INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY
ENVIRONMENTAL HEALTH CRITERIA 80
PYRROLIZIDINE ALKALOIDS
This report contains the collective views of an international group of
experts and does not necessarily represent the decisions or the stated
policy of either the World Health Organization or the United Nations
Environment Programme
Published under the joint sponsorship of
the United Nations Environment Programme
and the World Health Organization
World Health Organization
Geneva 1988
ISBN 92 4 154280 2
(c) World Health Organization 1988
Publications of the World Health Organization enjoy copyright
protection in accordance with the provisions of Protocol 2 of the
Universal Copyright Convention. For rights of reproduction or
translation of WHO publications, in part or in toto, application
should be made to the Office of Publications, World Health
Organization, Geneva, Switzerland. The World Health Organization
welcomes such applications.
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 letters.
CONTENTS
ENVIRONMENTAL HEALTH CRITERIA FOR PYRROLIZIDINE ALKALOIDS
PREFACE
INTRODUCTION - PYRROLIZIDINE ALKALOIDS AND HUMAN HEALTH
1. SUMMARY AND RECOMMENDATIONS
1.1. Summary
1.2. Sources and chemical structure
1.3. Mechanisms and features of toxicity
1.4. Effects on man
1.4.1. Nature and extent of health risks
1.5. Methods for prevention
1.6. Recommendations
1.6.1. General recommendations
1.6.2. Recommendations for research
2. PROPERTIES AND ANALYTICAL METHODS
2.1. Chemical structure and properties
2.2. Analytical methods
2.2.1. Extraction
2.2.1.1 Plant tissue
2.2.1.2 Biological fluids and tissues
2.2.2. Analysis for pyrrolizidine alkaloids
2.2.2.1 Thin-layer chromatography (TLC)
2.2.2.2 High-performance liquid chromatography
(HPLC)
2.2.2.3 Gas chromatography (GC) and mass
spectrometry (MS)
2.2.2.4 Nuclear magnetic resonance (NMR)
spectrometry
2.2.2.5 The Ehrlich reaction
2.2.2.6 Indicator dyes
2.2.2.7 Direct weighing
2.3. Determination of metabolites in animal tissues
3. SOURCES AND PATHWAYS OF EXPOSURE
3.1. Hepatotoxic pyrrolizidine alkaloids and their sources
3.2. Pneumotoxic and other toxic pyrrolizidine alkaloids
3.3. Pathways of exposure
3.3.1. Contamination of staple food crops
3.3.2. Herbal infusions
3.3.3. Use of PA-containing plants as food
3.3.4. Contaminated honey
3.3.5. Milk
3.3.6. Meat
3.3.7. Use of PAs as chemotherapeutic agents for cancer
4. METABOLISM
4.1. Absorption, excretion, and tissue distribution
4.1.1. Absorption
4.1.2. Excretion and distribution
4.2. Metabolic routes
4.2.1. Hydrolysis
4.2.2. N-oxidation
4.2.3. Conversion to pyrrolic metabolites
4.3. Effects of treatments affecting metabolism
4.4. Other factors affecting metabolism
4.5. Other metabolic routes
4.6. Metabolism of pyrrolizidine N-oxides
4.7. Metabolism in man
5. MECHANISMS OF TOXICITY AND OTHER BIOLOGICAL ACTIONS
5.1. Metabolites responsible for toxicity
5.1.1. Metabolic basis of toxicity
5.1.2. Isolation of pyrrolic metabolites
5.1.3. Chemical aspects of pyrrolic metabolites
5.1.3.1 Preparation
5.1.3.2 Chemistry associated with toxic actions
5.1.4. Possible further metabolites
5.2. Toxic actions of pyrrolic metabolites
5.2.1. Animals
5.2.1.1 Pyrrolic esters (dehydro-alkaloids)
5.2.1.2 Pyrrolic alcohols (dehydro-necines)
5.2.2. Cell cultures
5.2.3. Possible participation of membrane lipid
peroxidation
5.3. Chemical and metabolic factors affecting toxicity
5.3.1. Structural features of a toxic alkaloid
5.3.2. Activation and detoxication
5.3.3. Factors affecting the toxicity of active
metabolites
5.3.3.1 Reactivity of the metabolite
5.3.3.2 The number of reactive groups
5.4. Metabolites associated with the biological actions of
pyrrolizidine alkaloids
5.4.1. Acute hepatotoxicity
5.4.2. Chronic hepatotoxicity
5.4.3. Pneumotoxicity
5.4.4. Toxicity in other tissues
5.4.5. Carcinogenicity
5.4.6. Antitumour activity
5.5. Prevention and treatment of pyrrolizidine poisoning
5.5.1. Modified diets
5.5.2. Pre-treatment to enhance the detoxication of active
metabolites
5.5.3. Other treatments
6. EFFECTS ON ANIMALS
6.1. Patterns of disease caused by different plant genera and
of organ involvement in different species
6.2. Field observations - outbreaks in farm animals
6.3. Studies on farm animals
6.4. Experimental animal studies
6.4.1. Effects on the liver
6.4.1.1 Relative hepatotoxicity of different PAs
and their N-oxides
6.4.1.2 Factors affecting hepatotoxicity
6.4.1.3 Acute effects
6.4.1.4 Mechanism of toxic action
6.4.1.5 Chronic effects
6.4.2. Effects on the lungs
6.4.2.1 Acute effects
6.4.2.2 Chronic effects
6.4.2.3 Mechanisms of toxic action
6.4.3. Effects on the central nervous system
6.4.4. Effects on other organs
6.4.5. Teratogenicity
6.4.6. Fetotoxicity
6.4.7. Mutagenicity
6.4.7.1 Chromosome damage
6.4.8. Carcinogenesis
6.4.8.1 Purified alkaloids
6.4.8.2 Plant materials
6.4.8.3 Pyrrolizidine alkaloid metabolites and
analogous synthetic compounds
6.4.8.4 Molecular structure and carcinogenic
activity
6.4.9. Antimitotic activity
6.4.10. Immunosuppression
6.4.11. Effects on mineral metabolism
6.4.12. Methods for the assessment of chronic
hepatotoxicity and pneumotoxicity
6.5. Effects on wild-life
6.5.1. Deer
6.5.2. Fish
6.5.3. Insects
7. EFFECTS ON MAN
7.1. Clinical features of veno-occlusive disease (VOD)
7.2. Salient pathological features of veno-occlusive disease
7.3. Human case reports of veno-occlusive disease
7.4. VOD and cirrhosis of the liver
7.5. Differences between VOD and Indian childhood cirrhosis
(ICC)
7.6. Chronic lung disease
7.7. Trichodesma poisoning
7.8. Relationship between dose level and toxic effects
7.9. Pyrrolizidine alkaloids as a chemotherapeutic agent for
cancer
7.10. Prevention of poisoning in man
8. BIOLOGICAL CONTROL
9. EVALUATION OF HUMAN HEALTH RISKS AND EFFECTS ON THE ENVIRONMENT
9.1. Human exposure conditions
9.1.1. Reported sources of human exposure
9.1.2. Plant species involved
9.1.3. Modes and pathways of exposure
9.1.3.1 Contamination of grain crops
9.1.3.2 Herbal medicines
9.1.3.3 PA-containing plants used as food and
beverages
9.1.3.4 Other food contaminated by PAs
9.1.4. Levels of intake
9.2. Acute effects of exposure
9.2.1. Acute liver disease
9.3. Chronic effects of exposure
9.3.1. Cirrhosis of the liver
9.3.2. Mutagenicity and teratogenicity
9.3.3. Cancer of the liver
9.3.4. Effects on other organs
9.4. Effects on the environment
9.4.1. Agriculture
9.4.2. Wild-life
9.4.3. Insects
9.4.4. Soil and water
REFERENCES
APPENDIX I. PYRROLIZIDINE ALKALOIDS AND THEIR PLANT SOURCES
APPENDIX II.
