshly isolated primary hepatocytes for each experiment (Williams, 
1977).  Two further problems have been identified in the use of 
primary rat hepatocytes to screen chemicals for genotoxic activity 
(Lonati-Galligani et al., 1983).  First, hepatocytes show a high 
cytoplasmic background labelling because of the incorporation of 
radioactive thymidine into mitochondrial DNA.  Second, a large 
variation in the functional state of isolated hepatocytes affects 
the reproducibility of the system.  Whatever cell type is chosen, 
it is imperative that a control chemical, known to require 

metabolic activation in order to induce DNA damage, is included in 
each experiment in order to monitor the activity of the endogenous 
metabolizing enzymes in that particular cell population. 

    For the autoradiography method, cells are cultured on cover 
slips or glass slides, and, for scintillation counting, the cells 
are cultured in disposable petri dishes. 

2.3.4.  The elimination of semiconservative replication

    The minimizing of semiconservative DNA replication is an 
absolute prerequisite for the measurement of UDS by liquid 
scintillation counting.  Although this is less important for 
autoradiography, some research workers prefer to suppress the 
semiconservative process in actively-dividing cells, especially if 
the percentage of such cells is high, and if long repair times are 
to be studied.  In this last instance, the proportion of cells 
entering new rounds of DNA replication could be substantial. 

    Hydroxyurea, is commonly used at 10-3M in UDS studies to 
suppress semiconservative replication, and, at this dose, it has 
little effect on the amounts of UDS observed.  It should be noted 
that this drug may react with the microsomal activation mixture 
to produce DNA damage (Andrae & Greim, 1979) and that, at high 
concentrations, inhibitory effects on DNA repair have been reported 
(Collins et al., 1977).  Thus, unless its inclusion is essential, 
this drug is best avoided.  When hydroxyurea is used, the 
appropriate controls to monitor possible effects such as reaction 
with microsomal extract to produce DNA damage should be undertaken.  
Additional treatments, such as incubation with medium containing 
low serum, or, as mentioned, growing cells to a monolayer are 
also used, as both approaches result in a cell population that 
contains considerably fewer replicating cells.  Generally, when 
scintillation counting is the method selected for the detection of 
UDS, one or other of these two approaches is tried, before addition 
of hydroxyurea. 

2.3.5.  Chemical exposure

    Prepared cultures are exposed to a range of doses of the 
chemical to be tested with, if required, the addition of microsomal 
extract and appropriate cofactors.  The choice of a suitable dose 
range is governed by the toxicity of the compound.  Studies should 
usually be undertaken at doses that induce 50% or less cytotoxicity 
as measured by, for example, Trypan Blue exclusion.  In each 
experiment, it is imperative to include the appropriate control 
cultures to ensure that the system is functioning correctly.  Thus, 
background UDS in untreated cells in the presence or absence of 
activation systems and solvents should be included as well as an 
appropriate positive control known to be activated by the 
microsomal activation system used (e.g.,  n-acetylamino-fluorene).  
Cell cultures are usually exposed to five dose levels of the test 
compound and, ideally, each dose is duplicated.  Exposure times are 
usually of the order of one to a few hours, but longer periods have 
been studied. 

2.3.6.  Radiolabelling procedures

    A typical procedure involves adding to the culture medium 
(3H)thymidine of a specific activitya greater than 20 Ci/mmol at 
10 µCi/ml.  This is usually carried out immediately after the test 
compound is added so that both the radiolabel and compound are 
present simultaneously for the duration selected for the 

2.3.7.  Detection of UDS

    The autoradiographic method involves the removal of the medium 
containing the test compounds, followed by rinsing and fixation of 
the cells, coating the slides with autoradiographic emulsion and 
then drying them prior to developing.  Procedures will vary with 
the particular process used and are described in detail by Cleaver 
& Thomas (1981).  After developing, the cells are stained and the 
grains in the emulsion over the cell nuclei of control and treated 
samples are either observed microscopically and counted visually or 
with an electronic counter.  The data are expressed as grains per 

    In the liquid scintillation method, culturing and exposure 
procedures are similar to those for autoradiography, except that 
more cells and more replicate samples are usually analysed.  The 
data are expressed as disintegrations per min (dpm) of incorporated 
(3H) thymidine per µg of DNA.  Hence, not only the amount of 
radioactivity per sample needs to be determined, but also the 
amount of DNA.  This can be carried out by DNA extraction with 
perchloracetic acid  hydrolysis (Schmidt & Thannhauser, 1945), 
using one aliquot for reaction with diphenylamine to measure DNA 
concentration (Burton, 1956) and a second aliquot for scintillation 
counting to measure (3H) thymidine incorporation.  Alternative 
methods for estimating DNA concentrations are available (San & 
Stich, 1975). 

2.3.8.  Data processing and presentation

    For a compound to be accepted as positive with the UDS assay, 
there should be:  (a) a dose-related increase in UDS, and (b) a 
statistically-significant increase in UDS above that of a negative 
control.  Data are usually presented as grain counts per nucleus 
(often as histograms), or (3H) incorporation as dpm per µg DNA, as 
determined by scintillation counting.  The data should include the 
results from all treatments and controls.  At high concentrations 
of test agents, the amounts of UDS may plateau because of the 
saturation of repair mechanisms, or they may even decrease due to 
cytotoxicity.  This again emphasises the importance of undertaking 
experiments over a wide dose range, and then selecting a narrow 
range from the initial data to verify a potentially positive 

a The specific activity denotes how much of the thymidine is 
  actually radiolabelled.

    Various criteria have been used for the definition of a 
positive result.  Investigators have considered a compound positive 
when it induced at least 150% of the control levels of UDS as 
measured by liquid scintillation counting (San & Stich, 1975), or 
when it induced at least 6 grains per nucleus in excess of 
background levels with autoradiography (Williams, 1977).  In 
addition to such basic criteria, the data should also be subjected 
to statistical analysis to determine whether or not the increases 
are significant.  Cleaver & Thomas (1981) recommend that, when 40 - 
100 cells per slide are counted for several slides per treatment, 
the average grain number for each slide can be used as a measure of 
UDS, and the average and standard error of these averages would be 
the more suitable parameter for the amount and accuracy of the 
data.  When liquid scintillation counting is employed, the standard 
deviation or standard error of the mean should be included to 
describe the distribution of the data.  Additional analyses such as 
analyses of variance, non parametric comparisons of grain 
distribution, and estimates of the correlation between UDS and dose 
can be undertaken, and the selection of the most appropriate method 
will depend on the design of the experiment.  The t-test, used by 
some, increases the chances of obtaining false positives, whereas 
though the analysis of variance does not introduce this problem it 
does reflect cytotoxic effects.  Ideally, the analysis should be 
complemented with a contrast analysis that can distinguish between 
treatments giving negative, positive, or cytotoxic effects.  For a 
more complete review of the statistical analysis of UDS data, the 
Gene-tox report on UDS tests by Mitchell et al. (1983) can be 
referred to. 

2.3.9.  Discussion

    UDS is a relatively straightforward approach for measuring DNA 
repair and as such is extremely useful for examining compounds that 
are potentially genotoxic for mammalian cells.  Nevertheless, it is 
usually undertaken as part of a battery of screening tests.  In 
exceptional cases, it may be the only assay to provide a positive 
result and, in such cases,  in vivo tests should be undertaken to 
clarify this result.  It should also be noted that this assay 
detects the repair of DNA damage.  The assay would not detect a 
compound that induced an unrepairable lesion in DNA, though the 
same compound would be expected to induce genetic damage, e.g., 
mutations, in other test systems.  The usefulness of DNA repair 
assays in screening is more fully reviewed by Cleaver (1982). 

    It is obvious from the procedural discussion in this section 
that a number of different systems are currently used to measure 
UDS.  Thus, it would seem appropriate to list and briefly evaluate 
critical factors that can influence each type of assay, and which 
have been mentioned at various stages in the text.  Choice of cell line

    If possible, untransformed human fibroblasts or primary rat 
hepatocytes should be used, because semiconservative DNA 
replication is more readily suppressed in the former, whereas the 

latter are essentially non-dividing and can themselves activate 
proximate carcinogens.  Furthermore, a relatively large pre-
existing data base is available for both cell types, which enhances 
the possibility of making comparisons with other compounds tested.  
However, it should be remembered that primary rat hepatocytes 
exhibit high levels of incorporation of radioactivity into the 
cytoplasm.  It has been suggested that for autoradiographic 
estimates, instead of subtracting cytoplasmic grains from nuclear 
grains, as is usually done to account for non-nuclear 
incorporation, grains overlying the nucleus and a cytoplasmic area 
should be scored and plotted separately in these cells (Lonati-
Galliganai et al., 1983). The variation in the functional state of 
freshly isolated hepatocytes can be a problem and it is imperative 
to undertake adequate controls to verify their ability to activate 
a pro-carcinogen.  Choice of protocol

    If rat hepatocytes are used, autoradiography is the method of 
choice, as liquid scintillation counting (LSC) requires such a 
large number of cells.  Human fibroblasts can be analysed either by 
autoradiography or by LSC, but it should be borne in mind that the 
latter approach requires more cells per sample, more duplicate 
samples, and the addition of hydroxyurea to supress residual 
semiconservative replication.   At concentrations above 10-3 M, 
hydroxyurea can inhibit DNA repair and can also induce DNA damage 
by interacting with microsomal extract.  These facts have to be 
considered when interpreting data.  Where costs have to be kept to 
a minimum, it should be noted that a liquid scintillation counter 
is expensive compared with the cost of a microscope for 
autoradiography.  However, a counter can process many samples 
automatically, whereas microscopic analysis is more time consuming.  Method of activating proximate carcinogens

    The use of rat hepatocytes removes the necessity for the 
addition of microsomal extract.  Whatever kind of microsomal 
activation is used, it is important to include appropriate controls 
to verify that the activation system is functional. 

    Finally, it should be emphasized that, regardless of the system 
used, the most important features in undertaking UDS studies are a 
full understanding of, and extensive experience with, the test 

2.4.   In Vitro Cytogenetics and Sister-Chromatid Exchange

2.4.1.  Introduction

     In vitro cytogenetic tests are designed to demonstrate the 
induction of chromosome damage (aberrations), visible under the 
light microscope, in cultured cells (Fig. 3).  This usually 
involves examination at the metaphase stage of the cell cycle 
(Evans & O'Riordan, 1975; Savage, 1976).  Though other methods such 
as anaphase analysis and enumeration of micronuclei have been used, 

they are not generally considered suitable for routine testing in 
cultured cells.  A physical or chemical agent is classified as a 
clastogen if it produces an increase in the number of breaks in 
chromosomes over that found in control samples.  Cytogenetic tests 
therefore assess gross damage to the DNA involving at least one 
double-strand break.  A detailed discussion of the theoretical 
aspects of the development of chromosome aberrations is given in 
section 2.8. 


    Many agents only induce visible chromosome damage after the 
cells have undergone a round of DNA replication, and the test must 
be designed to allow enough time after treatment for aberrations to 
develop.  However, damaged cells may not survive for more than one 
or two cell cycles after aberrations have been induced, and it is 
essential that cells should be examined in their first metaphase 
after treatment (Evans, 1976). 
    Induction of chromosome aberrations involves major damage to 
chromosome structure, and thus to the DNA, and so clastogenic 
agents must be viewed as potentially harmful.  Although cells with 
visible chromosome aberrations are unlikely to have the potential 
to survive, repair of DNA damage may have occurred in apparently 
undamaged cells, and if this is error-prone, mutations could 
result.  Certain types of chromosome damage, such as some deletions 
and rearrangements (translocations, inversions), may not be lethal.  
The comparatively high level of chromosomal disorders in man 
emphasizes the importance of chromosome changes in human 
populations (DHSS, 1982). 

    The preparation of material for examination of the chromosomes 
is technically simple, and this has undoubtedly contributed, in 
part, to the widespread use of short-term cytogenic tests.  
However, accurate and reliable scoring of metaphase chromosomes for 
aberrations does require a high level of expertise. 


    As its name implies, sister chromatid exchange (SCE) involves 
an apparently symmetrical change between chromatids within a 
chromosome (i.e., between identical sequences of DNA).  SCEs are 
only visible under the microscope if sister chromatids can be 
distinguished (Fig. 4); this requires different culture methods 
from those used in the preparation of metaphases for the scoring of 
chromosome aberrations.  Because of the ease of preparation and 
scoring, it is a very widely-used test in the study of mutagens.  
SCE induction alone is not generally accepted as sufficient 
evidence to classify an agent as mutagenic.  The mechanism of SCE 
induction is not fully understood, though a number of models have 
been proposed (Wolff, 1982).  Some clastogens induce only a small, 
or no, increase in SCEs, X-irradiation being a particularly 
striking example.  There is a high background level of SCE compared 
with chromosome aberrations; it is rare to find many cells without 
SCEs in untreated samples.  This may be partly due to the 
5-bromodeoxyuridine (BrdUrd) that is added to the culture medium in 
order to visualise SCEs, since BrdUrd itself is known to induce 
SCEs (Latt et al., 1981).  This basal level also implies that cells 
with SCEs are capable of subsequent growth.  SCE evaluation may 

thus be a more valid indicator than chromosome breakage of events 
compatible with cell survival, hence its widespread use in 
mutagenicity screening programmes. 

2.4.2.  Procedure: chromosomal aberrations

    The techniques used in the study of chromosomal aberration have 
been described in detail by Evans (1976) and in the Gene-Tox 
reports of Latt et al. (1981) and Preston et al. (1981).  Cell types

    Two types of cells are used most widely for both tests.  These 
are CHO, an established fibroblast cell line, derived from Chinese 
hamster ovary, and human peripheral blood lymphocytes (mononuclear 
white blood cells).  The small number of chromosomes in CHO cells 
(modal number 22) makes scoring relatively straight forward.  It is 
an easy cell line to maintain, using standard tissue culture 
techniques, and, with a cell cycle time of 12 - 14 h, grows very 
rapidly (Latt et al., 1981; Preston et al., 1981; Dean & Danford, 

    Human peripheral lymphocytes do not divide spontaneously in 
culture, but can be stimulated to divide by treatment with a 
mitogen such as phytohaemagglutinin (PHA).  Cultures are initiated 
from fresh blood samples and are not maintained for more than a few 
cell divisions.  The first metaphase after the addition of the 
mitogen is not reached for about 36 - 40 h, after which the cells 
divide about every 18 h, with considerable variation, both within 
and between cultures.  Culture methods are described in detail by 
Dean & Danford (1975) and Evans & O'Riordan (1975). 

    Other cell lines have been used in cytogenetic assays, for 
example, lines with endogenous metabolizing capacity, such as rat 
liver epithelial cells (Dean & Hodson-Walker, 1979).  It is 
important for a cell line to be fully validated with a range of 
suitable chemicals before being used for routine testing. 

    Initially, information on the toxicity of a test agent is 
required, and subsequently, concentrations up to a level where some 
toxicity is observed are used in the cytogenetic assay.  Toxicity 
can be assessed, for example, by measurement of the mitotic index 
or by cell counts.  To obtain sufficient data from the chromosomal 
aberration assay, a minimum of 3 replicate cultures of each of 3 
doses, or 2 replicates of 4 doses is advisable, in addition to a 
negative control (solvent only) and a positive control.  With 
lymphocyte cultures, it is recommended to use blood from at least 
two different donors in each experiment.  Under rare circumstances, 
it may be possible to use a positive control structurally related 
to the test material; more commonly, the positive control will be a 
known clastogen.  A direct-acting clastogen or one requiring 
metabolic activation is used, as appropriate (see below).  The 
doses of the test agent should range from a concentration showing 
some toxicity down to 1/4 or 1/8 of this, or equivalent log doses.  

Since many clastogens only show effects close to the toxic dose, 
there is rarely any advantage in selecting lower concentrations for 
routine screening.  Culture methods

    Cell lines are grown either in small tissue-culture flasks or 
directly on sterile microscopic slides or cover slips until the 
cells are proliferating.  The agent under study is then introduced, 
preferably by replacing the culture medium with medium plus agent.  
Lymphocytes do not attach to culture flasks or slides, and are 
usually grown in small bottles as suspension cultures.  The test 
agent can be added to lymphocyte cultures when they are set up.  It 
is preferable, however, to allow time for the cells to leave the GO 
stage before adding the test agent, because toxic concentrations of 
the test agent may prevent the cells from entering the cell cycle.  
Thus, addition of the test agent after 24 - 36 h of culture is more 
likely to be effective. 

    Modifications to either system may be required; for example, 
volatile or gaseous compounds should be tested in a closed system.  
In some instances, components in the serum used in the culture 
medium may bind to the test agent, in which case it is desirable to 
treat the cells in serum-free medium for a few hours, and then 
continue culturing in normal medium.  Erythrocytes (red blood 
cells), present in lymphocyte cultures set up from whole blood, can 
also bind the test agent.  Various methods of separating out the 
lymphocytes are available (Boyum, 1968), though whole blood 
cultures are more widely used.  Unless the cell line has been shown 
to have intrinsic metabolizing capabilities, or there is strong 
evidence that the agent under test is direct-acting (requiring no 
metabolism), a metabolizing system, such as an S9 microsome 
fraction obtained from rat liver, must be included (section 2.1).  
When testing chemicals of unknown mutagenic activity, assays both 
with and without a metabolizing system are required.  If there is 
compelling evidence that the agent is either direct-acting or 
requires activation, then the first test can be carried out either 
without or with a metabolizing system.  If equivocal or negative 
results are obtained, further tests will be necessary. 

    The times of exposure differ between the two test systems, and 
are detailed below, but it is important to bear in mind that S9 can 
be toxic to mammalian cells.  Thus the time of exposure to the test 
agent may be limited to only 1/2 - 3 h in assays using S9.  In 
these cases, the medium containing the test material and S9 is 
replaced by normal medium, until the cells are harvested.  
Treatment at toxic doses may extend the cell cycle time, and the 
period between treatment with the test chemical and harvesting 
should be extended accordingly.  It may be necessary to use more 
than one sampling time; this is discussed in more detail in section 
2.4.5.  Chromosome assay

    Following the introduction of the test agent, the cells are 
incubated for between one and two cell cycles so that the majority 
of the mitotic cells are in the first metaphase after treatment, 
when harvested.  Before harvesting, a spindle poison such as 
colchicine is added to arrest cells at metaphase.  The cells are 
treated with a hypotonic solution, such as 0.56% (0.075 M) 
potassium chloride, and then fixed, usually with a freshly prepared 
3:1 mixture of methanol and glacial acetic acid.  Subsequently, the 
cells are stained with Giemsa, orcein, or other chromosome stain, 
and are then examined under the microscope.  The slides should be 
coded randomly and independently, and at least 100 metaphases 
scored from each replicate.  However, if the mitotic index is 
greatly reduced at the highest dose, it is not always possible to 
score 100 cells.  Aberration types are usually classified into 
chromatid (involving only one chromatid) and chromosomal or 
isochromatid (affecting both chromatids) and are clearly described 
by Evans & O'Riordan (1975), Savage (1976), ISCN (1978), and Scott 
et al. (1983).  Other classification systems have been devised, 
such as that of Buckton et al. (1962), in which chromosome 
aberrations are classified as stable (Cs), such as translocations, 
or unstable (Cu), such as dicentrics, depending on whether or not 
they can be maintained through successive cell cycles.  Banding 
techniques (Evans, 1976), commonly used for detailed chromosome 
analysis, are rarely used for routine screening. 

2.4.3.  Procedure: sister chromatid exchange

    Cultures to demonstrate SCEs are set up as for chromosome 
assays, but in addition to the test agent, BrdUrd is present, at 
concentrations of 10 - 25 µM, throughout the period from treatment 
to harvesting.  The cells must be allowed to pass through two 
rounds of DNA replication (S-phases) before harvesting.  BrdUrd is 
incorporated into the newly synthesized DNA in place of thymidine, 
and thus, at the first metaphase, all chromatids possess one DNA 
strand containing BUdR and one containing thymidine.  The 
chromatids separate at anaphase, and at the second S-phase, the DNA 
synthesized again contains BUdR.  Since one DNA template is 
unsubstituted and the other substituted with BUdR, the chromosomes 
now contain DNA with one chromatid completely (bifiliarly) 
substituted and the other, half (unifiliarly) substituted with 
BrdUrd.  These chromatids stain differentially following the 
treatment described below. 

    The cultures are harvested as usual but, thereafter, stained 
with Hoechst 33258 solution, exposed to a light source emitting 
long-wave UV radiation, and then stained with Giemsa.  The 
bifiliarly substituted chromatids stain to a much lesser degree 
than the unifiliarly substituted chromatids.  Where SCE has 
occurred, a change in the staining intensity can be seen on one 
chromatid, with a reciprocal change on the other (Perry & Wolff, 
1974).  These second metaphase chromosomes are referred to as 
harlequin chromosomes.  First metaphase cells have uniformly darkly 
stained chromatids, and third and subsequent metaphases have a 

mixture of pale-staining and harlequin chromosomes.  Since second 
division metaphases can be recognized by having entirely harlequin 
chromosomes, the problem of mitotic delay is not so great with SCE 
analysis as with assessment of chromosome aberrations.  Though 
there may be a reduction in the proportion of second-division 
metaphases at the highest doses, these can be recognized and 
scored.  A number of reviews of SCEs are available; Latt et al. 
(1977), Perry (1980), and Wolff (1982) include photographs of 
differentially stained chromosomes. 

    The slides should also be coded, and at least 30 (preferably 50 
or more) metaphases examined from each culture. 

2.4.4.  Procedure: scoring

    Accurate identification of chromosome aberrations and scoring 
of SCEs requires a high degree of skill, and should only be 
undertaken by suitably trained and experienced personnel.  
Descriptions of aberrations do not usually include details of 
potential artefacts, and it is essential to appreciate that there 
are a number of normal chromosome orientations which the 
inexperienced may score as aberrations.  In addition, the quality 
of the material must be sufficiently high for accurate assessment, 
and analysis of "fuzzy", overlapping, or highly-scattered 
chromosomes should not be attempted. 

    Results, including types of aberrations observed, should be 
recorded on a suitable score-sheet.  For chromosome aberrations, 
some record, either the vernier reading on the microscope stage or 
a photograph of each aberrant cells, is usually taken.  It is 
obviously important to ensure that the same metaphases are not 
scored twice.  For further details, see section 2.8. 

2.4.5.  Extent of testing

    The decision as to whether a chemical has been tested 
sufficiently to classify it as positive or negative in these test 
systems is not always clear cut.  If a clear positive response is 
seen, or if there is no increase above negative control levels at 
any dose, provided there is evidence of toxicity at the highest 
dose, no further testing is necessary.  If, however, the negative 
control gives unusually high values, or the positive control fails 
to induce the expected number of aberrations, this would suggest 
that the experiment should be repeated.  A weakly positive response 
will often need to be confirmed in an additional experiment, 
though, under some circumstances, adequate data may be obtained 
from simply scoring additional metaphases on slides already 
examined.  Otherwise, it may be necessary to repeat the experiment, 
either exactly as before or using different dose levels or exposure 
times.   A particularly important aspect of cell kinetics under the 
influence of toxic doses is the delay in the cell cycle time.  
Thus, an apparently negative dose-response can be obtained if cells 
at higher doses have not undergone a round of DNA replication 
between treatment and harvesting, when examining an S-dependent 
agent.  In these cases, a later sampling time is required in a 
repeat experiment. 

2.4.6.  Data processing and presentation

    When a clear dose-related increase in chromosomal aberrations 
is obtained, or when there is clearly no increase above control 
levels, the result may be obvious without the need for statistical 
analysis.  It is, however, always advisable that results should be 
subjected to an appropriate statistical analysis.  Metaphase 
analysis, particularly if gaps are excluded, can yield such small 
numbers of aberrations that objective interpretation is only 
feasible after such statistical tests as the Chi2 and Fisher's 
Exact test (Sokal & Rohlf, 1969). 

    The data are best presented in tabular form showing the 
results for each dose, and the positive and negative controls, 
and including details of the cell line, culture conditions, and 
slide codes.  The minimum data required are:  the doses, number 
of metaphases observed from each culture, and either the 
percentage of aberrant metaphases, including and excluding gaps, 
and the total number of aberrations, or the average number of SCEs 
per cell.  Further breakdown of chromosomal aberrations into 
chromatid/chromosomal-type aberrations, and classes within these 
groups (gaps, breaks, exchanges, etc.) is also essential.  Graphic 
representation of a dose-response can also be helpful. 

    Test agents are regarded as unequivocably positive if a dose-
related increase is observed over 3 or 4 doses (including the 
negative control), and/or 3 or 4 doses give aberration or SCE 
levels significantly higher than the negative control level.  
If no dose-related increase is observed, and no dose gives a 
significantly higher frequency of chromosome aberrations or SCEs 
than the control, then the data are interpreted as negative.  Weak 
or marginal findings usually require additional data or testing, 
but a reproducible dose response rising just above control levels, 
provided this is statistically significant, is the criterion for 
the designation of a weakly positive result. 

    Compounds should normally be tested up to concentrations that 
induce detectable toxicity or reduction in mitotic index.  Agents 
that show no evidence of toxicity in preliminary studies should 
be tested to the limit of solubility.  However, very high 
concentrations of some non-toxic chemicals may interfere with 
culture conditions, and the maximum dose level should be decided on 
a case by case basis. 

    Occasionally, a single high value is found at one dose or 
replicate culture.  This requires a repeat experiment, using a 
narrower dose range above and below that dose.  If the increase in 
aberrations or SCEs at the dose is not reproduced, then the 
isolated result can be discounted.  Another potential problem of 
interpretation is an increase in chromatid gaps at the highest 
dose.  The significance of gaps is discussed in section 2.8.  In 
the absence of an increase in other types of aberrations, it is 
possible that the increase in gaps is associated with the 
cytotoxicity of the test material and is not necessarily of 

genotoxic significance.  In such instances, it is particularly 
important to consider data from other test systems before 
evaluating the mutagenicity of the chemical. 

2.4.7.  Discussion  Critical factors

    The assessment of cytogenetic damage and SCEs depends crucially 
on accurate scoring.  It must be emphasized that this, in turn, 
depends on the training and experience of the scorer as well as the 
quality of the material.  Both under- and over-estimation can 
result from failure to meet these criteria. 

    It is also necessary for the cell line used to be suitably 
validated with known clastogens before it is used routinely.  Even 
where an apparently strongly positive response is obtained, the 
results may be viewed with question if, for example, a highly 
unstable cell line is used.  This emphasizes the need to use either 
the standard cell types (CHO or lymphocytes) or to establish an 
adequate data base for a new cell line, particularly with regard to 
its metabolizing capacity and the level of aberrations in untreated 
cells.  Experimental design and analysis

    If information is available on the chemical structure and 
metabolizing requirements of the test agent, the most suitable 
medium for treating the cells (complete or serum-free) can be 
assessed.  The value of using toxic doses is emphasised in cases 
where serum components are suspected to react with the test agent, 
as it serves to confirm that the agent has entered the cells. 

    A particularly important aspect is the influence of the test 
agent on the cell cycle time.  Mitotic delay is frequently found at 
toxic doses, and this may result in the examination of metaphases 
in which visible aberrations have not had time to develop.  
However, the next lowest dose should indicate whether or not this 
is the case; this can be further checked by testing additional 
doses and sampling times.  With SCEs, the problem still arises, but 
because second division metaphases can be recognised and are 
scored, it is possible to assess whether or not significant mitotic 
delay has occurred. 

    The system used to ensure that metabolic activation has 
occurred can lead to a number of problems.  The toxicity of S9 has 
already been mentioned, and cells retaining metabolic activation 
may vary in the extent and range of activation, and, furthermore, 
some may lose their activating ability following repeated 
subculturing.  A positive control structurally related to the test 
compound, in addition to a standard positive control, is ideal, but 
in practice this is rarely available. 

    An assessment of clastogenicity can only be accepted if the 
results from negative and positive controls are within the expected 
range.  The occurrence of a high value in the negative control 
cultures is by no means unusual, but, in a well-characterized line, 
its validity can be tested statistically.  Any indication that a 
positive control agent (especially one requiring activation) has 
not been detected would also require a repeat of the experiment. 

    Further studies are necessary to clarify unexpected or 
ambiguous results, in which case the use of additional sampling 
times will often provide confirmatory data.  If a clear negative or 
positive result is obtained, there is usually no need for 
verification in a second experiment.  If toxic doses cannot be 
achieved, a different solvent or a variation in the exposure time 
should be considered.  Unfortunately, the low tolerance of cultured 
mammalian cells to pH changes and organic solvents precludes more 
than slight variations in the culture conditions. 

2.4.8.  Conclusions

    The ability of a chemical to produce undoubted double-strand 
breaks, i.e., a strong clastogenic action, is an important finding 
as is strong evidence that there is no such activity.  Thus, it is 
important to try to resolve the ambiguous and weakly positive 
results.  These can usually be clarified by further testing.  An 
increase in gaps alone only implies a discontinuity in staining, 
and therefore may not be due to double-strand breaks.  Agents that 
induce gaps may cause disturbances in normal chromosome structure, 
but not necessarily chromosome breakage (section 2.8).  Since the 
precise mechanism of SCE induction is not fully understood, the 
genetic significance of SCEs cannot be determined, at present.  
However, they reflect a direct interaction of chemicals with DNA 
and represent a useful system for the detecion of genotoxic 
chemicals.  With certain exceptions, chromosome-aberration and SCE 
assays correlate well with other tests. 

2.5.   In Vitro Cell-Mutation Assays

2.5.1.  Principles and scientific basis of the assay

    The use of cultured mammalian cells, including human cells, for 
mutation studies can give a measure of the intrinsic response of 
the mammalian genome and its maintenance processes to mutagens, 
while offering rapidity of assay and ease of treatment compared 
with the use of whole animals. 

    Several forward and reverse mutation selection systems are 
available for use with cultured cells (Abbondandolo, 1977).  The 
basis of the majority is that the cells are cultured in a 
"selective medium" containing a toxic compound or anti-metabolite 
(called the "selective agent") which is toxic to all normal, non-
mutant cells, but in which rare, mutant cells can continue to grow 
to form colonies.  A major requirement for such assays is that 
evidence should be provided that the end-point of the measurement 
is a mutational event that occurs at a specific gene locus, that 

is, it should be consistent with the induction of a heritable 
alteration in the DNA sequence.  In general, detailed biochemical 
analysis of the gene product and cytogenetic study of the 
chromosomes points to the mutational origin of the selected 
colonies, though it is possible that some of the phenotypic changes 
observed may be the result of the kinds of non-mutational changes 
in gene expression that occur during normal development 
("epigenetic events") (Siminovitch, 1976). 

    One problem associated with the selection of mutants of 
mammalian cells is that two copies of each gene are present in a 
normal diploid cell, and, in many cases, the mutant gene product 
acts recessively:  that is, adequate gene product is transcribed 
from one (non-mutant) gene copy to fulfil the cell's needs.  In 
this case, mutation of both genes in a diploid cell (a very rare 
event) is necessary to detect the mutant phenotype.  Therefore, 
mutation of such genes in cultured mammalian cells is studied in 
the hemizygous or heterozygous state, using either X-chromosome 
genes (where only one X is present in male cells, or only one 
active X in female cells) or autosomal genes that have either been 
found, or deliberately selected to be hemizygous in some cells 

    An example of an X-chromosome-located gene that is the basis of 
a mutation selection system is the gene coding for the enzyme 
hypoxanthine-guanine phosphoribosyltransferase (HPRT).  HPRT is one 
of a number of "salvage" enzymes in which the function is to 
salvage the degradation products of nucleic acid synthesis (purines 
and pyrimidines), but which are not essential for the survival of 
the cultured cells, since these bases can be synthesized  de novo.  
HPRT catalyses the conversion of guanine and hypoxanthine to the 
corresponding nucleoside-5'-monophosphates.  In cells containing 
HPRT, toxic purine analogues, for example 6-thioguanine or 
8-azaguanine, are also incorporated, and this forms the basis of 
the selection of mutants.  Thus, cells with normal, non-mutant HPRT 
are killed when they are cultured in the presence of these 
selective agents, but mutants, with altered, or non-functional HPRT 
(or its complete absence) are able to survive and form colonies, 
because the toxic analogues are not incorporated into DNA or RNA.  
Purines continue to be made in the mutant cells by the  de novo  
pathway.  It is interesting to note that human beings with a rare 
sex-linked recessive disease, the Lesch-Nyhan syndrome, are mutant 
at the HPRT locus.  All cells from males with this disorder are 
HPRT deficient and are resistant to the toxic effects of 
6-thioguanine and 8-azaguanine. 

    Another gene coding for a "salvage" enzyme, this time on an 
autosomal chromosome, is the thymidine kinase (TK) gene.  This 
enzyme incorporates exogenously supplied thymidine, and its toxic 
analogues, into the cell.  In this case, both homozygous (TK+/+) 
and heterozygous (TK+/1) diploid cells contain sufficient 
thymidine kinase for the cells to be killed when they are cultured 
in the presence of toxic pyrimidine base analogues such as 
5-bromodeoxyuridine or trifluorothymidine.  Mutants that do not 
contain any functional TK (TK-/-) do not incorporate the analogues, 

and are therefore able to survive and form colonies in the presence 
of these selective agents.  Normal diploid cells contain two copies 
of the TK gene, and as simultaneous mutation of both genes is a 
very rare event, a heterozygous(TK+/-) cell line must first be 
constructed for mutation assays based on this enzyme to be 

     Since complete loss of the "salvage" enzymes HPRT and TK is 
not deleterious to cultured mammalian cells, all types of 
mutations, including base-pair substitution (which may result in 
altered gene product), frame-shifts, and deletions (which result in 
complete lack of enzyme) should be detected.  Evidence to support 
this has been presented in the considerable literature available on 
the HPRT locus (Caskey & Krush, 1979) and the TK locus (Hozier et 
al., 1981). 

    In both the above cases, the mutant gene product acts 
recessively.  A few mutation systems rely on the semi-dominant 
action of the mutant gene product, and in these cases mutation in 
only one of the two genes present in a diploid cell is necessary to 
detect the mutant phenotype.  For example, mutation to the semi-
dominant phenotype of ouabain-resistance involves an essential 
enzyme, the membrane-bound Na+/K+-dependent ATPase (Baker et al., 
1974).  Ouabain kills cells by binding to this enzyme and causing 
an imbalance in ion flow, but rare mutants can be found that fail 
to bind ouabain while retaining functional ATPase activity.  The 
range of mutations detected by ouabain resistance would be expected 
to be much more restricted than for the non-essential salvage 
enzymes, because, if a large mutagenic change, e.g., a deletion, 
occurred in the gene coding for the Na+/K+ ATPase, the essential 
enzyme function would be lost together with ouabain binding, and 
the mutant cell would die.  Some evidence has been presented to 
support this conclusion (Baker, 1979). 

2.5.2.  Relevance and limitations

    In the testing of potential chemical mutagens, a major 
limitation of cultured mammalian cell systems is the difficulty of 
simulating  in vitro the type and quantity of metabolic activation 
that may occur in different tissues  in vivo.  This is because the 
cultured cells lack the full range of enzymes required to activate 
the diverse range of potential mutagens and carcinogens encountered 
in the environment.  Thus, the experiments must be conducted in the 
presence of an "exogenous" metabolic activation system.  This may 
be supplied by the use of rat liver homogenates ('S9') or by co-
cultivating the tester cells with metabolically competent cells, 
such as freshly isolated rat hepatocytes.  The choice of the 
metabolizing system(s) and the way that it is applied in the assay 
has great importance for the efficiency of the test in predicting 
the mutagenic or carcinogenic potential of chemicals that require 
activation.  Considerable further work is required to determine the 
optimal conditions for the  in vitro activation of many chemicals. 

2.5.3.  Procedure  Outline of the basic technique

    A large population of cells is exposed to the test substance, 
with and without an exogenous metabolic activation system, for a 
defined period of time.  After removal of the test substance, the 
cytotoxicity is determined by measuring the colony-forming ability 
and/or the growth rate of the cultures after treatment.  Bulk 
cultures of the treated cells are maintained in a growth medium for 
a sufficient period of time to allow the newly induced mutations to 
be detected.  During this period, known as the expression time, 
the growth rate can be monitored and the cells sub-cultured if 
necessary.  The mutant frequency is then determined by seeding 
known numbers of cells at high density in a medium containing a 
selective agent to detect the number of mutant colonies, and at 
a lower density in a medium without selection to determine the 
cloning efficiency.  After a suitable incubation time, colonies 
are counted.  The mutant frequency per viable cell is derived by 
adjusting the number of mutant colonies in the selective medium by 
the estimate of viable, colony-forming cells obtained from the 
number of colonies in the non-selective medium. 


Cloning efficiency (CE)     Mean colonies per plate in non-
                                selective medium
                            Total cells seeded per plate in
                               non-selective medium

Mutant frequency (ME)       Mean mutant colonies per plate
                               in selective medium
                            Total cells seeded per plate
                               in selective medium

Mutant frequency per survivor = MF

    Each experiment contains "control" (untreated) cultures so that 
the background (spontaneous) mutant frequency can be determined.  Cell types and selective systems

    Many different cell types, including cells of human, rat, 
mouse, and hamster origin, and a wide variety of selective systems 
are available for potential gene mutation assay (Abbondandolo, 
1977; Holstein et al., 1979).  Many of these fulfil the criteria 
suggested for the use of mammalian cells in such assays, for 
example, a sound genetic basis for the system, high cloning 
efficiency, low spontaneous mutation frequency, and a demonstrated 
sensitivity to a variety of chemical mutagens.  However, in 
practice, only three cell lines, the V79 and CHO Chinese hamster-

derived cell lines, and L4178Y mouse lymphoma cells have been 
widely used for large-scale mammalian cell  in vitro assays.  In 
all three cases, HPRT and the Na+/K+ ATPase genes have been used as 
the genetic systems for the basis of mutant selection.  In 
addition, an L5178Y cell line, heterozygous at the TK locus (L5178Y 
TK+/-), has been developed by Clive and co-workers (Clive et al., 
1972), and has been extensively used for the selection of TK 
mutants in mutation assays.  More recently, a CHO cell line, 
heterozygous at the TK locus, has also been developed (Adair et 
al., 1980).  Several reviews are available which discuss the 
general principles of mammalian cell assays (Hsie et al., 1979; 
Fox, 1981) and the V79 (Bradley et al., 1981), CHO (Hsie et al., 
1981), and L5178Y (Clive et al., 1983) cell lines, in particular.  
Many of the critical factors involved are discussed in these papers 
and the extensive available literature is cited.  Considerable 
variations exist in the protocols for mutation assays using 
different mammalian cell lines and selective systems.  These should 
be carefully studied, preferably in close consultation with 
experienced investigators in this field. 