NOTE TO READERS OF THE CRITERIA DOCUMENTS
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.
* * *
ENVIRONMENTAL HEALTH CRITERIA FOR PYRROLIZIDINE ALKALOIDS
A WHO Task Group on Environmental Health Criteria for
Pyrrolizidine Alkaloids met in Tashkent, USSR, on 1 - 5 December
1986. Dr M. Gounar opened the meeting on behalf of the three
co-sponsoring organizations of the IPCS (UNEP/ILO/WHO). The Task
Group reviewed and revised the draft criteria document and made an
evaluation of the health risks of exposure to pyrrolizidine
alkaloids.
Access to the original papers on the subject published in the
USSR was made possible by PROFESSOR M. ABDULLAHODJAEVA. DR A.R.
MATTOCKS wrote the first drafts of the sections on Properties and
Analytical Methods, Metabolism, and Mechanisms of Toxicity and
Other Biological Actions. DR C.C.J. CULVENOR, assisted PROFESSOR
H.D. TANDON in the finalization of the document after the Task
Group meeting. Dr J. Parizek, who was originally the IPCS staff
member responsible for the preparation of the document, and was to
be Secretary of the Task Group, could not attend the meeting
because of sudden illness, and the Task Group was assisted in his
place by Dr M. Gounar, former IPCS staff member. Dr A. Prost was
responsible for the final version of the document.
The Secretariat acknowledge the help of both Professor H.D.
Tandon and Dr C.C.J. Culvenor. The Task Group meeting in Tashkent
was organized by the Centre of International Projects, USSR State
Committee for Science and Technology.
The efforts of all who helped in the preparation and
finalization of the document are gratefully acknowledged.
* * *
Partial financial support for the publication of this criteria
document was kindly provided by the United States Department of
Health and Human Services, through a contract from the National
Institute of Environmental Health Sciences, Research Triangle Park,
North Carolina, USA - a WHO Collaborating Centre for Environmental
Health Effects.
* * *
A comprehensive data base on pyrrolizidine alkaloids has been
made available by CSIRO Division of Animal Health, Private Bag
No. 1, Parkville, Vic. 3052, Australia. The data base consists of
alkaloid occurrence tables and keyworded bibliography readable by
SCI-MATE software system (Bibliographic Manager, Institute for
Scientific Information), but adaptable to other systems. It is
available from CSIRO on IBM - PC diskettes; price on application to
L.W. Smith.
PREFACE
A disease caused by the consumption of plants containing
pyrrolizidine alkaloids (PAs) has been recognized independently as
an endemic disease in certain parts of the West Indies and in
Uzbekistan in the USSR. Outbreaks of the disease have affected
significant segments of populations or large numbers of people in
geographically confined areas in Afghanistan, India, and
Uzbekistan. The outbreaks have been caused through contamination
of the staple food crops with the seeds of plants containing PAs,
growing among the crops; such plants are likely to thrive following
periods of drought.
It is notable that the same family of plants that caused
endemic disease and large-scale outbreaks in Uzbekistan also caused
another outbreak of the disease in adjacent Afghanistan, long after
the chemical etiology of the disease (through consumption of toxic
seeds in the food) had been identified in the USSR. This happened
because there was a lack of general awareness of the causal
relationship between the chemical present in the plant and the
disease. Sporadic cases continue to occur in different parts of
the world through the consumption of seeds or plant parts
containing toxic PAs, as home remedies, beverages, or food.
The IPCS recognized that this was a health problem that might
be lethal, and that it was entirely preventable, provided that it
was recognized in time. It was also recognized that the
dissemination of knowledge, about both the disease and the sources
of the chemicals involved, would be a critical step in its
prevention.
Accordingly, the IPCS invited Professor H.D. Tandon, who was
responsible for establishing such a causal relationship in the
outbreaks in Afghanistan and India, to prepare a draft criteria
document and to assist in its further development and finalization
after the Task Group meeting, which was held in Tashkent, USSR, on
1 - 5 December, 1986.
In most episodes of toxic human disease caused by PAs, the
liver has been the principal target organ, except for an outbreak
in the USSR caused by Trichodesma alkaloids, in which the symptoms
were mostly extra-hepatic. The Environmental Health Criteria
document provides comprehensive coverage of the hepatotoxic PAS,
but lack of relevant documentation prevented the Task Group from
analysing the role of Trichodesma alkaloids in detail.
INTRODUCTION - PYRROLIZIDINE ALKALOIDS AND HUMAN HEALTH
Pyrrolizidine alkaloids (PAs) are found in plants growing in
most environments and all parts of the world. The main sources are
the families Boraginaceae (all genera), Compositae (tribes
Senecionae and Eupatoriae), and Leguminosae (genus Crotalaria), and
the potential number of alkaloid-containing species is as high as
6000, or 3% of the world's flowering plants (Culvenor, 1980). They
have long been known to be a health hazard for livestock, at least
since 1902 (Schoental, 1963), and loss of livestock in various
parts of the world has been traced to their grazing on certain
plants growing in pastures, especially following periods of drought
or in arid climates. They have been found to be toxic for all
species of animals tested (Schoental, 1963), though some species,
notably the guinea-pig, are resistant (Chesney & Allen, 1973a;
White et al., 1973). Human disease caused by PA toxicity has been
known to be endemic in the central Asian republics of the USSR, at
least since the early thirties (Ismailov, 1948a,b; Mnushkin, 1949)
when several outbreaks occurred, and the cause was discovered to be
the seeds of plants of Heliotropium species (Dubrovinskii, 1947,
1952; Khanin, 1948), which contaminated the staple food crops. A
spate of reports followed, mostly from the West Indies, of acute
and chronic liver disease (Bras et al., 1954, 1961; Bras & Hill,
1956; Stirling et al., 1962), associated with the ingestion by
people of herbal infusions for the treatment of certain ailments.
Schoental (1961) and Davidson (1963) suggested that, in view of the
evidence of the hepatotoxicity of PAs, consumption of plants
containing them could be of etiological significance in human liver
disease, especially in developing countries where they are consumed
as food or herbal medicines. In spite of this, and the fact that
such an ubiquitous source of toxic material is capable of producing
animal and human disease and that there have been more recent
reports, the PAs have not attracted much attention in the world as
a health hazard. In fact, a recent handbook on naturally occurring
toxic agents in food (Rechicigl, Jr, 1983) refers to them only in
passing and makes no mention of human disease caused by them.
Veno-occlusive disease (VOD) (Bras & Hill, 1956), which is
characterized by the dominant occlusive lesion of the centrilobular
veins of the liver lobule and is caused by these alkaloids, has
since been reported from all parts of the world, in both man and
animals (Hill, 1960; Bras, 1973). It has been attributed to the
accidental contamination of food by toxic plant products or the
ingestion of herbal infusions. There have been reports of stray
cases and of small outbreaks from both developing and developed
countries. However, in the most recent studies from Afghanistan
(Tandon & Tandon, 1975; Mohabbat et al., 1976; Tandon, B.N. et al.,
1978; Tandon, H.D. et al., 1978) and India (Tandon, B.N. et al.,
1976; Tandon, R.K. et al., 1976; Krishnamachari et al., 1977;
Tandon, H.D. et al., 1977; Tandon, B.N. et al., 1978), the disease
has been reported to affect large masses of the population,
resulting in high mortality, and has been attributed to the
accidental contamination of their staple food crops by PA-
containing seeds of plants, following periods of drought.
There is conclusive evidence from studies on experimental
animals that the effects of a single exposure to PAs may progress
relentlessly to advanced chronic liver disease and cirrhosis
(Schoental & Magee, 1957, 1959; Nolan et al., 1966), following a
long interval of apparent well-being, and without any other latent
or provocative factor (Schoental & Magee, 1959). The lowest levels
of such alkaloids administered thus far to experimental animals,
e.g., 1 - 4 mg/kg diet, have produced chronic liver disease and
tumours (Hooper & Scanlan, 1977; Culvenor & Jago, 1979).
Pyrrolizidine alkaloids have also been shown to act synergistically
with aflatoxin, another environmental toxin present in agricultural
products, in causing cirrhosis and hepatoma in primates (Lin et
al., 1974). Though there is no conclusive evidence yet of a
carcinogenic role of PAs in man, such a possibility has been
suspected on the basis of experimental data (Hill, 1960; Williams
et al., 1967; IARC, 1976, 1983; Huxtable, 1980; Culvenor, 1983),
and experimental studies have demonstrated carcinogenicity in rats
given dosages equivalent to those reported to have been ingested in
human cases (Cook et al., 1950; Culvenor, 1983).
Alkaloids/toxic metabolites have been shown to be secreted in
the milk of lactating dairy cattle (Dickinson et al., 1976) and
rats, and the young of both sexes have been shown to suffer toxic
damage, even when suckled by mothers treated with retrosine, who
apparently are not affected themselves (Schoental, 1959). Such
suckling animals may also be in apparent good health while the
livers show toxic effects. Protein-deficient and young suckling
animals are particularly vulnerable (Schoental, 1959).
Chromosomal aberrations have been demonstrated in rats and
humans with veno-occlusive disease (Martin et al., 1972).
Alkaloids have been found in the honey secreted by bees feeding
on the toxic plants (Deinzer et al., 1977). According to Culvenor
and his co-workers, populations in some countries are exposed to
low levels of alkaloids in commonly used foodstuffs, e.g., honey in
Australia (Culvenor et al., 1981; Culvenor, 1983, 1985) and comfrey
in many countries (Culvenor et al., 1980a; Culvenor, 1985).
Human cases of acute disease following the brief ingestion of
the alkaloids have been known to progress to cirrhosis (Stuart &
Bras, 1957; Braginskii & Bobokhadzaev, 1965; Stillman et al., 1977;
Tandon, B.N. et al., 1977; Tandon, H.D. et al., 1977) in as short a
period as 3 months from the acute phase (Stuart & Bras, 1957). The
initial disease may be cryptic (Braginskii & Bobokhadzaev, 1965)
and may not be ascribed to herbal consumption, and yet may progress
to cirrhosis (Huxtable, 1980). Veno-occlusive disease was stated
to be the most common cause of cirrhosis in infants in Jamaica
(Bras et al., 1961) and has been believed to be a significant
etiological factor for adult cirrhosis, especially in developing
countries (Gupta et al., 1963).
Plants known or suspected to contain toxic alkaloids are widely
used for medicinal purposes as home remedies all over the world,
without systematic testing for safety (Schoental, 1963; Smith &
Culvenor, 1981) and some are even used as food (Schoental & Coady,
1968; Culvenor, 1980). There are several reports of the continued
use of such herbs for medicinal purposes in technically advanced
countries (Culvenor, 1980). Senecio jacobaea continues to be sold
at herbalists shops in the United Kingdom (Schoental, 1963; Burns,
1972), and Symphytum spp. (comfrey) are still used as a vegetables,
beverages, or remedies (Mattocks, 1980). Both these herbs are
known to be carcinogenic (IARC, 1976; Hirono et al., 1978). Young
flower stalks of Petasites japonicus Maxim, the pre-bloom flower
of coltsfoot, Tussilago farfara, the leaf and root of comfrey,
Symphytum officinale, and the young leaves and stalks of Farfugium
japonicum and Senecio cannabifolius, which are all used in Japan
as human food or herbal remedies, are known to be carcinogenic for
rats (Hirono et al., 1983). Symphytum x uplandicum Nyman (Russian
comfrey), which contains several toxic PAs (Culvenor et al., 1980b)
echimidine and 7 acetylycopsamine being the main constituents, is
used as a salad plant, green drink, and medicinal herb. It has
been estimated that the rate of ingestion of alkaloids from this
herb may, over a period of time, exceed the levels reported to have
been taken during the Afghan outbreak. There is a report of at
least one patient who developed toxic effects as a result of
consuming a comfrey preparation (Culvenor et al., 1980a; Ridker et
al., 1985). Arseculeratne et al. (1981) found that 3 of the 50
medicinal herbs commonly used in Sri Lanka contained PAs that had
been proved to be hepatotoxic for animals. They suggested that
consumption of such herbs might contribute to the high incidence of
chronic liver disease, including primary liver cancer, in Asian and
African countries, especially as they may act synergistically with
aflatoxin and hepatitis B virus. The risk of toxic effects due to
these alkaloids may be particularly high in children (Schoental,
1959; Jago, 1970) and protein malnutrition, which exists in some
countries, may potentiate them (Schoental & Magee, 1957). Recent
studies from Hong Kong (Kumana et al., 1985; Culvenor et al.,
1986), the United Kingdom (McGee et. al, 1976; Ridker et al.,
1985), and the USA (Stillman et al., 1977; Fox et al., 1978; Ridker
et al., 1985) report instances of human disease that have been
caused by the use of such herbs, resulting in fatality or the
development of cirrhosis, even in countries with well-developed
health services and among the higher economic and educated strata
of society. Indeed, Stillman et al. (1977), from the USA, called PA
toxicosis the "iceberg disease", implying that cases of this
disease might be more frequent than reported in the USA, especially
among populations of Mexican-American origin. In general, the use
of herbal remedies is not elicited in the clinical history and
patients do not volunteer this information themselves.