    Some of the problems associated with this assay have been 
highlighted and discussed in the recently conducted collaborative 
study on short-term  in vitro tests, by Ashby et al. (1985), and 
some recommendations for future improvements have been made.  Culture conditions

    Culture conditions should be well-defined, and the cells should 
be maintained under optimal growth conditions throughout the 
experiment.  Cultured cells require the presence of serum to 
maintain growth, and the serum batch may affect growth rate, 
cloning efficiency, and mutant frequency.  Batches of serum should 
therefore be carefully pretested and a large volume of a suitable 
batch stored frozen.  Medium, pH, temperature, humidity, and cell 
dispersion techniques are among the critical factors in mammalian 
cell culture techniques and should be carefully controlled if 
reproducible data are to be obtained.  Treatment

    To ensure that all stages of the cell cycle are exposed to the 
test substance, exponentially growing cells in tissue culture 
medium should normally be used.  Great care should be taken to 
standardize the treatment conditions.  Medium, pH, serum content, 
incubation conditions, and cell density during treatment should all 
be carefully controlled.  The cultures should be protected from 
light during treatment, and suspension cultures should be shaken.  
The test substance should be dissolved just before use, preferably 
in tissue culture medium.  Other vehicles may be used, for example, 
dimethyl sulfoxide, but each should be tested to be certain that 
its presence has no effect on cell viability or growth rate.  For 
initial toxicity studies, a wide range of molarity of the test 
substance should be used.  When the toxic response has been 
determined, the mutation experiment should cover several 
concentrations (usually a minimum of four) ranging from non-toxic 

(90 - 100% survival of the treated cells) to toxic (1 - 10% 
survival).  Greater levels of kill are not recommended (Bradley et 
al., 1981).  Treatment time is generally for 1 - 5 h, although 16 h 
(Bradley et al., 1981) or longer may be appropriate (Cole et al., 
1982).  Expression time

    After the cells have been exposed to a mutagen, they must be 
cultured in a non-selective medium for a period of time so that (a) 
the mutagen-induced damage can be "fixed" in the DNA, and (b) the 
constitutive level of the non-mutant enzyme, and its mRNA, can 
decrease to a negligible level.  The time required for a new 
mutation to be phenotypically expressed as a mutant enzyme (the 
"phenotypic expression period") will depend on the initial number 
of non-mutant enzyme (and mRNA) molecules, their half life under 
physiological conditions, and the rate of cell division.  The 
expression time varies with the cell line, the selective system, 
and possibly also the mutagen treatment.  After the maximum induced 
mutant frequency has been observed, there may be a plateau in the 
frequency of mutants, while in other cases, there may be a peak in 
the number of mutants followed by a fall in the frequency.  For 
example, it has been found that induced ouabain-resistant mutants 
are first observed within 24 h of mutagen treatment, reach a 
maximum 48 - 72 h after treatment, and later remain at an 
approximately constant value.  The thioguanine-resistant phenotype, 
however, requires a minimum of 6 - 7 days after treatment before 
new mutations are fully expressed, after which there is again a 
plateau in the mutant frequency.  In contrast to these 
observations, mutations at the TK locus reach a peak value 48 - 
72 h after treatment, and then there is a marked decline in 
frequency with time.  For quantitative mutation studies, it is very 
important that near-optimal phenotypic expression of induced 
mutation should be observed, and the shape of the expression time 
curve for newly-induced mutants must be carefully determined by 
experiment by each laboratory, under well-defined conditions, using 
a number of different mutagens.  Choice and concentration of selective agent

    The concentration of the selective agent is one of the most 
critical factors (Thompson & Baker, 1973).  The dose should be 
high enough for complete kill of non-mutant cells and the mutant 
frequency at the chosen concentration should not be affected by 
small variations that may occur in day-to-day culture conditions.  
6-Thioguanine is generally considered to be a more stringent 
selective agent for the selection of HPRT mutants than 
8-azaguanine, and is the agent of choice for mouse cells, as 
azaguanine is non-toxic to these cells.  Trifluorothymidine, which 
is recommended for the selection of TK mutants, is both heat and 
light unstable and must be handled with great care.  Stability of the spontaneous mutant frequency

    A high and variable spontaneous mutant frequency can cause 
considerable problems with data interpretation.  Several methods 
are available for maintaining a low, stable frequency: 

    (a) The cell line can be re-cloned to establish a suitable 
        sub-line.  A large frozen stock can then be stored in
        liquid nitrogen and one vial used for each experiment.

    (b) Cells regularly sub-cultured to maintain stocks should be 
        diluted to low density to remove pre-existing mutants.

    (c) For the TK and HPRT systems, pre-existing mutants lacking 
        these enzymes can be removed from the population by
        growing the cells in medium containing aminopterin.  This
        anti-metabolite blocks the  de novo purine and pyrimidine
        synthesis pathways.  If thymidine and hypoxanthine are also
        added to the medium (called "HAT" medium), non-mutant cells
        containing TK and HPRT continue to grow using the "salvage"
        pathway.   Mutant TK- or HPRT- cells die in HAT medium,
        because they are unable to use either the  de novo or the
        "salvage" pathways for nucleotide synthesis.  After "HAT"
        treatment, again, a large stock of cells can be stored in
        liquid nitrogen for future use.  Provision for metabolic conversion

    Three methods of supplying exogenous mammalian activation 
systems are available (Bartsch et al., 1982). 

    (a)  Rodent liver preparations ("S9") (see also section above)

    These can be prepared from untreated animals (usually rats) 
or from animals pre-treated with "inducing agent" (e.g., 
phenobarbital, 3-methylcholanthrene, or Aroclor(R) 1254) to induce 
high levels of the mixed-function oxidases that catalyse the 
metabolic activation steps.  Such preparations have been widely 
used with mammalian cells (Kuroki et al., 1977, 1979; Bartsch et 
al., 1979; Clive et al., 1979; Amacher & Turner, 1981, 1982a,b).  
These papers contain detailed methods for the preparation of both 
S9 and the NADPH energy generating systems "co-factor mix" for use 
with mammalian cell cultures. 

    A batch of S9 should be prepared, tested for sterility, and 
stored for up to 3 months at -70 °C, or in liquid nitrogen.  Both 
S9 and the co-factor mix should only be thawed immediately before 

    (b)  Cell-mediated metabolism

    In this case, the indicator cells (e.g., L5178Y or V79) are co-
cultivated with metabolically-competent cells, e.g., freshly-
isolated rat hepatocytes (Amacher & Paillet, 1983) or hamster cell 
lines such as BHK or SHE (Langenbach et al., 1981; Bartsch et al., 

1982).  The pro-mutagen or -carcinogen is metabolized to the active 
product by the competent cells and diffuses into the indicator 
cells, where it reacts with the DNA. 

    (c)  The host-mediated assay

    Finally, the cultured cells may be placed inside the body of an 
animal (usually a mouse) which is treated with the test substance.  
After a suitable period, the cells are withdrawn, and the mutant 
frequency determined.  For example, L5178Y cells can be grown in 
the peritoneal cavity of compatible mice (Fischer et al., 1974) or 
V79 cells in diffusion chambers in mice (Sirianni et al., 1979).  Controls and internal monitoring

    For each experiment, positive and negative controls are 
required.  A negative control is necessary to check the background 
mutant frequency.  It should consist of no treatment and/or the 
solvent as used to dissolve the test substance.  Two separate 
positive controls (to check that the assay is performing correctly) 
are necessary, one of which should require metabolic activation.  
It is an advantage if a positive control with a known dose-response 
is used, so that the sensitivity of the assay can be assessed in 
each experiment. 

    Cell cultures should be periodically checked for mycoplasma 
contamination (Russel et al., 1975) and can be periodically 
karyotyped to check chromosome stability.  Population size, replicates, and reproducibility 

    (a)  Population size

    The power and sensitivity of the test should be pre-determined, 
taking the toxic effect of the test substance and the mutant 
frequency in the untreated population into account.  The number of 
cells to be treated, sub-cultured, and exposed to selection should 
be sufficient for a particular increase over the control mean to be 
detected.  The precise numbers depend on the cell line and 
selective system, but as a general guide it has been suggested that 
ten times the inverse of the spontaneous mutant frequency should be 
used.  This means, for example, that if the spontaneous mutant 
frequency is 1 x 10-6, then 107 visible cells should be used for 
each treatment level.  If there is substantial initial toxicity, 
this number should be increased correspondingly.  Similar care 
should be taken over the numbers of cells sub-cultured during the 
expression period, to avoid sampling error.  The number of cells 
exposed to selection should be such that the numbers of mutant 
colonies observed on both control and test plates are sufficient 
for statistical analysis. 

    (b)  Replication

    One protocol recommends that duplicate samples should be 
treated, sub-cultured, and plated in every experiment (Clive et 
al., 1979).  Alternatively, single very large populations can be 
used for each treatment level. 

    (c)  Reproducibility

    The determinations should be quantitative and reproducible.  
The whole experiment should be carried out at least twice using 
freshly prepared test substance, though not necessarily over 
precisely the same dose range.  If both experiments give a positive 
or negative result, this could be considered acceptable.  However, 
for low or equivocal responses, further experimentation may be 

2.5.4.  Data processing and presentation  Treatment of results

    The test report should include precise details of all methods 
used in the test procedure.  All validation data should be provided 
and retained for further reference. 

    Data should be presented in tabular form.  All original data, 
including toxicity data, absolute cloning efficiency of the control 
cultures, and individual colony counts for the treated and control 
groups should be presented for both mutation induction and survival 
plates.  Survival and cloning efficiencies should be presented as a 
percentage of the controls.  Mutant frequency should be presented 
as per 106 clonable cells.  Possible toxicity of the vehicle should 
be indicated.  Evaluation of results

    Several criteria have been suggested for determining a positive 
result, one of which is a statistically significant, concentration-
related increase in the mutant frequency.  An alternative is based 
on the detection of a reproducible and statistically significant 
positive response for at least one concentration of the test 
substance.  The problem with such an approach is that, although 
several methods of statistical analysis have been published (Clive 
et al., 1979; Amacher & Turner, 1981; Snee & Irr, 1981; Tan & Hsie, 
1981), there is, at present, no general concensus as to the most 
appropriate method.  Further work is required on the optimum 
experimental design and statistical analysis of mammalian cell 

    A substance that does not produce either a reproducible 
concentration-related increase in mutant frequency or a 
reproducible significant positive result at any one test point is 
considered non-mutagenic in this test.  Ambiguous results

    Ideally, experimental design should be such that ambiguous 
results do not occur.  Examples of ambiguous results might be very 
early expression (day 0 or 1) of induced TK or HPRT mutants, marked 
variations in colony numbers at different expression times or an 
inverse concentration-related effect.  Repeat experiments, paying 
particular attention to growth conditions, stringent mutant 
selection, and all the critical culture conditions may be necessary 
to resolve ambiguous results. 

2.5.5.  Discussion

    Mammalian cell gene mutation assays have a sound genetic and 
biochemical basis.  Defined protocols have been developed for the 
three most commonly-used cell lines and reproducible results have 
been produced using a number of chemical mutagens.  A limited 
amount of testing has been done using carcinogens (mainly using the 
L5178Y TK system) and the role in predicting carcinogenicity 
requires further study.  Some systems are capable of determining 
multiple genetic end-points (Cole et al., 1982; Gupta & Singh, 
1982), and these are potentially advantageous as mutagen-screening 

    One of the most important factors influencing the validity of 
the data is that the investigator should have a thorough 
understanding of the particular cell system in use.  This includes 
the culture conditions that will support good cell growth and an 
awareness of the many possible causes of sub-optimal growth.  Slow 
growth rate may result in reduced incorporation of analogues and 
incomplete kill of wild-type cells.  Other factors that deserve 
particular emphasis are described below.  Mutant selection

    It is very important to ensure stringency of selection for each 
particular cell line.  Pool sizes differ between cell lines and 
under different growth conditions, and the relative affinity of 
salvage enzyme for analogue and natural substrate may differ 
markedly.  The kill curve of the selective agents used must be 
carefully checked, and the concentration chosen should not be 
within or close to the range in which exponential fall in survival 
occurs.  This is especially important if a high and variable 
spontaneous mutant frequency is found, as this makes data 
interpretation particularly difficult.  Expression time

    It is essential that a near-optimal expression time for the 
induction of mutants should be used if accurate data are to be 
obtained for the analysis of concentration-related effects.  The 
expression time should be carefully defined for each selective 
system using a number of mutagens.  The use of a single "standard" 
expression time may give misleading results and, ideally, at least 
two expression times should always be used so that it is clear that 

the peak has been observed.  This is particularly important if an 
unusual dose-response relationship is obtained, for example few 
mutants being induced with increased dose.  Cell numbers

    Experiments should be designed to maximize the possibility of 
statistical analysis of the data.  If small effects are to be 
detected, it is most important that the spontaneous mutant 
frequency should be borne in mind, and that sufficient cells should 
be exposed to treatment and cloned in selective medium to provide 
reasonable numbers of mutants as a basis for analysis.  Metabolic conversion

    This is a major area in mammalian cell assays requiring further 
research.  At present, no single experimental design is ideal for 
detecting all compounds that require metabolic conversion.  The 
factors requiring consideration are species and inducer used for 
tissue homogenate (S9) preparation, the correct final concentration 
of the homogenate, and the use of intact cells rather than 
homogenate.  These factors may make a considerable difference to 
the apparent mutagenicity of the test compound, and the laboratory 
conducting the test should be able to provide evidence that, using 
a clearly defined protocol for metabolic conversion, mutagens from 
different classes of chemicals requiring metabolic activation 
(e.g., benzo( a)pyrene,  N-nitrosodimethylamine, and  N-acetyl-2-
aminofluorene) induced mutations in a dose-dependent fashion.  
Flexibility is important as a single compromise protocol may not be 
appropriate in every case. 

2.5.6.  Conclusions

    Mammalian cell lines have been used in the study of chemically 
and physically induced specific locus mutations since 1968.  
Clearly defined methods for mutagenesis assays using cultured 
mammalian cells have been developed and a detailed examination of 
the genetic basis of the markers used has been made.  Although a 
number of areas requiring further study remain (Ashby et al., 
1985), criteria have been established for an evaluation to be made 
of the induction of specific locus mutations in mammalian cells, 
and of the role of such assays in predicting carcinogens. 

2.6.  The Use of Higher Plants to Detect Mutagenic Chemicals

2.6.1.  Introduction

    Many of the fundamental concepts of modern genetics were 
established in higher plants and the term "mutation" was introduced 
by the Dutch botanist, Hugo de Vries, in 1909, to describe a sudden 
hereditary change in  Oenothera lamarckiana.  Plant systems played a 
major part in early investigations of the genetic changes caused by 
radiation (Read, 1959; Revell, 1959) and a variety of plants have 
been used to study the mutagenic effects of chemicals at the gene 
and chromosome levels.  With the increasing concern over the 

genotoxicity of sophisticated techniques for studying mutations in 
bacteria, lower plants, insects, and mammalian cells, there has 
been a loss of interest in the testing of potentially mutagenic 
chemicals in higher plant systems.  This is surprising as plants 
appear to offer significant advantages over other organisms in 
certain circumstances, though they have, of course, important 

    Techniques for studying mutagenic chemicals have been developed 
in about 10 species of higher plants and a whole range of specific 
genetic end-points are available.  Mitotic chromosome alterations 
can be studied in the somatic cells from root tips, or pollen tubes 
in, for example, barley, the broad bean, or the onion.  Pollen 
mother cells from a number of species are suitable for detecting 
chemically-induced chromosomal aberrations in meiotic cells.  Gene 
mutations at specific loci can be investigated in maize or soybean 
plants and multilocus mutation systems are available in barley and 
maize.  The chromosome systems allow the observation of structural 
chromosome damage and effects on chromosome segregation and general 
mitotic function.  The chromosomes are morphologically similar, and 
appear to respond to treatment with mutagens in a similar way to 
those of mammals and other eukaryotes. 

    A survey of the literature prepared under the US Environmental 
Protection Agency Gene-Tox Program (Constantin & Owens, 1982) 
revealed that about 350 compounds, covering a wide range of 
chemical classes, had been tested for mutagenic activity in plants.  
The same authors also compared the results of testing eight model 
mutagens in plants with the results obtained in other systems.  
They claimed that the correlation between plant data and results 
from cultured mammalian cells was at least as good as that with 
data derived from bacteria and  Drosophila.  A comparison of the 
results of testing a series of pesticides in plant root tips and 
mammalian cells for chromosomal aberrations showed a remarkable 
qualitative similarity between the two sets of results.  However, 
the data on chromosome damage in mammalian cells for some of the 
pesticides was not truly representative of the literature on these 
chemicals.  Although a database representing more than 350 
compounds tested in plant systems has been assembled, a large 
proportion of the chemicals tested were shown to be mutagenic in 
one plant system or another, and there is a significant lack of 
information on non-mutagenic chemicals. 

    In spite of the above comments, it is apparent that plant 
assays possess some advantages over other systems that remain to be 
fully exploited in the area of genetic toxicology.  Chromosome 
assays on plants are rapid and inexpensive and do not require 
elaborate laboratory facilities, and a wide range of genetic end-
points is available.  However, before the full potential of plant 
systems can be exploited, some serious limitations have to be 
overcome.  There is a lack of knowledge concerning many of the 
critical molecular processes in plants, particularly those 
influencing the metabolism of foreign compounds; thus, it is 
difficult to assess the significance for mammals, including man, of 
data derived from plant experiments.  There are also fundamental 

differences in structure between plant and mammalian cells.  The 
rigid cellulose wall of plant cells almost certainly affects the 
penetration of certain chemicals and there may be selective 
differences between plant and mammalian cells in the kinds of 
molecules that can be absorbed.  However, the DNA of plants and 
animals appears to be similar in structure and function and the 
mechanism of protein synthesis seems to be the same.  Higher plants 
have more cytoplasmic (mitochondrial) DNA than animal cells and, in 
addition, the chloroplasts contain DNA. 

    Mitotic chromosome division in plants follows a similar course 
to that in mammalian cells, though meiosis and gametogenesis are 
very different.  In plants, cell division is accompanied by the 
formation of a plate that separates the daughter cells while in 
mammals, the cells divide by constriction of the cytoplasm. 

2.6.2.  Test systems

    Although about 25 different test-systems have been described in 
10 plant species, the following have been established as practical 
and useful for testing chemicals for mutagenic activity: 

    (a)  mitotic chromosomal damage;

    (b)  aberrations in meiotic chromosomes; and

    (c)  gene mutations at specific or miltiple loci.  Detection of mitotic chromosome damage

    Growing root tips of the broad bean,  Vicia faba (Ma, 1982b), 
the onion,  Allium cepa (Grant, 1982), the spiderwort,  Tradescantia 
 paludosa (Ma, 1982a), and of barley,  Hordeum vulgare (Constantin & 
Nilan, 1982) provide a readily available source of material for 
studying the damaging effects of chemicals on chromosomes. 

    (a)   Vicia faba

    The six pairs of chromosomes can be clearly observed at the 
metaphase stage of mitosis, and it is possible to identify all 
types of chromatid and chromosome aberrations.  In addition to 
conventional metaphase analysis, methods are also available for 
detecting chromosome damage by counting micronuclei and for 
recording sister chromatid exchanges.  The technique is most 
suitable for studying water-soluble chemicals, but by using organic 
solvents, e.g., dimethyl sulfoxide, other compounds can also be 
tested.  Normally, stock solutions of the test compound are added 
to the growth solution; appropriate buffers should be used to 
correct extremes of pH.  Freshly prepared solution should always be 
used.  The technique (Kihlman, 1971) is relatively simple, and 
requires only a minimum of laboratory equipment.  Seeds are 
softened by soaking in water for 6 - 12 h, then allowed to 
germinate in moist vermiculite or similar medium at a temperature 
of about 19 °C.  After germination (4 days), the growing shoot is 
removed and the seedlings transferred to a tank of water, which 

should be fully aerated.  After 24 h in the tank, primary root 
growth is sufficiently active for study.  It is important to 
control the pH and temperature of the water as both may affect the 
frequency of chromosome aberrations induced by a given chemical. 

    Treatment times may vary between 1 and 24 h, though short 
treatment times are preferable for the identification of the most 
sensitive mitotic stages.  However, it is conventional to 
incorporate two or three different treatment times, when testing 
chemicals of unknown mutagenic activity.  The mitotic cycle of 
 Vicia is between 18 and 22 h and, as the interphase stage is the 
most sensitive to the majority of chemicals, it is necessary to 
allow a recovery period of about 8 h in the absence of the test 
chemical.  This ensures that roots are fixed and processed at a 
stage where chromosomes damaged by the chemical will be in the 
first metaphase after treatment.  In some cases, an additional 
recovery period of 30 - 40 h may be used, so that chromosomes can 
be examined at the second metaphase.  Both treatment and recovery 
should take place in the dark.  Before the roots are fixed, they 
are transferred to a solution of 0.02 - 0.05% colchicine and 
agitated in this solution for 2 - 4 h.  This treatment blocks the 
cell cycle at the metaphase stage and leads to an accumulation of 
metaphase chromosomes that are suitable for analysis. 

    For most purposes, fixation in ethanol: acetic acid (3:1) 
gives satisfactory results.  This is best carried out at 4 °C 
(refrigerator temperature) for a minimum of 20 min; fixation from 
2 - 24 h is more effective for permanent preparations.  For 
preliminary analysis or when permanent preparations are not 
required, chromosomes can be stained using the aceto-orcein method.  
The Feulgen squash technique of Darlington & Lacour (1969) is 
preferable for permanent slides followed by rapid freezing, 
dehydration in alcohol, and mounting. 

    The various kinds of chromosomal aberrations can be scored in 
metaphase preparations and, for many chemicals, the scoring of 
chromatid-type aberrations in the first metaphase after treatment 
gives the most reliable measure of mutagenic activity.  However, 
some compounds, e.g., maleic hydrazide, produce a peak of activity 
during the second metaphase and this should be determined before a 
chemical is regarded as inactive.  Examination of anaphase 
chromosomes for fragments and bridges is a useful technique for 
rapid screening and for obtaining preliminary information on 
clastogenic (i.e., chromosome-breaking) activity, mitotic delay, 
and the absence of cell division.  Such information is useful for 
deciding treatment concentrations and times and recovery periods 
for subsequent metaphase studies.  For a detailed description of 
metaphase and anaphase aberrations, see Kihlman (1971). 

    Micronuclei resulting from chromosome fragments or lagging 
chromosomes can be scored at the interphase following treatment 
(Ma, 1982a), and a technique has been described for investigating 
sister-chromatid exchanges (SCE) in root tips (Kihlman & Andersson, 

    A sufficient number of root tips should be used for each of a 
wide range of concentrations of the test compound to give an 
adequate number of data for subsequent interpretation and, if 
necessary, statistical evaluation.  A minimum of 100 metaphase 
cells should be analysed from at least 10 roots for each 
experimental group.  Doses should be selected within half-log 
intervals and compounds should be tested up to obviously cytotoxic 
or inhibitory (i.e., reduction in mitotic index) concentrations.  
It is usual to conduct preliminary experiments to identify a 
suitable range of concentrations.  More than one exposure period 
and two or three recovery periods may be necessary to obtain the 
maximum incidence of chromosome damage and a major objective is to 
determine a dose-response relationship for chemicals that appear to 
be mutagenic.  Control experiments are needed for each assay and 
should include a negative control, consisting of roots cultured in 
the growth solution including any solvent used, and a positive 
control, consisting of roots treated with a known mutagen such as 
ethylmethane sulfonate. 

    Results are usually expressed as the number of aberrations per 
100 cells, per group and the number in each experimental group is 
compared with the values from the negative control group.  In most 
cases, positive results are so obvious that statistical analysis is 
unnecessary.  Where the number of aberrations is low, a simple 
t-test or Chi-squared test, using a significant level of 1% to 
determine positive results, is usually adequate. 

    (b)   Allium cepa

    Although a number of species of  Allium have been used for 
genetic studies, the common onion,  Allium cepa, has proved to be 
the species of choice for root-tip chromosome studies (Grant, 
1982).  Mitotic cells of  Allium contain 8 pairs of large 
chromosomes.  The technique for root-tip chromosome preparations is 
very similar to that described for  Vicia.  The outer scales are 
removed from young bulbs to expose the root primordia and they are 
then supported in a rack over a suitable tank containing water at 
20 °C.  Adequate root growth should be obtained in 2 - 4 days.  The 
roots are then ready for treatment with the test chemicals followed 
by processing and mounting as described above.  In an even simpler 
technique,  Allium cepa seeds are germinated on layers of paper 
towelling soaked with the test solution in a culture dish.  Primary 
roots are usually 0.5 - 1.0 cm long after 3 days, and they can then 
be processed for analysis. 

    (c)   Tradescantia paludosa

    Compared to  Allium and  Vicia, only a few chemicals have been 
tested for mitotic chromosomal aberrations in  Tradescantia, but it 
has the advantage that both meiotic and mitotic chromosomal damage 
and gene mutations can be tested in the same species.  Dividing 
cells in the root tip of  Tradescantia contain 12 large metacentric 
chromosomes.  A large number of roots can be obtained from cuttings 
from mature plants in about a week.  These rooted cuttings can then 
be used for chromosome studies in much the same way as those of 
 Allium or  Vicia (Ahmed & Grant, 1972). 

    (d)   Hordeum vulgare

    Both root-tip and shoot-tip cells can be used to investigate 
mitotic chromosome changes in barley.  The chromosomes are large, 
12 in number, and very suitable for the rapid scoring of 
aberrations.  The procedure is similar to that described above.  
Barley seeds are allowed to germinate while in contact with the 
test solution.  Five to seven primary roots develop from each seed 
and the roots are usually fixed between 24 and 48 h after 
germination and then processed for metaphase chromosome analysis.  
Squash preparations can be made from a number of growing points on 
the developing shoot and numerous cells are usually available for 
metaphase analysis.  Frequency of chromosome damage may vary 
between root tips and shoot preparations because of differences in 
the effectiveness of transport of different chemical molecules 
(Constantin & Nilan, 1982). 

    In general, the root tip procedures are relatively simple and 
sensitive assays for clastogenic chemicals.  The species described 
have small numbers of large chromosomes, which simplifies analysis, 
and aberrations can be scored at either metaphase or anaphase.  
They are more suitable for testing water-soluble compounds than 
those that are not easily soluble.  It should be emphasized that 
the metabolic pathways required for the activation of many 
chemicals have not been fully characterized in these plant systems.  
Thus, the relevance of these results for mammalian cells cannot be 
properly assessed, at present.  Detection of aberrations in meiotic chromosomes

    Although the processes of sexual reproduction in plants are 
greatly different from those in mammals, there are some 
similarities in meiotic cell division and chromosome behaviour.  
The induction of anomalies in the chromosomes of, for example, 
pollen mother cells, may be analagous to meiotic chromosome damage 
in mammalian reproductive cells, though convincing evidence for 
this is lacking.  A number of plants including  Vicia and  Hordeum  
offer relatively easy means of studying meiotic events including 
numerical (e.g., non-disjunction) as well as structural chromosome 
changes.  A method is also described for counting micronuclei in 
4-cell stages as a measure of chromosome breakage (Ma, 1982a).  The 
techniques are simple, involving fixing the flower buds in ethyl 
alcohol/acetic acid, staining the anthers using a squash technique, 
and then analysing the chromosomes in the pollen mother cells. 

    (a)   Vicia faba

    For the examination of meiotic chromosomes in  Vicia, it is 
necessary to raise the plants to maturity in either growth chambers 
or glasshouses.  This is fairly time-consuming and requires much 
more space than the root-tip assay.  Chemicals can be applied 
either by spraying in solution on the young flower buds or by 
exposing the buds to the chemical in the form of a gas or vapour in 
an appropriate chamber (Tomkins & Grant, 1976).  Suitable 
concentrations of the chemical and exposure times are determined 

from preliminary experiments and it is usual to allow a recovery 
period before processing the pollen tubes for the analysis of 
anaphase cells for bridges and fragments. 

    (b)   Tradescantia paludosa

    Strains of  T. paludosa that proliferate and propagate easily 
and quickly under local environmental conditions should be used.  A 
suitable clone should grow to maturity from cuttings in 40 - 60 
days.  Since the chromosomes of pollen mother cells are not of 
adequate quality for the detailed analysis of metaphase 
aberrations, a technique has been developed for detecting 
chromosome breakage on the basis of micronuclei at the tetrad 
stage.  In practice, the inflorescences are removed from the plant 
and the stems placed in solutions of the test chemical.  
Alternatively, the buds can be exposed to gaseous materials in a 
suitable chamber.  The optimum length of treatment is determined 
experimentally and a recovery period of 24 - 30 h is necessary to 
allow chromosome damage in early prophase 1 to reach the tetrad 
stage where micronuclei can be scored.  Micronuclei are assumed to 
be a result of either chromosome fragmentation or of whole 
chromosomes lost during meiosis and are therefore a measure of both 
structural damage and aneuploidy (or non-disjunction).  It is usual 
to score between 1000 and 1500 tetrads from each experimental group 
including both negative and positive controls. 

    (c)   Hordeum vulgare

    Chromatid and chromosomal aberrations can be investigated in 
pollen mother cells (microsporocytes), which are present in large 
numbers in the developing barley spike.  The spike is produced when 
the shoot apex undergoes a transition from a new leaf promordium to 
an inflorescence primordium.  The spike is collected for 
cytogenetic analysis at approximately the same time as the last 
leaf (i.e., the flag leaf) emerges.  As meiosis in the pollen 
mother cells is not synchronized, spikes can be used for testing 
over a period of up to 40 h during development.  Chemicals can be 
applied by spraying the spike or adjoining areas at selected times, 
before removing the spikes.  The entire spike is fixed in 
ethanol/acetic acid and processed in the normal way (Constantin & 
Nilan, 1982).  Detection of gene mutations at specific or multiple loci

    A specific locus is a region of a chromosome that controls the 
development of a phenotypic characteristic.  It is equivalent to 
the classical Mendelian gene and can mutate to a new allele with an 
associated change in phenotype.  Although there are a number of 
specific loci that are potentially useful for studying chemical 
mutagens, only a few systems are sufficiently well characterized to 
be used in practice.  An example of these is the waxy mutation as 
expressed in pollen grains of maize. 

    (a)   Waxy locus mutations in Zea mays

    Maize has a long history of use in genetic studies and hundreds 
of genotypically defined strains are available.  The pistillate 
flowers containing the female spores (megaspores) develop on a 
separate part of the plant to the characteristic tassels containing 
the male spores (microspores).  Tetrads of haploid microspores 
develop in the anthers through a process of meiosis and then, by 
mitotic division, the male gametophyte or pollen grain is formed.  
The haploid, female megaspore develops from the megasporocyte by 
meiotic division and, after a complex process of mitosis, the 
female gametophyte is produced. 

    The waxy locus assay is based on dominance or recessiveness in 
a gene that determines the presence of amylose in the kernel.  In 
the recessive genotype  (wx), the kernels have a waxy appearance 
and the starch of the endosperm contains only amylopectin.  The 
starch in the dominant  (Wx) form consists of a mixture of 
amylopectin and amylose. Kernels carrying the  Wx allele stain a 
dark blue-black when stained with iodine while  wx/wx kernels, 
which have no amylose, stain a red colour.  The waxy phenotype can 
also be detected in pollen grains using the iodine reaction and 
this forms the basis of the assay. 

    The assay can be conducted by the direct treatment of the 
tassels, which are harvested at an appropriate time and stored in 
70% ethanol.  Homozygous  Wx plants are exposed to the test 
chemical and forward mutations are detected by a lack of amylose in 
the iodine-treated pollen.  A reverse mutation assay using plants 
of the  Wx/wx genotype can be used in a similar technique. 

    It is usual to analyse some 250 000 pollen grains per tassel in 
5 - 10 plants.  The frequency of mutants in pollen from treated 
plants is compared with that from the untreated controls. 

    Further details of this and other mutation assays in
maize, and information on the application and interpretation
of these procedures are given in the review by Plewa (1982).

    (b)   Chlorophyll-deficient mutations in Hordium vulgare

    Chlorophyll synthesis and its control is governed by a large 
number of genes and a variety of recessive mutations can be 
detected after treatment of barley seed with mutagens or by 
exposure of the plant during its complete life cycle.  The 
procedure for detecting chlorophyll-deficient mutations is 
relatively time-consuming as they are observed in the second (M2) 
generation after treatment of the seed.  The system is reviewed by 
Constantin (1976).  The waxy pollen test can also be applied in 
barley (Sulovska et al., 1969). 

    (c)   Somatic mosaicism in Glysine max

    The induction of spots of contrasting colour in the leaves of 
soybean seedlings appears to have many attributes as a useful 
short-term test for mutagenic chemicals.  The spots result from a 
variety of genetic changes in either meiotic or mitotic cells, the 
assay can be completed in 4 - 5 weeks, and its requires a minimum 
of laboratory facilities.  The test is based on the Y11 locus and 
its mutation to y11.  The homozygous Y11 Y11 has dark green leaves 
that may show light green or very dark green spots, the 
heterozygous Y11 y11 has light green leaves showing dark green, 
yellow, or twin (dark green/yellow) spots.  Although cytological 
evidence of the genetic basis of the mosaicism is limited, it has 
been inferred from the phenotypic expression that the spots may be 
a result of somatic crossing over, non-disjunction, chromosome 
deletion, gene mutation, or somatic gene conversion. 

    In studies on the induction of leaf mosaics, seeds are treated 
with the test chemical during germination.  They are then planted 
in a non-nutritive medium and grown under controlled conditions in 
a glasshouse until the second compound leaf unfurls (4 - 5 weeks).  
The number and type of spots per leaf on each plant is recorded and 
the numbers of spots on treated plants compared with the untreated 
control values.  An appropriate positive control group (i.e., 
mitomycin C,  N-methyl- N-nitrosourea) is included in each assay.  
For a detailed review of the assay see Vig (1982). 

    (d)   Somatic gene mutations in Tradescantia

    The  Tradescantia assay, which involves a change in flower 
colour from blue to pink, is particularly suitable detecting 
mutagens in the atmosphere (Schairer et al., 1978).  The hybrid 
clone 4430 is heterozygous for a specific flower colour locus.  The 
dominant blue allele produces the phenotypically blue colour in the 
petals.  The recessive pink phenotype is only expressed by mutation 
or deletion at the blue allele.  The pink colour is detected as 
pink cells in the stamen or as sectors in the petals.  For 
laboratory studies, cuttings bearing a young inflorescence are 
treated with liquid or gaseous compounds for periods of a few hours 
to a number of days.  The cuttings are then transferred to growth 
chambers under standard conditions, until the necessary 
observations have been carried out.  Mutations are expressed as 
single pink cells or as strings of pink cells in the stamen hairs.  
Some 40 - 75 hairs can be obtained from each bud.  Details of the 
technique and its application for detecting gaseous mutagens in the 
environment are given by Van't Hof & Schairer (1982). 

2.6.3.  Discussion

    There are about 10 test systems in plants that can be used to 
investigate the mutagenic effects of chemicals and they cover a 
full spectrum of genetic end-points.  They range from the rapid and 
simple root-tip assays for structural chromosome damage to 
relatively complex tests for specific locus mutations.  Plant 
assays have been used extensively to test chemicals in solution and 

some of the systems are uniquely fitted for detecting low 
concentrations of atmospheric mutagens.  A test using homosporus 
ferns (Klewoski, 1978) is being developed for detecting water-borne 
mutagens in natural waters and effluents. 

    Although the literature on plant mutagenesis is extensive, 
there are few data comparing the results observed in plants with 
those in mammals, and extrapolation between the two remains 
somewhat tenuous.  Some mammalian carcinogens that are known to 
require metabolic conversion to reactive molecules are detected as 
mutagens in plant systems (e.g., some nitrosamines).  In the 
limited comparisons available, there is a positive correlation 
between mutagenicity in plants and mammalian cells.  However, there 
appear to be two serious limitations in the interpretations of the 
results of plant assays in terms of human hazard.  First, though 
there are data on up to a hundred chemicals in some systems, the 
majority of the chemicals tested have been mutagens (in some assays 
as many as 95% of chemicals tested).  Thus, many more data on the 
response of plants to chemicals shown to be non-mutagenic in other 
systems are required.  The second limitation is related to the 
fundamental differences in the metabolism of foreign compounds 
between plants and mammals, and information is lacking on the 
mutagenicity and metabolic mechanisms in plants for many of the 
major classes of mammalian carcinogens. 

    In spite of these reservations, it must be recognised that 
plant systems have many attributes in terms of cost and technical 
simplicity that recommend their use in specific circumstances for 
the initial screening of chemicals for mutagenic activity. 

2.7.  The  Drosophila Sex-Linked Recessive Lethal Assay (SLRL)

2.7.1.  Introduction

    The fruitfly  Drosophila melanogaster is a test organism in 
which it is possible to analyse  in vivo heritable mutations and 
chromosomal aberrations in the same population of treated germ 
cells.  Special strains (stocks) are available or can be 
constructed to study gonadal or somatic tissue for gene mutations, 
deletions, and for almost all possible types of chromosomal 
rearrangements.  In addition, special test protocols have been 
devised to detect aneuploidy resulting from nondisjunctional 
events.  Comparative investigations on the reliability of these 
different genetic end-points have clearly revealed that the 
X-linked recessive lethal test is by far the most sensitive and 
reliable assay in  Drosophila to screen compounds for heritable 
genetic damage.  One of the major reasons is that the phenomenon of 
"recessive lethality" can have different origins:  recessive 
lethals comprise point mutations (intragenic changes), deletions 
affecting more than one gene, and both small and large 
rearrangements (Auerbach, 1962a).  Thus, a mutagen that only 
produced gene mutations would not be detected in a test for 
translocations, but would still be picked up in the recessive 
lethal assay.  In this section, a brief outline of the performance 
and the most essential points of the recessive lethal method will 
be given. 

2.7.2.  Procedure  Test organism life cycle

     Drosophila melanogaster undergoes complete metamorphosis.  The 
egg produces a larva that undergoes two molts, so that the larval 
period consists of three stages (instars).  The third instar larva 
becomes a pupa which, in turn, develops into an imago, or adult.  
Depending on the temperature, this fly requires 9 - 20 days to 
complete one generation.  At 25 °C, the culture temperature 
preferred in most laboratories, the major stages in the life cycle 
are: embryonic development, 1 day; first larval instar, 1 day; 
second larval instar, 1 day; third larval instar, 2 days; prepupa, 
4 h; pupa, 4.5 days.  Thus, at 25 °C, one generation lasts only 
9 - 10 days.  Stock cultures

    Glass milk bottles of about 200 ml volume are used for stock 
cultures.  For smaller cultures, e.g., pair matings in the 
recessive lethal test, vials of about 40 ml are used.  The culture 
media most widely used are banana medium and cornmeal medium, i.e., 
74.3 g water, 1.5 g agar, 13.5 g molasses, 10.0 g cornmeal, and 
0.7 g methyl- p-hydroxybenzoate (to reduce growth of moulds).  List of nomenclature

    The book of Lindsley & Grell (1968) entitled "Genetic 
Variations of  Drosophila melanogaster" represents the exhaustive 
compilation of the mutants of  Drosophila.  This book gives the 
nomenclature used by  Drosophila geneticists, together with a 
detailed description of mutants, chromosomal aberrations, special 
balancer chromosomes, cytological markers, and wild-type stocks.  
This guide is indispensable when working with  Drosophila.  Equipment and laboratory techniques

    There are several detailed descriptions of mutation work on 
 Drosophila, including culture medium, equipment, stock culturing, 
and handling of flies (Abrahamson & Lewis, 1971; Demerec & Kaufman, 
1973; Würgler et al., 1977). 

2.7.3.  Principle of the recessive lethal assay

    Individual chromosomes of  Drosophila melanogaster have been 
labelled X, Y, 2, 3, and 4.  The female chromosomes consist of 
three pairs of autosomes (2, 3, 4) and one pair of rod-shaped X 
chromosomes.  The chromosome complex (2n) of the male has three 
pairs of autosomes, one X and one J-shaped Y chromosome.  The X and 
Y chromosomes, therefore, are called the sex chromosomes. 