Furthermore, the alkaloids are eliminated within 24 h (Huxtable,
1980) and, even though methods are available for their detection in
biological tissues and fluids, the suspicion cannot be confirmed,
as the symptoms may take several days or months to appear.
Contamination of food crops is particularly likely to occur in
parts of the world with arid climates, poor or uncertain rainfall,
poor irrigation facilities, and following periods of drought, all
of which promote the growth of the PA-containing plants that grow
as weeds among cultivated crops, as has been found in studies on
the outbreaks in Afghanistan, India, and the USSR (Terekhov, 1939;
Dubrovinskii, 1947; Ismailov, 1948a,b; Tandon & Tandon, 1975;
Mohabbat et al., 1976; Tandon, B.N. et al., 1976; Tandon, R.K. et
al., 1976; Tandon, H.D. et al., 1978) and in grazing pastures. The
use of traditional medicines is common in these countries and there
is insufficient awareness of this hazard, the disease condition,
and its diagnostic pathological picture. Furthermore, health
services are poorly developed. Thus, many of the cases or even
outbreaks may go unnoticed or unrecorded and may even be ascribed
to malnutrition (Lancet, 1984). Also, many of the reported cases
of so-called "Budd-Chiari syndrome", a condition associated with
obstruction of major hepatic veins and/or inferior vena cava, may
actually be cases of veno-occlusive disease (Sherlock, 1968), in
which only the central veins of the liver lobule or sublobular
veins are occluded.
Another type of PAs, Trichodesma alkaloids, has been known to
cause a human outbreak of disease in the USSR, through
contamination of the staple cereal with the seeds containing these
PAs; in this outbreak, the symptoms were principally extra-hepatic
(Ismailov et al., 1970).
This document is aimed at focusing on a health menace that is
insufficiently recognized, in order to evaluate the health risks on
the basis of published data, and to draft a set of recommendations
that would help in its recognition, prevention, and control.
1. SUMMARY AND RECOMMENDATIONS
1.1. Summary
The ingestion of pyrrolizidine alkaloids (PAs) in foods and
medicinal herbs results in acute and chronic effects in man,
affecting mainly the liver. Data from experimental animal studies
indicate that PAs represent a potential cause of cancer in man.
The alkaloids are produced by numerous plant species and occur
throughout the world. In the present document, the alkaloids and
their properties are described together with the sources of human
exposure and the diseases that they produce in man and animals.
The risks for human health are evaluated and recommendations are
made for reducing such risks.
1.2. Sources and Chemical Structure
The known pyrrolizidine alkaloids, most of which are
hepatotoxic, are produced by plant species within the following
families: Boraginaceae ( Heliotropium, Trichodesma, Symphytum, and
most other genera), Compositae ( Senecio, Eupatorium, and other
genera of the tribes Senecioneae and Eupatoriae), Leguminosae
(genus Crotalaria), and Scrophul-ariaceae (genus Castilleja).
These genera are mainly herbaceous and very widely distributed,
some species being found in most regions of the world. The
majority of the species within these genera have not yet been
investigated, but are expected to contain pyrrolizidine alkaloids.
The hepatotoxic alkaloids have a 1,2-double bond in the
pyrrolizidine ring and branched chain acids, esterifying a
9-hydroxyl and preferably also the 7-hydroxyl substituent. Modified
seco-pyrrolizidine alkaloids, in which the central bond between the
N and C8 atoms is broken, are also hepatotoxic. Some Senecio
species contain non-basic derivatives that are 5-oxopyrroles. The
toxicity of these derivatives may be similar to that of the
alkaloids, but this aspect has not been investigated. The
alkaloids occur as free bases and N-oxides. The latter are
reduced to the free bases in the gastrointestinal tract of animals
and have a similar toxicity when ingested orally.
Suitable analytical procedures are available for screening
plant species, including a simple field test for toxic alkaloids.
Thin-layer chromatography (TLC), high-performance liquid (HPLC),
gas chromatography (GC), and gas chromato-graphy-mass spectrometry
(GC-MS) have been applied for separating, characterizing, and
quantifying the alkaloids present. Effective use of these
procedures requires authentic alkaloids for standards, few of which
are available. Improved analytical methods are required for the
determination of very low levels of alkaloids in some foodstuffs.
1.3. Mechanisms and Features of Toxicity
The toxic effects of pyrrolizidine alkaloids are due to
activation in the liver. Metabolism of the alkaloids by mixed-
function oxidases leads to pyrrolic dehydro-alkaloids, which are
reactive alkylating agents. Reaction of initial metabolites with
constituents of the liver cell in which they are formed are
probably the main cause of liver cell necrosis. Metabolites are
released into the circulation and are believed to pass beyond the
liver to the lung causing vascular lesions characteristic of
primary pulmonary hypertension, especially when alkaloids, such as
monocrotaline, are administered to animals.
In experimental animals, PAs are quickly metabolized and are
almost completely excreted in 24 h, so that no residual products
are detectable in the biological fluids or body tissues after this
period.
The rate of formation of pyrrolic metabolites is influenced by
the induction or inhibition of the mixed-function oxidases in the
liver, but the relationship between the rate of metabolism and
expression of toxicity is uncertain.
Several pyrrolizidine alkaloid-derivatives and related
compounds are known to cause chromosome aberrations in plants,
leukocyte cell cultures of the marsupial (Potorus tridactylus),
and in hamster cell lines. Some pyrrolizidine alkaloids induce
micronuclei formation in erythrocytes in the bone marrow and fetal
liver in mice, sister chromatid exchanges in a Chinese hamster cell
line and human lymphocytes in vitro, and repair DNA synthesis in
rodent hepatocyte cell cultures. Chromosome aberrations have been
reported in the blood cells of children suffering from veno-
occlusive disease VOD, presumably caused by fulvine.
A number of pyrrolizidine alkaloids have been shown to be
mutagenic in the Salmonella typhimurium assay, after metabolic
activation. The carcinogenic activity of pyrrolizidine alkaloids
appears to parallel their mutagenic behaviour, but not their
hepatotoxicity.
Heliotrine at doses of 50 mg/kg body weight or more,
administered to rats during the second week of gestation, has been
shown to induce several abnormalities in the fetus. Doses of
200 mg/kg body weight resulted in intrauterine deaths or resorption
of fetuses. Dehydroheliotridine, the metabolic pyrrole derivative of
heliotrine, was 2.5 times more effective on a molar basis than its
parent PA in inducing teratogenic effects.
The ability of PAs to cross the placental barrier in the rat
and to induce premature delivery or death of litters has been
demonstrated. The embryo in utero appears to be more resistant to
the toxic effects of pyrrolizidine alkaloids than the neonate. PAs
are known to have passed through the mother's milk to the
sucklings.
Megalocytosis, the presence of enlarged hepatocytes containing
large, hyper-chromatic nuclei, is a characteristic feature of
pyrrolizidine alkaloid-induced chronic hepatotoxicity in
experimental animals. The enlarged hepatocytes arise through the
powerful antimitotic action of the pyrrole metabolites of
pyrrolizidine alkaloids. This change has not been observed in the
human liver, though human fetal liver cells in vitro culture
become enlarged when exposed to PAs, indicating susceptibility to
the antimitotic effect of the alkaloids.
In experimental animals, protein-rich and sucrose-only diets
have given some measure of protection against the effects of the
alkaloids, as has pre-treatment of animals with thiols, anti-
oxidants, or zinc chloride.
PAs are noted mainly for the poisoning of livestock due to the
animals grazing on PA-containing toxic weeds, and large-scale
outbreaks have been recorded. Such episodes have been reported
from most parts of the world, including those with temperate or
cold climates. Studies carried out on a wide variety of farm and
laboratory animals have revealed generally common features of
toxicity with some species variations. The liver is the principal
target organ. In small laboratory animals, doses approaching a
lethal dose produce a confluent, strictly zonal haemorrhagic
necrosis in the liver lobule, within 12 - 48 h of administration of
PAs. Simultaneously in non-human primates, or after a short time in
the rat, chicken, and swine, changes begin to occur, and later
become organized, in the subintima of the central or sublobular
veins in the liver resulting in their occlusion. The reticulin
framework in the central zone of the lobule collapses following
necrosis leading to scarring. Repeated administration of suitable
doses leads to chronic liver lesion characterized by megalocytosis,
and increasing fibrosis, which may result in cirrhosis. Chronic
liver disease including cirrhosis has been shown to develop in the
rat following administration of a single dose of a PA. In a number
of animal species, the lungs develop vascular lesions
characteristic of primary pulmonary hypertension with secondary
hypertrophy of the right ventricle of the heart. In rats,
appropriately low repeated doses of several alkaloids have been
shown to induce tumours, mainly in the liver. In some studies, a
single dose has been carcinogenic.
The central nervous system is the target organ of the toxic PAs
contained in Trichodesma, which produce spongy degeneration of the
brain.
1.4. Effects on Man
In man, PA poisoning is usually manifested as acute veno-
occlusive disease characterized by a dull dragging ache in the
right upper abdomen, rapidly filling ascites resulting in marked
distension of the abdomen, and sometimes associated with oliguria,
and massive pleural effusion. It can also manifest as subacute
disease with vague symptoms and persistent hepatomegaly. Children
are particularly vulnerable. Many cases progress to cirrhosis and,
in some cases, a single episode of acute disease has been
demonstrated to progress to cirrhosis, in spite of the fact that
the patient has been removed from the source of toxic exposure and
has been given symptomatic treatment. Mortality can be high with
death due to hepatic failure in the acute phase or due to
hematemesis resulting from ruptured oesophageal varices caused by
cirrhosis. Less severely affected cases may show clinical, or even
apparently complete, recovery. The Task Group was not aware of any
substantiated report of primary pulmonary hypertension resulting
from PA toxicity. However, in view of the evidence in experimental
animals and circumstantial evidence in one case report, the
possibility of the development of toxic pulmonary disease in man
cannot be ruled out. There is a report of an outbreak of
Trichodesma poisoning in the USSR in which the symptoms were mainly
neurological.
1.4.1. Nature and extent of health risks
The two main sources of pyrrolizidine alkaloid poisoning
reported in human beings are the consumption of cereal grain
contaminated by weeds containing the alkaloids and the use of
alkaloid-containing herbs for medicinal and dietary purposes. A
third form of exposure, with the potential to affect large
populations is the possible low-level contamination of some
foodstuffs, such as honey and milk, but the Task Group was not
aware of any cases of human toxicity having been caused through the
contamination of these foods.
Liver disease caused by the contamination of cereal grains has
been reported in rural populations in Afghanistan, India, South
Africa, and the USSR. A contributing factor appears to be
abnormally dry weather, resulting in the growth of an exceptionally
high proportion of the alkaloid-containing weeds in the crops, the
seeds of which contaminate the cereal grain on harvesting. The
weeds responsible for known outbreaks have been Heliotropium,
Trichodesma, Senecio, and Crotalaria species. Mortality in such
outbreaks has been reported to be high. In the largest reported
outbreak in northwestern Afghanistan, an estimated 8000 people were
affected in a total population of 35 000 with 1600 - 2000 deaths.
Human poisoning through the medicinal use of herbs containing
pyrrolizidine alkaloids has been reported from all parts of the
world. PAs were responsible for a common liver disease in children
in Jamaica, and individual cases in Ecuador, Hong Kong, India, the
United Kingdom, and the USA. The plants involved were species of
Crotalaria, Heliotropium, Senecio, Symphytum, and Gynura.
Symphytum-containing preparations present a particular hazard
because of their widespread use and the generally high levels of
individual exposures. The use of herbs is almost universal in
traditional folk medicine and is increasing in developed countries.
Some of the herbs used contain pyrrolizidine alkaloids and have a
long-term toxicity that is unsuspected by the people taking them.
Knowledge of the species used in herbal medicine and the frequency
of such use is very limited in the scientific literature. About 40
such species are listed in this report, about one-third of which
are in use in developed countries. They are often prescribed by
herbalists, naturopaths, and other non-orthodox practitioners. The
extent of the contribution to acute and chronic liver disease
cannot be accurately assessed. It may also constitute an
etiological factor in cirrhosis of the liver and, once this stage
is reached, it may not be possible to identify the cause as a PA.
PAs are known to be transmitted from the feed of dairy animals
into milk and to cause toxic damage in the suckling young. One
instance of large-scale contamination of honey is known to have
been caused by a common weed rich in PAs, which was the source of
nectar and pollen for the honey-secreting bees. No reports of
cases of acute toxicity caused by consumption of contaminated dairy
products or honey were available to the Task Group. Furthermore,
no information is available on the possible presence of PAs or
their metabolites in the meat of animals fed toxic weeds before
slaughter; however, the possibility of toxic disease being caused
through this medium is considered to be low.