    The recessive lethal test can be readily designed to detect the 
induction of heritable genetic lesions in a large part of the 
 Drosophila genome.  Two generations are required for the detection 
of recessive lethals on the X-chromosome, which represents about 

20% of the entire genome.  It is estimated that about 700 - 800 of 
the 1000 loci on the X-chromosome can mutate to give rise to 
recessive lethal mutations. 


    The most relevant features of the X-chromosomal recessive 
lethal test (also referred to as sex-linked recessive lethal assay) 
are illustrated in Fig. 5.  Males from a wild-type laboratory 
strain are treated (or kept untreated as controls) and are then 
mated (P1) with virgin females that are homozygous for the X-linked 
markers  B (Bar, semi-dominant; eye restricted to a narrow vertical 
bar in male and in homozygous female.  Heterozygous female has a 
number of facets intermediate between homozygous female and wild-
types) and " wa" (white-apricot, recessive; eye colour yellowish 
pink), affecting the shape and colour of the eyes (Lindsley & 
Grell, 1968).  This  "Basc" balancer X-chromosome also carries an 
inversion to prevent crossing-over of a lethal from the treated 
(paternal) X-chromosome to its homologue in the heterozygotes (F2-
P2).  Thus, the two "marker genes"  B and " wa" serve to distinguish 
"treated" (paternal) from "untreated" (maternal) chromosomes.  The 

F1-P2 generation is intercrossed.  In the F2, which splits into 
four genotypes that can easily be identified by their different 
phenotypes, it is possible to distinguish the two classes of flies 
carrying copies from a treated chromosome (left side) from those 
that do not (right side).  If a complete recessive lethal mutation 
is induced in an X-bearing germ cell of the treated P1 male, all 
the somatic cells of the resulting F1 female will be heterozygous 
for this mutation, and also 50% of its eggs will carry it.  Half of 
the F2 males will be hemizygous carriers for it and will therefore 
die.  But this can be seen only when single-mating is conducted in 
the F1, which is an absolute prerequisite for the proper 
performance of the assay. 

    Female treatment is not recommended in routine testing 
procedures, because the females may contain pre-existing lethals 
that have to be crossed out before starting an experiment.  The 
major advantages of the recessive lethal test are: 

    (a) The criterion used to decide whether a mutation is 
        present or not is very objective.  The decision is
        based on whether, in the F2-generation, one entire
        class of males is absent or not (Fig. 6).  Therefore,
        personal bias is reduced to a minimum.

    (b) Lethals are much more frequent than other types of
        genetic lesions, i.e., viable visible mutations or
        large structural aberrations.

    (c) A representative part of the  Drosophila genome is
        covered by this multi-locus procedure.


2.7.4.  Metabolic activation

    The organism itself has a complex metabolic system (Vogel et 
al., 1980).  The presence in  Drosophila of cytochrome P-450-
dependent oxygenase, cytochrome b5, aryl hydrocarbon hydroxylase, 
and other components of the xenobiotic-metabolizing enzymes has 
been demonstrated.  There is substantial experimental evidence 
supporting the conclusion that  Drosophila has the enzymic 
potential for converting a wide array of pro-mutagens/pro-
carcinogens (about 80 pro-carcinogens to date) into genetically-
active species. 

2.7.5.  Test performance  Treatment procedures

    Chemicals are most commonly administered to  Drosophila, either 
by injection into the body cavity or by feeding, at the adult or 
larval stage.  Other methods of treating flies include treatment 
through inhalation or using aerosols.  Adult males are recommended 
for testing purposes, since females are more readily sterilized by 
chemicals and have, so far, proved more refractory to the induction 
of heritable genetic changes. 

    Experience with several classes and types of mutagens indicates 
the importance of a flexible protocol, when using the recessive 
lethal assay.  There are several examples in which the route of 
administration has been shown to have a profound effect on the 
mutagenicity detected.  Injection seems to be more reliable for the 
detection of highly reactive mutagens such as the unstable beta-
propionylactone and chloroethylene oxide.  Adult feeding is more 
effective in cases where a single injection (pulse treatment) is 
highly toxic, as has been demonstrated with the carcinogen 
 N-nitrosodiethylamide (DEN).  Several solvents (ethanol, Tween 60, 
Tween 80, special fat emulsions) can be used to dissolve or 
emulsify chemicals of low water solubility.  The use of DMSO and 
DMF should be avoided.  These solvents have recently been shown to 
inhibit chemical mutagenesis in some pro-carcinogens by blocking 
their metabolic activation (Zijlstra et al., 1984).  Toxicity tests

    Pilot studies should give concentration-mortality relationships 
to express the general biological reactivity of the chemical under 
test.  The availability of such toxicity data aids the adequate 
design of the genetic studies and, if the compound under test is a 
mutagen, provides the condition that will produce the maximum yield 
of mutations without killing the animal as a result of lethal 
overdose.  Thus, pilot studies should give an approximate idea of 
the possible toxicity (LD50) of the test compound.  Technical 
aspects of toxicity tests are described by Würgler et al. (1977).  
Actual results of toxicity tests with a series of monofunctional 
alkylating agents were reported by Vogel & Natarajan (1979). 
Regarding the general testing strategy, the highest possible 
concentrations that can be used for testing for recessive lethal 

mutations should be used first.  Acute toxicity, reduced fertility, 
and solubility problems may then be the limiting factors.  Use of 
two different dose levels is recommended, i.e., the MTD (maximum 
tolerated dose) and 1/4 or 1/5 of it.  Brooding

    It is well known that chemical mutagens often exhibit stage 
specificity, i.e., show more or less pronounced mutagenic effects 
at different stages in germ-cell development.  It is, therefore, 
essential to analyse the progeny from treated spermatozoa, late and 
early spermatids, and spermatocytes.  Analysis of offspring from 
treated spermatid stages is of particular importance because there 
is considerable evidence, derived from experiments with 
alkaryltriazenes, nitrosamines, and other pro-mutagens, that 
release of ultimate mutagenic metabolites from the parent pro-
mutagen takes place directly in metabolically-active spermatid 
stages, whereas spermatocytes are highly susceptible to killing.  
On the other hand, there seems no need to include in the test the 
analysis of spermatogonia, because there are only few cases of 
mutagens that affect spermatogonia, but are not active in meiotic 
or postmeiotic cells (Auerbach, 1962b). 

    With the brooding technique, the spatial pattern of 
spermatogenesis is translated into a temporal pattern of successive 
broods.  Treated wild-type males are therefore re-mated at regular 
intervals of 2 - 3 days with fresh virgin Basc females.  An excess 
of 3 - 5 females per male serves to sample all germ cells that are 
in the mature stage.  A total sampling period of 7 - 9 days (3 to 4 
broods) is considered sufficient for mutagen testing.  Control and replicate experiments: sample size

    Würgler et al. (1975) prepared sample-size tables that are very 
helpful for adequately planning recessive lethal tests. The most 
significant points in this respect are: 

    (a) the dependence of the outcome of the genetic test on
        the number of chromosomes tested;

    (b) the dependence of the result on the frequency of
        spontaneous mutations; and

    (c) on statistical grounds, the optimal number of tests
        to be performed.

    It is essential that, before starting a study, particular 
attention should be paid to these statistical questions.  To give 
an example, with a spontaneous mutation frequency of 0.2% lethals 
(10 lethals in 5000 progeny) and a sample size of 4500 in the 
treated group, 0.47% was the lowest value to prove statistically 
that a mutagenic effect was observed (Fig. 6).  It is also obvious 
from Fig. 6 that an increase in the number of tested chromosomes 
above 5000 does not really help to improve the resolving power of 

the assay.  The most effective way of planning a study in order to 
achieve results of statistical significance is to test about equal 
numbers of chromosomes in the control and the treated groups. 

    Two studies, consisting of three successive broods each, can be 
carried out easily within one week, with the aid of one technician.  
If 600 - 800 cultures are set up for each brood, there is a testing 
capacity of 1800 - 2400 chromosomes per study.  One to two 
replicate studies should be sufficient to classify a given test 
substance.  Complicated cases with mutation frequencies slightly 
higher than the spontaneous background, will need further studies.  
On the average, 80 man-hours are involved in testing an unknown 

    Experience with several tester stocks has shown that the 
spontaneous mutability (about 0.1 - 0.2%) remains fairly 
constant over the years.  Initially, control studies should 
be run concurrently; after more experience, concurrent control 
studies are not mandatory, if the recessive lethal test is 
either clearly positive (> 1 - 2% lethals) or negative (lethal 
frequency < historical controls from the same laboratory).  If the
percentage of lethals falls between the historical control and 1%, 
concurrent control runs are obligatory.  At least one replicate 
study should be conducted in all cases.  Literature

    Performance of, and possible pitfalls in, the recessive lethal 
test have been extensively described by Auerbach (1962a), 
Abrahamson & Lewis (1971), and Würgler et al. (1977). 

2.7.6.  Data processing and presentation

    The experimental data to be reported should include the strains 
and mating schedule used, the number of chromosomes tested, and 
both the number and percentage of lethals.  The aim of a study is 
to find out whether the mutation frequency obtained from the 
treated group is significantly higher than the spontaneous 
background.  After subtracting from the total number of F2 cultures 
those that are sterile, the experimental data consist of the 
following numerical values: 

    Nc =  number of tested X-chromosomes (number of progeny) in the 
          control group; 

    Mc =  number of recessive lethals in the control group;

    Nt =  number of tested X-chromosomes (number of progeny) in the 
          treated group; and 

    Mt =  number of recessive lethals found in the treated group.

    The basic question to be answered is: is the mutation frequency 
pt (%) = Mt/Nt x 100, determined for the treated group, 
significantly higher than pc = Mc/Nc x 100 for the control?  For 

statistical consideration of the data, the simple significance test 
developed by Kastenbaum & Bowman (1970) should be applied.  
Statistical analysis of mutagenicity data from the recessive lethal 
assay is further described by Würgler et al. (1975, 1977). 

2.7.7.  Discussion  Disadvantages of the recessive lethal test

    The scoring of induced recessive lethal mutations is a highly 
objective method of exploring the mutagenic potential of a 
chemical.  Nevertheless, there are a few cases of misclassification 
through incorrect performance of the test.  It may, therefore, be 
profitable to summarize some of the obvious problems that can arise 
in the design of such studies: 

    (a) As a standard rule, single-mating (one male and 3 - 5 
        females) should be applied to identify the very rare 
        cases of spontaneous clusters, i.e., mutants of common 
        origin.  It is then possible to keep track of the F1 
        family of cultures derived from each P1 culture.  
        Clusters will tend to appear in families.  If, in 
        postmeiotic broods, large clusters of lethals are 
        observed among F2 progeny derived from the same P1 
        male, it is recommended that these should be eliminated 
        from the final score, because they may reflect 
        spontaneous mutations that arose in dividing 
        spermatogonia during the development of that particular 
        P1 male.

    (b) Great care must be exercised in the scheme to ensure 
        the use of virgin females in the P-generation.  Thus, 
        all F1 females must be heterozygous for the treated 
        X-chromosome and the  Basc balancer chromosome, and at 
        least three of the four different phenotypes must be 
        present in the F2 generation.  Weak mutagens and non-mutagens

    Relatively large sample sizes are needed to discriminate 
between weak mutagens and non-mutagens.  It is possible to use 
either concurrent negative controls, or historical controls.  In 
the latter case, at least 10 000 control tests (chromosomes) should 
exist for a particular tester strain and each particular solvent 
(e.g., Tween 80/ethanol).  Gocke et al. (1982) reported a very 
extensive set of historical controls, collected over many years.  
It has to be stressed that results obtained with a large number of 
tests, but with only one type of exposure or application, are only 
informative with regard to that particular set of experimental 
conditions.  Recent instances of weak mutagenic activity in the 
recessive lethal test are provided in an extensive study on some 
carcinogenic polycyclic hydrocarbons and aromatic amines (Vogel et 
al., 1983).  No single technique (injection; feeding) was found to 
be suitable for all the carcinogens investigated; hence, very 
extensive experiments had to be carried out, using a flexible test 
protocol.  Data base

    The recessive lethal assay has been well developed and 
calibrated against a wide array of direct-acting agents and pro-
mutagens.  Interlaboratory variability has not been a problem with 
this assay.  According to a report of the US EPA Gene-Tox Program, 
421 compounds have been tested in the recessive lethal assay (Lee 
et al., 1983).  Of these, 198 compounds were found to be positive 
and 46 negative, at the highest concentration tested.  A third 
group containing as many as 177 compounds, was not classified as 
either positive or negative, because the very rigid criterion used 
was the test of at least 7000 chromosomes in both the control and 
the treated groups (per dose level), with a spontaneous frequency 
of 0.2%.  With the fulfillment of this criterion, it would be 
possible to detect a doubling of the recessive lethal frequency 
(Lee et al., 1983).  The problem with this approach is that 
flexibility is diminished, and that too much weight is put on one 
experimental condition.  The alternative procedure, which seems 
more realistic in view of the fact that  Drosophila constitutes a 
very complex metabolic system, would be to use a flexible test 
protocol in studies with weak mutagens.  Reliance should not be 
placed on a large number of chromosomes tested at only one 
concentration.  A variety of experimental conditions (e.g., 
injections versus feeding) can be used to identify optimal 
experimental conditions for a given genotoxic agent.  A good 
example of the latter approach is the demonstration by Zijlstra & 
Vogel (1984) that 7,12-dimethyl-benz( a)anthracene, methyltosylate, 
and  nor-nitrogen mustard are strongly mutagenic, weakly mutagenic, 
or even ineffective in the recessive lethal assay, depending on the 
route of administration used.  Correlation with mammalian carcinogenicity data

    In the Gene-Tox report by Lee et al. (1983), there were 62 
compounds that could be classified as positive or negative for both 
carcinogenesis in mammals and mutagenesis in the recessive lethal 
assay.  Of the 62 compounds, there was agreement between 
carcinogenic activity and mutagenesis classification in 56 cases 
(50 positive and 6 negative), i.e., 90% would have been correctly 
classified as to carcinogenicity using only the SLRL test.  The 
data were derived from a list in which 198 National Cancer 
Institute (NCI) bioassays were evaluated for carcinogenicity 
(Griesemer & Cueto, 1980). 

    In another comparative analysis (Vogel et al., 1980), 85 out of 
107 carcinogens (79%) were found to be mutagenic in  Drosophila.  Of 
the remaining compounds, 17 were negative and another 5 were not 
sufficiently tested to reach meaningful conclusions.  The 
documentation by Vogel et al. (1980) is based predominantly on the 
142 chemicals considered in the IARC Monographs volumes 1 - 20 for 
which there is "sufficient evidence of carcinogenicity" in 
experimental animals, according to evaluations by expert committees 
(IARC, 1979).  A second source for the documentation of 
carcinogenicity data was a list prepared by the US EPA (1976).  Recent developments

    The recessive lethal assay is a relatively time-consuming 
method compared with systems using bacteria or lower eukaryotes.  
This disadvantage may, however, be offset in the future when, in 
addition to overall metabolic considerations, attention is directed 
to differences in metabolism existing between somatic and gonadal 
tissue, as was recently demonstrated for the inducibility of AHH 
(aryl hydrocarbon hydroxylase) activity.  Thus, somatic assay 
systems might be particularly valuable as a complement to recessive 
lethal tests on the germ line.  One system is based on eye-colour 
markers (Becker, 1966), and another on wing-hair markers (Garcia-
Bellido et al., 1976).  Both systems are currently under validation 
in several laboratories (Graf et al., 1983; Vogel et al., 1983).  
With the white/white-coral system (Becker, 1966), which has been 
calibrated against 35 reference mutagens, it is possible to test 
about 4 - 6 chemicals in 2 weeks.  Moreover, tests based on the 
detection of genetic changes in somatic cells have the advantage 
that they can be performed within one generation. 

2.8.   In Vivo Cytogenetics: Bone Marrow Metaphase Analysis and 
Micronucleus Test

2.8.1.  Introduction

     In vivo bone marrow tests, which include metaphase chromosome 
analysis, and the micronucleus assay are used to identify 
clastogenic compounds, that is, those that are capable of inducing 
structural changes in chromosomes.  Chromosomal aberrations are 
analysed in mitotic metaphases from proliferating tissue, such as 
bone marrow samples from laboratory animals.  In the micronucleus 
test, clastogenic effects can be measured indirectly by counting 
small nuclei in interphase cells formed from acentric chromosome 
fragments or whole chromosomes. 

    Both tests are widely used, and they are regarded as of 
particular importance by many regulatory authorities, because, in 
the whole animal, the obvious deficiencies in artificial metabolic 
activation systems used in  in vitro systems are avoided.  Current understanding of the formation of chromosomal 

    Chromosomal aberrations occur because of lesions in the DNA 
that lead to discontinuities in the DNA double helix.  The primary 
lesions, which include single- and double-strand breaks, base 
damage, DNA-DNA and DNA-protein crosslinks, alkylations at base or 
phosphate groups, intercalations, thymine dimers, apurinic and 
apyrimidinic sites, are recognized by DNA-repair processes.  
Therefore, the lesions may be corrected or transformed, to 
restitute the original base sequence or produce chromosomal 
aberrations and/or gene mutations. 

    The breakage-reunion hypothesis (Sax, 1938) implies that a 
discontinuity in the DNA may be stabilized to appear as a break at 
metaphase.  Alternatively, the discontinuity may be restituted by 
repair processes to the original state, whereby the chromosome does 
not exhibit visible structural changes.  Two DNA discontinuities in 
temporal and spatial proximity may interact in the reunion of the 
broken ends, thus forming exchange configurations.  The exchange 
hypothesis (Revell, 1959) postulates that all aberrations are the 
result of exchange processes that involve interaction between two 
local instabilities in close proximity.  Experimental data have 
been provided in support of both hypotheses. 

    Recent experimental results support the breakage-first 
hypothesis.  The evidence for double-strand breaks being the 
ultimate DNA lesion for chromosomal aberrations has been summarized 
by Obe et al. (1982).  Double-strand breaks lead directly to 
chromosomal aberrations; all other primary lesions require 
transformation to double-strand breaks by DNA replication and/or 
repair processes. 

    Double-strand breaks can be induced directly by ionizing 
radiation and S-independent chemicals, e.g., bleomycin.  Double-
strand breaks may lead to an immediate fixation of the aberration 
by misrepair.  Depending on the time of induction within the cell 
cycle, the types of aberrations observed at the succeeding 
metaphase are of a chromosome (from G1) or chromatid (from G2) 
nature, i.e., involve both or only one chromatid (Evans, 1962).  
Most chemical mutagens do not cause double-strand breaks directly.  
The preliminary lesions are transformed in S-phase, and the 
aberrations observed at metaphase are of the chromatid type (Evans 
& Scott, 1964).  A classification of chemicals according to their 
mode of action during the different phases of the cell cycle was 
given by Bender et al. (1974) and is still valid (Brewen & Stetka, 
1982).  Classification of chromosomal aberrations

    Aberrations are divided into chromatid-type and chromosome-
type, the first involving only one chromatid, the latter, both 
chromatids at identical sites.  Furthermore, breaks can be 
distinguished from exchange configurations by their physical 
appearance at metaphase rather than by their mode of formation.  
Breaks are true discontinuities with clearly dislocated fragments 
and also include fragments without obvious origin.  They should not 
be confused with achromatic lesions (gaps), which do not represent 
true discontinuity in the DNA.  It is generally assumed that gaps 
are sites of despiralization in the metaphase chromosome that 
render the DNA non-visible under light microscopy.  It has been 
proposed that an achromatic lesion may actually be a single-strand 
break in the DNA double helix as a result of incomplete excision 
repair and, thus, may represent a point of possible instability 
(Bender et al., 1974).  Therefore, gaps are always noted but 
reported separately from true chromosomal aberrations. 

    Exchange configurations can be subdivided into intrachanges, 
i.e., exchange within one chromosome, and interchanges, i.e., 
exchange between two chromosomes.  The classification into intra- 
and interchanges applies to chromatid- as well as to chromosome-
type aberrations.  Depending on the location of the original 
discontinuities and the ways of reunion, further classification of 
exchanges is possible such as symmetrical or asymmetrical, complete 
or incomplete.  The terminology is complex but has been clearly 
reviewed by Savage (1976) and Scott et al. (1983). 

    The majority of aberrations observed at first metaphase after 
exposure are lethal to the cell that carries them or to the 
daughter cells.  Whenever acentric fragments are formed, genetic 
imbalance will result.  In the case of chromosome-type aberrations, 
both daughter cells will die, since both chromatids are affected.  
In some cases of chromatid-type aberrations, only one chromatid is 
affected and only one of the daughter cells may die.  Only 
symmetrical forms of chromosome or chromatid exchanges, with no 
loss of genetic material, will survive cell division and be 
transmitted to future cell generations.  Chromatid-type aberrations 
that survive the first division are converted to derived 
chromosome-type aberrations.  The balanced "stable" types of 
aberrations are reciprocal translocations and inversions.  Since 
there is no rule without exceptions, the occasional balance of 
genetic material will allow cell survival.  Chromosomal syndromes 
in human diseases such as Cri du Chat (deletion of chromosome 5) or 
Down's syndrome (trisomy 21) are due to structural or numerical 
chromosomes changes, i.e., genetic imbalance. 

    Some more unspecific chromosomal changes should be mentioned 
for the sake of completeness.  So-called sub-chromatid aberrations 
have been primarily observed at anaphase, most typically as "side 
arm bridges", after exposure of cells to ionizing radiation during 
prophase of cell division.  A model of sub-chromatid aberrations 
has been discussed by Klasterska et al. (1976).  However, another 
phenomenon, namely chromosome stickiness, cannot be discriminated 
from subchromatid aberrations, when cells are scored at metaphase 
rather than anaphase. 

    Chromosome shattering can be seen at metaphase, similarly 
chromosome pulverisation has been described.  There is no clear 
distinction between these two phenomena, which simply represent 
different degrees of damage inflicted on the chromosomes.  In the 
case of shattering, chromosomes appear to have been broken up into 
many small pieces of various lengths.  Sometimes just a few, 
sometimes all, chromosomes are shattered but usually intact 
chromosomes or conventional chromatid-type aberrations remain 
recognizable.  In cells with pulverisation, the chromosomes can be 
reduced to masses of fragments.  The phenomenon of premature 
chromosome condensation (PCC) is sometimes confused with shattering 
or pulverisation.  However, its appearance is quite different.  Thin 
chromosomal fragments of various lengths lie among the debris.  The 
PCC phenomenon was shown to arise from the virus-mediated fusion of 
a cell in division to a cell in S-phase, which brings about 
visualization (condensation) of chromosomes in the process of 

duplication (Johnson & Rao, 1970).  PCC was also described in 
Chinese hamster bone marrow after chemical treatment (Kürten & Obe, 
1975).  Here, it was explained as condensation of chromatin in 
micronuclei, induced by mitotic condensation of the chromosomes in 
the main nuclei while the micronuclei were still in S-phase. 

    Polyploidy and endoreduplication are frequently described in 
cultured cells, though they are less often seen in bone-marrow 
material.  Although it is important to assess aneuploidy, the 
routine scoring of numerical aberrations in bone-marrow metaphases 
is not recommended, because deviations in chromosome number often 
arise as preparational artifacts.  The basis for micronucleus formation

    Micronuclei originate from chromosomal material that has lagged 
in anaphase.  In the course of mitosis, this material is 
distributed to only one of the daughter cells.  It may be included 
in the main nucleus or form one or more separate small nuclei, 
i.e., micronuclei.  The micronuclei mainly consist of acentric 
fragments as demonstrated by DNA content measurements (Heddle & 
Carrano, 1977).  They may also consist of entire chromosomes and 
may result from non-disjunction due to malfunction of the spindle 
apparatus.  These larger micronuclei are formed by spindle poisons 
(Yamamoto & Kikuchi, 1980).  Micronuclei can be observed in any 
cell type of proliferating tissue.  They are, however, most easily 
recognized in cells without the main nucleus, namely erythrocytes. 

    The scoring of micronuclei in bone-marrow cells was proposed as 
a screening-test by Boller & Schmid (1970) and Heddle (1973).  The 
frequency of micronuclei can be evaluated most readily in young 
erythrocytes, shortly after the main nucleus is expelled.  The 
young ones are termed polychromatic erythrocytes (PCEs), the mature 
ones normochromatic erythrocytes (NCEs).  With conventional 
staining techniques, PCEs stain bluish to purple because of the 
high content of ribonucleic acid in the cytoplasm.  NCEs stain 
reddish to yellow.  The PCEs are also slightly larger than the 

    In mouse-bone marrow, the maturing erythroblasts go through six 
or seven cell divisions with a cell-cycle length of about 10 h 
(Cole et al., 1979).  About 10 h after the last mitotic division, 
the expulsion of the main nucleus is completed and the resulting 
PCE remains in the bone marrow for another 10 h.  Treatment-induced 
micronuclei derived from chromosomal fragements produced during the 
preceeding cell cycle will thus appear in PCEs not earlier than 
10 h after injection of the animal with the test chemical.  
Experience with known chemical mutagens has shown that, in fact, 
micronuclei appear much later than this.  Jenssen & Ramel (1978) 
demonstrated by simultaneous treatment of mice with methyl 
methanesulfonate (MMS) and 3H-thymidine labelling that nuclear 
expulsion was delayed by about 9 h.  Even though an increase in 
micronuclei levels compared with controls was already seen after 
12 h, the curve rose steeply between 18 and 24 h, corresponding to 
the MMC-induced delay of the last cell cycle up to nuclear 

2.8.2.  Procedure

    Detailed experimental procedures are described by Adler (1985).  Experimental animals

    Bone-marrow studies can be carried out with most laboratory 
mammals.  Chinese hamsters may be preferred for metaphase analysis 
because of their low chromosome number (2n = 22) and their readily-
distinguishable chromosomes.  Mice (2n = 40) or rats (2n = 42) are 
also frequently used.  For the micronucleus test, the use of rats 
is less convenient.  Rat tissue is rich in mast cells.  In the 
course of bone marrow preparation, these cells shed granules 
containing heparin, which stain in a similar manner to micronuclei, 
and thus, make scoring of true micronuclei rather difficult. 

    The animals used in bone-marrow studies should be young adults.  
The high proliferative activity and the low fat content of bone 
marrow in young animals favour the quality of the preparations.  
Each group of test animals should consist of equal numbers of males 
and females to allow for sex differences in response to the 
treatment.  Treatment and sampling

    The treatment should generally comprise a single application 
of the test compound, followed by multiple sampling of groups 
of animals at different times (Preston et al., 1981).  The most 
commonly used routes of application are intraperitoneal 
injection and oral intubation.  Other routes of application, e.g., 
inhalation,  are possible.  The time of maximum response may vary 
from chemical to chemical, depending on the sensitive cell-cycle 
stage and the influence of the chemical on cell-cycle length.  
Moreover, absorption, distribution, and metabolism may influence 
the optimum interval between treatment and sampling.  Therefore, a 
single, generally applicable sampling time cannot be recommended.  
The central sampling interval after dosing is usually 24 h for 
chromosome analysis and 30 h for the micronucleus test.  In 
addition, one earlier and one later sampling interval should be 
used, e.g., between 12 - 18 and 36 - 48 h for metaphase analysis 
and 12 - 18 and 60 - 72 h for the micronucleus test. 

    Earlier publications on bone-marrow cytogenetics and the 
micronucleus test (Matter & Schmid, 1971; Schmid et al., 1971) 
recommended two treatments with a 24-h interval between the two.  
Other authors have used 5 daily applications of the test compound.  
The single treatment schedule is preferable for the following 

    (a)   Cell killing

    Chromosomal aberrations do not accumulate over successive cell-
cycles, because, in most cases, they are cell-lethal.  Thus, true 
clastogens also kill cells.  In repeated treatment schedules, the 

first dose kills off the most sensitive cells leaving a changed 
cell population of more resistant cells for the following 

    (b)   Analysis of first post-treatment mitosis

    Chromosomal aberrations can only be assessed quantitatively if 
scored at the first post-treatment mitosis.  If multiple treatments 
are applied at 24-h intervals, the cells damaged by the first 
application will have gone through one or more cell divisions.  
Aberrations that are cell-lethal will have been lost by death of 
the cells.  Scorable chromatid-type aberrations, if the cells have 
not been rendered non-viable, will have transformed into derived 
chromosome-type aberrations.  These can only be recognized with 
banding techniques and karyotyping of each cell, a procedure that 
is too laborious for screening purposes.  Thus, the chromatid-type 
aberrations scored after long-term treatments will represent only 
the effect of treatment on the penultimate cell cycle. 

    In theory, micronuclei may accumulate after treatment of two or 
more cell cycles, since they are scored in a cell type that does 
not undergo further cell division (Salamone et al., 1980).  
However, as with metaphase analysis, the cell-killing effect will 
adversely influence the micronucleus yields after repetitive 
treatment of the proliferating precursor cells.  In order to 
measure the accumulation of damage under conditions of low cell-
toxicity, the spacing of treatment should be governed by the length 
of the cell-cycle.  Cole et al. (1981) recommend that for studies 
using the bone marrow of adult mice, 10-h intervals should be used 
and erythrocytes should be sampled 25 h after the last treatment.  
But, even if cell killing does not occur at low doses, cell-cycle 
delay caused by the test chemical, as described in the previous 
section for MMS, may defeat the purpose of the study. 

    Repeated treatment schedules can occasionally be justified for 
pharmacological reasons.  For example, if a compound requires 
metabolic activation by an enzyme that is induced by the chemical 
itself, it can be argued that, to establish the necessary enzyme 
level, the compound should be administered several times.  However, 
7,12-dimenthylbenz( a)anthrazene and benzo( a)pyrene, which are 
metabolized by self-induced enzyme systems, readily induced 
micronuclei after a single application and required relatively late 
sampling (36 h or 48/72 h) (Salamone et al., 1980; Kliesch et al., 
1982).  An increase in micronuclei was not observed after benzo
( a)pyrene was administered in a 5-day treatment schedule (Bruce & 
Heddle, 1979).  Thus, for the micronucleus test, the advantage of 
repeated treatments is questionable. 

    Because of cell-killing effects and the necessity to analyse 
first post-treatment mitoses, long-term treatments are not suitable 
for chromosome studies with proliferating tissues.  If, for 
whatever reason, prolonged treatment is required, e.g., in a 
feeding study, non-proliferating cells such as peripheral 
lymphocytes should be sampled from the treated animals.  These 

cells can be stimulated to cycle  in vitro so that it is possible 
to score first mitoses after treatment or to count micronuclei in 
second interphase cells.  Dose levels

    The choice of test dose levels is based on the maximum 
tolerated dose (MTD) of the compound in the species used for the 
test.  The MTD is defined by the cellularity of the bone marrow and 
the yield of analysable metaphases or PCEs.  A rule of thumb is to 
use the MTD as the highest dose.  To accept a positive result, it 
is usually necessary to demonstrate an increase in effect with 
increasing dose.  If the test results are negative, the conclusion 
is only acceptable when two or three dose levels have been tested 
or the test has been repeated.  Testing with additional dose levels 
can be restricted to the interval of maximum effect with the 
highest dose level or to the central sampling interval in case of 
negative results with the highest dose, bearing in mind the fact 
that cell-cycle delay is related, not only to the nature of the 
chemical, but also to the dose level.  Number of cells scored per animal

    The number of metaphases scored or PCEs counted per animal is 
governed by the number of animals in each group and the statistical 
procedure used for planning and evaluating the study.  At least 500 
metaphases or 4000 PCEs should be scored for a single-dose group.  Positive and negative controls

    A negative vehicle-control (solvent) is an essential part of 
each study.  A positive control is generally required.  The 
positive control is only meaningful if it demonstrates the 
test sensitivity with the lowest positive dose of a known 
clastogen, e.g., 0.16 mg/kg of mitomycin C or 3.1 mg/kg of 
procarbazine (Kliesch et al., 1982).  When the chemical under test 
is given in repeated treatments, the positive control has to 
demonstrate that treatment with a known clastogen at low dose 
levels produces an effect.  Preparation procedure for bone-marrow metaphases

    The preparation procedure has been described in detail by Dean 
(1969).  Animals are injected with colchicine or Colcemid solutions 
prior to bone marrow sampling, in order to accumulate metaphases.  
Other spindle poisons can also be used.  Dose levels and timing 
depend on the animal species.  For mice, 4 mg colchicine/kg body 
weight is usually given, 1 - 1.5 h prior to sacrifice.  Chinese 
hamsters require a longer period of colchicine treatment.   In vitro 
colchicine treatment is also possible, after collection of bone 
marrow (Tjio & Wang, 1965). 

    Bone marrow is flushed from the femur into a neutral medium 
such as 2.2% sodium citrate, Hank's balanced salt solution (HBSS).  
After the sampling of bone marrow from all animals into individual 

centrifuge tubes is completed, the cells are centrifuged for 5 min 
at 100 x 6.  The supernatant is discarded completely and a 
hypotonic solution is slowly added while agitating the tube to 
disperse the pellet.  The hypotonic medium can be 1% sodium 
citrate, 0.56% potassium chloride, or the medium diluted with 
distilled water (1:1).  The duration of the hypotonic treatment 
depends on the animal species and ranges from 15 to 30 min at room 
temperature.  Hypotonic effects may be intensified at 37 °C; 
however, clumping of cells due to collagens and fat in the bone 
marrow is also increased.  After centrifugation, at the end of 
hypotonic treatment, the cells are fixed by the addition drop-wise 
of freshly prepared cold methanol/acetic acid mixture (3:1) to the 
resuspended pellet.  The fixative is changed 3 times.  In between, 
the cell suspensions should be stored in the refrigerator and can 
remain there overnight before slide making.  It is essential that 
fresh fixative be prepared just prior to fixation; it cannot be 
kept overnight because ester formation will weaken the fixation 

    For slide making, the most crucial factor is that the glass 
slides are absolutely clean and grease-free.  They can be kept in 
70% alcohol (overnight), used wet, and then flame-dried which 
facilitates chromosome spreading.  Other methods of slide making 
include cooling the slides in an ice-box before use or storing them 
in cold distilled water and using them wet.  Shortly after the cell 
suspension has been applied (2 - 3 drops per slide), they can be 
dried on a warm plate. 

    Slides are usually stained for 10 min in a 5% Giemsa solution 
(pH 6.8).  The staining solutions have to be filtered, immediately 
before use.  The stained slides are washed in distilled water, air-
dried, cleared in xylene, and cover-slipped using a suitable 
mounting medium.  Staining with 2% acetic orcein for 30 min is also 
suitable.  Preparation procedure for micronuclei

    The method, which has been described in detail by Matter & 
Schmid (1971), Heddle (1973), and Schmid (1976), includes the 
following basic steps.  Bone marrow is flushed from the femur into 
fetal calf serum, the cells are centrifuged for 5 min at 100 x g, 
and the supernatant is discarded as completely as possible.  The 
pellet is resuspended and care should be taken to prevent the loss 
of any material into the wide part of the pasteur pipette.  One 
drop of the bone-marrow suspension is placed on one end of a clean, 
grease-free slide, and pulled behind a glass cover slip to produce 
a cone of bone-marrow smear.  The slides are air-dried before 
staining, possibly overnight, and double-stained with May-Grünwald 
and Giemsa as described by Schmid (1976).  The only change adapted 
by various laboratories is the replacement of distilled water by 
phosphate buffer (pH 6.8) for the Giemsa solution.  Microscopic analysis

    Slides are coded before scoring and only decoded after scoring 
of the entire study is completed. 

    (a)   Chromosomal aberrations at metaphase

    Slides are screened for analysable metaphases under low-power 
magnification (16 or 25 x objective).  High magnification (oil 
immersion objective) is used for examination of each individual 
metaphase.  Only cells with the complete number of centromeres are 
included.  Each aberration is noted separately on a scoring sheet.  
Vernier readings can be taken for all cells or only for those that 
carry an aberration. 

    Selection of analysable metaphases may pose a certain bias.  
Criteria for rejecting a metaphase include:  incomplete number of 
centromeres; loss of chromatid alignement and/or centromere 
splitting due to extended colchicine treatment; extensive overlap 
of chromosomes; and poor fixation of the chromosomes. 

    The mitotic index, which is the fraction of cells in a given 
population that undergo mitosis at a given time, indicates cell 
proliferation activity.  The number of mitoses should be determined 
for each animal by counting 500 nuclei.  Changes in the mitotic 
index reflect the cytotoxic effect of the treatment. 

    (b)   Micronuclei in polychromatic erythrocytes (PCEs)

    Polychromatic erythrocytes are counted in each field of high-
power magnification (oil immersion objective), and the number of 
those with micronuclei is determined.  The ratio of PCEs to NCEs is 
established for each animal by counting a total of 1000 
erythrocytes.  Changes in the ratio of PCEs to NCEs reflect the 
cytotoxic effect of the treatment.  The number of NCEs with 
micronuclei is also noted. 

2.8.3.  Data processing and presentation  Chromosomal aberrations

    From the raw data on the scoring sheets, two ways of tabulating 
the individual animal data should be used, i.e., number of 
aberrations per cell, and number of cells with aberrations.  These 
express the severity of damage to the affected cell, and to the 
cell population, respectively.  In the individual animal data 
sheets, results from each animal within one experimental group are 
listed, and the various types of aberrations are recorded.  In the 
summary reporting sheets, mean values and standard deviations over 
all animals are given for the various experimental groups.  While 
gaps and breaks are kept separate in the summary reports, the 
various forms of exchange can be summarized, but should be 
separated according to chromatid- or chromosome-type.  Two columns 
should give the mean of all aberrations per cell (including and 

excluding gaps) and the average percent of cells with aberrations 
(including and excluding gaps).  Changes in the mitotic index 
should be reported separately.  Micronuclei

    Individual animal-report sheets are compiled from the raw data.  
They contain the total number of PCEs counted, the ratio of PCEs to 
NCEs, and the number of PCEs with micronuclei.  From the individual 
animal reporting sheets, the summary report is compiled by giving 
the average frequency of micronucleated PCEs (for each experimental 
group), the average ratio of PCEs/NCEs, and the average 
micronucleated NCEs.  Statistical evaluation

    Katz (1978) stresses the point that the best designed studies 
are those with approximately equal numbers of individuals in the 
experimental and control groups.  He also points out that the 
minimum number of animals in the study is governed by the 
spontaneous incidence and the required sensitivity. 

    A statistical design for the micronucleus test was described by 
Mackey & MacGregor (1979).  They used a sequential sampling 
strategy and based the statistical analysis on the negative 
binomial distribution or the binomial distribution.  The number of 
animals in their design was not fixed and the number of PCEs per 
animal was arbitrarily chosen to be either 500 or 1000.  Decision 
limits were given on the basis of the spontaneous micronucleus 
incidence of 2/1000, a required 3-fold increase for a positive 
result and an 0.01 probability of error to both sides.  Sampling 
and treatment of animals in this design have to be continued as 
long as the cumulative micronucleus counts fall between the given 

    Equally as important as the number of animals, is the number of 
cells scored.  This again depends on the spontaneous frequency of 
the parameter under test and the required increase for a positive 
result.  Grafe & Vollmar (1977) published a table that related 
these two factors to the minimum number of cells required in the 
micronucleus test.  The table was based on the assumption of 
binomial distribution and a probability of error of 0.05.  
According to the table, a sample of 15 700 cells would be necessary 
to recognize an increase by a factor of 2 over the spontaneous 
micronucleus frequency of 2/1000.  This approach is useful when the 
total number of cells is scored with no regard for interanimal 
variation.  However, because there is usually a lack of homogeneity 
between treated animals, it is the number of animals in the 
experimental group that determines the precision of the statistical 
procedure rather than the total number of cells scored. 