There are no substantial, long-term follow-up data to assess
whether exposure to PAs results in increased incidence of chronic
liver disease or cancer in man. Available clinical and
experimental data suggest that a single episode of PA toxicity and
possibly also a long-term low level exposure may lead to cirrhosis
of the liver. PAs could also be possible carcinogens in man, since
a number of them have been demonstrated to induce cancer in
experimental animals, the main target organ being the liver. These
include some which have caused episodes of human toxicity, and some
others which are found in herbs traditionally used as items of
food. Also, in several instances of human toxicity, the reported
daily rates of intake of such PAs were in close range of those
known to induce tumours in rats. However, these risks cannot be
adequately assessed on a quantitative basis. There are indications
that PA intoxications leading to liver disease are more prevalent
than the reported frequency of cases would seem to indicate.
Because of their known involvement in human poisoning and their
possible carcinogenicity, exposure to pyrrolizidine alkaloids
should be kept as low as practically achievable. The setting of
regulatory tolerance levels for certain food products may be
required in some situations.
1.5. Methods for Prevention
The only known method of prevention is to avoid consumption of
the alkaloids. In the USSR, a set of agricultural (or
agrotechnical) legislative, phyto-sanitary and educational measures
has prevented new outbreaks of poisoning due to Heliotropium and
Trichodesma, since 1947.
1.6. Recommendations
1.6.1. General recommendations
1. Cereal crops should be assessed throughout the world for
possible contamination by weeds likely to contain pyrrolizidine
alkaloids. Appropriate grain inspection systems are desirable
in order to achieve near-zero levels of contamination by such
weeds.
2. There is a need to create awareness, among the general
population and those responsible for the delivery of health
services, with regard to the hazards of consuming such plants
as contaminants in food or as food, or for medicinal purposes.
Advice on hazards should include mention of possible increased
risks, if the alkaloid intake is associated with drug
treatment, (e.g. phenobarbitone) or foods which increase the
level of liver metabolizing enzymes.
3. Ethnobotanical and taxonomic studies are required in many
countries to provide specific information on the use of plant
species containing pyrrolizidine alkaloids for medicinal and
dietary purposes. There may be a need to control the sale of
some species, and their prescription by herbalists and other
practitioners of traditional systems of medicine.
4. Honey and dairy products, both local and bulk supplies, should
be assayed for pyrrolizidine alkaloids in all regions where a
risk of contamination of these foodstuffs has been identified.
1.6.2 Recommendations for research
1. Long-term follow-up studies of the survivors of both alkaloid
poisoning in human beings and animal outbreaks are required, in
order to determine the possible development of chronic liver
disease or cancer. Similar studies are also desirable on
individuals who regularly consume comfrey or other PA-
containing herbs over a substantial period of time.
2. Epidemiological studies should be carried out in countries with
a high incidence of primary liver cancer, in order to determine
whether there is an association with the intake of herbs
containing pyrrolizidine alkaloids.
3. A network of reference laboratories is needed to assist member
states in identifying plants and their seeds suspected of
producing toxic effects and for the assay and identification of
PAs. Provisions may be made for the easy availability of pure
alkaloids for use as reference standards for assays.
4. It is necessary to develop improved assay procedures, suitable
for the purposes of recommendation (4) in section 1.6.1,
particularly using fluorescence and immunochemical methods.
5. There is a need for further toxicological studies, such as
studies on the carcinogenicity of echimidine and the toxicity
of the 5-oxopyrrole constituents of Senecio species, and for
studies that would provide more quantitative information on the
various adverse biological effects of PAs. A study of the
carcinogenicity of the alkaloids in the pig is also indicated,
since the pig exhibits a high sensitivity to acute and subacute
toxicity similar to that seen in man.
6. Study is required of the possible alkaloid content of the meat,
organs, and fat of animals that have recently consumed plants
containing pyrrolizidine alkaloids.
7. Experimental studies are needed on the influence of nutritional
status on the metabolism, and acute and chronic effects of PAs.
8. Further metabolic studies are required to define more
specifically the enzymes involved in the microsomal activation
and detoxification of PAs, to determine whether organelles
other than microsomes are involved, and to explore further,
quantitative relationships between different routes of
metabolism.
9. The maximum no-observed-adverse-effect dose levels for repeated
long-term administration in the rat and the pig need to be
determined.
10. Experimental studies should be conducted to determine:
(a) whether pyrrolizidine alkaloid N-oxides may be
metabolized directly into the pyrrolic dehydroalkaloid
in mitochondria, especially in tumour cells; and
(b) which P450 enzymes are involved in the activation and
N-oxidation of PAs and thence in the selective
induction of N-oxidation enzymes.
11. A study might be conducted of human variability and its genetic
aspects in relation to factors that influence susceptibility to
PAs; for example, the study of mixed-function oxidase levels
in the liver by metabolism of appropriate test substances
recognized as harmless.
2. PROPERTIES AND ANALYTICAL METHODS
2.1 Chemical Structure and Properties
The chemical structure of PAs in relation to their toxic
effects has been reviewed recently by Mattocks (1986). The
pyrrolizidine alkaloids with which this document is concerned are
those that have previously been called "hepatotoxic" or
"nucleotoxic". Here it is proposed to refer to them as "toxic"
PAs, because of the weight of evidence now available that they
produce damage in other organs as well as the liver, and the need
to avoid a restrictive term. There are other types of
pyrrolizidine alkaloids, such as those that occur in the plant
family Orchidaceae, which are not toxic and are not discussed here.
The toxic PAs are esters of the amino-alcohols derived from the
heterocyclic nucleus. The pyrrolizidine molecule is made up of two
5-membered rings inclined to each other as shown in Fig. 1 so that
geometric isomerism is possible, and which share a common nitrogen
at position 4.
Most hepatotoxic alkaloids are esters of molecules similar to
that shown in Fig. 1(b) (1-hydroxymethyl-1:2-dehydro-
pyrrolizidine). However, a few hepatotoxic alkaloids are esters of
the amino-alcohol otonecine, e.g., petasitenine (Fig. 2, No.7).
The unsaturated pyrrolizidine nucleus itself is not toxic, but
esters of branched-chain acids are. Ester linkages may be at
positions 9, 7, or (rarely) 6. Some esters have an "open"
molecule, e.g., heliotrine, whereas others are macrocyclic
diesters, e.g., monocrotaline and retrosine. Examples of some
pyrrolizidine alkaloid structures are shown in Fig. 2.
The ring nucleus contains a double bond at the 1:2 position,
which is essential for the toxic effects of the alkaloid, but not
for unrelated effects.