    The inadequacy of the currently available statistical 
approaches lies in the fact that the number of cells per animal or 
the number of animals per group is chosen arbitrarily.  So far, 
none of the published recommendations for the statistical planning 

and evaluation of cytogenetic  in vivo tests (including the 
micronucleus test) has dealt with the problem of interanimal and 
within-animal variability in treated groups.  The distribution of 
cytogenetic variables remains debatable and difficult to determine. 

    Until more satisfactory statistical models become available, it 
may be prudent to use a non-parametric statistical procedure to 
determine whether or not two samples, one drawn from the control 
group and one from the group of treated animals, belong to the same 
population (null-hypothesis).  The rank tests by Mann & Whitney 
(1947), based on the so-called Wilcoxon Test, seem to be the 
methods of choice.  Correction for tied ranks is possible (Walter, 
1951).  If more than two independent samples are to be compared, 
the test described by Kruskal & Wallis (1952) can be used.  These 
tests require that the per animal sample size of cells is constant.  
From the tables of critical values of the test statistic U, it can 
be deduced that at least 4 animals in both groups are required 
before a test ( P = 0.05) can be applied. 

    Data from control animals should be tested for inhomogeneity by 
means of the binomial dispersion test (Snedecor & Cochran, 1967) 
and should only be accepted if homogeneity is obtained.  In 
experimental groups, individual animal response to the treatment 
may vary such that inhomogeneity is obtained.  Therefore, standard 
deviations should be calculated with the sample size given 
according to the number of animals and not the total number of 
cells scored. 

    For micronuclei in mouse bone marrow, the spontaneous rate in 
the controls is 0.2% and fairly constant.  Scoring 2000 
polychromatic erythrocytes per animal, 1000 per slide, has proved 
practical, but is still an arbitrary value.  Likewise, the 
spontaneous frequency of breaks in mouse bone-marrow metaphases is 
0.5 - 1.0%, and the arbitrarily chosen number of cells to be scored 
per animal is 125.  Practical experience with known clastogens has 
shown that, using these samples with 4 animals per group, 
significant differences can be recognized by the rank test if the 
experimental value is twice as high as the control value, i.e., the 
spontaneous rate is doubled. 

2.8.4.  Discussion  Possible errors in microscopic evaluation

    Microscopic evaluation of metaphase chromosomes for chromosomal 
aberrations is somewhat subjective.  The criteria for the 
discrimination between gaps and breaks, for instance, has been 
discussed frequently, and two general opinions exist: 

    (a) A gap is an unstained region in the chromatid that is
        smaller than the width of the chromatid.  If the
        unstained region is larger, it is termed a break.

    (b) An unstained region in the chromatid is termed a gap
        if no dislocation of the fragment is recognizable.
        The length of the unstained region is not important.
        The dislocation characterizes the break.

    As discussed in the introduction, it is generally believed 
that, unlike breaks, gaps do not represent true discontinuities 
with DNA.  This view suggests that the second criterion should be 
chosen.  Opinions differ, however; for example, Scott et al. (1983) 
recommend the first criterion. 

    Another subjective element is the choice of analysable 
metaphases, particularly as the loss of chromatid alignment can 
contribute to failures in the recognition of aberrations of the 
chromatid type. 

    Mouse chromosomes carry a heterochromatic region near to the 
centromere.  Often, the centromeres appear separated from the rest 
of the chromatids, and this separation may be more pronounced in 
one chromatid than in the other.  This phenomenon should not be 
mistaken for a gap.  Quite frequently, two acrocentric mouse 
chromosomes may appear in close juxtaposition at their centromeric 
ends and, thus, mimic a Robertsonian translocation.  True 
centromeric fusions to Robertsonian translocations are very rare 
events.  Chromatid breaks in the centromeric region lead to whole 
arm exchanges, which should be clearly distinguished from short-arm 
association due to preparational artifacts. 

    It is important that only cells with the complete number of 
centromeres are included in the scores.  If cells carrying an 
aberration, but lacking some of the chromosomes, were included in 
the scoring, it would be biased in favour of aberration-carrying 
cells.  If normal cells with one or two fewer chromosomes were 
included in the scores, it would be biased against aberrations, 
since the lost chromosomes might have been aberrant. 

    Artifacts can also obscure the micronucleus scores.  Granules 
shed by mast cells have already been described.  If they lie on 
PCEs, they can be mistaken for micronuclei.  Granular or fibrillar 
structures in or on PCEs can be discriminated from micronuclei by 
their irregular appearance.  True micronuclei are round or, on rare 
occasions, oval or half-moon shaped, but always with a sharp 
contour and evenly stained.  Most artifacts can be recognized as 
such by focusing up and down.  If the particles show a ring of 
reflection when out of focus, they are artifacts.  Comparison of test sensitivity

    The two methods, metaphase analysis and micronucleus test, are 
described, in most cases, as if they were equally sensitive.  
Accumulating evidence supports the theoretical expectation that 
metaphase analysis is more sensitive (Kliesch & Adler, 1983).  It 
is certainly more time consuming.  However, it should also be more 
exact, because all types of aberrations are scored.  As stated 
earlier, micronuclei reflect basically acentric fragments.  

Equating the frequency of acentric fragments at metaphase with the 
frequency of micronuclei in simultaneous experiments demonstrated 
that, for an expected frequency of micronuclei, the observed 
micronucleus frequencies were always below the expected, but 
without any particular pattern of reduction.  Thus, it has to be 
assumed that not every fragment forms a micronucleus.  Other 
possibilities are that it can be maintained in the main nucleus, or 
that it can be lost from observation because of cell death.  
Furthermore, the micronucleus, as such, can be expelled with the 
main nucleus.  On the other hand, the micronucleus test can reveal 
chemicals that disturb the function of the spindle, such as 
colchicine or related spindle poisons, or chemicals that 
predominantly act on tubular proteins rather than DNA, thus 
inducing aneuploidy instead of structural chromosomal aberrations.  
For the reasons mentioned above, the two tests cannot be regarded 
as true alternatives.  Thus, even though metaphase analysis 
requires a higher degree of skill and is more time consuming than 
the micronucleus test, the extra effort would seem justified.  Application of the method to other tissues

    Chromosomal aberrations can be evaluated in most tissues of 
treated experimental animals, whether somatic or germinal, e.g., 
lymphocytes, spleen, or liver (after stimulation by partial 
hepatectomy or  in vitro) (Dean, 1969), ascites tumours (Adler, 
1970), early cleavage stages of embryos (Brewen et al., 1975), 
embryonic tissues (Adler, 1981), spermatogonial mitoses (Adler, 
1974), and meioses of spermatocytes and oocytes (Adler, 1982; Adler 
& Brewen, 1982).  However, because of the increased effort needed 
to examine the preparations, the use of these tissues is, at 
present, confined to special studies. 

    Similarly, micronuclei can be counted in the cells of various 
tissues, despite the presence of the main nucleus.  Reports on the 
application of the micronucleus test to rat liver hetaptocytes 
(Tates et al., 1980), mouse embryonic liver and blood erythrocytes 
(Cole et al., 1979), and rat spermatids (Lähdetie & Parvinen, 1981) 
have been published. 

2.8.5.  Conclusions

    It is frequently stated that  in vivo methods lack the 
necessary sensitivity compared with  in vitro studies.  This 
argument is more than offset by the considerable advantage that an 
 in vivo study is much closer to the human situation, which is the 
ultimate concern of these studies.   In vivo metabolic processes 
such as activation and detoxification have to be stimulated  in 
 vitro by relatively crude enzymatic preparations.  Thus, negative 
results  in vivo may be more relevant than positive results  in 
 vitro, even with mammalian cell preparations.  The resolution of 
these discrepancies requires careful additional studies on the 
metabolism of the chemical.  Whether the problem is resolved 
depends on an understanding of whether or not the test compound or 
one of its metabolites reaches a target organ in significant 
amounts and what is the half life of the molecule at the target 

site.  It is also necessary to establish that the same metabolite 
occurs in the  in vitro and  in vivo situation.  If similar 
information can be obtained from studies on human beings, then 
sound grounds for a decision on whether or not a particular agent 
poses a genotoxic danger have been established. 

2.9.  Dominant Lethal Assay

2.9.1.  Introduction

    The term dominant lethal is used to describe a genetic change 
in a gamete that kills the conceptus early in development.  Any 
induced changes that affect the viability of the germ cells 
themselves or render the gametes incapable of participating in 
fertilization are excluded.  The pioneering studies of Hertwig 
(1935), Brenneke (1937), and Schaefer (1939) had already indicated 
that litters sired during the pre-sterile period of irradiated male 
mice were found to be of reduced size.  Since there was no effect 
on sperm mobility and since the number of fertilized eggs was 
normal, it was concluded that the reduced litter size was due to 
death of embryos after ferilization.  The observation of various 
nuclear and chromosomal abnormalities in fertilized ova led to the 
conclusion that embryonic death was caused by chromosomal 
abnormalities, induced by irradiation in spermatozoa.  The dominant 
lethal assay was used as an indicator of radiation-induced 
mutations by Kaplan & Lyon (1953), W.L. Russel et al. (1954), and 
recommended for mutagenicity testing by Bateman (1966). 

    To study the cytogenetic basis of chemically-induced dominant 
lethals, Brewen et al. (1975) collected fertilized ova from females 
mated to young adult male mice after treatment with methyl 
methanesulfonate (MMS).  The ova were collected from day 1 to day 
23 after injection.  The types of aberrations observed were 
predominantly double fragments (presumably isochromatid deletions), 
chromatid interchanges, and some chromatid deletions, as well as a 
shattering effect of the male chromosomal complement at 100 mg 
MMS/kg body weight during the peak sensitivity of dominant lethal 
induction.  These data strongly suggest that chromosomal 
aberrations observed at the first cleavage division of zygotes are 
the basis of MMS-induced dominant lethality.  In general, it can be 
concluded that most dominant lethals probably result from multiple 
chromosomal breaks in the germ cells. 

    A basic problem of chemical mutagenesis is that it is not 
possible to measure directly the genetic effects of chemicals in 
human germ cells.  Therefore, there is no alternative to using the 
data obtained in studies with mammals, particularly the mouse, to 
predict the induction of mutations in human beings.  The general 
belief that the human gonads are well protected (Neel & Schull, 
1958) was one important reason for a 20-year delay between the 
discovery of the induction of mutations by chemicals and the 
development of research programmes for mutagenicity testing. 

    The effectiveness of the blood-testis barrier (Setchell, 1970) 
was sometimes wrongly used as an argument to conclude that the 
dominant lethal assay is insensitive.  In fact, the dominant lethal 
assay is one of the few test systems that provides information 
about compounds that are able to cross the blood-testis barrier.  
Such information is of great importance in the assessment of the 
possible mutagenic hazard of a chemical. 

    When a chemical substance has penetrated the blood-testis 
barrier, it might be subjected to the enzymatic activation 
processes in the various tissues of the gonad.  The compound can 
also be detoxified.  After these modifying processes, the chemicals 
or their metabolic products may interact with the DNA.  The 
resulting damage may be subjected to different DNA repair processes 
and, finally, a sperm develop that may or may not carry a mutation.  
In addition, Generoso et al. (1979) demonstrated that the yield of 
chemically-induced dominant lethal mutations in male mice depends 
on the genotype of the female, and this difference between females 
of different strains may be caused by differences in the activity 
of the repair enzymes.  Though the possibility cannot be ruled out 
that the induction of a repair enzyme was stimulated by the 
chromosome lesions themselves, it seems more likely that the repair 
enzyme existed in the egg prior to sperm entry. 

2.9.2.  Procedure for male mice

    The method consists essentially of sequential mating between 
treated or untreated male mice and untreated females.  Mating 
usually occurs at night, and conceptions can be recognized the 
following morning by the presence of a vaginal plug.  This plug is 
a convenient means of timing a pregnancy.  Pregnant females are 
sacrificed on the 14 - 16th day of pregnancy.  The corpora lutea, 
representing the number of ova shed, are counted.  The uterine 
contents are scored for early and late deaths and living fetuses.  
The induction of dominant lethals is determined by the increase in 
pre- and postimplantation loss of zygotes in the experimental group 
over the loss in the control group.  This simple procedure is an 
essential advantage of the dominant lethal assay. 

    The testis of the adult male contains the complete sequence of 
maturation stages of the germ cells, from the stem-cell 
spermatogonia to the spermatozoa passing out of the testis into the 
epididymis.  The timing of this sequence has been determined for 
the mouse by Oakberg (1960).  The earliest time at which specified 
cells reach the ejaculate is: 

    1 - 7 days         spermatozoa
    8 - 21 days        spermatids
    22 - 35 days       spermatocytes
    36 - 41 days       differentiated spermatogonia
    more than 42 days  AS (stem cell) spermatogonia.

    During the first day after conception, fertilized eggs remain 
aggregated in the cumulus of follicle cells.  By the 4th day, a 
blastocyst has developed.   The blastocyst passes from the oviduct 
into the uterus.  It is the most advanced stage to which a 
fertilized egg can develop without implantation.  Implantation 
occurs on the 5th day.  By the 10th day, organogenesis is complete 
and, during the following 10 days, the existing structures 
differentiate, and the embryo completes its development (Bateman & 
Epstein, 1971).  Dominant lethals include the loss of fertilized 
eggs before and after implantation. 

    The various germ-cell stages have different sensitivities to 
the induction of dominant lethals by chemical mutagens, and the 
germ-cell stage assayed depends on the interval between treatment 
and mating.  The frequency of dominant lethals can change 
drastically in a 24-h mating interval (Ehling et al., 1968).  
Therefore, it is essential to use a sequential mating schedule of 
only a few days.  An overview of the differential induction of 
dominant lethals has been made by Ehling (1977). 

    The recommended test systems for mutagenicity screening are 
generally based on the expertise and the facilities of a given 
laboratory.  The procedure given here for the dominant lethal assay 
is based on a collaborative study involving 9 laboratories (Ehling 
et al., 1978).  The comparative testing of substances using the 
same method in several laboratories was designed to identify 
criteria that are critical for the optimal conduct of the dominant 
lethal assay.  From these studies, it was concluded that, while 
certain test conditions could be standardized for improving the 
reproducibility of results obtained in different laboratories, 
other test conditions were matters for establishment in individual 
laboratories, depending on preferences and conditions.  Standard and test conditions

    (a) The mating period should be short enough to provide 
        information about the action of a chemical mutagen on 
        a specific germ cell-stage.  For screening purposes, 
        where high fertilization rates are expected, a 4-day 
        mating period is recommended.
        The reason for this recommendation is the 4-day 
        estrous cycle in female mice.  In addition, the
        incidence of conceptions should be approximately 
        equally distributed over all days, and only a few 
        females conceive on days 5, 6, and 7 during a weekly 
        mating period.

    (b) The total test period should cover the whole
        spermatogenic cycle, i.e., at least 12 consecutive
        mating intervals of 4 days each.  Limiting of the
        dominant lethal assay to certain "critical" mating
        periods of high sensitivity is only permissible in
        repeat tests, or if the parts of gametogenesis
        concerned, e.g., spermatogoniogenesis or
        spermatocytogenesis, are known from previous studies.

        The reason for this recommendation is that chemicals 
        of unknown mutagenic action can induce mutations in a 
        very specific stage of gametogenesis.  After long-term 
        treatment, however, either a 4-day or a 7-day mating 
        interval following treatment can be used (Anderson et 
        al., 1983).

    (c) The preferred ratio of mating is 1 female to each male.

        The reason for this recommendation is the high
        conception rate of females when mated 1:1.  In
        addition, using a 1:1 mating mode, the results of
        each female can be directly associated with a certain
        male.  However, other mating modes, e.g., 2 females
        to each male, are widely used and acceptable,
        providing a highly fertile strain of mice is used
        (Anderson et al., 1983).

    (d) Dose levels should be calculated in terms of mg/kg 
        body weight.  The dose volume is adjusted according 
        to the weight of the animals, and administered by 
        the appropriate route.

    (e) The allocation of the animals to the various treatment 
        groups must be based on a statistically randomized 

    (f) Results obtained from sick animals or those that died 
        during the course of the trial should not be included 
        in the evaluation, but should be reported.

    (g) The sensitivity of the chosen mouse strains should be 
        regularly checked using a standard dose of a known 
        mutagen (section, recommendation (b)).  The 
        results of these studies should be documented.  Test conditions to be established by each investigator

    (a) The following test conditions are matters for the 
        individual laboratory and are based on the experience 
        of the investigator:  animal strain, age, and housing 
        conditions.  A preliminary mating to check the 
        fertility of the animals may be necessary, and vaginal 
        plug evidence is useful for this purpose.  The 
        spontaneous postimplantation losses depend, not only 
        on the genome and the housing conditions, but also on 
        the age of the females.  The optimum age for the 
        genotype of females to be used must be determined.  
        This age is characterized by a maximum number of 
        corpora lutea and a minimum postimplantation loss.

    (b) The suitability of an animal strain for the dominant 
        lethal assay has to be confirmed by using known mutagens 
        that produce specific effects at different germ-cell 
        stages.  According to experience gained in a coordinated 

        study, the induction of dominant lethals following the 
        intraperitoneal injection of 20 mg MMS/kg body weight or 
        40 mg cyclophosphamide/kg body weight is an indicator of
        the suitability of a particular mouse strain.

    (c) If vaginal plug data are not used to determine the timing 
        of the pregnancy, the autopsy of the females is best 
        carried out a fortnight after the middle day of the mating 

2.9.3.  Dominant lethals in female germ cells

    Females are less suitable than males for the screening of 
potential mutagens.  The treatment of a female with a chemical 
mutagen could interfere with the hormonal status and thereby the 
competence of the animal to carry pregnancies to full term.  The 
treatment could also interfere with the implantation or affect the 
cytoplasm of the ovum in such a way that the chances of it being 
fertilized are reduced or the cleavage divisions interfered with.  
The mutagen could also affect the ovulation rate.  All these 
factors are extremely important for the interpretation of dominant 
lethal studies in the female (Bateman & Epstein, 1971).  There is 
also the simple technical point that, while the mutational response 
of a male can be analysed by mating it with several females at 
different times, the response of a female can be studied only in a 
single pregnancy. 

    However, for some compounds, it may be desirable to test the 
induction of dominant lethals in female mice.  This will 
necessitate proving that the ova were fertilized.  On the basis of 
this knowledge, Generoso (1969) developed a method for the 
calculation of the dominant lethal frequency in females. 

2.9.4.  Data processing and presentation

    The number of animals necessary for mutagenicity testing with 
the dominant lethal assay depends on the genotype of mice and the 
quality of the animal husbandry.  In simulation runs on a computer, 
the sample sizes have been determined for NMRI-Kisslegg and 
(101xC3H)F1 mice.  The data for simulation runs were taken from a 
total of 7000 untreated control animals.  If a type 1 error of 
a = 0.05 is assumed, together with an equally large type 2 error of 
B = 0.05, then the sample sizes required for different alternative 
hypotheses are given in Table 1 (Vollmar, 1977). 

    The presentation of the data should contain all the information 
required to assess the test design and to evaluate the results.  
The following items should be stated: 

    number of females paired (absolute);
    number of females with implantations (absolute and in %);
    number of implantations (absolute and per female);
    live implants (absolute and per female); and
    dead implants = postimplantation loss (absolute and per

Table 1.  Samples sizes for dominant lethal assay 
in the male mousea
Mutagenic effectb             Genotype         
(%)                NMRI-Kisslegg  (101 x C3H)F1   
10                 70             45
15                 27             19
20                 22             15
a From:  Vollmar (1977).
b Lowering of the probability that a live implant 
  will arise from an ovulated oocyte by %.

    In some strains, corpora lutea counts pose difficulties.  Since 
the knowledge of the number of corpora lutea is not absolutely 
necessary for the evaluation of induced mutagenicity in the 
dominant lethal assay on male animals, this count can be dispensed 
with, even though, depending on the quality of the counts, this 
entails a loss of information.  However, for certain methods of 
evaluation, the corpora lutea count is indispensable.  It is also 
necessary for the dominant lethal assay on female mice, because, 
for example, of the possibility of induced superovulation (Russell 
& Russell, 1956). 

    If the corpora lutea count has been obtained, it should be 
stated (absolute and per female) as should the preimplantation loss 
derived from the difference between the number of corpora lutea and 
the number of implantations (absolute and per female).  Although 
the postimplantation loss is the most important criterion for the 
dominant lethal assay, calculations of the frequency of dominant 
lethals should not be based on the rate of postimplantation losses 
only.  Otherwise, a sterile phase, induced by a highly potent 
mutagen as a result of cytotoxic effects or a 100% preimplantation 
loss due to a genetic cause (Kratochvilova, 1978), could be 

    The following formal relationship exists between the 
postimplantation and the preimplantation losses: 

    DI < I, with I = CL - PL

(DI = dead implants = postimplantation loss; I = number of 
implantations; CL = corpora lutea; PL = preimplantation loss).  
This means that after a high preimplantation loss, the maximum 
possible postimplantation loss decreases automatically, since the 
number of corpora lutea can be assumed to be fixed when the male 
mice were treated. 

    The calculation of dominant lethals comprises principally the 
pre- and postimplantation losses.  These losses are expressed as 
the mean number of live implants per female.  A good approximation 
for the induced dominant lethal frequency can be obtained by using 

the formula of Ehling et al. (1968), which is based on the number 
of live implants: 

    Frequency of dominant lethals (FL) =

          live implants per female of the test group
    1 -   ---------------------------------------------
          live implants per female of the control group

    or the percentage of the dominant lethals (FL%) = 

          live implants per female of the test group
    1 -   ---------------------------------------------  x 100
          live implants per female of the control group

    This formula has the advantage that the spontaneous lethal 
rate, which is independent of the treatment and is specific for 
each mouse strain, is eliminated so that the proportion of lethals 
induced by the treatment is directly evident.  In addition, if the 
treatment is effective, this calculation should give a zero value, 
but the sample size will result in statistically-insignificant 
deviation from zero in positive and negative directions.  The minus 
deviations are a good indicator of the biological variability of 
the sample. 

    A drawback of this formula is that it is not derived from the 
individual values of the females and does not incorporate 
interindividual variability, which is obligatory for statistical 
analysis.  Furthermore, this kind of computation presupposes a 
target model with Poisson distribution, which cannot be assumed in 
every case.  It should be noted that the proposed calculation can 
only be used for the evaluation of treated male mice.  The 
calculation of the induction of dominant lethals in females is 
based on the numbers of corpora lutea (Russell & Russell, 1956). 

    For statistical evaluation, it must be decided which biological 
model is to be used and which statistical criteria are appropriate.  
For biometric analysis, 5 variables are available at 3 levels 
(Table 2).  All these variables are integer frequencies and must be 
rated as discrete variates. 

    The female should be chosen as the sample unit.  Prior to the 
actual analysis for mutagenic effects, the death rate of the male 
animals (if required) and the fertilization rate of the females 
(obligatory) must be ascertained.  If there is a significant 
difference between groups, the analysis of any mutagenic action 
will be limited, and ambiguous conclusions may result. 

    The Wilcoxon test, modified by Krauth and recommended by 
Vollmar (1977) is an appropriate procedure for the statistical 
analysis of the dominant lethal assay.  According to Vollmar, 
fertilization rate is tested by the exact Fisher-Yates test and the 
quotients I/CL, LI/CL or DI/CL by a separate linear rank test for 
each mating interval. 

Table 2.  Variables for biometric analysis
Level  Variable                                     Abbreviation

I      Number of corpora lutea per female           CL
II     Number of implantations per female           I
       Number of preimplantation losses per female  PL
III    Number of live implants per female           LI
       Number of dead implants per female           DI

    Haseman & Soares (1976) have compared different statistical 
test procedures by computer simulations of the dominant lethal 
assay.  They concluded that the Chi-squared test, frequently used 
for the analysis of the dominant lethal data, may seriously 
exaggerate the level of significance and should not be used.  The 
inappropriateness of the underlying Poisson or binominal models 
appears to have little effect on the validity of analysis of 
variance procedures based on transformed fetal death data.  It can 
be concluded that, until a satisfactory parametic model can be 
established, nonparametic procedures are to be preferred. 

2.9.5.  Discussion

    Two different approaches are used to determine the frequency of 
dominant lethal mutations in male mice.  One is based on 
postimplantation death, the other on both pre- and postimplantation 
loss.  The index of dominant lethality based on postimplantation 
death alone was advocated by Bateman (1958), Epstein & Shafner 
(1968), and Searle & Beechy (1974).  Support for this index comes 
from irradiation experiments in which Searle & Beechey (1974) found 
that the decrease in implantations per female was mainly due to 
failure of fertilization.  This finding cannot, however, be 
generalized to apply to chemically-induced dominant lethals.  
Results of MMS studies clearly demonstrate that 100% 
preimplantation death of fertilized ova can be observed in the 
mating interval, 9 - 12 day post-treatment (Kratochvilova, 1978). 
The postimplantation index underestimates even the frequency of 
radiation induced dominant lethal mutations, as has already been 
pointed out by L.B. Russell (1962).  Certainly, for the detection 
of a chemical mutagen, possible underestimation of the mutation 
frequency should be avoided. 

    Calculations based on comparisons of treated and control groups 
with respect to the ratio of living plus recently dead embryos to 
corpora lutea, used by W.L. Russell et al. (1954), or with respect 
to the number of live embryos per females (Ehling et al., 1968), 
include the pre- and postimplantation losses.  The disadvantage of 
such calculations of dominant lethal frequency is that the formula 
does not differentiate between preimplantation loss and 
unfertilized ova.  However, no formula, by itself, can achieve such 
a differentiation.  For the exact determination of dominant lethal 
frequency, it is necessary to determine the rate of fertilization 

of ova (Kratochvilova, 1978).  It should be mentioned that a 
decreased frequency of fertilization is also an indication of a 
possible hazard. 

    A critical review of the mutagenicity of 20 selected chemicals 
in the dominant lethal assay was published by Dean et al. (1981).  
Approximately 300 publications were scrutinized, and data from 130 
of these were selected for inclusion in the review.  This report 
contains a concise tabulation of the most relevant data and a 
detailed review of each individual chemical including the lowest 
dose to induce dominant lethals and the highest dose with no 
significant dominant lethality.  The review also contains data for 
the induction of dominant lethals in rats. 

    The protocol described in this paper is based on the experience 
with the mouse.  However, the factors that are important for the 
optimal test procedure are likewise essential for testing dominant 
lethals in other species.  Species differences for estrous cycle 
and embryonic development have to be taken into account for the 
adoption of this protocol for other species. 


3.1.  Introduction

    In order to conduct the procedures described in section 2 
according to acceptable scientific standards, certain minimum 
levels of laboratory facilities and equipment are essential.  The 
design of the facilities and their complexity varies with the type 
of study to be performed but, in all cases, they are governed by 
the need to ensure adequate control of cleanliness and sterility, 
safety, and accuracy and reproducibility of experimental results. 

    The need for national and international authorities to regulate 
the manufacture, transport, and use of chemical substances has led 
to the introduction of legislation to control these activities.  
Included in this legislation are requirements for the toxicological 
testing of chemicals that, in most cases, include testing for 
possible mutagenic and carcinogenic hazards.  In order to 
establish acceptable standards of quality and reliability of the 
toxicological data submitted to the regulatory authorities, various 
bodies have published codes of Good Laboratory Practice.  The 
application of Good Laboratory Practice (GLP) in genetic toxicology 
testing laboratories is described in section 3.3. 

3.2.  Laboratory Facilities and Equipment

    Regardless of whether laboratories are designed for conducting 
only the minimum of  in vitro mutagenicity studies or for carrying 
out an extensive programme of  in vitro and  in vivo genetic 
toxicology testing, the same basic principles of laboratory design 
apply.  For example, separate areas should be provided for 
microbiology, tissue culture, cytogenetics and, where necessary, 
for  Drosophila testing and plant studies.  They will usually be 
situated in the same building and should be conveniently served by 
a common glassware washing and sterilising facility.  Chemistry and 
biochemistry laboratories should be available nearby to provide 
analytical support, e.g., for confirming the stability, purity, 
etc., of test compounds or for conducting associated metabolic 
studies.  Where  in vivo studies are to be carried out, animal 
facilities should either be housed in a different building or, at 
least, have a separate entrance to that of the microbiology and 
tissue culture laboratories.  Animal-holding rooms may be required 
in the main laboratory complex for cytogenetic studies and for 
providing material for microsomal enzyme preparations. 

    Particular attention should be paid to environmental conditions 
within the laboratory.  Temperature and humidity should be 
controlled within strictly-defined limits appropriate to the 
techniques being carried out.  Ventilation should be adequate with 
a given number of air changes, e.g., 8 - 10/h, but draughts and 
direct-intake of external unfiltered air should be avoided to 
minimize the introduction of dust and contaminating microorganisms. 

    The design of individual laboratories should be based on the 
provision of adequate bench space for the number of staff to be 
employed and adequate room for the equipment and storage of 
materials.  The safety of staff should be a prime consideration.  
Access to areas where studies are conducted should be limited to 
those directly involved in the testing.  Staff should be fully 
aware of the hazards of working with carcinogenic and mutagenic 
chemicals, particularly with the safe disposal of chemical waste, 
and appropriate safety cabinets, protective clothing, and washing 
facilities should be provided.  Such practices as mouth-pipetting 
should be prohibited, and an area should be specifically designated 
for weighing mutagens and carcinogens and for the preparation of 
stock solutions.  A high standard of cleanliness should be 
encouraged in all working areas, and the design of working 
surfaces, storage areas, ventilation systems, etc., should be aimed 
at making clean and sterile working practices an easily-attainable 

    Certain items of equipment are common to most types of testing 
laboratories.  Refrigerators should be capable of safe storage of 
flammable solvents, while deep freezes are required to be suitable 
for the long-term storage of materials at low temperature (-70 °C).  
In areas where the main electricity supply is unreliable, some form 
of emergency power generation is advisable.  A competent glassware 
washing and sterilizing facility is essential for experimental work 
of acceptable quality.  In addition to properly-trained personnel, 
a supply of high-quality distilled water and a suitable non-
residual detergent are necessary to provide clean glassware.  
Equipment such as autoclaves and hot-air sterilizers should be 
of a design appropriate to the types of materials being sterilized. 

3.2.1.  Microbial laboratories

    Two important factors in the design of laboratories for 
bacterial or yeast assays are the prevention of contamination of 
cultures by other microorganisms and the protection of staff 
against exposure to hazardous test chemicals.  Experimental 
procedures should be conducted in appropriate biological safety 
cabinets in which a curtain of filter-sterilised air protects the 
worker from chemical exposure and the cultures from contamination.  
Air from the cabinets should be extracted outside the building 
through appropriate filters to prevent environmental contamination. 
Incubators should have precise temperature control, and those used 
for testing purposes should be in an area where the ventilation 
system removes any hazardous vapours from volatile test chemicals, 
when incubator doors are opened.  Culture media may be either 
purchased as ready-poured plates or prepared in the laboratory from 
basic ingredients.  In the latter case, a clean working area must 
be available for pouring and drying plates.  Either manual or 
electronic devices are available for counting bacterial colonies.  
A safe means of disposal of cultures should be provided, e.g., they 
should be sealed in plastic or paper sacks in the laboratory and 
then incincerated. 

3.2.2.  Tissue culture laboratories

    Laboratories involved in cell and tissue culture are even more 
dependent on sterile procedures and working conditions than 
microbial laboratories.  Even a small initial microbial 
contamination can rapidly spread to other cultures and easily 
destroy a number of experiments and many weeks work.  Although the 
incorporation of antibiotics into the culture medium serves to 
limit many bacterial infections; contamination with yeast and fungi 
presents serious problems in inadequate working conditions.  The 
incidence of contamination can be significantly reduced by 
conducting all manipulations of cell cultures in appropriate 
biological cabinets.  Culture media can be purchased in a ready-
prepared form or can be prepared and filter-sterilized in the 
laboratory.  Liquid nitrogen storage flasks are necessary for 
keeping stocks of cell lines.  For many types of experiments, cells 
are cultured in Petri-type dishes and incubators in which a 5% 
carbon dioxide (CO2) atmosphere can be maintained are required.  
The safety precautions described above (section 3.2, 3.2.1) for 
handling and disposing of material containing hazardous chemicals 
also apply to tissue culture procedures. 

3.2.3.  Facilities for other procedures

    The main requirement for cytogenetic studies on mammalian cells 
or plant material is for high magnification and resolution 
microscopes with good quality lenses.  Microphotography equipment 
is useful for recording purposes, and a dark-room facility is 
required for unscheduled DNA synthesis studies. 

    One of the major advantages of test systems using plants is 
that many of them can be performed with relatively simple 
facilities and equipment.  For example, root tip chromosome studies 
in  Allium bulbs require little more than a thermostatically 
controlled water bath, basic glassware, and a suitable microscope, 
while seeds of  Vicia, Hordeum, and  Allium can be germinated on 
moist filter paper or in a suitable aqueous medium (section 2.6).  
Specific locus studies in species such as  Tradescantia, Zea, and 
 Hordeum can be conducted either in controlled environmental 
chambers, conventional greenhouses or, where climatic conditions 
are suitable, outdoors.  In these cases, some form of control over 
temperature, light conditions, and humidity is important. 

    In general,  Drosophila studies, also, only require relatively 
simple equipment.  Stock cultures are maintained in glass bottles, 
smaller vials are used for the experimental procedures, and they 
need to be kept in a constant temperature room.  A means of 
anaesthetising the flies is required; examinations are carried out 
using a low power microscope, and it is useful to have a separate 
area for the preparation of  Drosophila medium. 

    The maintenance of laboratory animal facilities for breeding 
and experimentation purposes is an extremely expensive part of a 
toxicology laboratory, and it is probably uneconomical to maintain 
breeding colonies of rodents wholly for use in  in vivo  

genotoxicity studies.  They are usually available from adjoining 
conventional toxicology laboratories or from commercial suppliers.  
An animal-holding area to house animals during dosing and 
dissection is useful for cytogenetic studies, but should be 
isolated from laboratories undertaking sterile procedures.  In 
addition, the area used for dosing and holding animals during 
studies should be of a design suitable for housing in safety, 
animals dosed with genotoxic chemicals.  Dominant lethal studies 
require more extensive animal facilities and should be conducted in 
a conventional toxicology animal unit. 

3.3.  Good Laboratory Practice in Genetic Toxicology

3.3.1.  Origins and nature of GLP

    In the years following 1970, several national governments 
enacted legislation to control the way in which chemical substances 
were manufactured, traded, and used.  In general, the laws placed 
the responsibility of testing chemicals to detect potential hazards 
for man and the environment on the manufacturers, as a prerequisite 
of their being allowed to market the chemicals.  The present-day 
codes of Good Laboratory Practice (GLP) have arisen as a result of 
these enactments. 

    The reliability of the data was of crucial importance to the 
governmental agencies charged with administering these regulations, 
and with assessing the risks on the basis of the experimental 
evidence submitted to them.  The US Food and Drug Administration 
(US FDA) reacted to finding that some of the data being presented 
were of poor quality and unreliable by publishing a code of 
practice (Federal Register, 1978), to which all laboratories 
generating data for submission to this authority were expected to 
adhere.  Conformity with the code was to be monitored by inspectors 
from the US FDA.  Other regulatory authorities, similarly placed, 
followed in promulgating codes of their own.  These authorities 
included the US Environmental Protection Agency, concerned 
separately with industrial chemicals and with pesticides, and a 
number of departments of other national governments. 

    Trade in chemicals is international and the proliferation of 
national and departmental codes reflecting different criteria for 
the acceptability of data would have hindered it.  Thus, the work 
of the Organisation of Economic Cooperation and Development (OECD) 
in developing its Principles of Good Laboratory Practice in 
May 1981 (OECD, 1982) was significant as one element in a programme 
to enable the mutual acceptance of experimental data among the 
member nations.  The OECD Principles are generic, i.e., they are 
intended to be applied irrespective of the end-use proposed for the 
material being tested (e.g., drug, pesticide, or industrial 
chemical) and of the nature of the data being sought (e.g., 
determination of mammalian toxicity, or of ecotoxicity).  Mutual 
acceptability of study data would also depend on the studies having 
been conducted in accordance with the OECD Test Guidelines 
(standards relating to the scientific content of the tests) and on 
the testing laboratory having been subject to inspection, in 

accordance with OECD recommendations, by its national inspectorate.  
Thus, data generated in one country in compliance with the OECD 
principles and also subject to the other criteria mentioned, should 
be accepted by all the member states. 

    Acceptance and recommendation of the Principles of GLP 
established by the Council of the OECD is not legally binding for 
the member countries.  The various regulatory authorities, within 
their own countries, are subject to different legislative and 
administrative systems.  The OECD Principles have therefore to be 
seen as guidelines to be incorporated by each country into its own 
legislation and adapted according to its special needs.  To date, 
this has been the case when countries or regulatory bodies have 
issued definitive codes of GLP.  However, differences in 
requirements reflected in existing codes of GLP are not serious, 
and there is a consensus that adherence to the OECD Principles 
would ensure the integrity of experimental data.  The US FDA rules 
(US FDA, 1979, 1982) remain the most comprehensively documented of 
the existing codes. 

    GLP sets a standard for the management of scientific 
experimentation so that laboratory management and scientists are 
able to assure, and regulatory authorities to assess, the integrity 
and quality of the data generated.  As defined by the OECD, it is 
concerned "with the organisational processes and environmental 
conditions under which laboratory studies are planned, performed, 
monitored, recorded and reported".  Owing to their origin, codes of 
GLP relate only to studies made for regulatory submission.  
Although several codes of GLP exist, the underlying concepts are 
common to them all; moreover, scientists will recognize that the 
majority of the requirements are those that have traditionally been 
regarded as the basis of all sound scientific investigation. 

    The objectives of GLP are met by ensuring that: 

    (a) responsibilities of personnel are properly delineated
        and assigned;

    (b) appropriate standards are defined for all resources
        (staff, facilities, equipment, materials) and for the
        conduct of work, and appropriate plans are defined
        for studies; adherence to the set standards and plans
        is monitored; and

    (c) all relevant aspects of the running of a testing
        laboratory or the conduct of a study are documented
        and the records are kept so that a study can be
        reconstructed and assessed in retrospect.

    It is worth emphasising the importance of adequate 
documentation in GLP. 

    A testing laboratory generating data for submission to 
regulatory authorities must therefore establish formal systems for 
fulfilling the requirements of GLP and must document these systems.  

The salient points of GLP are described and commented on in the 
following section.  More detailed accounts can be found in the 
Federal Register (1978, 1983), US FDA (1979, 1982), and OECD 

3.3.2.  GLP requirements

    (a)   Roles and responsibilities of personnel

    Three principal and distinct roles can be discerned in relation 
to the conduct of studies under GLP. 

    The Management of a testing laboratory carries the ultimate 
responsibility for running the laboratory and ensuring that 
compliance with GLP is maintained throughout.  Management defines 
the appropriate standards for all necessary resources including 
laboratory facilities, equipment and supplies, personnel, and 
methodologies, and ensures their timely and adequate provision.  A 
particular obligation of management is to appoint a Study Director 
for each study, before the study starts, and to replace him 
promptly, if necessary. 