1. Echimidine
Chemical structure:
Chemical formula: C20H31NO7
Relative molecular mass: 397
CAS registry number: 520-68-3
2. Heliotrine
Chemical structure:
Chemical formula: C16H27NO5
Relative molecular mass: 313
CAS registry number: 303-33-3
3. Indicine- N -oxide
Chemical structure:
Chemical formula: C15H25NO6
Relative molecular mass: 315
CAS registry number: 41708-76-3
4. Jacobine
Chemical structure:
Chemical formula: C18H25NO6
Relative molecular mass: 351
CAS registry number: 6870-67-3
5. Lasiocarpine
Chemical structure:
Chemical formula: C21H33NO7
Relative molecular mass: 411
CAS registry number: 303-34-4
6. Monocrotaline
Chemical structure:
Chemical formula: C16H23NO6
Relative molecular mass: 325
CAS registry number: 315-22-0
7. Petasitenine
Chemical structure:
Chemical formula: C19H27NO7
Relative molecular mass: 381
CAS registry number: 60132-19-6
8. Retrorsine (retrosine N -oxide = isatidine)
Chemical structure:
Chemical formula: C18H25NO6
Relative molecular mass: 351
CAS registry number: 480-54-6
9. Senecionine
Chemical structure:
Chemical formula: C18H25NO5
Relative molecular mass: 335
CAS registry number: 130-01-8
10. Symphytine
Chemical structure:
hemical formula: C20H31NO6
Relative molecular mass: 381
CAS registry number: 22571-95-5
11. Trichodesmine
Chemical structure:
Chemical formula: C18H27NO6
Relative molecular mass: 353
CAS registry number: 548-90-3
12. Incanine
Chemical structure:
Chemical formula: C18H27NO5
Relative molecular mass: 337
CAS registry number: 480-77-3
As the Task Group met in Tashkent, it is of historical interest
to recall that the structures of heliotrine and lasiocarpine, the
main alkaloids of Heliotropium lasiocarpum, were worked out by
Dr G.P. Men'shikov and associates in Moscow in the 1930s. This
work included determining the structure of heliotridine, the parent
compound of the amino-alcohol, heliotridane. Dr Men'shikov's
studies were carried out at essentially the same time, but
independently of studies by English and American authors on
retronecine-based alkaloids.
The alkaloids in plants are often found together with their
N-oxides, which are also toxic, when ingested orally. The
pyrrolizidine alkaloids acquire their toxic properties only through
the toxic pyrrolic intermediates (the general structure of which is
shown in Fig. 3) formed by the mixed-function oxidases of the
hepatocytes. To form these pyrrolic derivatives, the alkaloid
molecule should have:
(a) a double bond at the 1:2 position of the ring nucleus;
(b) esterified hydroxyl groups in the nucleus at the C 9
and/or C 7 positions; and
(c) a branched carbon chain in at least one of the ester side-
chains (McLean, 1974).
Substitution at the a position of the acid and esterification of
the C-7 hydroxy group both enhance the toxicity of the alkaloid
(Robins, 1982).
A group of related alkaloids, isolated from Senecio species by
Bohlmann et al. (1979), have non-basic pyrrolic structures similar
to those of toxic pyrrolizidine alkaloid metabolites, but they are
chemically deactivated by the presence of a carbonyl group at
position 3 of the pyrrolizidine nucleus, e.g., senaetnine (Fig. 4).
Senaetnine does not possess the acute hepatotoxic characteristics
of basic pyrrolizidine alkaloids. However, it had a direct
irritant action on tissues near the site of intraperitoneal
administration and caused damage to pulmonary vascular tissue when
given intraveinous to rats (Mattocks & Driver, 1987).
The alkaloids are fairly stable chemically, but the ester
groups may undergo hydrolysis under alkaline conditions. Some
alkaloids in plant material may decompose during drying (Bull et
al., 1968), but others appear to be stable under similar conditions
(Pedersen, 1975; Birecka et al., 1980). The N-oxides of
unsaturated pyrrolizidines are more readily decomposed by heat than
the basic alkaloids, especially when dry. However, the stability
of the alkaloids and N-oxides in hot water as, for example, in
cooking, is not known.
Some pyrrolizidine alkaloids have a limited water solubility,
unless neutralized with acid; but others (e.g., indicine), and all
the N-oxides, are readily soluble.
2.2 Analytical Methods
When analysing for PAs, it is important to recognize that this
group consists of many different compounds (section 2.1) and that
these often occur as mixtures in plants or in materials of plant
origin. They may vary in structure, relative molecular mass,
response to analytical procedures, and toxicity. Both basic
alkaloids and corresponding N-oxides may be present at the same
time. Thus, where such mixtures are present, analyses will
inevitably be approximate, unless the individual components are
separated and identified.
Nevertheless, such estimates can be useful. In particular, all
hepatotoxic PAs are unsaturated in the sense that they possess a
1:2-double bond in the pyrrolizidine nucleus, and analytical
methods that are specific for this structure can be of value in
screening for potential toxicity. A simple qualitative field test
for screening plant materials for the presence of such alkaloids
and their N-oxides, without the need of high technology equipment,
is described in section 2.2.2.5.
2.2.1 Extraction
2.2.1.1 Plant tissue
Pyrrolizidine alkaloids are usually extracted from dried,
milled plant material with hot or cold alcohol. The alcohol is
evaporated, the bases taken up in dilute acid, and fats extracted
with ether or petroleum. It is usual, at this stage, to reduce any
N-oxides present to the corresponding basic alkaloids with zinc,
before making the solution alkaline and extracting the alkaloids
with chloroform (Koekemoer & Warren, 1951). Alternatively, alcohol
can be continuously circulated through the plant material and then
cation exchange resin, and the alkaloids subsequently eluted from
the resin (Mattocks, 1961; Deagen & Deinzer, 1977). PAs can also
be extracted by soaking plant material in dilute aqueous acid
(Briggs et al., 1965; Craig et al., 1984).
2.2.1.2 Biological fluids and tissues
Pyrrolizidine alkaloids have been extracted for analytical
purposes from honey (Deinzer et al., 1977), milk (Dickinson et al.,
1976), blood-plasma (Ames & Powis, 1978; McComish et al., 1980),
urine (Mattocks, 1967a; Jago et al., 1969; Evans et al., 1979), and
bile (Jago et al., 1969; Lafranconi et al., 1985).
When attempting to isolate PAs from animal tissues, it must be
appreciated that the toxic alkaloids are often metabolized very
rapidly in animals, so that the amounts that are recoverable
(except from urine), only a few hours after alkaloid ingestion, may
be extremely small. Various methods have been used to separate
PAs, but some mixtures are extremely difficult to separate. On the
analytical scale, the most useful methods are thin-layer
chromatography (TLC), high-performance liquid chromatography
(HPLC), and gas chromatography (GC) (section 2.2.2).