    The Study Director is the "chief scientist" in overall charge 
of the study.  GLP requires this single point of control to avoid 
ambiguities and conflicting instructions that might arise from a 
diffusion of the responsibility among more than one individual.  
The Study Director must agree to the approved protocol or plan of 
the study and, thereafter, must ensure that the study is carried 
out in accordance with the protocol and with GLP.  He must obtain 
authorization for, and document, all necessary deviations from the 
protocol.  He is responsible overall for the technical conduct of 
the study and the recording, interpretation, and reporting of the 
observations including unanticipated responses or relevant 
unforeseen circumstances.  The scientist appointed Study Director 
must have training and experience appropriate to, and commensurate 
with, his role in the study. 

    The third principal role identified in GLP is that of the 
Quality Assurance Unit (QAU).  This is a concept adopted from the 
more traditional and familiar function of product quality control 
in industry.  The "product" of a testing laboratory consists of the 
data that arise from its studies and the concern of the QAU is with 
the authenticity and integrity of these data. 

    The primary responsibility of the QAU is to monitor, by direct 
observation, the operation of the testing laboratory and the 
conduct of studies, to ensure that these comply with the approved 
standards and with GLP and, in the case of the study, that the 
protocol is being followed.  Thus, the QAU must carry out periodic 
inspections of the facilities and equipment, the operation of the 
relevant administrative systems and the actual conduct of the 
experimental work.  Surveillance of a study by the QAU must include 
an audit of the final report.  The findings of the QAU inspections 
or audits are reported to Laboratory Management and, in respect of 
studies, also, to the Study Director so that necessary corrective 

action can be taken.  It has been the policy of regulatory 
authorities, when monitoring laboratories for GLP compliance, not 
to look at the reports of QAU inspections in order to promote the 
frankness and hence the effectiveness of the laboratories internal 
monitoring.  The QAU must maintain a "master schedule" giving 
details of the identity and current status of studies conducted at 
the laboratory. 

    Personnel of the QAU report to management.  They must be 
entirely independent of staff engaged in carrying out the study 
that is being monitored, i.e., they may not participate technically 
and may not be subordinates of the Study Director (it follows that 
the chief manager of a testing laboratory may not himself be a 
Study Director).  However, the QAU staff need not be dedicated 
solely to the quality assurance role, e.g., a scientist 
participating in one study can be assigned quality assurance 
responsibilities in respect of another.  If this expedient is used, 
the quality assurance documentation must still be kept in the one 
place in the testing laboratory.  QAU personnel must be familiar 
with GLP requirements and also be knowledgeable about the tests 
being monitored, though the quality assurance function does not 
extend to a scientific appraisal of the study and its results. 

    All staff concerned with studies must have qualifications, 
training, and experience appropriate to the function to which they 
have been assigned.  This should not only be in respect of the 
scientific discipline within which their contribution is made, but 
also in respect of the requirements of GLP.  When necessary, 
appropriate training must be given, the level of which should take 
into consideration the degree of supervision of the staff member.  
The management of a testing laboratory must maintain a current job 
description for every staff member together with a summary of any 
training and experience received in relation to the job. 

    (b)   Facilities

    Laboratory accommodation must be of adequate size and be 
suitably constructed and located for the experimental work that is 
to be performed.  In genotoxicity work, the safe containment of 
hazardous materials, both chemical and biological, is an important 
consideration but, generally, the requirements of personnel safety 
are prescribed in legislation other than that of GLP.  Two criteria 
are of particular relevance to GLP.  First, similar materials from 
different studies must be sufficiently separated so that no 
confusion between them can occur.  Disciplined working methods, 
e.g., adequate labelling of containers, will complement, but cannot 
replace, adequate provision of bench space or incubator capacity to 
achieve this.  However, once laboratory facilities have been 
provided, foresight is essential to avoid having too many studies 
at the same stage simultaneously. 

    Different phases of studies must also be adequately separated 
to prevent interference between them.  Thus, areas devoted to 
chemistry and formulation, where the test and control substances 
are handled at high concentrations, should be separate from areas 

where these materials are encountered at only low concentrations.  
In the construction of laboratories, appropriate attention should 
be given to ventilation and access so that the likelihood of cross-
contamination is minimized. 

    Similarly, separate areas are necessary for the storage of 
contaminated glassware pending its disposal or cleaning.  GLP also 
requires appropriate areas to be made available for administration, 
e.g., for "writing up", and further areas for the general 
convenience of personnel such as for changing into and from 
protective clothing.  The requirement for archive space is noted 

    (c)   Equipment

    Use of equipment that is inappropriate, inadequately designed, 
or faulty can lead to the generation of unreliable data.  
Furthermore, equipment must be properly maintained and, where 
relevant, calibrated.  It should, therefore, be easily accessible 
for inspection, cleaning, and servicing.  These rules apply both to 
equipment used to generate data (laboratory instrumentation) and 
equipment used to maintain special environmental conditions. 

    Written instructions, i.e., Standard Operating Procedures 
(SOPs) (see below), must be provided concerning procedures for the 
use, cleaning, and routine care of the equipment.  These should 
include schedules for the items of routine maintenance and the name 
of the person responsible for seeing that they are carried out.  
Appropriate action in the event of breakdown must also be covered.  
Much of this information will be available in the manufacturer's 
literature, but, where necessary, the handbook must be 
supplemented, e.g., to cover variations in technique peculiar to 
the laboratory.  This documentation must be freely accessible in 
the area where the equipment is used. 

    Written records must be kept of all routine maintenance of the 
equipment and also of non-routine events such as repairs after 
breakdown.  In the latter case, the nature and circumstances of the 
defect and the remedial action taken must be recorded.  One 
instance of equipment malfunction could conceivably affect the data 
from more than one study.  Therefore, while routine equipment 
maintenance records can be regarded as non-study specific and filed 
chronologically, individual study records should enable equipment 
used in the study to be identified and unscheduled events to be 

    (d)   Standard operating procedures (SOP)

    Under GLP, the routine methods of a testing laboratory, as well 
as some administrative procedures that relate to the conduct of 
studies, have also to be documented in the form of "standard 
operating procedures" (SOPs).  Written SOPs serve to ensure that 
all staff are familiar with, and use, the same working methods; 
thus, errors or loss of data arising from variability between 
individuals is minimized.  They can serve also as a documented 
specification of the laboratory's procedures, helping evaluation of 
study methods, or the monitoring of compliance for quality 
assurance purposes.  Documentation of the following procedures is 
usually regarded as the minimal requirement: 

    Test and reference substances   receipt, identification,
                                    labelling, handling, sampling, 
                                    storage; confirming homogeneity 
                                    of test formulations, stability 
                                    under test conditions;

    Equipment                       use, routine maintenance;

    Records                         coding, indexing, or labelling 
                                    of studies and study-related 
                                    material; collection, handling, 
                                    storage, retrieval of data; 
                                    report preparation; 

    Laboratory operations           special environmental 
                                    conditions, laboratory 
                                    techniques, preparation of 

    Quality assurance               conduct and reporting of
                                    inspections/audits; record 

    Working methods should be presented in SOPs in sufficient 
detail to ensure the integrity of study data, judgement of their 
adequacy being the prerogative of management.  The degree of detail 
should also be such that the SOPs can be understood and followed by 
trained staff.  The instructions should cover any work necessary, 
preliminary to the main procedure, e.g., methods of sampling prior 
to the application of a given test, and should extend to procedures 
for the handling of data and records.  Citation of published 
literature is permissible to supplement the text of an SOP.  Copies 
of both the SOP and any supplementary document cited must be freely 
available in the area where the procedure is carried out. 

    Adoption of a given standard procedure by a testing laboratory 
and any changes made have to be approved by the laboratory 
management and all changes must be formally documented.  The 
laboratory must then retain copies of the superseded SOPs with a 
record of the dates of their implementation and replacement.  It 
has been claimed that too rigid documentation of standard methods 
can present difficulties in scientific areas, such as genetic 
toxicology, where techniques are undergoing rapid evolution, but 
regulatory authorities have not excluded genotoxicity studies from 
the requirement.  Deviations (as distinct from significant changes) 
from a documented SOP in the course of a study can be made on the 
authorisation of the Study Director, provided that such departures 
are noted in the experimental record, if not already anticipated in 
the study plan. 

    (e)   Planning conduct and monitoring of studies

    Sound experimentation requires clear objectives and a 
definition of how the objectives are to be attained.  The design 
and methods of the study can then be evaluated in relation to its 
objectives to ensure that attainment of the latter is within the 
proposed scope of the study.  Few studies are entirely within the 
compass of a single scientific discipline and involvement of all 
the relevant professionals at the design stage is highly desirable.  
For these reasons, the protocol or plan of a study under GLP must 
be drawn up in writing before the study starts.  The codes of GLP 
itemize the information that the protocol must contain but, 
fundamentally, it must state the objectives and, in detail, all of 
the experimental design and methods that are to be used.  Citation 
of readily available documents such as SOPs is in order.  The 
protocol must be formally approved by the laboratory management 
and, where appropriate, by the sponsor of the study and must be 
agreed to, and signed by the Study Director.  Where changes in the 
protocol become necessary during the course of the study, these 
must be justified and documented in a formal protocol amendment, 
signed by the Study Director.  Except as provided for in protocol 
amendments, the conduct of the study must then follow the approved 
protocol and any inadvertent derviations must be documented in the 
experimental record. 

    The batch or sample indicator, as well as the chemical identity 
of the test substance, will be given in the study protocol and 
records, but it is also necessary to characterize the test 
substance and authenticate the sample by appropriate chemical 
tests.  In this connection, the stability of the test substance 
itself has to be known; moreover, if its stablity under the test 
conditions, i.e., in the formulation media, is not known, exposure 
of the test organism to the intended challenge cannot be assured.  
The homogeneity, concentration, and stability of the test 
formulation must therefore be determined. 

    In GLP, considerable importance is attached to the authenticity 
of the records.  Observations and data must be recorded directly 
and indelibly, and any changes made to the original record must not 
obscure the superseded entry.  Each entry or change must show the 
identity of the originator and the date, and the justification for 
the changes must be included.  The originals of data records, 
termed the "raw data", have a special importance in GLP in that 
they have to be preserved, though exact copies of the originals are 
also acceptable (not, however, expurgated records transcribed from 
the originals into clean notebooks.) 

    The conduct of studies under GLP calls, therefore, for a 
disciplined approach to the making and recording of observations 
and measurements.  Responsibility for the accuracy and completeness 
of the data resides formally with the Study Director, though he is 
not expected to validate individual entries; rather, he should 
assure that recording methods are adequate.  Careful attention to 
the design of the data-collection methods, e.g., the provision of 
printed data sheets that incorporate prompts to help ensure 

completeness of the entries and display them for easy review, can 
complement valuably the proper training and supervision of the 
operators.  However, the correct recording of unanticipated events 
or observations must not be overlooked. 

    The role of the QAU in monitoring the operation of the 
laboratory facilities and the conduct of studies has already been 
indicated.  For repetitive, short studies such as genotoxicity 
studies, the formal GLP requirement to inspect each phase of every 
study can be satisfied by inspection on a random sample basis. 

    (f)   Reports

    The final report of a study must be a complete presentation of 
its objectives, the results of the observations made, and the 
conclusions drawn from them.  To enable accurate evaluation of the 
conclusions to be made, the conditions under which the work was 
done must be correctly described, including unplanned occurrences. 

    An additional requirement under GLP is that the report must be 
audited against the raw data by the QAU.  This is to ensure that 
the methods and conditions of the study are correctly given and 
that the results reported are an accurate reflection of the raw 
data recorded.  An evaluation of the interpretations placed on the 
results by the scientists is beyond the remit of the QAU.  The 
report must include a statement signed by the QAU listing the dates 
on which the inspections and audits of the quality assurance 
programme (including the audit of the report itself) were made and 
when the findings were presented to management. 

    The Study Director must sign and date the final report; this 
signifies the termination of the study.  Subsequent changes to the 
report must be made formally, the amendment and its justification 
being signed and dated by the person responsible. 

    (g)   Archives

    Full evaluation of a completed study and its report could 
necessitate reconstruction of part or all of the study and this 
would require access to the experimental records.  A suitable 
archive for the preservation of the records must, therefore, be 
provided.  At the end of a study, the Study Director must ensure 
the transfer to the archives of all the raw data and other relevant 
documentation as well as the approved protocol and the final 
report.  Records that are not specific to any study, such as 
records of staff training, of equipment maintenance, and superseded 
SOPs, will also be stored in the archive. 

    Control and care of the archive and its contents must be vested 
in a named individual, with access to the archive restricted to 
authorised personnel.  Storage of the material must be orderly with 
appropriate indexing to facilitate its retrieval.  The period of 
time for which the records of a given study are retained should be 
in accordance with the rules of the regulatory authority to whom a 
submission based on the study will be made.  This period cannot 

easily be predicted but a requirement in excess of ten years should 
be anticipated.  Reasonable precautions, having regard to the 
prevailing risks and climatic conditions, must be taken to ensure 
that the stored records remain viable for this period. 

3.3.3.  Summary of resources and records needed

    The resources and mechanisms that must be established and the 
records that must be kept to comply with GLP are summarized in 
Table 3. 

Table 3.  Resources and mechanisms that must be established and records
that must be kept to comply with GLP
Resource        Provision                          Documentation
Personnel       Adequate appropriate staff         Responsibilities
                Study Director                     Training, experience
                Quality Assurance Unit

Facilities      Laboratories (biological,          Special conditions,
                chemical)                          procedures

                Archives                           procedures, indexing

Equipment       Adequate capacity,                 Methods of use and
                appropriate design                 Records of maintenance

Methodologies/  Scientific techniques
administrative  Protocol development, approval
systems         Test substance handling,           SOPs indexing
                Data collecting handling, storage
                Quality Assurance

 Records to be preserved

                Personnel responsibilities, training and experience

                Equipment maintenance

                SOPs: superseded editions, with dates

                Experimental records

                 "raw data"

                Master schedule of studies

                Quality assurance reports


4.1.  Introduction

    The procedures described in section 2 of this guide are those 
that are generally accepted as suitable for testing chemicals for 
mutagenic and putative carcinogenic activity.  Some are more widely 
used than others, and the purpose of this section is to offer 
practical guidance on the use and interpretation of these tests on 
the basis of current knowledge, experience, and acceptance.  It 
must be emphasized that there is no universal agreement on the best 
test or combination of tests for a particular purpose, though there 
have been attempts to harmonise the selection of the most 
appropriate assays by national and international bodies such as the 
Organisation for Economic Cooperation and Development.  An expert 
committee of the International Commission for Protection against 
Evironmental Mutagens and Carcinogens (ICPEMC) is also considering 
the question of the most effective combination of short-term tests 
to detect mutagenic and carcinogenic chemicals.  In addition, the 
International Programme on Chemical Safety has organized a series 
of international collaborative studies aimed at assessing the 
performance of short-term tests.  The results of the latest of 
these, the IPCS Collaborative Study of Short-term Tests for 
Genotoxicity and Carcinogenicity (CSSTT), has recently been 
published (Ashby et al., 1985). 

    The objective of testing chemicals in these short-term 
procedures is to provide an assessment of the possible mutagenic 
and carcinogenic hazards associated with the release of the 
chemicals into the human environment.  No single test has yet been 
devised that can achieve this objective with certainty.  By the 
judicious selection of a combination of assays, however, and by 
strict adherance to certain minimum technical and scientific 
criteria in their conduct, the possible genotoxic hazard of many 
groups of chemicals can be assessed with a useful degree of 
confidence.  Such assessments are inevitably subject to errors, 
varying in magnitude, that are influenced by, among other factors, 
the suitability of the chosen assays for a particular class of 
chemicals.  Furthermore, assays that detect genotoxic activity do 
not usually detect tumour promotors, hormones, and various other 
factors that affect tumour formation. 

    The possible adverse consequences of human exposure to a 
specific chemical will rarely be assessed from short-term tests 
alone.  Rather, the judgment is made from a total toxicology data 
package that may include, depending on the nature of the chemical 
or product, short- and long-term animal studies including tests for 
reproductive effects, irritancy, sensitisation, neurotoxicity, and 
data on the absorption, distribution, metabolism, and excretion of 
the chemical.  Data from some of these studies may also help 
achieve a proper understanding of the significance of the results 
of short-term tests. 

    Following the overall assessment of the possible human hazard 
from exposure to a chemical, the potential risk associated with 
human exposure is estimated by a regulatory process called risk 
management.  In this process, the potential hazard is balanced 
against the likely extent of human exposure, the perceived benefit 
of using the chemical, and other considerations.  Risk management 
involves non-scientific as well as scientific considerations and 
should not be confused with hazard assessment, which is based on 
the scientific evaluation of toxicological data. 

4.2.  Selection of Assays

    Some 20 - 30 diffferent assays are referred to in the eight 
subsections of section 2.  The selection of the most appropriate of 
these to meet a particular requirement is governed by a number of 
factors.  These include the type of genetic change to be detected, 
the metabolic capability of the procedure in relation to the 
structure of the chemical to be tested, the predictive value of the 
assay in terms of mutagenicity and carcinogenicity, the available 
expertise and facilities and, when appropriate, the legislative 
requirements of regulatory authorities. 

4.2.1.  Detection of the major types of genetic damage

    Chemicals that interact with DNA produce lesions that, after 
the influence of various repair processes, may lead to genetic 
changes at the gene level, e.g., gene or point mutations, small 
deletions, mitotic gene conversion (e.g., in yeast), or various 
microscopically-visible chromosome changes; assays are available to 
investigate each of these events (section 1).  Gene mutations

    The most widely used and most fully validated assays for 
detecting chemically induced gene mutations are those using 
bacteria (section 2.1).  They are relatively simple to perform, 
reproducible, and give reliable data on the ability of a chemical 
to interact with DNA and produce mutations.  It should be 
remembered, however, that bacteria are very simple organisms, and 
that a positive result in a bacterial assay does not necessarily 
indicate that the compound will induce similar effects in animal 
cells or other eukaryotes.  Likewise, a negative result does not 
invariably mean that the compound lacks mutagenic activity in 
eukaryotic cells or in intact mammals. 

    In order to generate data on gene mutations in eukaryotic 
cells, a choice of screening test systems is available, including 
certain procedures with yeasts (section 2.2), cultured mammalian 
cells (section 2.5),  Drosophila (section 2.7) and, to a lesser 
extent, some plant systems (section 2.6).  Each of these has 
certain advantages and disadvantages that will be further discussed 
later in this section.  Chromosomal damage

    As discussed in section 1, chromosomal aberrations are changes 
in the structure of eukaryotic chromosomes.  The simplest assays 
for investigating clastogenic (i.e., chromosome-breaking) effects 
are those involving either cultured mammalian cells (section 2.4) 
or plant root tips (section 2.6).  These tests can identify 
chemicals capable of inducing chromosome damage,  per se.  In order 
to investigate the ability of a chemical to produce chromosome 
damage in the whole mammal, two well-established  in vivo  
procedures are available.  Clastogenicity in somatic cells can be 
studied in the bone-marrow cells of rodents dosed with the suspect 
chemical, either by counting micronuclei in polychromatic 
erythrocytes or by analysing chromosomes in metaphase cells 
(section 2.8).  Alternatively, chemicals that cause chromosome 
damage in germ cells can be detected using the dominant lethal 
assay (section 2.9). 

    There is increasing evidence that chemically-induced numerical 
chromosome changes (i.e., aneuploidy) as well as being the cause of 
much inherited disease, are associated with the carcinogenic 
process.  Among assays for detecting such chemicals, a system using 
yeast is described in section 2.2. It is not yet clear, however, 
how predictive this test is for effects on mammals.  DNA damage

    Three of the procedures described in section 2 are generally 
accepted as assays that respond to chemically-induced DNA damage.  
One cellular response to such damage is the initiation of enzymatic 
repair of the damage, which involves the synthesis of a new, 
relatively short, strand of DNA.  Such repair, referred to as 
"unscheduled DNA synthesis" or UDS (to differentiate it from the 
synthesis occurring during normal cell replication), is the basis 
of the UDS assay in cultured mammalian cells (section 2.3).  
Mitotic gene conversion in the yeast  Saccharomyces cerevisiae, 
involves the accurate transfer of small segments of DNA between 
homologous chromosomes and is also regarded as a useful indicator 
of primary DNA damage.  The investigation of sister chromatid 
exchange (SCE) in cultured mammalian cells also falls within this 
category.  Although the molecular mechanism of SCE formation has 
still to be fully elucidated, it has been shown to be different 
from the mechanism leading to chromosome breakage, and the SCE 
assay is a useful method for detecting chemicals that interact with 
and damage DNA. 

4.2.2.  Scientific validity

    Before a short-term test can be used with confidence, it must 
be shown to be a valid procedure for the purpose of detecting 
genotoxic chemicals.  The target cell of the assay, whether it is a 
bacterium, a yeast, or an animal cell, must be fully characterized 
both genetically and biologically to ensure that it will respond in 
the expected fashion in the experimental system in which it is 
being used.  The second important factor is the experimental system 

itself.  The system must be capable of maintaining the target cell 
in optimum experimental conditions while ensuring that the test 
chemical has every opportunity of reaching the molecular target 
(e.g., DNA) in the cell in its most reactive form.  Third, the 
assay must be shown to be "robust", i.e., it should be fully 
reproducible so that data generated in different laboratories are 
comparable.  Genetic basis

    The genetic basis of the target cells used in the bacterial and 
yeast assays is described comprehensively in sections 2.1 and 2.2, 
respectively.  Guidance on the maintenance of the genetic integrity 
of the suggested strains is also given.  The details given in these 
sections must be followed faithfully to ensure that the genetic 
make-up of the test organisms meets the requirements of the 
particular assay.  For example, the  Salmonella strains usually 
used in bacterial assays respond to different types of mutagens, 
and the range of strains selected must be capable of detecting 
these different mutagens, e.g., frame-shift mutagens; base-pair 
substitution mutagens.  Similarly, the yeast strains described in 
section 2.2 have been specially selected to respond to genetic 
events and it is essential to confirm that the correct strains are 
used.  Similar principles also apply to  Drosophila tests in which 
properly maintained colonies of the correct strains must be used 
(section 2.7). 

    With mammalian-cell assays, the situation is slightly 
different.  The cell types described in section 2 are, in general, 
selected by tissue culture cloning techniques so that they meet the 
requirements of the genetic change being investigated.  For 
example, CHO cells, used for cytogenetic assays (section 2.4) are 
selected and cultured in a way that maintains the integrity of the 
chromosome complement.  Cells used in gene-mutation assays (section 
2.5) must be sensitive to a particular type of induced mutation 
(e.g., at the HGPRT locus), and cultures with a low spontaneous 
mutation frequency are maintained.  Metabolic capability

    Many carcinogenic/mutagenic chemicals are not able to interact 
with DNA until they have undergone some degree of enzyme-mediated 
biotransformation.  In animals, including man, foreign chemicals 
are subject to a series of modifying enzymic and non-enzymic 
reactions aimed at detoxifying the chemical and altering it to 
water-soluble forms suitable for elimination from the body.  These 
enzymic reactions are also capable of activating certain chemicals 
to reactive molecules that can interact with DNA to produce 
potentially harmful damage (section 1).  The appropriate enzyme 
systems are usually partially or completely inactive or absent in 
bacteria, yeasts, and cultured mammalian cell systems and are 
introduced in the form of an enzyme-rich, cell-free fraction of 
mammalian liver (section 2.1). 

    An acceptable  in vitro assay must, therefore, be shown to
have a metabolic capability appropriate to the chemical class
being studied and the experimental conditions must be designed
to allow the metabolic activation system to operate at an
optimum rate.  Guidance on these factors is provided in
section 2.  Although most mammalian cell types used in  in
 vitro tests retain some endogenous enzyme activity, it is
usually too low to activate the majority of carcinogens and
such tests are supplemented as described above.  Some
carcinogens, however, have been shown to be poorly metabolized
by the conventional rat liver (S9) microsomal enzyme system,
but, under appropriate experimental conditions, can be
activated by enzymes endogenous to the cultured cell.
Technical modifications necessary to detect this type of
chemical usually involve a longer than usual incubation period
that allows the compound to be available to endogenous enzymes
for 18 - 30 h (Dean, 1985).

4.2.3.  Predictive value

    The ultimate goal of the short-term tests described in this 
guide is to identify, with an useful degree of confidence, 
chemicals that may be hazardous.  As will be discussed later in 
this section, such a goal is approached through a series of stages, 
each of which leads to an assessment of the activity of the 
chemical at that stage.  Proving the safety of a compound is a much 
more difficult undertaking, rarely performed, and never based on 
short-term tests alone.  In general, given the limited resources 
available, it is usual to accept a substance as safe in practice in 
the absence of evidence to suggest otherwise.  Mutagenic activity

    The first stage in the evaluation of a chemical is to 
investigate the ability of the chemical to interact with DNA and 
produce a detectable change in the genetic material.  Bacterial, 
yeast, plant,  Drosophila,  and  in vitro mammalian cell assays are 
designed for this purpose.  They have a high value in predicting 
whether a chemical is a bacterial mutagen, is active in yeasts or 
plants, or can induce genetic damage in insect tissues or isolated 
animal cells.  It cannot be predicted with a high degree of 
certainty, from these assays alone, whether a chemical will produce 
mutations in a mammal such as man.  To provide an insight into the 
activity of the chemical in the whole animal,  in vivo procedures 
are used in which the chemical is given to test animals by an 
appropriate route and some means of detecting genetic changes is 
applied (e.g., chromosome study in bone-marrow cells, dominant 
lethal assay, or detection of mutagenic excretory products).  The 
predictive value of these  in vivo assays is fairly high when 
positive results are observed.  For example, if a chemical produces 
chromosome damage in rodent bone marrow it is usually assumed that 
the chemical could present a human hazard, under particular 
exposure conditions.  Negative findings in a properly conducted 
chromosome study in rodents are also frequently regarded as 
indicating a low or negligible human hazard, even with a chemical 

that induces chromosome changes in cultured cells (de Serres & 
Ashby, 1981; ICPEMC, 1983a).  However, certain chemicals, the 
effects of which are confined to specific tissues, such as the 
liver or gut, may not be detected using a bone-marrow assay.  For 
such chemicals, tissue-specific assays are being developed, e.g., 
unscheduled DNA synthesis in rodent liver (Mirsalis & Butterworth, 
1980), and an assay for nuclear anomalies in gut tissue (Heddle et 
al., 1982).  There are few practicable procedures for investigating 
gene mutations in animals and a negative bone-marrow study may have 
poor predictive value for chemicals that have been shown to induce 
only gene mutations in  in vitro studies.  These problems of 
interpretation are discussed further in section 4.5. 

    The hazards associated with exposure to genotoxic chemicals 
differ according to the cell type in which the genetic damage is 
induced.  Mutations in somatic cells are generally regarded as 
presenting a hazard (e.g., carcinogenic) only to the individual in 
which they occur.  Germ-cell mutations, however, may have far-
reaching effects in future generations and it is important to be 
able to predict whether a mutagenic chemical may present this 
hazard.  Unfortunately, the only practicable procedures currently 
available for studying germ-cell genetic changes directly are 
limited to chromosome damage.  It is usually assumed, in the 
absence of evidence to the contrary, that chemicals shown to induce 
chromosome damage in mammalian germ cells may be able to cause 
mutations in human germ cells.  Negative results in such assays 
with a chemical shown to be a clastogen in other tests may also be 
highly relevant (section  4.5).  Existing procedures, with the 
exception of large-scale experimental animal studies such as 
specific locus tests (section 1) cannot be extrapolated directly to 
the induction of gene mutations in human germ cells.  Carcinogenic activity

    As outlined in the Introduction (section 1), carcinogenesis 
induced by genotoxic agents is a multi-stage process that includes 
transport and metabolism of the chemical, interaction with the 
critical target molecule (e.g., DNA), DNA repair and replication of 
the lesion, and progressive development of the fixed lesion to form 
a malignant cell.  Long-term studies for carcinogenicity in 
experimental animals do not necessarily reflect a realistic 
situation as they only measure the ability of a test compound to 
function as a complete carcinogen.  In reality, a person may be 
exposed to a combination of agents acting on different stages of 
the carcinogenic process.  Current  in vitro tests cannot, of 
course, mimic all these stages and are frequently assumed to detect 
only the event leading to the initiation phase, i.e., the ability 
to induce a mutagenic or clastogenic DNA lesion.  The main value of 
short-term tests, therefore, lies in their ability to identify 
chemicals that may, under certain exposure conditions, either cause 
cancer by a predominantly genotoxic mechanism or induce the initial 
phase of the carcinogenic process.  Carcinogenesis enhancers 
(Clayson, 1981), including the so-called tumour promotors, will 
usually escape detection in the conventional DNA-based assays.  It 
is apparent, from the complexity of the carcinogenic process 

compared with the relative simplicity of short-term  in vitro  
assays, that, although such assays provide useful qualitative 
information, considerable caution is required in their 
interpretation in terms of human carcinogenicity.  The predictivity 
of the tests for detecting potential carcinogens is usually derived 
from their performance in validation studies in which a range of 
established carcinogenic and non-carcinogenic chemicals are tested 
under controlled conditions (the designation "carcinogen" or "non-
carcinogen" is obtained from long-term studies in laboratory 
animals or, more rarely, from human epidemiology data).  For 
example, the predictive value of the  Salmonella/microsomal 
bacterial assay (section 2.1) has been evaluated on a number of 
occasions, and the accuracy with which it differentiates between 
carcinogens and non-carcinogens varies between about 60 and 90% 
depending, among other factors, on the nature of the chemicals 
selected for the study (Rinkus & Legator, 1979).  Technical factors 
also contribute towards its accuracy and it is important when 
considering data from bacterial assays (and other tests) that the 
experimental protocol was appropriate to the chemical type being 
tested (section 2.3.3).  However, a properly-conducted bacterial-
mutation assay can give results of a useful predictive value, when 
considered together with results from other tests. 

    The induction of structural chromosome damage is also a 
property common to many carcinogenic chemicals (de Serres & Ashby, 
1981), and recent studies have shown that some carcinogens that do 
not induce mutations in bacterial systems are capable of causing 
chromosomal damage in cultured mammalian cells (Dean, 1985).  Thus, 
as will be discussed in section 4.5, a combination of a bacterial 
mutation test and a chromosome assay in cultured cells is 
considered, by many, to have a higher predictive value for 
carcinogenic activity than either test alone (e.g., Ashby et al., 

    Assays that indicate DNA damage, such as the yeast assay for 
mitotic gene conversion (section 2.3), SCE in animal cells (section 
2.4), and unscheduled DNA synthesis have also provided valuable 
predictive information on carcinogenic potential.  Positive results 
in these assays generally indicate that a chemical can interact 
with DNA in a eukaryotic cell, though they do not prove whether or 
not the lesion induced is capable of progression to a true somatic 
mutation or a carcinogenic initiation event.  Such assays, however, 
provide useful supplementary evidence in constructing an overall 
genotoxic profile of the possible adverse effects of a chemical. 

    Because of the physiological and genetic differences between 
bacterial and eukaryotic cells, it is inevitable that some 
chemicals will induce gene mutations in bacteria but not in 
eukaryotes (and occasionally, vice versa).  Assays for gene 
mutations in mammalian cells (section 2.5), yeasts (section 2.2), 
and recessive lethal mutations in  Drosophila (section 2.7) are 
extremely useful for detecting some classes of chemical 
carcinogens.  Relevance to chemical class

    The experimental protocols outlined in section 2 were designed 
to provide optimum experimental conditions for testing most types 
of chemicals.  Because of variation in chemical structure and 
reactivity, it must be accepted that such protocols, particularly 
for the  in vitro tests, are, in reality, compromises, and that, in 
many cases, the conditions are not necessarily the best for the 
particular chemical being studied.  The nature and rate of the 
enzymic reactions that transform a pro-carcinogenic chemical to its 
ultimate reactive form are dependant on the structure of the 
chemical.  The enzymes provided by the microsomal fraction usually 
incorporated into  in vitro tests, i.e., predominantly mixed-
function oxidases, are capable of activating most pro-carcinogens.  
Experimental conditions, including the source (e.g., species, 
tissue) and quantity of the microsomal fraction and the proportion 
of co-factors may need to be adjusted to provide near-optimum 
conditions for a particular chemical class.  For example, the 
standard  Salmonella/microsomal assay can detect most aromatic 
amines, polycyclic hydrocarbons, mono- and bi-functional alkylating 
agents and mycotoxins, but must be modified to respond to some 
nitrosamines, metallic salts, and many other compounds.  A few 
compounds, such as 1,2-dichloroethane, are activated by conjugation 
with glutathione, in which case an exogenous metabolic system may 
not provide optimum conditions for activation.  In addition, 
compounds activated by enzymes not active in the liver microsomal 
preparation, e.g., those provided by intestinal flora, will not be 
detected.  These factors apply to all assays that are enriched with 
exogenous metabolizing enzymes.  Other confounding factors may also 
influence the biotransformation of a pro-carcinogen.  For some 
chemicals, the residual endogenous enzymes in cultured mammalian 
cells are more active than the added microsomal enzyme mixture 
(section  Under appropriate conditions, the yeast 
 Saccharomyces cerevisiae contains stage-dependant mixed-function 
oxidase activity that may also be more suitable for the activation 
of some chemicals. 

    Although, as already mentioned, the standard protocols for the 
 in vitro assays represent compromises in experimental design, in 
practice, reliable results can be obtained with most chemicals 
using these protocols.  For the interpretation of results, however, 
an awareness of the possible influence of the structure of the 
chemical is an important factor in deciding the adequacy of an 
experimental protocol for that particular chemical. 

4.2.4.  Available expertise and facilities

    Most short-term test data are generated in government, 
academic, industrial, or contract laboratories by personnel having 
considerable experience with the techniques (i.e., section 3).  The 
selection of assays to meet a particular need may, in some 
respects, be based on the specific areas of expertise and 
facilities available in a specific institute.  For example, a 
laboratory with a long history of research on  Drosophila or fungal 
genetics may choose to select these organisms in preference to, for 

example, mammalian cells, for the study of gene mutation.  Indeed, 
in such situations more reliable data might be obtained from these 
organisms, at least intially, if experience in, or appropriate 
facilities for, mammalian cell culture techniques were lacking.  
Similarly, a life-long experience with a particular set of 
bacterial strains, e.g.,  Escherichia coli, may lead to their use 
in preference to  Salmonella typhimurium.  Although, in principle, 
some tests have distinct advantages over others for detecting the 
same type of genetic change, a solid background of experience and 
an adequately equipped laboratory is essential before attempting to 
generate data from a "new" (i.e., to the laboratory) assay for 
hazard assessment purposes. 

4.3.  Application of Assays

    Under ideal circumstances, short-term tests are applied in such 
a way that, beginning with an initial battery of two to four 
assays, tests are selected in order to accumulate data on the 
activity of a compound until a point is reached where an assessment 
of the probable genotoxic hazard can be made with an acceptable 
degree of confidence. 

4.3.1.  The phased approach

    Some 80 - 90% of chemicals shown to be carcinogenic in 
laboratory animals are capable of interacting with DNA and, under 
appropriate experimental conditions, the majority of assays for 
mutation will respond to most genotoxic chemical carcinogens.  Some 
assays of genetic damage respond better than others to various 
classes of chemical carcinogens and mutagens; some carcinogens give 
consistently negative results in standard assays for mutation-
induction, and other chemicals, shown to be non-carcinogenic in 
laboratory animals are fairly strong mutagens.  Since no single 
assay has proved capable of detecting animal carcinogens with an 
acceptable level of precision and reproducibility, it is usual 
practice to apply the assays in "packages" or "batteries".  For 
practical purposes, testing is usually divided into two or three 
phases or tiers (for review see Williams, 1980), though in many 
cases, data from the first phase of testing provides sufficient 
information for a provisional assessment of the genotoxicity of a 
compound.  The first phase, i.e., the basic screen, consists of a 
battery of two to four assays designed to detect genetic activity 
in the test material.  The second and third phases consist of 
supplementary assays selected to complement the phase 1 tests, to 
establish whether genetic damage is induced  in vivo and to provide 
a basis for making an assessment of possible human hazard 
associated with exposure to the material.  Phase 1 - the basic screen

    The initial battery (i.e., the basic screen or the "base set") 
consists of tests with an established broad data base generated 
from extensive validation studies.  One fairly comprehensive 
package consists of a bacterial mutation assay, an assay for 
chromosome changes, a test for DNA damage, and a eukaryotic gene-

mutation assay.  Final selection may be influenced by the nature of 
the material, e.g., drug, pesticide, industrial chemical, the 
extent of its eventual distribution and use (section 4.3.2), the 
objective of testing the material and, in some cases, the available 
technical expertise in the testing laboratory. 

    Although a package containing four assays is sometimes 
recommended for chemicals where extensive human exposure is 
anticipated (Draper & Griffin, 1980; DHSS, 1981), the assessment of 
a chemical is often begun with data from an initial battery of just 
two assays (OECD, 1982b, 1984).  These are usually a bacterial 
assay using a range of tester strains of  Salmonella typhimurium  
(section 2.1) and a test for the induction of structural chromosome 
aberrations.  The latter may be a micronucleus test in rodent bone-
marrow cells (section 2.8) or, more often, a chromosome assay in 
cultured mammalian cells (section 2.4).  Providing full 
consideration is paid to the physical and chemical properties and 
the metabolic behaviour of test chemicals, few potential mutagens 
or genotoxic carcinogens will escape detection in a combination of 
a  Salmonella/microsomal activation assay and an  in vitro  
mammalian cell chromosome assay (Ishidate & Odashima, 1977; 
Ishidate, 1981; Ashby et al., 1985).  It must be noted, however, 
that it is inherent in the concept of  in vitro screening that some 
potentially harmful molecules will slip through the screen and that 
some molecules active in the  in vitro system will prove to be 
inactive  in vivo. 

    A recent international collaborative study sponsored by the 
International Programme on Chemical Safety, was designed to 
identify the assay most suitable to be used in parallel with the 
 Salmonella/microsomal activation assay in a two-test battery for 
the detection of genotoxic chemicals.  Eight carcinogenic 
chemicals, chosen for the ambiguity of their results in bacterial 
mutation assays, together with 2 carefully-chosen non-carcinogens, 
were tested in a variety of  in vitro assays (Ashby et al., 1985).  
A number of assays performed extremely well in differentiating 
between carcinogens and non-carcinogens in the group of ten.  In a 
final analysis, however, an  in vitro mammalian cell culture assay 
for chromosomal aberrations was selected as the most suitable 
partner for bacterial tests on the basis of (a) performance in the 
collaborative study, (b) their advanced state of technical 
development and wide usage, and (c) a generally internationally 
accepted broad data base. 

    It will be evident from the previous paragraph that a reliable 
bacterial mutation test is widely regarded as a virtually 
indispensible component of the first phase of testing and that the 
second test will usually investigate chromosome changes.  In some 
cases, it may be appropriate and more convenient to study the 
effects of a chemical on chromosome structure  in vivo rather than 
in cultured cells, and assays such as the micronucleus tests 
(section 2.8) can be used in parallel with the bacterial assay.  
Though the published data base for  in vivo assays is not as 
extensive as that for cell-culture procedures, the micronucleus 
test (or a metaphase analysis of bone-marrow cells) has the 

advantage of using an intact animal and this may provide a sounder 
basis for hazard assessment.  In laboratories where cell culture 
facilities and laboratory animals are not readily available, it may 
be necessary to generate chromosome data from plant material.  
Although well-established techniques are available and a limited 
number of studies have demonstrated some correlation between plant 
chromosome changes and mammalian genotoxicity, additional 
validation of plant systems is essential, before they can be used 
to assess the potential effects of a chemical on man with any 
degree of confidence (section 2.6). 

    As described above, data from a two-test base set can provide 
reliable detection of most genotoxic chemicals.  It must be 
emphasized, however, that not all chemicals that provoke a positive 
response in one or both of the tests are necessarily hazardous for 
man (section 4.5.1).  In addition, negative results in these assays 
do not prove conclusively that the compound lacks genotoxic 
activity in the intact mammal, including man.  Supplementary tests

    Supplementary tests are conducted to complement, verify, or 
assist in the interpretation of the results of the initial battery.  
They may simply involve repetition of one of the initial assays 
under different experimental conditions or, in other cases, a 
completely different type of test. 

    The results of the basic screen provide information on the 
ability of the test chemical to induce genotoxic effects in a 
limited number of assays, e.g., mutation in bacteria and chromosome 
aberrations in eukaryotic cells.  In many instances, these may 
provide sufficient data and, indeed, may be the only available data 
on which to make a preliminary hazard assessment.  Where, however, 
a bacterial mutagen does not produce chromosome damage in the 
mammalian cell assay, it may be useful to know if the genetic 
activity is confined to bacterial cells, or if the chemical is also 
active in eukaryotic cells.  A variety of systems can be used to 
answer this question including gene conversion or mutation in 
yeasts (section 2.2), gene mutation (section 2.5), sister chromatid 
exchanges (section 2.4) or unscheduled DNA synthesis (section 2.3) 
in mammalian cells, or mutation in  Drosophila (section 2.7).  The 
next stage in the assessment may be to determine whether a chemical 
shown to be genotoxic in eukaryotic cells is also active in  a 
whole animal.   In vitro clastogens can be investigated using 
chromosome studies on the bone-marrow cells of rodents after dosing 
with the suspect chemical.  The  in vivo activity of bacterial 
mutagens can be further evaluated in studies of mutagenic products 
in urine or body fluids from animals treated with the compound 
(section 2.1) (Combes et al., 1984) or in the mouse coat Spot Test 
(Fahrig, 1977; Russell, 1978).  Techniques are also available for 
investigating DNA repair (Waters et al., 1984) and sister chromatid 
exchanges (Perry & Thomson, 1984) in treated animals.  In some 
cases, it may be appropriate to study the effect of a compound on 
mammalian germ cells using either the dominant lethal assay 
(section 2.8) or chromosome analysis of rodent germ cells (Adler, 
1982; Brewen & Preston, 1982; Albanese et al., 1984). 

    In summary, an assessment of the possible genotoxic hazard is 
generally carried out after each phase of testing, attempting to 
answer the questions, a) is the compound mutagenic? (Phase 1); is 
it active in mammalian cells? (Phases 1 and 2); is it active  in 
 vivo? (Phases 2 and 3); does it present a hazard for man? (may be 
considered after Phase 1, 2 or 3, depending on the nature of the 
chemical) (sections 4.3.2, 4.5.2). 

4.3.2.  Nature and extent of potential human exposure

    The selection of assays and, in particular, the extent of 
testing required before assessing the potential hazard, depends on 
the nature and eventual use of the chemical or product.  Limited or negligible distribution

    There are chemicals for which environmental distribution is 
severely limited and the chance of human exposure is unlikely, or 
limited to small groups of people, the levels of exposure being 
very low.  In these cases, data from a base set of two assays are 
often the only data available to those who have to determine how 
such chemicals should be handled.  For example, manufacturing 
intermediates are usually handled by trained personnel using 
established safe handling procedures and information on the 
genotoxicity of such chemicals is of value in the design of safe 
manufacturing processes.  Specialized chemicals, that are usually 
produced in small volumes and supplied for specific industrial or 
research applications, may be tested in order to assess the safety 
requirements in their transport or use.  For materials of this 
type, data from both base set assays are normally provided.  Only 
in rare cases is information from a single assay, e.g., a bacterial 
mutation test, considered adequate, as for example, when screening 
candidate pharmaceuticals, or dyestuffs, food additives, etc.  The 
results from simple bacterial assays may then be sufficient to 
identify mutagenic structures in a series of analogues and thus set 
priorities for further testing or further product development.  Medium distribution, limited exposure potential

    Chemicals in this group are those to which some degree of human 
exposure may be possible, but where environmental distribution is 
restricted and only specific groups of individuals may be 
inadvertantly exposed.  Examples include solvents, paints, 
adhesives, oil products, some pesticides, and other materials that 
will generally be used in an industrial or commercial environment 
and to which the general population is unlikely to be exposed.  
They are chemicals that may be fairly widely used in an environment 
where potential exposure can be controlled, but do not include 
materials for domestic use, or those that are present in products 
available to the general population or are released into the 

    For assessing the genotoxic hazard of such chemicals, results 
from two-base set tests together with information on their 
structural relationship with known carcinogenic and non-

carcinogenic chemicals and other basic toxicity data are normally 
available.  Such an assessement may require data from supplementary 
assays, for example, to confirm negative results provided by the 
base set or to determine if the genetic damage identified in the  in 
 vitro assays can be detected in animals.  However, this should not 
be regarded as a complete evaluation of genetic toxicity.  Extensive distribution, intentional or unavoidable exposure

    This group contains chemicals, materials, and products that 
may be widely distributed in the environment and to which human 
beings will almost certainly be exposed.  Examples include:  
pharmaceutical products, both those used for very specialised 
treatments and those used by a relatively high proportion of the 
population; chemicals that are an integral part of foodstuffs or 
may appear as residues or contaminants in food; domestic and 
agricultural pesticides; domestic chemicals of all kinds; 
environmental contaminants including naturally occurring chemicals 
in plants, soil etc.; combustion products; industrial effluents; 
and many more. 

    Because human exposure to these materials is generally to be 
expected, the objective of short-term tests (and, indeed, of all 
toxicity testing) is to ensure, as far as possible, that exposure 
does not present a potential or actual hazard.  It is important to 
keep in mind that the eventual assessment of genotoxicity is aimed 
at deciding if the chemical presents either a carcinogenic hazard 
or an adverse effect on germ cells with the possibility of 
producing heritable genetic damage. 

    Initially, results of the base set are assessed, but data from 
these assays are rarely sufficient for hazard assessment for 
chemicals in this group (section 4.5.1).  It is usual to conduct 
supplementary assays to investigate other genotoxic effects such as 
the induction of gene mutations or unscheduled DNA synthesis in 
mammalian cells, and, when this genetic profile has been completed 
(section, to assess the activity of the material  in vivo.  
At this stage, and with the help of data on the absorption, 
distribution, metabolism, and excretion of the chemical and other 
toxicological data, it is possible to conduct a reasonable 
assessment of the potential of a substance for mutagenicity and 
genotoxic carcinogenicity.  The final assessment of carcinogenic 
potential, however, usually requires the provision of data from 
long-term cancer studies in laboratory animals. 

4.3.3.  Regulatory requirements

    Many countries require the submission of testing data before 
approving the marketing of certain types of products.  However, the 
requirements differ considerably between countries, and various 
national and international bodies have attempted to harmonise 
mutagenicity testing requirements by preparing guidelines.  
Organizations such as the Organisation for Economic Cooperation and 
Development (OECD) and the European Economic Community (EEC) have 
published regulatory requirements or guidelines for mutagenicity 

testing, though many countries continue to have their own 
individual requirements.  The specific assays required by 
individual countries are often unclear and usually depend on the 
nature of the chemical or product and the outcome of discussions 
between the marketing company and the competent authority of the 
country (ICPEMC, 1983b). 

    Two bodies whose authority extends beyond national boundaries 
are the OECD (representing some 24 countries) and the EEC.  Their 
guidance and regulations, respectively, on mutagenicity testing are 
very similar and, in practice, apply to the marketing of new 
substances rather than existing chemicals.  Both authorities relate 
the extent of testing to the perceived degree of exposure and 
distribution.  For example, testing is only required on chemicals 
that will be produced or imported in quantities of one tonne or 
more per annum.  Many types of product are excluded by these 
authorities; medicinal and food products are often regulated by 
individual countries rather than international bodies. 

    The OECD and EEC require a base-set of two tests for 
mutagenicity, i.e., assays for bacterial mutation and for 
chromosome damage, on all products that are not excluded from their 
authority, with a requirement for supplementary assays when 
exposure of relatively large numbers of people is unavoidable or 

4.4.  Acceptability and Reliability of Data

    It is essential that data used to assess the genotoxicity of a 
chemical are derived from studies designed to meet predefined 
minimum technical and scientific criteria.  One of the purposes of 
the descriptions of assays contained in section 2 of this guide is 
to define these criteria.  The reliability of data, therefore, can 
be confirmed by ensuring that the experimental protocol used to 
produce the data conforms to the requirements for an acceptable 
assay.  In addition, evidence must be provided to show that the 
investigator faithfully adhered to the protocol by accurate and 
full recording of all experimental procedures, raw data, 
mathematical calculations, etc., so that every step of each assay 
can be audited by an independent observer.  Such monitoring 
procedures are described in detail in section 3.3. 

4.5.  Interpretation of Results and Significance for Human Hazard 

    Results of short-term tests are assessed with two distinct 
types of hazard in mind:  the carcinogenic activity of the chemical 
and the possibility that the chemical may affect human germ cells 
to produce heritable genetic changes. 

4.5.1.  General principles

    Much of the data used in the assessment are generated from 
relatively simple tests, often consisting of cultures of single 
cells, and it must be emphasized at the outset that the behaviour 

of a particular chemical may be dramatically different in a complex 
organism such as man.  In a simple bacterial mutation assay, the 
chemical may be metabolically transformed by an auxillary 
microsomal enzyme system and the reactive molecule thus generated 
has simply to penetrate the bacterial cell wall to be readily 
available to interact with DNA to produce genetic changes.  In the 
whole animal, however, the same chemical must be absorbed into the 
body across a number of chemical and physical barriers, and must be 
transported to the site where the appropriate metabolizing enzymes 
are situated, where it may be activated or detoxified, before it is 
in a form that can interact with DNA (some metabolism may also 
occur in the gut).  Even then, the DNA lesion is subject to 
protective devices, such as DNA repair, before genetic changes are 
expressed, and these protective factors differ between the 
bacterial system and animal cells.  Thus, in animals, a range of 
physiological and biochemical factors that are different from those 
in simple assays may influence the ultimate fate of the chemical, 
either inhibiting or enhancing its potential toxicity. 

    The structure of the chemical and its possible fate in animals 
are important factors in the interpretive process.  Data from 
studies on absorption, distribution, metabolism, and excretion are 
generally only available for chemicals for which the possibility of 
human exposure is relatively high, e.g., drugs, foods, and many 
pesticides, and, even then, detailed metabolic data may not be 
available.  For other materials, the possible  in vivo fate may be 
inferred from the chemical structure by analogy with related 
chemicals for which more information is available. 

    A further factor in the interpretation of short-term tests lies 
in the correlation between positive and negative results in a 
particular assay and known carcinogenic and non-carcinogenic 
activity.  This correlation is obtained from validation studies in 
which the activity of known animal carcinogens and non-carcinogens 
is established in the short-term test (Purchase et al., 1978).  The 
assays commonly used in the initial test batteries, e.g., bacterial 
mutation tests and  in vitro chromosome assays, are selected 
because they have performed well in validation studies and 
currently have a good predictive value for animal carcinogenicity 
with many classes of chemicals (de Serres & Ashby, 1981; Ashby et 
al., 1985). 

    Thus, the important factors to be considered when interpreting 
the findings of short-term tests are:  (a) the predictive value of 
the assays as demonstrated by their correlation with known 
carcinogens and non-carcinogens; (b) the structure of the chemical 
in relation to chemicals of known genotoxicity; (c) the known or 
probable metabolic route of the chemical in the whole animal; and 
(d) data from other toxicity studies. 

    In section, the significance of the results of 
individual assays is discussed, but it must be emphasised that 
assessment of human hazard should be based on combinations of 
assays rather than on data from single tests.  The interpretation 

of data from batteries of short-term assays is described in section, together with the application of the phased approach to 
assessing genotoxic hazard using supplementary tests.  Results of individual assays

    For many years, assays using a range of tester strains of 
bacteria have been the cornerstone of short-term tests for 
genotoxicity.  Positive results indicate, primarily, that the 
chemical or one of its metabolites is capable of interacting with 
DNA to produce mutations.  In spite of the fact that many genotoxic 
carcinogens produce mutations in bacteria, not all bacterial 
mutagens are animal carcinogens and the interpretation of bacterial 
assays in isolation, in terms of human hazard is not an acceptable 
procedure.  Data from at least two base set assays are usually 
available before even a preliminary extrapolation is attempted, 
for example, when establishing safe working practices in a 
manufacturing plant. 

    There is increasing evidence, particularly from a recent 
collaborative study sponsored by the International Programme on 
Chemical Safety (Ashby et al., 1985), that some carcinogens that 
are negative or difficult to detect in bacteria induce genetic 
changes in eukaryotic systems, such as cultured mammalian cells, 
yeasts, or  Drosophila, in the form of structural chromosomal 
aberrations or gene mutations.  Again, these assays are not usually 
interpreted in isolation but only as part of an expanding data 

    Studies on whole animals are usually considered to be more 
relevant to man than  in vitro assays and, as a general rule, a 
chemical that gives clear, unequivocal positive results in an  in 
 vivo assay, such as chromosome damage in rodent bone-marrow cells 
or the dominant lethal assay, is usually regarded as a possible 
human mutagen or carcinogen.  Results from combinations of assays

    It is usual practice to begin the assessment of the 
genotoxicity of a chemical on the results of an initial battery of 
at least two assays.  One of these is almost invariably a bacterial 
mutation assay and, in a two-test battery, the second is usually an 
 in vitro or  in vivo chromosome assay.  Data from other tests, 
e.g., yeasts, UDS, etc., may also be available, and, where other 
toxicity or pharmacokinetic studies have been conducted, the data 
base is considered as a whole during the assessment of possible 

    The following sequence of assessment procedures is based, as an 
example, on data available, initially, from a bacterial mutation 
assay and a chromosome assay in cultured mammalian cells.  For the 
purpose of this exercise, it is assumed that the data have been 
generated from reliable and acceptable protocols (section 4.4). 

    A.  Chemicals clearly positive in both assays

    Such findings demonstrate unequivocally that the chemical or a 
metabolic derivative is capable of interacting with DNA to produce 
genetic damage in both eukaryotic and prokaryotic cells.  It is 
thus classified as a genotoxic chemical and, unless, and until, 
data from other toxicity studies or supplementary short-term tests 
show that it is unlikely to be active  in vivo, it is prudent to 
regard it as potentially hazardous for man. 

    In some instances, chromosome data may be presented from plant 
systems rather than mammalian cells, and the same principles apply.  
However, it may be appropriate to confirm the plant data in a 
mammalian-cell assay or other eukaryote (e.g.,  Drosophila or 
yeast) at an early stage. 

    It may be prudent to designate a chemical as potentially 
hazardous on the basis of these assays and this may indicate that 
distribution and human exposure will be restricted.  The potential 
value of the chemical may be such that further testing designed 
either to confirm this assessment or to determine the extent of the 
potential hazard may be worthwhile.  By no means will all such 
chemicals be shown to be hazardous in subsequent testing.  
Additional testing may be aimed at elucidating:  (a) the nature of 
the genetic change induced by the chemical in mammalian cells; (b) 
the dependance of the chemical on metabolic enzyme activation for 
its genetic activity; and (c) the behaviour of the chemical and the 
genetic damage it induces in the intact mammal. 

    (i)   Positive results only after metabolic activation

    Such results indicate that reactive metabolites are generated 
by microsomal enzymes, i.e., the compound is an "activation-
dependent" mutagen.  As the chemical has been proved to be a 
clastogen  in vitro, the next step may be to carry out a chromosome 
study in bone-marrow cells in rodents after dosing with the 
chemical by an appropriate route (e.g., oral, intraperitoneal).  
Either the micronucleus test or analysis of metaphase chromosomes 
can be used.  If the chemical is shown to produce chromosome damage 
 in vivo, there is little to be gained by any further testing and 
it is usually regarded as having mutagenic or carcinogenic 
potential for man.  In rare cases, for example, mutagenic anti-
tumour agents, the benefits of using the drug may outweigh this 
potential hazard and it may be useful to conduct a dominant lethal 
assay or a cytogenetic analysis of germ cells in rodents to assess 
the induction of heritable genetic changes. 

    Negative results in a properly conducted  in vivo study may 
alleviate most concern regarding the potential hazard of a chemical 
and, with many chemicals, such negative results suggest that the 
adverse chromosome effects shown  in vitro are unlikely to occur in 
the intact animal.  (It should be remembered that, occasionally, a 
negative result may be obtained in a bone-marrow chromosome study 
because the compound or its reactive metabolite(s) did not reach 
the target cell in the bone marrow.)  However, the chemical is 

still a mutagen and the decision to release it into the general 
environment will usually be measured very cautiously against its 
possible benefits; further testing may be judicious.  Since sister 
chromatid exchange or unscheduled DNA synthesis (UDS) are 
mechanistically unrelated to chromosome breakage, it may prove 
useful to establish the activity of the chemical in these  in vitro  
assays.  If either yields a positive result,  in vivo activity can 
be investigated by conducting assays for SCE in bone-marrow cells 
or UDS in hepatocyte cultures from treated rodents. 

    A detailed pharmacokinetic profile of the chemical may be 
available and could provide evidence on the generation of reactive 
metabolites  in vivo.  Such evidence may support the assumption that 
the chemical is or is not putatively genotoxic for mammals. 

    (ii)   Positive results in the absence of metabolic activation

    Chemicals that produce mutations in bacteria without the need 
for exogenous metabolic activation are either "direct-acting 
mutagens" or, in rare cases, are activated by bacterial enzymes.  
Most eukaryotes are capable of some degree of endogenous activation 
(i.e., without the use of an auxillary metabolizing system) and 
direct-acting chemicals are usually classified as such on the 
results of bacterial assays.  If there is an indication from the  in 
 vitro assays that the incorporation of a metabolic activation 
system eliminates or significantly reduces the mutagenic activity, 
then it is possible that the microsomal enzymes serve to detoxify 
the chemical.  Some confirmation of this can be obtained from an  in 
 vivo chromosome study, from the results of one of the other  in 
 vivo tests for mutagenic activity (A(i)), or from the results of a 
study of the metabolism of the chemical  in vivo.  Negative results 
from a properly conducted  in vivo investigation of a direct-acting 
mutagen usually indicate that it is unlikely to pose a serious 
carcinogenic hazard. 

    B.  Chemicals that produce gene mutations in bacteria but not 
        chromosome aberrations in mammalian cells

    A chemical that produces mutations in bacteria with negative 
results in the eukaryotic-cell test is classified as a bacterial 
mutagen.  The question then arises as to whether the mutagenic 
activity is specific to bacterial cells.  This may be investigated 
by applying one or more of the other tests described in section 2.  
Where mutagenic activity is established in a eukaryotic cell 
system, interpretation of the data and the need for additional 
tests follows the procedure outlined in A(i). 

    Occasionally, chemicals are encountered that produce mutations 
in bacteria, but are clearly negative in other  in vitro tests.  
The assessment of such findings presents a number of difficulties.  
Interpretation may be helped by conducting an  in vivo cytogenetic 
assay (in spite of the fact that the  in vitro chromosome assay was 
negative, the absence of chromosome aberrations in a bone marrow 
study is valuable confirmatory evidence), by testing urine from 

treated animals for mutagenic activity and by consideration of the 
pharmacokinetics, the relationship of the structure of the chemical 
to known genotoxins, and other toxicity data.  However, the 
observation of bacterial mutagenicity may be the only evidence that 
the compound is genotoxic and a great deal of effort can be 
expended in trying to elucidate its significance to human hazard.  
Where other  in vitro and  in vivo tests fail to reveal mutagenic 
activity, and where there is no evidence from pharmacokinetic and 
conventional toxicity studies to suspect possible adverse effects, 
then the finding of bacterial mutagenic activity in isolation, 
particularly at high test concentrations, may not constrain the use 
and distribution of most materials.  Certain drugs, food chemicals, 
and ubiquitous materials have been exempt from this view and their 
use restricted pending long-term cancer studies in laboratory 

    C.  Chemicals that produce chromosome aberrations in mammalian 
        cells but not mutations in bacteria

    This pattern of results raises three important questions: 

    (a) has the chemical been tested in bacteria under an
        appropriate range of experimental conditions, e.g.,
        using a preincubation technique, variable levels of
        metabolic activation (i.e., S9), and using an
        adequate range of tester strains of bacteria;

    (b) can the chemical induce mutations or UDS in cultured
        cells; and

    (c) can the  in vitro chromosome damage be reproduced  in

    Some chemical carcinogens (e.g., hexamethylphosphoramide) give 
negative results in bacteria but are clearly mutagenic when tested 
in mammalian cells (Ashby et al., 1985).  It is often useful, 
therefore, to check the mutagenicity in a mammalian-cell system and 
if shown to be active, it can be assessed as described in A(i).  
For small-volume chemicals (section, the information that 
a chemical induces chromosome aberrations may be all that is 
available to formulate guidance on its use and distribution.  Where 
it is important to determine its activity  in vivo, a bone-marrow 
cytogenetic assay using micronuclei counts or metaphase analysis is 
the next logical step.  Positive results in such a test confirm 
genotoxic activity  in vivo and, in most cases, are interpreted as 
suggesting a possible hazard for man.  Where the induction of 
chromosome damage in cultured cells is the only indication of 
genotoxicity, i.e. where other short-term tests in cultured 
mammalian cells are negative, where there is no evidence of 
chromosomal damage in animal studies and other toxicity studies do 
not show any adverse effects, the chemical is unlikely to pose a 
genotoxic hazard for man.  As indicated in section B, certain drugs 
and food chemicals may be exempt from this view and long-term 
animal studies are desirable before the material is released for 
human use. 

    D. Negative results in both assays

    In assessing the significance of negative results in the basic 
screen, it is essential to determine whether the results are a true 
indication of lack of genotoxic activity by confirming that the 
protocols used were appropriate for the type of material being 
tested.  For example, if the chemical is volatile, assays must be 
conducted in sealed vessels to prevent erroneous negative results 
caused by evaporation of the test chemical.  Some chemicals, e.g., 
nitrosamines, require special experimental conditions to detect 
mutagenic activity.  The physical and chemical properties of the 
test agents and the influence of protocol variables on the 
performance of the assays are taken into consideration when 
evaluating the significance of negative results in the basic 
screen.  For many chemicals of limited distribution and exposure, 
the provision of reliable evidence of the absence of mutagenic 
activity in the two initial assays is frequently considered to be 
sufficient grounds for regarding the chemical as non-genotoxic.  
Because of the existence of a small class of genotoxic agents that 
are not detected in the two initial assays, it must be accepted 
that a decision to permit widespread use and unlimited distribution 
on the basis of negative results in these tests carries a 
significant risk, and further testing in other assays may be of 

    E.  Non-genotoxic carcinogens

    The majority of carcinogenic chemicals have demonstrable 
genotoxic activity.  There are certain classes of chemicals, 
including, for example, some metals, organochlorine compounds, and 
estrogens, that are known to be carcinogenic in animals but fail to 
elicit a positive response in assays for genotoxicity.  There are 
other compounds that are not in themselves complete carcinogens, 
which are able to exacerbate certain stages of the carcinogenic 
process (ICPEMC, 1982).  Collectively, these latter compounds are 
referred to as carcinogen enhancers (Clayson, 1981).  At present, 
there is no short-term test that has been sufficiently well 
validated to be used with confidence to detect non-genotoxic 
carcinogens and enhancers.  Evidence is emerging that some of these 
chemicals can induce numerical chromosome changes in eukaryotic 
cells and, in some cases, structural chromosomal aberrations (Ashby 
et al., 1985) and that they can be detected in modified forms of 
certain assays for neoplastic transformation (Meyer, 1983).  
However, these findings require confirmation in further validation 
studies and it must be accepted that a proportion of this class of 
chemical will escape detection in current toxicological practice. 

    F.  Complex mixtures

    The principles outlined in this guide apply to chemicals that 
are pure compounds or relatively simple mixtures, formulations, or 
solutions.  The application of  in vitro assays to more complex 
mixtures, e.g., foods, crude industrial products, etc., may give 
results that are unreliable because of the influence of such 
factors as, for example, competition between components in the 

mixture for enzyme sites in the activation system, presence in the 
mixture of cytotoxic components that limit adequate testing, and 
uncertainties regarding the concentrations of mutagenic components.  
 In vitro assays, therefore, must be used and interpreted with 
caution and, where it is not possible to isolate and identify 
mutagenic components, the emphasis should be on  in vivo testing 
(although it must be realized that problems similar to those 
mentioned above may beset whole-animal assays). 

4.5.2.  Influence of the extent of exposure and distribution

    The amount of toxicity testing that a material undergoes before 
it can be released for specific or general applications or into the 
environment is decided, to a large extent, by its perceived 
distribution and the expected pattern of human exposure.  For each 
of the broad groups of materials considered here, evaluation of 
genotoxic activity is based on the phased approach (section 4.3.1) 
using as a starting point, data from a two-test base set.  It must 
be remembered that mutagenicity tests provide only part of an 
overall package of toxicity data that should be available before 
making a final assessment of human hazard. 

    This guide is concerned only with hazard assessment and it is 
not within the objectives of the guide to describe methods for 
estimating, in quantitative terms, the risk of adverse effects in 
man following exposure to specific genotoxic chemicals.  However, 
in order to avoid confusion between the terms "hazard" and "risk", 
brief definitions of risk estimation and risk management are 

    (a)   Hazard assessment

    The assessment of the possible hazard associated with exposure 
to a genotoxic chemical is a purely scientific process and involves 
an appraisal of experimental data in order to attempt to predict 
the possibility and the nature of any adverse effect in man. 

    (b)   Risk estimation

    This is the second stage in the evaluation of a product or 
chemical and is an attempt to derive a quantitative estimate of the 
risk resulting from the use or release of the material, i.e., the 
number or frequency of individuals in a population of a given size 
who may exhibit a given adverse effect (e.g., cancer or heritable 
mutations) under certain exposure conditions.  Risk estimation is 
almost always an uncertain undertaking.  Quantitative data derived 
from screening tests cannot be used as a basis for predicting the 
potency of carcinogenic activity in animals or man (For additional 
details see Bridges et al, 1979; Ehrenberg, 1979; Brusick, 1980; 
Sankaranarayanan, 1982; ICPEMC, 1983d). 

    (c)   Risk management

    While hazard assessment and risk estimation are scientific 
processes, risk management is a non-scientific decision-making 
procedure (US NAS, 1983).  Risk management attempts to balance the 

perceived benefit of using or distributing the chemical or product 
with the risk of adverse effects to individuals or to populations, 
i.e., a risk-benefit equation.  Where the benefits are regarded as 
great, for example with a new, unique and valuable drug or 
pesticide, an identifiable and measurable risk may be considered 
acceptable.  In situations where less toxic alternatives with 
similar benefits are available, the risk associated with the 
introduction of the new material would normally be unacceptable.  Pharmaceutical compounds

    With few exceptions, pharmaceutical compounds that show 
unequivocal mutagenic activity are identified and discarded by the 
drug company at an early stage of development.  Most major 
companies involved in the development and manufacturing of drugs 
use a three-stage development regime that includes some toxicity 
testing at each stage.  The first stage is an in-house toxicity 
screen that is used primarily for the selection of candidate 
compounds, i.e., those that show promising pharmacological 
activity, for further development.  Only very limited mutagenicity 
testing is likely to be conducted at this time, and, in most cases, 
will consist of a bacterial mutation test.  Compounds that show 
mutagenic activity will frequently be discarded at this early 
stage.  There are exceptions to this rule, for example, where a 
compound or a group of compounds with unique pharmacological 
activity may be considered of great potential benefit.  Promising 
candidates eventually reach a stage of development where it is 
necessary to test their efficacy in human beings.  Before these 
clinical trials, the compounds undergo a second phase of testing, 
i.e., the pre-clinical trial toxicity screen, the objective of 
which is to assess the safety of the drug for use in small groups 
of human volunteers.  Assuming that a bacterial assay was conducted 
in the primary toxicity screen, mutagenicity testing usually 
consists, initially, of either a test for the induction of 
chromosome aberrations in cultured cells or an  in vivo assay in 
rodent bone-marrow cells for micronuclei or metaphase chromosome 

    The interpretation of the findings from the initial 
mutagenicity assays is greatly influenced by the pharmacological 
and pharmacokinetic data generated during the development of the 
drug.  If clear negative results are obtained with a compound that 
also shows no indication of interaction with macromolecules such as 
DNA, then it may be regarded as safe enough, from the point of view 
of genotoxicity, to proceed to clinical trials.  Where there is any 
doubt about the pharmacokinetics of the drug or its metabolites (as 
is often the case at this stage of development), i.e., where the 
possibility of DNA interaction cannot be excluded, then additional 
testing is indicated.  The supplementary tests are aimed at filling 
in the gaps in the genotoxicity profile.  For example, if an 
analysis of metaphase chromosomes in bone-marrow cells from treated 
rodents was not part of the initial testing, then this is usually 
also carried out.  Although the chemical failed to induce mutations 
in bacteria, it may be tested in a mammalian cell assay for gene 
mutations, or, perhaps, for recessive lethal mutations in 

 Drosophila or gene mutations in yeasts.  Other tests that indicate 
the induction of DNA damage may also provide useful data.  The 
pharmacological data available may also indicate the testing of 
urine and other body fluids from treated animals for mutagenic 
activity, using bacterial assays.  Where the structure of the 
chemical indicates that nitrosation products may be formed in the 
human stomach, tests for the formation of such products and their 
mutagenicity may be conducted (Kirkland et al., 1984). 

    Thus, before a pharmaceutical chemical undergoes clinical 
trials in human volunteers, it is normally shown at least to be 
incapable of inducing mutations in bacteria and chromosome damage 
in mammalian cells.  Where the structure or the pharmacokinetics of 
the chemical suggest that genotoxic interactions are conceivable, 
additional testing will usually include analysis of chromosome 
aberrations  in vivo, eukaryotic assays for primary DNA damage and 
gene mutations and, where indicated, for chromosome changes in germ 
cells (e.g., a dominant lethal assay) and for mutagenic metabolites 
in urine or body fluids of treated animals.  Assuming negative 
findings in these assays and after considering the data from other 
toxicity studies, the drug may then undergo clinical trials. 

    Except in the case of, for example, cytostatic or cytotoxic 
drugs used in the treatment of serious, life-threatening diseases, 
compounds that are shown to be genotoxic will rarely undergo 
efficacy studies in man.  Where there is sound evidence that a 
drug, shown to be mutagenic  in vitro, is rapidly detoxified in 
intact animal studies, limited clinical trials may occasionally be 

    The first two stages of toxicity evaluation are conducted as 
part of the development programme.  After successful clinical 
trials, a final toxicological evaluation is undertaken before the 
drug is submitted for registration prior to marketing.  
Registration is usually a responsibility of a government department 
in the country in which a marketing permit is sought.  The 
genotoxicity data required for registration vary considerably 
between individual countries though, in most cases, the assays 
conducted before clinical studies in human beings comprise a 
package that is acceptable for registration purposes.  Some 
authorities require data from a very specific series of assays, 
but, in general, a package that includes properly conducted assays 
for mutations in bacteria, chromosome aberrations and gene 
mutations in mammalian cells, and an  in vivo test for chromosome 
aberrations (in somatic and/or germ cells) should satisfy most 
authorities of the absence of a potential mutagenic hazard 
providing that there is no contradictory evidence from other 
toxicity studies. 

    This package of assays also provides some indication of the 
carcinogenic hazard.  However, for registration and marketing 
purposes with a new drug, carcinogenicity is almost invariably 
assessed from animal studies rather than predicted from  in vitro  
assays.  Chemical compounds in food

    Although this section is primarily concerned with chemicals 
that are added to natural food products to improve their keeping 
properties, palatability or appearance, etc., it is pertinent to 
summarize some other factors that contribute towards the mutagenic 
activity of food (For review, see Knudson, 1982).  Some edible 
plants and their fruits contain compounds, e.g., pyrrolizidine 
alkaloids, flavonoids, etc., that are mutagenic in  in vitro  
assays.  A small number are also carcinogenic, but the majority 
have not yet been tested for carcinogenicity.  They are often 
present in only minute quantities in the plant material and are 
often destroyed when the plants are cooked or completely detoxified 
by the gut flora.  The contribution of mutagenic food components to 
human cancer is not known.  Another source of potential 
genotoxicity is the fungal contamination of foods, e.g., aflatoxins 
in mouldy groundnuts.  Food may also contain residues of 
pesticides, compounds absorbed from packaging materials, and other 
chemicals.  An additional contribution to genotoxic activity can 
occur during cooking, and it has been demonstrated that pyrolysis 
products, formed during the cooking of meat and fish at certain 
temperatures, have significant mutagenic activity.  Although 
dietary factors are known to contribute towards the overall 
incidence of cancer in man, the part played by naturally occurring 
mutagens and pyrolysis products in human disease has yet to be 

    Most foods are complex mixtures of many hundreds of compounds 
and evaluating their genotoxicity is far more difficult than 
evaluating that of pure chemicals.  Because of this, attempts to 
investigate the genotoxicity of whole foods are usually undertaken 
using intact animals, including  Drosophila.  However, the 
fractionation of foods for mutagenicity testing purposes is 
currently being explored (Rowland et al., 1984). 

    Artificial food additives include chemicals that either enhance 
the natural flavour of foods, improve colour or appearance, or are 
preservatives added to prevent bacterial spoiling or oxidative 
degradation of food.  Artificial flavouring materials are usually 
identical in chemical structure to naturally occurring flavourings 
and are either synthesized or purified extracts from natural 
sources.  A large number of natural and synthetic dyes have been 
used to improve the appearance of food.  Many synthetic dyes have 
been removed from national and international lists of permitted 
food colourants because of their mutagenic or carcinogenic 
activity.  Compounds commonly used to preserve foods include sodium 
nitrite, a weak mutagen in  in vitro tests, and antioxidants such 
as butylated hydroxytoluene (BHT).  Although the mutagenicity of 
nitrite itself is unlikely to present a human hazard, it is able to 
react with secondary amines in conditions found in the human 
stomach to form carcinogenic nitrosamines. 

    The application of short-term tests for genotoxicity to food 
additives follows the principles outlined earlier.  Chemicals 
proposed for use in foods are usually tested initially in a base 

set of two assays and those that, for example, induce mutations in 
bacteria or chromosome aberrations in mammalian cells are very 
carefully evaluated before being used as either flavouring or 
colouring materials.  Chemicals that give negative results in these 
assays usually undergo a second phase of tests in eukaryotic cells 
for the induction of gene mutations, and, possibly, for the 
induction of DNA damage, and for the induction of chromosome damage 
in rodent bone marrow.  Completely negative results in these assays 
frequently allay concern regarding mutagenic potential with most 
chemicals.  However, additives that are structurally related to 
known mutagens or carcinogens, and, in particular, chemicals 
containing a secondary amine structure may be candidates for 
additional  in vivo testing, e.g., germ cell chromosome studies, 
dominant lethal assays, body fluid mutation tests, etc.  Negative 
results also suggest that the chemical is unlikely to be 
carcinogenic, but few new food additives are currently released for 
general use without evidence of the absence of carcinogenic 
activity in long-term animal studies.  Domestic chemical compounds

    Cosmetics such as perfumes, hair dyes, sun screen oils, etc., 
household detergents and cleaning fluids, and a variety of other 
chemical mixtures are considered under the general heading of 
domestic chemical compounds.  Because of their diverse nature, 
there have been wide differences in the amount of toxicological 
information available on these materials and the following examples 
illustrate the need for caution when considering their safety in 
domestic use.  Several hair dyes of the substituted 
phenylenediamine type have been shown to be mutagenic in  in vitro  
assays, and some of these have produced cancers in experimental 
animals.  Tris(2,3-dibromopropyl)phosphate is not strictly a 
domestic chemical but enters the home in the form of a flame 
retardent in clothing.  Widely used to reduce the flammability of 
children's clothing in particular, the compound was detected as a 
bacterial mutagen initially, and was eventually shown to be 
carcinogenic in long-term rodent studies.  Fortunately, these are 
relatively infrequent instances and would have been detected in the 
base set of two assays now widely used to assess the genotoxicity 
of new products. 

    Because these materials are sold for use in an environment 
where human exposure is to be expected or intended, a complete 
genotoxicity assessment is usual, and may begin with data from  in 
 vitro assays in bacteria and mammalian cells.  The sequence of 
assessment phases described in section is then followed.  
New chemicals intended for domestic use will normally give 
unequivocal negative results in tests for gene mutation in both 
prokaryotic and eukaryotic cells, and for chromosome aberrations  in 
 vitro.   Where direct contact with the chemical is perceived, data 
from an  in vivo assay for chromosome breakage are usually 
available.  Following the principles developed earlier, positive 
results in any of these assays may prevent the release of a 
chemical for domestic use.  Evidence from  in vivo mutation 

studies, pharmacokinetic data, or long-term animal studies may, 
however, remove the concern caused by an isolated positive result 
in an  in vitro assay.  Pesticides

    Exposure to pesticides may occur in a variety of different ways 
including exposure of workers during manufacture, exposure during 
the transport, formulation, or application of pesticides, and 
exposure to residues in edible crops, soil, and water.  Adverse 
effects on man may result from either the compound itself, its 
mammalian metabolites, plant and soil metabolites and, possibly, 
from breakdown products in the environment.  Unlike the chemicals 
described previously as medicinal, food, and domestic chemicals, 
pesticides are often dispersed widely in the environment and stable 
materials, such as DDT, may remain as virtually permanent 
contaminants at minute, though detectable concentrations. 

    Because of this potential for ubiquity, detailed information on 
the toxicity, stability, and fate of pesticides in the environment 
is mandatory in many countries, before they can be registered and 
released for use.  The use of pesticides, however, is virtually 
indispensible for the successful production of most major crops, 
and for the control of certain major insect-born diseases of man 
and domestic animals.  This, together with the fact that pesticides 
are highly biologically-active molecules, requires a fine balance 
to be set between the benefits accrued by using the pesticide and 
its possible hazard to man or the environment. 

    Tests for mutagenicity form only a small part of the overall 
package of data accumulated before a pesticide is released for use.  
Short-term tests are usually carried out in parallel with the 
development of a new pesticide.  For example, bacterial mutation 
data are normally available before the first limited field trials 
to test the efficacy of candidate compounds are carried out so that 
safe handling procedures can be formulated for both laboratory and 
field researchers.  The next stage in the development is usually a 
more extensive field trial on the target crop grown under 
commercial conditions, and another phase of toxicity testing, 
including an assay for chromosome aberrations in mammalian cells, 
precedes this stage.  Unless there is evidence from other toxicity 
studies or from chemical structure/pharmacokinetic considerations 
that the chemical may be genotoxic, negative results in base set 
assays frequently allow the pesticide to proceed through the 
developmental and evaluation stages.  Where, however, potential 
genotoxicity is still suspected, supplementary tests, including 
assays for gene mutations or primary DNA damage in eukaryotic cells 
may be considered at this time. 

    The final phase of toxicity testing is carried out after the 
development stage is completed and field evaluation has 
demonstrated a potentially successful product.  These tests are 
usually designed to complete the toxicity package required by most 
authorities responsible for the licencing of pesticides for use.  
Results from a battery of short-term tests including the bacterial 

and chromosome assays conducted during the development phase, an 
assay for gene mutations in mammalian cells, and an analysis of 
metaphase chromosomes of bone-marrow cells from rodents dosed with 
the chemical, meet the requirements of most regulatory authorities.  
However, different countries have different requirements, and 
additional tests, for example, for aneuploidy or primary DNA 
damage, may sometimes be required. 

    The finding of mutagenic activity in either of the two initial 
short-term tests need not necessarily indicate that development of 
a pesticide should be abandoned, though this is often the case.  If 
the potential value of the pesticide merits further development, it 
is usually treated as a highly toxic material and handled 
accordingly in subsequent field trials.  Additional testing to 
characterise the mutagenic activity and to determine its activity 
 in vivo may then be initiated.  An assessment of the hazards 
associated with a mutagenic pesticide will depend on data from  in 
 vivo studies (e.g., bone-marrow cytogenetics and either germ-cell 
cytogenetics or a dominant lethal assay), the metabolic profile of 
the chemical, and data on its stability and rate of elimination or 
degradation from the crop and the immediate environment.  A final 
decision on whether to continue large-scale development and 
evaluation of a mutagenic pesticide may be delayed until data from 
other biochemical and toxicological studies, including long-term 
animal cancer studies are available. 

    Pesticides are often supplied and used in a variety of 
formulations and in mixtures with other pesticides.  It is usual, 
therefore, to consider both the pure material and the specific 
formulation, when testing pesticides and assessing the significance 
of toxicity data. 

    The assessment of the hazards of residues of pesticides in 
plants, soil, and water is usually based on analytical chemical 
data.  However, some pesticides, e.g., some atrazines, are 
metabolized by plant enzymes to mutagenic products.  Although these 
metabolites can be analysed chemically, their mutagenic activity 
can be detected by testing extracts of plants exposed to pesticides 
in bacterial mutation assays. 

    Pesticides as a class contain two widely quoted examples of 
ambiguity between mutagenic activity and carcinogenicity. 
Dichlorvos (2,2-dichlorovinyl dimethyl phosphate), an 
organophosphate insecticide, is a confirmed bacterial mutagen.  
However, results from  in vitro mammalian cell assays are either 
negative or equivocal, and it does not produce mutations  in vivo.  
Comprehensive long-term cancer studies indicate that dichlorvos is 
not a carcinogen.  Pharmacokinetic and other biochemical studies 
suggest that this compound is efficiently detoxified in animals, so 
that, in spite of being a bacterial mutagen, it is still marketed 
as a domestic and agricultural insecticide.  The other example is a 
class of insecticides including DDT and dieldrin known collectively 
as organochlorine compounds.  Both these chemicals induce tumours 
in liver tissue in mice after prolonged exposure.  Both have also 
been subjected to comprehensive  in vivo and  in vitro mutagenicity 

tests, and although isolated positive results appear in the 
literature, a detailed analysis of the data suggests that these two 
organochlorines are not genotoxic, i.e., do not cause adverse 
effects as a direct result of a DNA lesion.  The primary 
carcinogenic growth appears to be confined to rodent liver, and 
although the potential hazard has been debated at length for many 
years, the significance of these findings for human health remains 

    Both these examples are given to illustrate the complexity of 
the extrapolation of  in vitro data to animal data to human hazard 
and serve to emphasise the caution needed in some cases in the 
assessment of genotoxic hazard from the results of short-term 
tests.  Chemical compounds used in industry

    Most industries use chemical compounds in one form or another.  
The function of the chemical industry itself is to manufacture, 
from primary sources such as oil, coal, and ore, chemicals that are 
valuable commodities in everyday life or that are necessary 
components in the manufacture of other products.  The principal raw 
materials undergo a series of processes to convert them initially 
to base chemicals (e.g., inorganic compounds such as alkalis and 
acids, and organic compounds such as olefin and aromatic 
compounds), then to intermediates and finally to the finished 
chemical product.  These may be consumer products such as solvents, 
etc., or are used by other industries in the manufacture of, for 
example, paints, adhesives, drugs, and plastics. 

    The output of the chemical industry is enormous in both 
quantity and diversity and the management of safety in the industry 
is based on the principle of identifying and assessing the hazards 
of exposure to particular chemicals, and then taking steps to 
reduce or eliminate human exposure.  It should be accepted that 
many of the chemicals used in industry are dangerous to man and a 
great deal of effort is expended in ensuring the safety of workers 
by the introduction of safe handling procedures, protective 
clothing, and enclosed industrial processes. 

    Exposure to chemicals is possible during manufacture, during 
transport of material from one industry to another, and as a result 
of environmental contamination.  The amount of toxicity data 
necessary to provide a sound assessment of the possible hazard of a 
chemical is governed primarily by the extent of human exposure and 
environmental distribution.  For many of the chemicals used in 
industry, human exposure is minimal and data from base set assays 
are often regarded as providing sufficient information on the 
potential mutagenicity or carcinogenicity of such chemicals to 
allow the small groups of workers involved to be protected 
accordingly.  Materials that are produced in larger volumes and 
that are transported in bulk or widely used in other industries may 
require additional testing.  Bulk products that are non-mutagenic 
in the initial battery of tests may need to have these findings 
confirmed in, for example, a eukaryotic assay for gene mutation or 

primary DNA damage and an  in vivo test for chromosome  
aberrations, before assessing the genotoxic hazard.  With large-
scale chemicals that are shown to be mutagenic in the base set, the 
genotoxicity may need to be further characterised in supplementary 
 in vitro and  in vivo assays.  Further testing may involve detailed 
studies of mutagenic activity in laboratory animals and long-term 
cancer studies may be necessary before the potential hazard can be 
fully evaluated and safe working conditions established. 

    Many chemicals used in industry are volatile and present a 
different sort of hazard, for not only can such chemicals present 
atmospheric contamination in the workplace, they may also escape 
into the surrounding environment.  In the modern chemical industry, 
the hazards associated with toxic vapours are well recognised and 
safe working practices are, in general, fully implemented, though 
the toxicity of vapours is still a real hazard in some cottage 
industries.  When assessing the mutagenicity of volatile chemicals, 
it is important to ensure that the experimental conditions were 
appropriate, i.e.,  in vitro tests require the use of sealed 
vessels to eliminate the loss of test material by evaporation, and, 
ideally, an inhalation exposure regimen should be used in  in vivo  

    The manufacture, use, and transport of chemicals used in 
industry is strictly regulated by national and international bodies 
responsible for industrial and environmental health.  The role of 
toxicity testing is described in detail in the guidelines of the 
appropriate authorities such as the Organisation for Economic 
Cooperation and Development (OECD). 


Acentric                      chromosomal fragment lacking a 

Acrocentric                   chromosome with the centromere close 
                              to one end of the chromatids

Allele                        one of two or more alternate forms of 
                              a gene at a specific locus on a 
                              particular chromosome

Anaphase                      stage of mitosis in which the 
                              centromere divides and the chromatids 
                              migrate towards poles of the cell

Aneuploidy                    addition or loss of one or more 
                              chromosomes from the haploid (i.e., 
                              meieosis) or diploid (i.e., in 
                              mitosis) number, i.e., 2n + 1, 
                              2n - 2, etc.

Autosome                      any chromosome other than the sex 
                              (i.e., X and Y) chromosomes

Banding                       techniques that result in 
                              differentially-stained bands along a 
                              chromosome, the pattern of banding 
                              being characteristic for a particular 
                              species and for specific chromosomes; 
                              banding techniques are commonly used
                              to identify exchange of material 
                              between chromosomes, e.g., 

Break                         damage to a chromatid or isochromatid 
                              involving a discontinuity of the 
                              chromosome greater than the width of 
                              a chromatid

Bromodeoxyuridine (BrdUrd)    base analogue that is incorporated 
                              into DNA in place of thymidine and, 
                              using suitable techniques, makes 
                              possible the observation of sister
                              chromatid exchanges

Budding and Fission           morphological features of cell 
                              division in yeast species

Centriole                     cellular component that divides into 
                              two prior to mitosis allowing the two 
                              daughter centrioles to migrate to
                              opposite ends of the cell forming 
                              points of origin of the spindle

Centromere                    region at which sister chromatids are 
                              held together; also known as the 
                              kinetochore, it is the structure by 
                              which chromosomes are attached to the 
                              spindle; the centromere splits 
                              longitudinally at anaphase allowing 
                              the chromatids to move to opposite 

Chromatid                     unreplicated chromosome or one half 
                              of a complete chromosome with the 
                              identical copy being its sister

Chromatid aberration          structural aberration affecting only 
                              one of the two chromatids of a 

Chromosomal aberration        structural aberration affecting both 
                              chromatids of a chromosome; also 
                              referred to as an isochromatid 

Clastogen                     a physical or chemical agent that 
                              induces chromosome breakage

Cross links                   covalent bonds between bases in 
                              parallel DNA strands

Deletion                      chromatid or isochromatid aberration 
                              in which part of a chromosome is 
                              missing as a result of a break; the 
                              deletion may be from the end of the 
                              chromatid, i.e., terminal, or from 
                              the middle of the chromatid, i.e., 

Dicentric                     a chromosome with two centromeres

Diploid                       the normal chromosome number of the 
                              somatic cells of most higher 
                              organisms; referred to as "2n", where 
                              n = the haploid number

DNA                           deoxyribonucleic acid

Dominant mutant               term applied to any mutant the effect 
                              of which is detectable in the 
                              heterozygous condition

Double-strand breaks          rupture of both strands of the DNA 
                              double helix at the same site

Endo-reduplication            chromatid alignment is maintained in 
                              a cell in which the chromosomes have 
                              duplicated but the cell has failed to
                              cleave; a form of polyploidy

Erythroblast                  proliferating precurser of red blood 
                              cells (erythrocytes)

Extrachromosomal gene         gene carried on an element outside 
                              the nucleus, e.g., a mitochondrial 

Gap                           non-staining region of chromatid not 
                              larger than the width of the 

Gene conversion               recombination event within a gene 
                              producing non-reciprocal product

Giemsa stain                  chromosome-staining solution 
                              containing the dyes azure, eosin, and 
                              methylene blue

Haploid                       chromosome number in the gametes; a 
                              single set of the chromosomes; 
                              referred to as the "n" number of 

Hemizygous                    occurrence of genes in a haploid 
                              condition in a normally diploid cell 
                              or organism; as on the X-chromosome 
                              of  Drosophila males

Heterozygote                  a zygote derived from the union of 
                              gametes, dissimilar in respect of the 
                              quality, quantity, or arrangement of

Heteroallele                  diploid cell carrying two non-
                              identical alleles of a gene

Heterozygote                  diploid cells contain two complete 
                              sets of chromosomes; the pairs of 
                              equivalent chromosomes are called
                              "homologous" and are considered to be 
                              structurally identical, at equivalent
                              loci, along the chromsome, alleles of 
                              a gene occur which, in homologues, 
                              serve the same function; sometimes,
                              the pairs of alleles are not 
                              identical and, in such cases, the 
                              cell is described as "heterozygous" 
                              for the gene at that locus

Hoechst 33258(R)              fluorescent dye used to demonstrate 
                              chromosomes in which the DNA has been
                              treated with bromodeoxyuridine, 
                              making observation of sister 
                              chromatid exchanges under a 
                              fluorescent microscope possible; 
                              subsequent staining with Giemsa 
                              allows observation of sister 
                              chromatid exchange under a light

Homoallele                    diploid cell carrying two identical 
                              alleles of a gene

Homologous                    see heterozygote

Homozygote                    a zygote derived from the union of 
                              gametes identical in respect of the 
                              quality, quantity, and arrangement of

Hyperdiploidy                 aneuploidy in which the chromosome 
                              number is greater than 2n

Hypotonic                     solution with an ionic strength lower 
                              than that of the cell contents; when 
                              cells are placed in a hypotonic 
                              solution, there is a net uptake of 
                              water resulting in swelling of the 
                              cell; hypotonic treatment of cells at
                              metaphase improves spreading of 
                              chromosomes for microscopic 

Idiogram (Karyogram)          the arrangement of chromosomes (i.e., 
                              from a photograph or drawing) into 
                              pairs and groups of pairs, usually in 
                              order of decreasing size

Instars                       periods in larval development in 
                               Drosophila; the larvae undergoes two 
                              moults so that the larval period 
                              consists of three stages: the first,
                              second, and third instars

Intercalation                 insertion of a molecule, e.g., 
                              adriamycin, between adjacent bases in 
                              the DNA molecule

Interchange                   exchange of material between two 
                              chromatids from different chromosomes

Intrachange                   exchange of material between sister 
                              chromatids, i.e., on the same 
                              chromosome, or exchange within one 

Inversion                     chromosome rearrangement in which a 
                              region between two breaks has been 
                              inverted; "paracentric": the inverted
                              region is within one chromatid arm; 
                              "pericentric": the inverted region 
                              includes the centromere

Isochromatid aberration       chromosome aberration affecting both 
                              chromatids; chromosomal aberration

Karyotype                     the chromosome complement of a cell 
                              or of a particular species

Lethal gene                   a gene the substitution of which, for 
                              its normal allele, converts a viable 
                              into a non-viable gamete or zygote;
                              may be dominant or recessive

Mating type                   in yeasts, mating occurs between 
                              strains of opposite mating type, 
                              i.e., a and alpha strains in 
                               S. cerevisiae and h+ and h- in 
                               S. pombe; the genetic event that
                              changes mating type from a to alpha 
                              and vice-versa is called a "mating 
                              type switch"

Meiosis                       cell division in germinal cells 
                              resulting in cells with the haploid 
                              number of chromosomes

Metacentric                   chromosome with the centromere 
                              approximately at the midpoint; 
                              "submetacentric": centromere between 
                              the centre and one end of the 
Metaphase                     stage of mitosis at which the 
                              chromosomes are condensed and aligned 
                              on the equator of the spindle

Micronucleus                  small fragment of chromosome material 
                              visible during interphase outside and 
                              separate from the main nucleus; may 
                              occur as a result of a chromosome 
                              fragment or a whole chromosome that 
                              detached from the spindle during 

Minute                        very small fragment or minute ring of 
                              chromosome material; may occur singly 
                              or in pairs

Mis-sense                     a mutation producing a gene product 
                              with a substituted amino acid

Mitogen                       an agent that stimulates resting 
                              (interphase) cells to divide and 

Mitotic index                 the proportion, usually expressed as 
                              a percentage, of dividing cells in a 

Mitosis                       stage of the cell cycle at which the 
                              chromosomes condense, thus becoming 
                              discrete structures when observed
                              microscopically; the chromosomes 
                              align on the spindle and then 
                              separate into chromatids that migrate 
                              to opposite poles of the cell before
                              the cell cleaves to form two daughter 

Mosaic                        a state in which a single individual 
                              has cells of two or more different 

Non-disjunction               failure of chromosomes to separate 
                              during mitosis or meiosis resulting 
                              in daughter cells with additional and
                              lost chromosomes

Nonsense                      mutation producing a messenger RNA 
                              molecule with a triplet not coding 
                              for an amino acid, e.g., "amber" and
                              "ochre" are nonsense mutations

Normochromatic erythrocytes   mature erythrocytes staining red-
                              yellow with Giemsa stain

Orcein                        chromosome-staining solution

Polychromatic erythrocytes    young or immature erythrocytes 
                              staining blue-red with Giemsa stain

Polyploidy                    cell containing more than the diploid 
                              number (2n) of chromosomes in exact 
                              multiples of the haploid number (n), 
                              e.g., triploid = 3n, tetraploid = 4n, 

Recessive mutant              term applied to any mutant the effect 
                              of which is detectable in the 
                              homozygous or hemizygous condition

Ring                          chromosome rearrangement in which 
                              fusion of ends of a chromosome 
                              results in a ring structure either 
                              with (centric) or without (acentric) 
                              a centromere

Sex chromosomes               chromosomes that determine the gender 
                              of an individual; in mammals, the X 
                              chromosome signifies female gender, 
                              and the Y chromosome indicates males; 
                              diploid cells in normal females are 
                              XX and XY in normal males

Single-strand breaks          breakage of only one of the two 
                              molecules (strands) in the DNA double 

Sister chromatid exchange     an apparently symmetrical exchange of
 (SCE)                        material between sister chromatids

S-phase                       phase in the cell cycle during which 
                              normal DNA synthesis occurs

Spermatogenesis               development of the sperm from its 
                              precurser cell; successive stages in 
                              spermatogenesis are spermatogonia
                              (pre-meiotic), spermatocytes (meiotic 
                              stages), spermatids, and spermatazoa 

Spindle                       polymerized tubulin, radiating from 
                              the centrioles formed early in 
                              mitosis; chromosomes attach to the
                              central point (equator) of the 
                              spindle at their centromeres and, 
                              subsequently, move along the spindle 
                              fibres during anaphase

Spindle poison                agent such as colchicine, colcemid, 
                              and vinblastine that prevents tubulin 
                              polymerization and thus, chromosome 
                              migration, resulting in an 
                              accumulation of cells at metaphase; 
                              used to arrest cells at metaphase for 
                              chromosome examination

SLRL                          Sex-linked Recessive Lethals:
                              recessive lethal mutations located on 
                              sex chromosomes, i.e., the 
                              X-chromosome of  Drosophila

Suppressor mutation           second site mutation that eliminates 
                              the phenotype produced by a previous

Telocentric                   chromosome with the centromere at the 
                              end of the chromatids

Telophase                     stage of mitosis during which the 
                              cell cleaves to give two daughter 

Translocation                 isochromatid rearrangement resulting 
                              from an exchange of material between 
                              two chromosomes

Transposition                 transformation of genetic information 
                              from one chromosome location to 
                              another, e.g., in yeast cells

Vernier reading               location of an object, e.g., a cell, 
                              on a microscope slide given as values 
                              on two scales (the X- and Y-axis) of 
                              the microscopic stage


ABBONDANDOLO, A.  (1977)  Prospects for evaluating genetic 
damage in mammalian cells in culture.  Mutat. Res., 42: 279-298.

ABRAHAMSON, S. & LEWIS, E.B.  (1971)  The detection of 
mutations in  Drosophila melanogaster.  In: Hollaender, A., ed. 
 Chemical mutagens: principles and methods for their detection, 
New York, London, Plenum Press, pp. 461-487.

ADAIR, G.M., CARVER, J.H., & WANDRES, D.L.  (1980)  
Mutagenicity testing in mammalian cells. I. Derivation of a 
Chinese hamster ovary cell line heterozygous for the adenine 
phosphoribosyl transferase and thymidine kinase loci.  Mutat. 
 Res., 72: 187-205.

ADLER, I.-D.  (1970)  Cytogenic analysis of Ascites tumour 
cells in mice in mutation research. In: Vogel, F. & Röhrborn, 
G., ed.  Chemical mutagenesis in mammals and man, Berlin, 
Heidelberg, New York, Springer-Verlag, pp. 251-259.

ADLER, I.-D.  (1974)  Comparative cytogenetic study after 
treatment of mouse spermatogonia with mitomycin C.  Mutat. 
 Res., 23: 369-379.

ADLER, I.-D.  (1981)  Transplacental cytogenetic effects by 
TEM, mitomycin C, and benzo( a)pyrene.  Mutat. Res., 85: 299-300.

ADLER, I.-D.  (1982)  Male germ cell cytogenetics. In: Hsu, 
T.C., ed.  Cytogenetic assays of environmental mutagens, 
New York, Allanheld Osmun & Co., pp. 249-276.

ADLER, I.-D.  (1985)  Cytogenetic tests in mammals. In: 
Venitt, S. & Parry, J.M., ed.  Mutagenicity testing: a 
 practical approach, Oxford, IRL Press, pp. 275-306.

ADLER, I.-D. & BREWEN, J.G.  (1982)  Effects of chemicals on 
chromosome aberration production in male and female germ 
cells. In: Hollaender, A. & de Serres, F.J., ed.  Chemical 
 mutagens: principles and methods for their detection,  
New York, London, Plenum Press, Vol. 7, pp. 1-35.

AHMED, M. & GRANT, W.F.  (1972)  Cytological effect of the 
pesticides Phosdrin and Bladex on  Tradescantia and  Vicia 
 faba. Can. J. Genet. Cytol., 14: 157-165.

C.  (1984)  Mammalian germ cell cytogenetics. In: Dean, B.J., 
ed.  UKEMS Sub-Committee on Guidelines for Mutagenicity 
 Testing. Part II. Supplementary Tests, Swansea, United Kingdom 
Environmental Mutagen Society.

AMACHER, D.E. & PAILLET, S.C.  (1983)  The activation of 
procarcinogens to mutagens by cultured rat hepatocytes in the 
L5178Y/TK mutation assay.  Mutat. Res., 113: 77-88.

AMACHER, D.E. & TURNER, G.N.  (1981)  Promutagen activation by 
rodent-liver post-mitochondrial fractions in the L5178Y/TK 
cell mutation assay.  Mutat. Res., 74: 485-501.

AMACHER, D.E. & TURNER, G.N.  (1982a)  Mutagenic evaluation of 
carcinogens and non-carcinogens in the L5178Y/TK assay 
utilizing post-mitochondrial fractions (S9) from normal rat 
liver.  Mutat. Res., 97: 49-65.

AMACHER, D.E. & TURNER, G.N.  (1982b)  The effect of liver 
post-mitochondrial fraction concentration from Arochlor 
1254-treated rats on promutagen activation in L5178Y cells. 
 Mutat. Res., 97: 131-137.

AMES, B.N., DURSTEN, W.E., YAMASAKI, E., & LEE, F.W.  (1973)  
Carcinogens are mutagens: a simple test system combining liver 
homogenates for activation and bacteria for detection.  Proc. 
 Natl Acad. Sci. (USA), 70: 2281-2285.

AMES, B.N., MCCANN, J., & YAMASAKI, E.  (1975)  Methods for 
detecting carcinogens and mutagens with the  Salmonella/ 
mammalian microsome mutagenicity test.  Mutat. Res., 31: 

ANDERSON, D., BATEMAN, A., & MCGREGOR, D.  (1983)  Dominant 
lethal mutation assays. In:  Report of the UKEMS Sub-Committee 
 on Guidelines for Mutagenicity Testing, Swansea, 
United Kingdom Environmental Mutagen Society, pp. 143-164.

ANDRAE, U. & GREIM, H.  (1979)  Induction of DNA repair 
replication by hydroxyurea in human lymphoblastoid cells 
mediated by liver micromes and NADPH.  Biochem. biophys. Res. 
 Commun., 87: 50-58.

MARGOLIN, B.H, MATTER, B.E., & SHELBY, M.D., ed.  (1985)  
 Evaluation of short-term tests for carcinogens. Report of the 
 International Programme on Chemical Safety's Collaborative 
 Study on In vitro Assays, Amsterdam, Oxford, New York, 
Elsevier Science Publishers (Progress in Mutation Research, 
Vol. 5).

AUERBACH, C  (1962a)   Mutation. An introduction to research on 
 mutagenesis. Part 1. Methods, Edinburgh, London, Oliver and 
Boyd, 176 pp.

AUERBACH, C.  (1962b)  The production of visible mutations in 
 Drosophila by chloroethyl methylsulfonate (CB 1506).  Genet. 
 Res., 3: 461-466.

BAKER, R.M.  (1979)  Nature and use of ouabain-resistant 
mutants. In: Hsie, A.W., O'Neill, J.P., & McElheny, V.K., ed. 
 Mammalian cell mutagenesis: the maturation of the test 
 systems, New York, Cold Spring Harbor Laboratory, pp. 237-247 
(Banbury Report No. 2).

Ouabain-resistant mutants of mouse and Chinese hamster cells. 
 Cell, 1: 9-21.

BARTSCH, H., CAMUS, A.-M., & MALAVEILLE, C.  (1976)  
Comparative mutagenicity of  N-nitrosamines in a semi-solid and 
liquid incubation system in the presence of rat or human 
tissue fractions.  Mutat. Res., 37: 149-162.

T., DREVON, C., PICOLLI, C., & MONTESANO, R.  (1979)  Validation 
and comparative studies on 180 chemicals with  S. typhimurium 
strains and V79 Chinese hamster cells in the presence of various 
metabolizing systems.  Mutat. Res., 76: 1-50.

T., DREVPM, C., PICCOLI, S., & MONTESANO, R.  (1980)  
Bacterial and mammalian mutagenicity tests: validation and 
comparative studies on 180 chemicals. In: Montesano, R., 
Bartsch, H., & Tomatis, L., ed.  Molecular and cellular aspects 
 of carcinogen screening tests, Lyons, International Agency for 
Research on Cancer, pp. 179-241 (IARC Scientific Publications 
No. 27).

(1982)  Metabolic activation systems  in vitro for 
carcinogen-mutagen screening tests. In: de Serres, F.J. & 
Hollaender, A., ed.  Chemical mutagens, New York, London, 
Plenum Press, Vol. 7, pp. 95-161.

BATEMAN, A.J.  (1958)  Mutagenic sensitivity of maturing germ 
cells in the male mouse.  Heredity, 12: 213-232.

BATEMAN, A.J.  (1966)  Testing chemicals for mutagenicity in a 
mammal.  Nature (Lond.), 210: 205-206.

BATEMAN, A.J. & EPSTEIN, S.S.  (1971)  Dominant lethal 
mutations in mammals. In: Hollaender, A., ed.  Chemical 
 mutagens: principles and methods for their detection, 
New York, London, Plenum Press, Vol. 2, pp. 541-568.

BECKER, H.J.  (1966)  Genetic and variegation mosaics in the 
eye of  Drosophila. Curr. Topics dev. Biol., 1: 155-171.

BENDER, M.A., GRIGGS, H.G., & BEDFORD, J.S.  (1974)  
Mechanisms of chromosomal aberration production. III. 
Chemicals and ionizing radiation.  Mutat. Res., 23: 197-212.

BOLLER, K. & SCHMID, W.  (1970)  [Chemical mutagenesis in 
mammals. The bone marrow of the Chinese hamster as an  in vivo 
test system.]  Haematologische Befunde nach Behandlung mit 
 Trenimon. Humangenetik, 11: 35-54 (in German).

BOYCE, R.P. & HOWARD-FLANDERS, P.  (1964)  Release of 
ultraviolet light-induced thymine dimers and DNA in  E. coli 
K-12.  Proc. Natl Acad. Sci. (USA), 51: 293-300.

BOYUM, A.  (1968)  Isolation of mononuclear cells and 
granulocytes from human blood.  Scand. J. clin. Lab. Invest., 
21(Suppl. 97): 77-89.

PETERSON, A., & HUMBERMAN, E.  (1981)  Mutagenesis by chemical 
agents in V79 Chinese hamster cells: a review and analysis of 
the literature: a Report of the Gene-tox program.  Mutat. Res., 
87: 81-142.

BRENNEKE, H.  (1937)  [Radiation damage to mouse and rat 
sperm, observed from the early development of the ova.] 
 Strahlentherapie, 60: 214-238 (in German).

BREWEN, J.G. & PRESTON, R.J.  (1982)  Cytogenic analysis of 
mammalian oocytes in mutagenicity studies. In: Hsu, T.C., ed. 
 Cytogenic assays of environmental mutagens, New York, 
Allanheld Osmun & Co., pp. 277-287.

BREWEN, J.G. & STETKA, D.G.  (1982)  Cytogenetic events in 
vivo. In: Heddle, J.A., ed.  Mutagenicity: new horizons in 
 genetic toxicology, New York, London, Academic Press, 
pp. 351-384.

(1975)  Studies on chemically-induced dominant lethality. I. 
The cytogenetic basis of MMS-induced dominant lethality in 
post-meiotic male germ cells.  Mutat. Res., 33: 239-250.

Cigarette smoking - does it carry a genetic risk?  Mutat. Res. 
65: 71-81 (ICPEMC Publication No. 3).

BRUCE, W.R. & HEDDLE, J.A.  (1979)  The mutagenic activity of 
61 agents as determined by the micronucleus,  Salmonella and 
sperm abnormality assays.  Can. J. Genet. Cytol., 21: 319-335.

BRUSICK, D.  (1980)  Genetic risk estimation. In:  Principles 
 of genetic toxicology, New York, London, Plenum Press, 
pp. 109-135.

(1962)  A study of the chromsome damage persisting after X-ray 
therapy for ankylosing spondolitis.  Lancet, 2: 676-682.

BURTON, K.  (1956)  A study of the conditions and the 
mechanism of the diphenylamine reaction for the colorimetric 
estimation of DNA.  Biochem. J., 62: 315-323. 

CALLEN, D.F. & PHILPOT, R.M.  (1977)  Cytochrome P-450 and the 
activation of promutagens in  Saccharomyces cerevisiae. Mutat. 
 Res., 45: 309-324.

CASKEY, C.T. & KRUSH, G.D.  (1979)  The HPRT locus.  Cell, 16: 

CLAYSON, D.B.  (1981)  Carcinogens and carcinogen enhancers. 
 Mutat. Res., 86: 217-229 (ICPEMC Working Paper No 2/3).

CLEAVER, J.E.  (1982)  DNA damage, DNA repair, and DNA 
replication in short-term tests for exposure to mutagens. In: 
Kora, K.C., Douglas, G.R., & Nestmann, E.R., ed.  Mutagenesis, 
 human population monitoring, and genetic risk assessment, 
Amsterdam, Oxford, New York, Elsevier Science Publishers, 
pp. 111-124 (Progress in Mutation Research, Vol. 3).

CLEAVER, J.E. & THOMAS, G.H.  (1981)  Measurement of 
unscheduled synthesis by autoradiography. In: Friedberg, E.C. 
& Hanawalt, P.C., ed.  DNA repair: a laboratory manual of 
 research procedures, New York, Basel, Marcel Dekker, pp. 

(1972)  A mutational assay system using the thymidine kinase 
locus in mouse lymphoma cells.  Mutat. Res., 16: 77-87.

BROWN, M.M.M.  (1979)  Validation and characterization of the 
L5178Y TK+/- mouse lymphoma mutagen assay system:  Mutat. Res., 
59: 61-108.

K.H.  (1983)  Report of the Gene-Tox program on L5178Y.  Mutat. 
 Res., 115: 225-251.

COLE, J., ARLETT, C.F., LOWE, J., & BRIDGES, B.A.  (1982)  The 
mutagenic potency of 1,8-dinitropyrene in cultured mouse 
lymphoma cells.  Mutat. Res., 93: 213-220.

COLE, R.J., TAYLOR, N.A., COLE, J., & ARLETT, C.F.  (1979)  
Transplacental effects of chemical mutagens detected by the 
micronucleus test.  Nature (Lond.), 277: 317-318.

COLE, R.J., TAYLOR, N.A., COLE, J., & ARLETT, C.F.  (1981)  
Short-term tests for transplacentally active carcinogens. I. 
Micronucleus formation in fetal and maternal mouse 
erythroblasts.  Mutat. Res., 80: 141-157.

COLLINS, A.R.S., SHOR, S.L., & JOHNSON, R.T.  (1977)  The 
inhibition of repair in UV-irradiated human cells.  Mutat. 
 Res., 42: 413-432.

S.  (1984)  The detection of mutagens in urine, faeces, and 
body fluids. In: Dean, B.J., ed.  UKEMS Subcommittee on 
 Guidelines for Mutagenicity Testing. Part II. Supplementary 
 Tests, Swansea, United Kingdom Environmental Mutagen Society.

CONSTANTIN, M.J.  (1976)  Mutations for chlorophyll-deficiency 
in barley: comparative effects off physical and chemical 
mutagens. In: Gaul, H., ed.  Barley genetics. III, Munich, Karl 

CONSTANTIN, M.J. & NILAN, R.A.  (1982)  Chromosome aberrations 
in barley  (Hordeum vulgare): a report of the US Environmental 
Portection Agency Gene-Tox Program.  Mutat. Res., 99: 13-36.

CONSTANTIN, M.J. & OWENS, E.T.  (1982)  Introduction and 
perspectives of plant genetic and cytogenetic assays: a report 
of the US Environmental Protection Agency Gene-Tox Program. 
 Mutat. Res., 99: 1-12.

DARLINGTON, D.C. & LACOUR, L.F.  (1976)   The handling of 
 chromosomes, 6th ed., London, Allen and Unwin, 201 pp.

DAVIES, P.J., TIPPINS, R.S., & PARRY, J.M.  (1978)  Cell-cycle 
variation in the induction of lethality and mitotic 
recombination after treatment with UV and nitrous acid in the 
yeast  Saccharomyces cerevisiae. Mutat. Res., 51: 327-328.

DEAN, B.J.  (1969)  Chemically-induced chromosome damage. 
 Lab. Anim., 3: 157-174.

DEAN, B.J.  (1985)  Summary report on the performance of 
cytogenetic assays in cultured mammalian cells. In:  Ashby, 
J., de Serres, F.J., Draper, M., Ishidate, M., Jr, Margolin, 
B., Matter, B., & Shelby, M.D., ed.  Evaluation of short-term 
 tests for carcinogens, Amsterdam, Oxford, New York, Elsevier 
Science Publishers, pp. 69-83 (Progress in Mutation Research, 
Vol. 5).

DEAN, B.J. & DANFORD, N.D.  (1985)  Assays for the detection 
of chemically-induced chromosome damage in cultured mammalian 
cells. In: Venitt, S. & Parry, J.M., ed.  Mutagenicity testing: 
 a practical approach, Oxford, IRL Press, pp. 187-232.

DEAN, B.J. & HODSON-WALKER, G.  (1979)  An  in vitro chromosome 
assay using cultured rat-liver cells.  Mutat. Res., 64: 329-337.

DEAN, B.J., ANDERSON, D., & SRAM, R.J.  (1981)  Mutagenicity 
of selected chemicals in the mammalian dominant lethal assay. 
In: de Serres, F.J. & Shelby, M.D., ed.  Comparative chemical 
 mutagenisis, New York, London, Plenum Press, pp. 487-538.

TAYLOR, N.A.  (1983)  Framework of supplementary testing 
procedures. In: Dean, B.J., ed.  UKEMS Sub-Committee on 
 Guidelines for Mutagenicity Testing. Part 1, Swansea, 
United Kingdom Environmental Mutagen Society.

DEBAUN, J.R., MILLER, E.C., & MILLER, J.A.  (1970)  
 N-hydroxy-2 acetylaminefluorene sulfotransferase: its probable 
role in carcinogenesis and in protein (methionyl-S-yl) binding 
in rat liver.  Cancer Res., 30: 577-595.

DEMEREC, M., & KAUFMAN, B.P.  (1973)   Drosophila guide: 
 introduction to the genetics and cytology of Drosophila 
 Melanogaster, Washington DC, Carnegie Institute of Washington, 
45 pp.

DE SERRES, F.J. & ASHBY, J., ed.  (1981)   Evaluation of 
 short-term tests for carcinogens, Amsterdam, Oxford, New York, 
Elsevier Science Publishers (Progress in Mutation Research, 
Vol. 1).

DHSS  (1981)   Guidelines for testing chemicals for 
 mutagenicity, London, Committee on Mutagenicity of Chemicals 
in Food, Consumer Products, and the Environment, Department of 
Health and Social Security (Report on Health and Social 
Subjects No. 24).

DHSS  (1982)   Guidelines for the testing of chemicals for 
 mutagenicity. V. Genetic and partly-genetic diseases of man: 
 types, frequencies, and mutation rates, London, UK Department 
of Health and Social Security (Report on Health and Social 
Subjects No. 24).

DRAPER, M.H. & GRIFFIN, J.P.  (1980)  Draft guidelines on 
mutagenicity testing of new drugs issued by the CPMP. A four 
test screen.  Arch. Toxicol., 46: 9-19.

EHLING, U.H.  (1977)  Dominant lethal mutations in male mice. 
 Arch. Toxicol., 38: 1-11.

EHLING, U.H., CUMMING, R.B., & MALLING, H.V.  (1968)  
Induction of dominant lethal mutations by alkylating agents in 
male mice.  Mutat. Res., 5: 417-428.

WIEMANN, H.  (1978)  Standard protocol for the dominant lethal 
test on male mice.  Arch. Toxicol., 39: 173-185.

EHRENBERG, L.  (1979)  Risk assessment of ethylene oxide and 
other compounds. In: McElheny, V.K. & Abrahamson, S., ed. 
 Assessing chemical mutagens: the risk to humans, New York, 
Cold Spring Harbour Laboratories, pp. 157-190 (Banbury Report 
No. 1).

EPSTEIN, S.S. & SHAFNER, H.  (1968)  Chemical mutagens in the 
human environment.  Nature (Lond.), 219: 385-387.

EVANS, H.J.  (1962)  Chromosome aberrations induced by 
ionizing radiation.  Int. Rev. Cytol., 13: 221-321.

EVANS, H.J.  (1976)  Cytological methods for detecting 
chemical mutagens. In: Hollaender, A., ed.  Chemical mutagens: 
 principles and methods for their detection, New York, London, 
Plenum Press, Vol. 4, pp. 1-29.

EVANS, H.J. & O'RIORDAN, M.L.  (1975)  Human peripheral blood 
lymphocytes for the analysis of chromosomal aberrations in 
mutagen tests.  Mutat. Res., 31: 135-148.

EVANS, H.J. & SCOTT, D.  (1964)  Influence of DNA synthesis on 
the production of chromatid aberrations by X-rays and maleic 
hydrazide in  Vicia faba. Genetics, 49: 17-38.

FAHRIG, R.  (1975)  Development of host-mediated mutagenicity 
tests: yeast systems. Recovery of yeast cells out of testes, 
liver, lung, and peritoneum of rats.  Mutat. Res., 31: 381-394.

FAHRIG, R.  (1977)  The mammalian spot test with mice.  Arch. 
 Toxikol., 38: 87-98.

FEDERAL REGISTER  (1978)  Non-clinical laboratory studies: 
good laboratory practice regulations.  Fed. Reg., 43(247): 

FEDERAL REGISTER  (1983)  Toxic substances control: good 
laboratory practice standards.  Fed. Reg., 48(230): 53922-53969.

FINK, G.R. & LOWENSTEIN, R.  (1971)  Simplified method for 
testing mutagens in  Saccharomyces. J. Bacteriol., 100: 

FISCHER, G.A., LEE, S.Y., & CALABRESI, P.  (1974)  Detection 
of chemical mutagens using a host-mediated assay (L5178Y) 
mutagenesis system.  Mutat. Res., 26: 501-511.

FOX, M.  (1981)  Some quantitative aspects of the response of 
mammalian cells  in vitro to induced mutagenesis. In: 
Marchalonis, J.J. & Hanna, M.G., ed.  Cancer Biology Review, 
New York, Basel, Marcel Decker Inc, Vol. 3, pp. 23-62.

Developmental compartmentalization in the dorsal mesothatic 
disc of  Drosophila. Dev. Biol., 48: 132-147.

HANSON, R.S.  (1971)  Formation of a factor lethal for  S. 
 typhimurium TA 1530 and TA 1531 on incubation of aflatoxin B1 
with rat liver microsomes.  Biochem. biophys. Res. Commun., 45: 

GENOROSO, W.M.  (1969)  Chemical induction of dominant lethals 
in female mice.  Genetics, 61: 461-470.

GENOROSO, W.M., CAIN, K.T., KRISHNA, M., & HUFF, S.W.  (1979)  
Genetic lesions induced by chemicals in spermatozoa and 
spermatids of mice are repaired in the egg.  Proc. Natl. Acad. 
 Sci. (USA), 76: 435-437.

GOCKE, E., ECKHARDT, K., KING, M.-T., & WILD, D.  (1982)  Some 
statistical aspects of spontaneous lethal mutations in 
 Drosophila. Mutat. Res., 104: 239-242.

(1983)  A pilot study on a new  Drosophila spot test.   Mutat. 
 Res. Lett., 120: 233-239.

GRAFE, A. & VOLLMAR, J.  (1977)  Small numbers in mutagenicity 
tests.  Arch. Toxicol., 38: 27-34.

GRANT, W.F.  (1982)  Chromosome aberration assays in Allum: a 
report of the US Environmental Protection Agency Gene-Tox 
Program.  Mutat. Res., 99: 273-291.

GRIESEMER, R.A. & CUETO, C., Jr  (1980)   Toward a 
 classification scheme for degrees of experimental evidence for 
 the carcinogenicity of chemicals for animals. In: Montesano, 
R., Bartsch, H., & Tomatis, L., ed.  Molecular and cellular 
 aspects of carcinogen screening tests, Lyons, International 
Agency for Research on Cancer, pp. 259-281 (IARC Scientific 
Publication No. 27).

GUPTA, R.A. & SINGH, B.  (1982)  Mutagenic responses of five 
independent genetic loci in CHO cells to a variety of 
mutagens: development and characteristics of a mutagen 
screening system based on selection for multiple 
drug-resistant markers.  Mutat. Res., 94: 449-466.

(1979)  DNA repair in bacteria and mammalian cells.  Ann. Rev. 
 Biochem., 48: 783-836.

HASEMAN, J.K. & SOARES, E.R.  (1976)  The distribution of 
fetal death in control mice and its implications on 
statistical tests for dominant lethal effects.  Mutat. Res., 
41: 277-288.

HEDDLE, J.  (1973)  A rapid  in vivo test for chromosomal 
damage.  Mutat. Res., 18: 187-190.

HEDDLE, J.A. & CARRANO, A.V.  (1977)  The DNA content of 
micronuclei induced in mouse bone marrow by y-irradiation: 
evidence that micronuclei arise from acentric chromosomal 
fragements.  Mutat. Res., 44: 63-69

NEWMARK, H., WARGOVICH, M.J., & BRUCE, W.R.  (1982)  
Micronuclei and related anomalies as a short-term assay for 
colon carcinogens. In: Bridges, B.A., Butterworth, B.E., & 
Weinstein, I.B., ed.  Indicators of genotoxic exposure,  
New York, Cold Spring Harbor Laboratory, pp. 367-375 (Banbury 
Report No. 13).

HERTWIG, P.  (1935)  [Sterility phenomena in mice irradiated 
with X-rays.]  Z. indukt. Abstamm. - u. VererbLehre, 70: 
517-523 (in German).

(1979)  Short-term tests for carcinogens and mutagens.  Mutat. 
 Res., 64: 133-226.

(1981)  Cytogenic analysis of the L5187Y TK+1- TK1 mouse 
lymphoma mutagenesis assay system.  Mutat. Res., 84: 167-181.

HSIE, A.W., O'NEILL, J.P., & MCELHENY, V.K., ed.  (1979)  
 Mammalian cell mutagenesis: the maturation of test systems, 
New York, Cold Spring Harbor Laboratory (Banbury Report No. 2).

O'NEILL, J.P., & WHITFIELD, B.L.  (1981)  The use of Chinese 
hamster ovary cells to quantity specific locus mutation and to 
determine mutagenicity of chemicals, A report of the Gene-tox 
program.  Mutat. Res., 86: 193-214.

IARC  (1979)   Chemicals and industrial processes associated 
 with cancer in humans, Lyons, International Agency for 
Research on Cancer (IARC Monographs Supplement No. 1).

IARC  (1980a)  Mutagenesis assays with bacteria. In:  Long-term 
 and short-term screening assays for carcinogens: a critical 
 appraisal, Lyons, International Agency for Research on Cancer, 
pp. 85-106 (IARC Monographs on the Evaluation of the 
Carcinogenic Risk of Chemicals to Humans, Supplement 2).

IARC  (1980b)  In: Castegnaro, M., Hunt, D.C., Sansone, E.B., 
Schuller, P.L., Siriwardana, M.G., Telling, G.M., van Egmond, 
H.P., & Walker, E.A., ed.  Laboratory decontamination and 
 destruction of aflatoxins B1, B2, G1, G2,  in laboratory wastes, 
Lyons, International Agency for Research on Cancer (IARC 
Scientific Publications No. 37).

ICPEMC  (1982)  Mutagenesis testing as an approach to 
carcinogenesis (Committee 2 Report).  Mutat. Res., 99: 73-91.

ICPEMC  (1983a)  Screening strategy for chemicals that are 
potential germ cell mutagens in mammals (Committee 1 Report). 
 Mutat. Res., 114: 117-117.

ICPEMC  (1983b)  Regulatory approaches to the control of 
environmental mutagens and carcinogens (Committee 3 Report). 
 Mutat. Res., 114: 179-216.

ICPEMC  (1983c)  Estimation of genetic risk and increased 
incidence of genetic disease due to environmental mutagens 
(Committee 4 Report).  Mutat. Res., 115: 255-291.

ICPEMC  (1983d)  Estimation of genetic risk and increased 
incidence due to environmental mutagens (Committee 4 Report). 
 Mutat. Res., 115: 255-291.

IRR, J.D. & SNEE, R.D.  (1982)  A statistical method for 
analysis of mouse lymphoma L5178Y cell TK locus forward 
mutation assay: comparison of results among three 
laboratories.  Mutat. Res., 97: 371-392.

ISCN  (1978)   Original Article Series 14(8), Basel, S. Karger 
AG, International System for Human Cytogenetic Nomenclature 
Birth Defects (Also in:  Cytogenet. Cell Genet., 21 309-404).

ISHIDATE, M., Jr  (1981)  Application of chromosomal aberration 
tests  in vitro to the primary screening for chemicals 
with carcinogenic and/or genetic hazards. In: Short-terms 
tests for carcinogens: quo vadis? In:  Proceedings of a 
 Symposium held at Montpelier, France (Excerpta Medica),  
pp. 58-59.

ISHIDATE, M., Jr & ODASHIMA, S.  (1977)  Chromosome tests with 
134 compounds in Chinese hamster cells  in vitro:  a screening 
test for chemical carcinogens.  Mutat. Res., 48: 337-354.

JENSSEN, D. & RAMEL, C.  (1978)  Factors affecting the induction 
of micronuclei at low doses of X-rays, MMS, and dimethylnitrosamine 
in mouse erhythroblasts.  Mutat. Res., 58: 51-65

JOHNSON, R.T. & RAO, P.N.  (1970)  Mammalian cell fusion: 
introduction of premature chromosome condensation in 
interphase myclei.  Nature (Lond.), 226: 717-722.

KAPLAN, W.D. & LYON, M.F.  (1953)  Failure of mercaptoethylamine 
to protect against the mutagenic effects of radiation. II.
Experiments with mice.  Science, 118: 777-778.

KASTENBAUM, M.A. & BOWMAN, K.O.  (1970)  Tables for 
determining the statistical significance of mutation 
frequencies.  Mutat. Res., 9: 527-549.

KATZ, A.J.  (1978)  Design and analysis of experiments on 
mutagenicity. I. Minimal sample size.  Mutat. Res., 50: 301-307.

KELLY, D. & PARRY, J.M.  (1983a)  Metabolic activation by 
cytochrome P-450/P-448 in the yeast  Saccharomyces cerevisiae. 
 Mutat. Res., 108: 147-158.

KELLY, S.L. & PARRY, J.M.  (1983b)  The induction of mutation 
and recombination following UV irradiaion during meiosis in 
 Saccharoymces cerevisiae. Mutat. Res., 108: 109-120.

KIHLMAN, B.A.  (1971)  Root-tips for studying the effects of 
chemicals on chromosomes. In: Hollaender, A., ed.  Chemical 
 mutagens, New York, London, Plenum Press, Vol. 2, pp. 489-514.

KIHLMAN, B.A. & ANDERSSON, H.C.  (1982)  Sister chromatid 
exchanges in plants. In: Wolff, S., ed.  Sister chromatid 
 exchange, New York, John Wiley and Sons, pp. 243-265.

WATKINS, P.  (1984)  Bacterial mutation assays with 
nitrosation products. In: Dean, B.J., ed.  UKEMS Sub-Committee 
 on Guidelines for Mutagenicity Testing. Part II. Supplementary 
 Tests, Swansea, United Kingdom Environmental Mutagen Society.

interpretation of the origin of subchromatid aberrations and 
chromosome stickiness as a category of chromatid aberrations. 
 Hereditas, 83: 153-162.

KLEWOSKI, E.J., Jr  (1978)  Detection of mutational damage in 
fern populations: an  in situ bioassay for mutagens in aquatic 
ecosytems. In: Hollaender, A. & de Serres, F.J., ed.  Chemical 
 mutagens, New York, London, Plenum Press, Vol. 5, pp. 77-99.

KLIESCH, U. & ADLER, I.-D.  (1983)  Chromosome analysis and 
micronucleus test in mouse bone marrow are not equally 
sensitive as shown by results with 5 chemical mutagens.  Mutat. 
 Res., 113: 340-346.

KLIESCH, U., ROUPOVA, I., & ADLER, I.-D.  (1982)  Induction of 
chromosome damage in mouse bone marrow by benzo (a)pyrene. 
 Mutat. Res., 102: 265-273.

KNUDSON, I.  (1982)  Natural, processed, and artificial 
mutagens in food: significance and consequences. In: Sorsa, M. 
& Vainio, H., ed.  Mutagens in our environment, New York, 
Alan R. Liss, pp 315-328.

KRATOCHVILOVA, J.  (1978)  Evaluation of pre-implantation loss 
in dominant-lethal assay in the mouse.  Mutat. Res., 54: 47-54.

KRUSKAL, W.H. & WALLIS, W.A.  (1952)  Use of ranks in 
one-criterion variance analysis.  J. Am. Stat. Assoc., 47: 

KUNZ, B.A., BARCLAY, B.J., & HAYNES, R.H.  (1980)  A simple, 
rapid-plate assay for mitotic recombination.  Mutat. Res., 73: 

KUROKI, T., DREVON, C., & MONTESANO, R.  (1977)  Microsome- 
mediated mutagenesis in V79 Chinese hamster cells by various 
nitrosamines.  Cancer Res., 37: 1044-1050.

M., & SELKIRK, J.K.  (1979)  Critical importance of microsome 
concentration in mutagenesis assays with V79 Chinese hamster 
cells.  Mutat. Res., 63: 259-272.

KURTEN, S. & OBE, G.  (1975)  Premature chromosome 
condensation in the bone marrow of Chinese hamsters after 
application of bleomycin  in vivo. Mutat. Res., 27: 285-294.

LAHDETIE, J. & PARVINEN, M.  (1981)  Meiotic micronuclei 
induced by X-rays in early spermatids of the rat.  Mutat. Res., 
81: 103-105.

KUSZYNSKI, C.  (1981)  Lung and liver cell-mediated 
mutagenesis systems: specificities in the activation of 
chemical carcinogens.  Carcinogenesis, 2: 852-858.

(1977)   In vitro and  in vivo analysis of sister chromatid 
exchange formation. In: Kilbey, B.J., ed.  Handbook of mutagen 
 testing, Amsterdam, Oxford, New York, Elsevier Science 
Publishers, pp. 275-291.

B., & WOLFF, S.  (1981)  Sister-chromatid exchanges: a report 
of the Gene-Tox Programme.  Mutat. Res., 87: 17-62.

WURGLER, F.E., & ZIMMERING, S.  (1983)  The sex-linked 
recessive lethal test for mutagenesis in  Drosophila 
 melanogaster:  a report of the US EPA Gene-Tox Program.  Mutat. 
 Res., 123: 183-279.

G.  (1976)  A continuous tumour-cell line from a human lung 
carcinoma with properties of type II alveolar epithelial 
cells.  Int. J. Cancer, 17: 62-70.

AMES, B.N.  (1982)  A new  Salmonella tester strain (TA102) 
with A.T base pairs at the site of mutation detects oxidative 
mutagens.  Proc. Natl Acad. Sci. (USA), 79: 7445-7449.

LINDSLEY, D.L. & GRELL, E.H.  (1968)   Genetic variations of 
 Drosophila melanogaster, Washington DC, Carnegie Institute of 
Washington, 472 pp (Publication No. 627).

The validity of the autoradiographic method for detecting DNA 
repair synthesis in rat hepatocyte primary culture.  Mutat. 
 Res., 113: 145-160.

LOPRIENO, N.  (1981)  Screening of coded carcinogenic 
non-carcinogenic chemicals by a forward mutation system with 
the yeast  Schizosaccharomyces pombe.  In: de Serres, F.J. & 
Ashby, J., ed.  Evaluation of short-term tests for carcinogens, 
Amsterdam, Oxford, New York, Elsevier Science Publishers 
(Progress in Mutation Research, Vol. 1).

ROSELLINI, D., & ROSSI, A.M.  (1976)  Mutagenicity of 
industrial compounds: styrene and its possible metabolite 
styrene oxide.  Mutat. Res., 40: 317-324.

R.C.  (1983)  Testing of chemicals for mutagenic activity with 
 Schizosaccharomyces pombe:  a report of the US Environmental 
Protection Agency Gene-Tox Program.  Mutat. Res., 115: 215-233.

MA, T.H.  (1982a)   Tradescantia cytogenetic tests (root-tip 
mitotis, pollen mitosis, pollen mother-cell meiosis): a report 
of the US Environmental Protection Agency Gene-Tox Program. 
 Mutat. Res., 99: 293-302.

MA, T.H.  (1982b)   Vicia cytogenetic test for environmental 
mutagens: a report of the US Environmental Protection Agency 
Gene-Tox Program.  Mutat. Res., 99: 257-271.

MCCANN, J. & AMES, B.N.  (1976)  Detection of carcinogens as 
mutagens in the  Salmonella/microsome test: assay of 300 
chemicals: discussion.  Proc. Natl Acad. Sci. (USA), 73: 

MCCANN, J.E., CHOI, E., YAMASAKI, E., & AMES, B.N.  (1975)  
Detection of carcinogens as mutagens in the  Salmonella/microsome 
test: assay of 300 chemicals.  Proc. Natl  Acad. Sci. (USA), 
72: 5135-5139.

MACKEY, B.E. & MACGREGOR, J.T.  (1979)  The micronucleus test: 
statistical design and analysis.  Mutat. Res., 64: 195-204.

MCMAHON, R.E., CLINE, J.C., & THOMPSON, C.Z.  (1979) Assay of 
855 test chemicals in ten tester strains using a new 
modification of the Ames test for bacterial mutagens.  Cancer 
 Res., 39: 682-693.

MALLING, H.V.  (1971)  Dimethylnitrosamine: formation of 
mutagenic compounds by interaction with mouse liver 
microsomes.  Mutat. Res., 13: 425-429.

MANN, H.B. & WHITNEY, D.R.  (1947)  On a test of whether one 
of two random variables is stochastically larger than the 
other.  Ann. Math. Stat., 18: 50-60.

MARON, C. & AMES, B.N.  (1984)  Revised methods for the 
 Salmonella mutagenicity test. In: Kilbey, B.J., Legator, M., 
Nichols, W., & Ramel, C., ed.  Handbook of mutagenicity test 
 procedures, Amsterdam, Oxford, New York, Elsevier Science 
Publishers, pp. 93-141.

Comparability of organic solvents with the 
 Salmonella/microsome test.  Mutat. Res., 88: 343-350.

MARTIN, C.N., MCDERMID, A.C., & GARNER, R.C.  (1978)  Testing 
of known carcinogens and non-carcinogens for their ability to 
induce unscheduled DNA synthesis in HeLa Cells.  Cancer Res., 
38: 2621-2627.

(1981)  Reverse mutation test on H2-coded compounds with the 
 E. coli WP2 system. In: de Serres, F.J. & Ashby, J., ed. 
 Evalutation of short-term tests for carcinogens, Amsterdam, 
Oxford, New York, Elsevier Science Publishers, pp. 287-395 
(Progress in Mutation Research, Vol. 1).

MATTER, B. & SCHMID, W.  (1971)  Trenimon-induced chromosomal 
damage in bone-marrow cells of six mammalian species, 
evaluated by the micronucleus test.  Mutat. Res., 12: 417-425.

MEHTA, R.D. & VON BORSTEL, R.C.  (1981)  Mutagenic activity of 
42 encoded compounds in the haploid reversion assay strain 
XV185-14C. In: de Serres, F.J. & Ashby, J., ed.  Evaluation of 
 short-term tests for carcinogens, Amsterdam, Oxford, New York, 
Elsevier Science Publishers, pp. 414-423 (Progress in Mutation 
Research, Vol. 1).

MEYER, A.L.  (1983)   In vitro transformation assays for 
chemical carcinogens.  Mutat. Res., 115: 323-338.

MILLER, E.C. & MILLER, J.A.  (1966)  Mechanisms of chemical 
carcinogenesis: nature of proximate carcinogens and 
interactions with macromolecules.  Pharmacol. Rev., 18: 805-838.

MILLER, E.C. & MILLER, J.A.  (1971)  The mutagenicity of 
chemical carcinogens: correlations, problems, and 
interpretation. In: Hollaender, A., ed.  Chemical mutagens: 
 principles and methods for their detection, New York, London, 
Plenum Press, pp. 83-119.

MIRSALIS, J.C. & BUTTERWORTH, B.E.  (1980)  Detection of 
unscheduled DNA synthesis in hepatocytes from rats treated 
with genotoxic agents: an  in vivo - in vitro assay for 
potential carcinogens and mutagens.  Carcinogenesis, 1: 621-625.

SAN, R.H.C., WILLIAMS, G.M., & VON HALLE, E.S.  (1983)  
Unscheduled DNA synthesis tests: a report of the US Environmental 
Protection Agency Gene-Tox Program.  Mutat. Res., 123: 363-410.

MITELMAN, F.  (1983)  Catalogue of chromosome aberrations in 
cancer.  Cytogenet. cell Genet., 36: Nos. 1-2.

(1979)   Handling chemical carcinogens in the laboratory: 
 problems of safety, Lyons, International Agency for Research 
on Cancer (IARC Scientific Publications No. 33).

MRC  (1981)   Guidelines for work with chemical carcinogens in 
 Medical Research Council establishments, London, Medical 
Research Council.

NCI  (1981)   Safety standards for research involving chemical 
 carcinogens, National Cancer Institute, US Department of 
Health, Education and Welfare (US DHEW Publication No. (NIH) 

NEEL, J.V. & SCHULL, W.J.  (1958)   Human heredity, Chicago, 
The University of Chicago Press.

OAKBERG, E.F.  (1960)  Irradiation damage to animals and its 
effect on their reproductive capacity.  J. dairy Sci., 
43(Suppl.): 54-67.

OBE, G., NATARAJAN, A.T., & PALITTI, F.  (1982)  Role of DNA 
double-strand breaks in the formation of radiation-induced 
chromosomal aberrations. In: Natarajan, A.T., Obe, G., & 
Altmann, H., ed.  DNA repair, chromosome alterations and 
 chromatin structure, Amsterdam, Oxford, New York, Elsevier 
Science Publishers, pp. 1-9 (Progress in Mutation Research, 
Vol. 4).

OECD  (1982a)   Good laboratory practice in the testing of 
 chemicals, Paris, Organisation for Economic Cooperation and 
Development (ISBN 92-64-12367-9).

OECD  (1982b)   Decision of the Council concerning the minimum 
 premarketing set of data in the assessment of chemicals, 
Paris, Organisation for Economic Cooperation and Development.

OECD  (1984)   Data interpretation guides for initial hazard 
 assessment of chemicals (provisional), Paris, Organisation for 
Economic Cooperation and Development, pp. 58-60.

PARRY, J.M. & PARRY, E.M.  (1983)  The detection of induced 
chromosome aneuploidy using strains of the yeast  Saccharomyces 
 cerevisiae.  In: Kilbey, B.J., ed.  Handbook of mutagenicity 
 test procedures, Amsterdam, Oxford, New York, Elsevier Science 
Publishers, Vol. 2.

PARRY, J.M. & SHARP, D.C.  (1981)  Induction of mitotic 
aneuploidy in the yeast strain D6 by 42 coded compounds. In: 
de Serres, F.J. & Ashby, J., ed.  Evaluation of short-term 
 tests for carcinogens, Amsterdam, Oxford, New York, Elsevier 
Science Publishers, pp. 468-480 (Progress in Mutation 
Research, Vol. 1).

PARRY, J.M. & WILCOX, P.  (1982)  The genetic toxicology in 
fungi of 4CMB, 4HMB, and BC. Survey of the results of the 
UKEMS collaborative genotoxicity trial 1981.  Mutat. Res., 100: 

PARRY, J.M. & ZIMMERMANN, F.K.  (1976)  The detection of 
monosomic colonies produced by mitotic non-disjunction in the 
yeast  Saccharomyces cerevisiae. Mutat. Res., 36: 49-66.

PARRY, J.M., TWEATS, D.J., & AL-MOSSAWI, M.A.J.  (1976)  
Monitoring the marine environment for mutagens.  Nature 
 (Lond.), 264: 538-540.

PARRY, J.M., PARRY, E.M., & BARRETT, J.C.  (1980)  Tumour 
promotors induce mitotic aneuploidy in yeast.  Nature (Lond.), 
294: 263-265.

PERRY, P.E.  (1980)  Chemical mutagens and sister-chromatid 
exchange. In: de Serres, F.J. & Hollaender, A., ed.  Chemical 
 mutagens: principles and methods for their detection,  
New York, London, Plenum Press, Vol. 6, pp. 1-39.

PERRY, P.E. & THOMSON, E.J.  (1984)  Sister-chromatid exchange 
methodology. In: Kilbey, B.J., et al., ed.  Handbook of 
 mutagenicity test procedures, 2nd ed., Amsterdam, Oxford, 
New York, Elsevier Science Publishers, pp. 495-529.

PERRY, P.E. & WOLFF, S.  (1974)  New Giemsa method for the 
differential staining of sister chromatids.  Nature (Lond.), 
251: 156-158.

PLEWA, M.J.  (1982)  Specific-locus mutation assays in  Zea 
 mays:  a report of the US Environmental Protection Agency 
Gene-Tox Program.  Mutat. Res., 99: 317-377.

PMAA  (1976)   Current good laboratory practices in drug safety 
 evaluation in laboratory animals: guidelines, Washington DC, 
Pharmaceutical Manufacturers Association of America.

(1981)  Mammalian  in vivo and  in vitro cytogenetic assays: a 
report of the US EPA's Gene-Tox Program.  Mutat. Res., 87: 

evaluation of six short-term tests for detecting organic 
chemical carcinogens.  Br. J. Cancer, 37: 873-959.

RASSMUSEN, R.E. & PAINTER, R.B.  (1964)  Evidence for repair 
of ultraviolet damaged deoxyribonucleic acid in cultured 
mammalian cells.  Nature (Lond.), 203: 1360-1362.

READ, J.  (1959)   Radiation biology of Vicia Faba in relation 
 to the general problem, Springfield, Illinois, 
Charles C. Thomas.

REGAN, J.D. & SETLOW, R.B.  (1974)  Two forms of repair in the 
DNA of human cells damaged by chemical carcinogens and 
mutagens.  Cancer Res., 34: 3318-3325.

REVELL, S.H.  (1959)  The accurate estimation of chromatid 
breakage and its relevance to a new interpretation of 
chromatid aberrations induced by ionizing radiation.  Proc. 
 Royal Soc. Lond. Ser. B, 150: 563-589.

RINKUS, S.J. & LEGATOR, M.S.  (1979)  Chemical 
characterization of 465 known or suspected chemical 
carcinogens and their correlation with mutagenic activity in 
the  Salmonella typhimurium system.  Cancer Res., 39: 3289-3318.

ROWLAND, I.R., RUBERY, E.D., & WALKER, R.  (1984)  Bacterial 
assays for mutagens in food. In: Dean, B.J., ed.  UKEMS 
 Sub-Committee on Guidelines for Mutagenicity Testing. Part II. 
 Supplementary Tests, Swansea, United Kingdom Environmental 
Mutagen Society.

RUSSELL, L.B.  (1962)  Chromosome aberrations in experimental 
mammals. In: Steinberg, A.G. & Bearn, A.G., ed.  Progress in 
 medical genetics, New York, London, Grune and Stratton, Vol. 
2, pp. 230-294.

RUSSELL, L.B.  (1978)  Somatic cells as indicators of germinal 
mutations in the mouse.  Environ. Health Perspect., 24: 113-116.

RUSSELL, L.B. & RUSSELL, W.L.  (1956)  The sensitivity of 
different stages in oogenesis to the radiation induction of 
dominant lethals and other changes in the mouse. In: Mitchell, 
J.S., Holmes, B.E., & Smith, C.L., ed.  Progress in 
 radiobiology, Edinburgh, Oliver and Boyd, pp. 187-192.

RUSSELL, W.L., RUSSELL, L.B., & KIMBALL, A.W.  (1954)  The 
relative effectiveness of neutrons from a nuclear detonation 
and from a cyclotron in inducing dominant lethals in the 
mouse.  Am. Nat., 88: 269-286.

simple cytochemical technique for demonstration of DNA in 
cells infected with mycoplasmas and viruses.  Nature (Lond.), 
253: 461-462.

SALAMONE, M., HEDDLE, J., STUART, E., & KATZ, M.  (1980)  
Towards an improved micronucleus test. Studies on 3 model 
agents mitomycin C, cyclophosphamide, and dimethylbenzanthracene. 
 Mutat. Res., 74: 347-356.

SAN, R.H.C. & STICH, H.F.  (1975)  DNA repair synthesis of 
cultured human cells as a rapid bioassay for chemical 
carcinogens.  Int. J. Cancer, 16: 284-291.

SANKARANARAYANAN, K.  (1982)  Determination and evaluation of 
genetic risk to humans from exposure to chemicals. In: Bora, 
K.C., Douglas, G.R., & Nestmann, E.R., ed.  Chemical 
 mutagenesis, human population monitoring and genetic risk 
 assessment, Amsterdam, Oxford, New York, Elsevier Science 
Publishers, pp. 289-321 (Progress in Mutation Research, Vol. 

SAVAGE, J.R.K.  (1976)  Annotation: Classification and 
relationships of induced chromosomal structural changes. 
 J. med. Genet., 13: 103-122.

SAX, K.  (1938)  Chromosome aberrations induced by X-rays. 
 Genetics, 23: 494-516.

SCHAEFER, H.  (1939)  [The fertility of male mice after 
irradiation with 200r.]  Z. mikrosk. -anat. Forsch., 46: 
121-152 (in German).

SERRES, F.J.  (1978)  Exploratory monitoring of air pollutants 
for mutagenic activity with the  Tradescantia stamen hair 
system.  Environ. Health Perspect., 27: 51-66.

SCHMID, W.  (1976)  The micronucleus test for cytogenetic 
analysis. In: Hollaender, A., ed.  Chemical mutagens: 
 principles and methods for their detection, New York, London, 
Plenum Press, Vol. 4, pp. 31-53.

(1971)  Chemical mutagenesis. The Chinese hamster bone marrow 
as an  in vivo test system. I. Cytogenetic results on basic 
aspects of the methodology, obtained with alkylating agents. 
 Humangenetik, 11: 103-118.

SCHMIDT, G. & THANNHAUSER, S.  (1945)  A method for the 
determination of deoxyribonucleic acid, ribonmucleic acid, and 
phosphoproteins in animal tissues.  J. Biol. Chem., 161: 83-89.

RICHARDSON, C.R.  (1983)  Chromosome aberration assays in 
mammalian cells  in vitro.  In: Dean, B.J., ed.  Report of the 
 UKEMS Sub-Committee on Guidelines for Mutagenicity Testing, 
Swansea, United Kingdom Environmental Mutagen Society, 
pp. 43-64.

SEARLE, A.G.  (1975)  The specific locus test in the mouse. 
 Mutat. Res., 31: 277-290.

SEARLE, A.G. & BEECHEY, C.V.  (1974)  Sperm-count, 
egg-fertilization, and dominant lethality after X-irradiation 
of mice.  Mutat. Res., 22: 63-72.

SETCHELL, B.P.  (1970)  Testicular blood supply, lymphatic 
drainage, and secretion of fluid. In: Johnson, A.D., Gomes, 
W.R., & VanDemark, N.L., ed.  The testis, New York, London, 
Academic Press, Vol. 1, pp. 101-239.

SETLOW, R.B. & CARRIER, W.L.  (1964)  The disappearance of 
thymine dimers from DNA: an error correcting mechanism.  Proc. 
 Natl Acad. Sci. (USA), 51: 226-231.

SHARP, D.C. & PARRY, J.M.  (1981)  Induction of mitotic gene 
conversion by 41 coded compounds using the yeast culture JD1. 
In: de Serres, F.J. & Ashby, J., ed.  Evaluation of short-term 
 tests for carcinogens, Amsterdam, Oxford, New York, Elsevier 
Science Publishers, pp. 491-501 (Progress in Mutation 
Research, Vol. 1).

SIMINOVITCH  (1976)  On the nature of hereditable variation in 
cultured somatic cells.  Cells, 7: 1-11.

SIRIANNI, S.R., FURUKAWA, M., & HUANG, C.C.  (1979)  Induction 
of 8-azagucnine and ouabina-resistant mutants by 
cyclophosphamide and 1-(pyridyl-3)-3, d-dimehytyltriazie in 
Chinese hamster cells cultured in diffusion chambers in mice. 
 Mutat. Res., 64: 259-267.

SNEDECOR, G.W. & COCHRAN, W.G.  (1967)   Statistical methods, 
6th ed., Ames, Iowa State University Press, pp. 240-242.

SNEE, R.D. & IRR, J.D.  (1981)  Design of a statistical method 
for the analysis of mutagenesis at the hypocanthine-guanine 
phosphoribosyl transferase locus of cultured Chinese hamster 
ovary cells.  Mutat. Res., 85: 77-93.

SOKAL, R.R. & ROHLF, F.J.  (1969)   Biometry, San Franciso, 
California, W.H. Freeman and Co.

SORA, S., LUCCHINI, G., & MAGNI, G.E.  (1982)  Meiotic diploid 
progeny and meiotic non-disjunction in  Saccharomyces 
 cerevisiae. Genetics, 101: 17-33.

STADLER, L.J.  (1932)  On the genetic nature of induced 
mutations in plants.  Proc. VI Int. Congr. Genet., 1: 274-294.

(1969)  The mutagenic effect of low concentrations of ethylene 
oxide in air.  Hereditas, 62: 264.

TAN, E.-L. & HSIE, A.W.  (1981)  Mutagenicity and cytotoxicity 
of haloethanes as studied in the CHO/HGPRT system.  Mutat. 
 Res., 90: 183-191.

(1980)  A micronucleus technique for detecting clastogenic 
effects of mutagens/carcinogens (DEN, DMN) in hepatocytes of 
rats  in vivo. Mutat. Res., 74: 11-20.

THOMPSON, L.H. & BAKER, R.M.  (1973)  Isolation of mutants of 
cultured mammalian cells. In: Prescott, D.M., ed.  Methods in 
 cell biology. IV, New York, London, Academic Press, pp. 209-28.

TJIO, J.H. & WHANG, J.  (1965)  Direct chromosome preparations 
of bone marrow cells. In: Yunis, J., ed.  Human chromosome 
 methodology, New York, London, Academic Press, pp. 51-56.

TOMKINS, D.J. & GRANT, W.F.  (1976)  Monitoring natural 
vegitation for herbicide-induced chromosomal aberrations. 
 Mutat. Res., 36: 73-84.

UNIVERSITY OF BIRMINGHAM  (1980)   Rules and notes of guidance 
 for the use of chemical carcinogens in the university, 
Birmingham, The University of Birmingham.

US EPA  (1976)   An ordering of the NIOSH suspected carcinogens 
 list based only on data contained in the list BP-251851, 
Washington DC, US Environmental Protection Agency.

US FDA  (1979)   GLP regulations: management briefings, 
Washington DC, US Food and Drug Administration 
(Post-conference Report).

US FDA  (1982)   GLP regulations: questions and answers, 
Washington DC, US Food and Drug Administration.

US NAS  (1983)   Risk assessment in the federal government: 
 managing the process, Washington DC, National Academy of 
Sciences, National Research Council.

VAN'T HOF, J. & SCHAIRER, L.A.  (1982)   Tradescantia assay 
systems for gaseous mutagens: a report of the US Environmental 
Protection Agency Gene-Tox Program.  Mutat. Res., 99: 303-315.

VENITT, S. & CROFTON-SLEIGH, C.  (1981)  Mutagenicity of 42 
coded compounds in a bacterial assay using  Escherichia coli  
and  Salmonella typhimurium.  In: de Serres, F.J. & Ashby, J., 
ed.  Evaluation of short-term tests for carcinogens, Amsterdam, 
Oxford, New York, Elsevier Science Publishers, pp. 351-360 
(Progress in Mutation Research, Vol. 1) 

VENITT, S., FORSTER, R., & LONGSTAFF, E.  (1983)  Bacterial 
mutation assays. In: Dean, B.J., ed.  Report of the UKEMS 
 Sub-Committee on Guidelines for Mutagenicity Testing. Part 1. 
 Basic Test Battery; Minimal Criteria; Professional Standards; 
 Interpretation; Selection of Supplementary Assays, Swansea, 
United Kingdom Environmental Mutagen Society, pp. 5-40.

VIG, B.K.  (1982)  Soybean (Glycine max (L) merrill) as a 
short-term assay for study of environmental mutagens: a report 
of the US Environmental Protection Agency Gene-Tox Program. 
 Mutat. Res., 99: 339-347.

VOGEL, E. & NATARAJAN, A.T.  (1979)  The relation between 
reaction kinetics and mutagenic action of mono-functional 
alkylating agents in higher eukaryotic systems. I. Recessive 
lethal mutations and translocations in  Drosophila. Mutat. 
 Res., 62: 51-100.

J.A.  (1980)  Some current perspectives of the application of 
 Drosophila in the evaluation of carcinogens: In: Williams, 
G.M., Kroes, R., Waaijen, H.W., & van de Poll, K.W., ed.  The 
 predictive value of short-term screening tests, Amsterdam, 
Oxford, New York, Elsevier Science Publishers, pp. 125-147.

VOGEL, E.W., ZIJLSTRA, J.A., & BLIJLEVEN, W.G.H.  (1983)  
Mutagenic activity of selected aromatic amines and polycyclic 
hydrocarbons in  Drosophila melanogaster. Mutat. Res., 107: 

VOLLMAR, J.  (1977)  Statistical problems in mutagenicity 
tests.  Arch. Toxicol., 38: 13-25.

WALTER, E.  (1951)  [A few nonparametric test procedures.] 
 Mathem. Statist., 3: 31-44 (in German).

C., & PENMAN, M.  (1984)  Unscheduled DNA synthesis. In: Dean, 
B.J., ed.  UKEMS Sub-Committee for Guidelines on Mutagenicity 
 Testing. Part II. Supplementary Tests, Swansea, United Kingdom 
Environmental Mutagen Society.

WATSON, J.D. & CRICK, F.H.C.  (1953)  Genetical implications 
of the structure of DNA.  Nature (Lond.), 171: 964-967.

WILCOX, P. & PARRY, J.M.  (1981)  The genetic activity of 
dinitropyrenes in yeast: unusual dose-response curves for 
induced mitotic gene conversion.  Carcinogenesis, 2: 1201-1208.

WILCOX, P., DANFORD, N., & PARRY, J.M.  (1982)  The genetic 
activity of dinitropyrenes in eukaryotic cells. In:  Mutagens 
 in our environment, New York, Alan R. Liss, pp. 249-258.

WILKIE, D. & GOONESKERA, S.  (1980)  The yeast mitochondrial 
system in carcinogen testing.  Chem. Ind., 21: 847-850.

WILLIAMS, G.M.  (1977)  Detection of chemical carcinogens by 
unscheduled DNA synthesis in rat liver primary cell culture. 
 Cancer Res., 37: 1845-1851.

WILLIAMS, G.M.  (1980)  Batteries of short-term tests for 
carcinogen screening. In: Williams, G.M., Kroes, R., Waaijers, 
H.W., & van de Poll, K.W., ed.  The predictive value of 
 short-term screening tests for carcinogenicity evaluation, 
Amsterdam, Oxford, New York, Elsevier Science Publishers, 
pp 327-346.

WOLFF, S., ed.  (1982)   Sister chromatid exchange, New York, 
John Wiley and Sons.

WURGLER, F.E., GRAF, U., & BERCHTHOLD, W.  (1975)  Statistical 
problems connected with the sex-linked recessive lethal test 
in  Drosophila melanogaster.  I. The use of the Kastenbaum- 
Bowman test.  Arch. Genet., 48: 158-178.

WURGLER, F.E., SOBELS, F.H., & VOGEL, E.  (1977)   Drosophila: 
as assay system for detecting genetic changes. In: Kilbey et 
al., ed.  Handbook of mutagenicity test procedures, Amsterdam, 
Oxford, New York, Elsevier Science Publishers, pp 335-373.

T., & OKADA, M.  (1977)  Mutagenicities of  N-nitrosamines on 
 Salmonella. Mutat. Res., 48: 121-130.

YAMAMOTO, K. & KIKUCHI, Y.  (1980)  A comparison of diamaters 
of micronuclei induced by clastogens and by spindle poisons. 
 Mutat. Res., 71: 127-131.

ZIJLSTRA, J.A. & VOGEL, E.W.  (1984)  Mutagenicity of 
7,12-dimethylbenz( a)anthracene and some aromatic mutagens in 
 Drosophila melanogaster. Mutat. Res., 125: 243-261.

ZIJLSTRA, J.A., VOGEL, E.W., & BREIMER, D.D.  (1984)  
Strain-differences and inducibility of microsomal oxidative 
enzymes in  Drosophila melanogaster flies.  Chem.-biol. 
 Interact., 48: 317-338.

ZIMMERMANN, F.K.  (1975)  Procedures used in the induction of 
mitotic recombination and mutation in the yeast  Saccharomyces 
 cerevisiae. Mutat. Res., 31: 71-86.

ZIMMERMANN, F.K. & SCHEEL, I.  (1981)  Induction of mitotic 
gene conversion in strain D7 of  Saccharomyces cerevisiae by 42 
coded chemicals. In: de Serres, F.J. & Ashby, J., ed. 
 Evaluation of short-term tests for carcinogens, Amsterdam, 
Oxford, New York, Elsevier Science Publishers, pp. 481-490 
(Progress in Mutation Research, Vol. 1).

ZIMMERMANN, F.K., KERN, R., & ROSENBERGER, H.  (1975)  A yeast 
strain for the simultaneous detection of induced crossing- 
over, mitotic gene conversion, and reverse mutation.  Mutat. 
 Res., 28: 381-388.

ZIMMERMANN, F.K., MAYER, V.W., & PARRY, J.M.  (1982)  Genetic 
toxicology studies using  Saccharomyces cerevisiae. J. appl. 
 Toxicol., 2: 1-10.

N.  (1984)  Testing of chemicals for genetic activity with 
 Saccharomyces cerevisiae:  report of the US Environmental 
Protection Agency Gene-Tox Program.  Mutat. Res., 133: 199-244.

    See Also:
       Toxicological Abbreviations