2.2.2 Analysis for pyrrolizidine alkaloids
2.2.2.1 Thin-layer chromatography (TLC)
For TLC, silica plates are usually used, eluted with chloroform:
methanol:aqueous ammonia mixtures (Sharma et al., 1965; Chalmers
et al., 1965); solvents suitable for the N-oxides, which
are more water-soluble, have been described by Mattocks (1967b)
and Wagner et al. (1981). The most sensitive methods for
detecting PAs on TLC are those using Ehrlich reagent
(4-dimethylaminobenzaldehyde) (Mattocks, 1967b). The unsaturated
alkaloids are best visualized by spraying the plates first with a
solution of orthochloranil, then with Ehrlich reagent, heating
after each spray (Molyneux & Roitman, 1980). The N-oxides of
unsaturated pyrrolizidines are detected by spraying a solution of
acetic anhydride, heating the plate, and then spraying Ehrlich
reagent (Mattocks, 1967b).
Pyrrolizidine alkaloids with a saturated base moiety must be
detected in other ways (which are not specific for pyrrolizidines),
e.g., by exposing the dried plates to iodine vapour, or by spraying
with an iodobismuth (Dragendorff) reagent (Munier, 1953).
2.2.2.2 High-performance liquid chromatography (HPLC)
Analytical or preparative scale HPLC separation of
pyrrolizidine alkaloids has been described by Segall (1979a,b) and
Dimenna et al. (1980), and an improved method has been reported by
Ramsdell & Buhler (1981). Alkaloids from Symphytum officinale
(comfrey) have been separated on an analytical scale by Tittel et
al. (1979), and partially separated on a preparative scale by
Huizing et al. (1981). UV detectors are usually used for the HPLC
of pyrrolizidine compounds (Mattocks, 1986).
2.2.2.3 Gas chromatography (GC) and mass spectrometry (MS)
The GC characterization of PAs using packed columns has been
described by Chalmers et al. (1965) and Wiedenfeld et al. (1981).
Mixtures of alkaloids from comfrey ( Symphytum sp.), normally hard
to separate, were resolved by Culvenor et al. (1980a) and Frahn et
al. (1980) by GC of the methylboronate derivatives.
Gas chromatography combined with mass spectrometry (GC-MS) has
become a valuable and highly sensitive means for both the
identification and the quantitative determination of pyrrolizidine
alkaloids. Thus, alkaloids extracted from honey were separated and
identified by Deinzer et al. (1977) and (as butylboronate
derivatives) by Culvenor et al. (1981). Deinzer et al. (1978)
described a method for the recognition (but not the individual
identification) of retronecine-based pyrrolizidine alkaloids, by
hydrolysing them to retronecine (the amino alcohol moiety) followed
by GC-MS of its bis-trifluoroacetate. The use of capillary GC has
greatly improved the sensitivity of pyrrolizidine alkaloid
analysis, especially when used with MS (Luthy et al., 1981). The
MS of pyrrolizidine compounds has been reviewed (Bull et al., 1968;
Mattocks, 1986).
Pyrrolizidine N-oxides generally undergo thermal decomposition,
when subjected to GC, but they can first be reduced to the
corresponding basic alkaloids (Koekemoer & Warren, 1951).
Alternatively they may be derivatised. Thus, trimethylsilylation
of indicine N-oxide or heliotrine N-oxide can lead either to the
trimethylsilyl (TMS) derivative of the parent alkaloid or to the
TMS derivative of the dehydro-alkaloid (pyrrolic derivative),
depending on the reagents used, and these products will run
successfully on GC-MS (Evans et al., 1979, 1980).
2.2.2.4 Nuclear magnetic resonance (NMR) spectrometry
A convenient, but relatively insensitive, method, specifically
for the determination of unsaturated PAs, has been described by
Molyneux et al. (1979). The basic alkaloids are extracted, then
subjected to NMR spectrometry along with an internal standard
( p-dinitrobenzene). This enables quantitative measurements to be
made of the signal(s) representing the H2 proton(s) in unsaturated
pyrrolizidines, and thus the alkaloid(s) can be determined.
Quantitative NMR analysis of pyrrolizidine alkaloid mixtures from
Senecio vulgaris has been described by Pieters & Vlietinck (1985)
and compared with an HPLC method by the same authors (1986).
Qualitative aspects of the NMR spectrometry of pyrrolizidine
alkaloids have been reviewed by Bull et al. (1968) and Mattocks
(1986).
2.2.2.5 The Ehrlich reaction
This method (Mattocks, 1967a, 1968b) is specific for
unsaturated pyrrolizidine alkaloids and is not suitable for other
alkaloids. Thus, it is the most useful colorimetric method for
potentially hepatotoxic pyrrolizidine compounds. The procedure
converts the alkaloid into its N-oxide, using hydrogen peroxide.
The product reacts with acetic anhydride to form a pyrrolic
derivative (dehydro-alkaloid) that gives a magenta colour with a
specially modified Ehrlich reagent. The latter contains boron
trifluoride to give maximum sensitivity. As little as 5 µg of most
unsaturated pyrrolizidines can be measured by this method. If the
oxidation stage is omitted, only the unsaturated pyrrolizidine
N-oxides can be determined. The determination of pyrrolizidine
N-oxides has also been discussed by Mattocks (1971b).
A simplification of the above colorimetric procedure was
described by Mattocks (1971d) to provide a qualitative test that
could be used to screen large numbers of plant samples for the
presence of unsaturated pyrrolizidine alkaloid N-oxides. An
improved version of this field test is now available (Mattocks &
Jukes, 1987). It is suitable for any plant parts, such as leaves,
stems, flowers, seeds, or roots, or materials of plant origin, such
as cereals or herbal teas, but has not yet been applied to cooked
food.
The plant material (0.2 - 1 g) is extracted by grinding it with
aqueous ascorbic acid (5%) and a small amount of sand. The
solution is filtered and divided into two equal portions ("test"
and "blank"). An aqueous solution (0.2 ml) of sodium nitroprusside
(5%) containing sodium hydroxide (10-3 mol) is added to the "test"
sample. Both portions are heated for approximately 1 min at 70 -
80 °C; then Ehrlich reagent is added and heating is continued for
1 min. The Ehrlich reagent contains 4-dimethylaminobenzaldehyde
(5 g) dissolved in a mixture of acetic acid (60 ml), water (30 ml),
and 60% perchloric acid (10 ml). A magenta colour in the "test"
compared with the "blank" indicates the presence of an unsaturated
PA N-oxide. The "blank" may show a colour if the plant contains
compounds, such as indoles or pyrroles, which can themselves give a
colour with Ehrlich reagent. The intensity of colour in the
"sample" compared with the "blank" can give a rough idea of the
amount of alkaloids present, and indicate whether further chemical
or toxicological testing of the plant material is adviseable.
In practice, the majority of PA-containing plants contain
enough alkaloid in the N-oxide form (often a large proportion) to
react positively in this test. The main exceptions are some seeds
(Crotalaria), which may contain much alkaloid base, but little or
no N-oxide. These (and any other sample not containing
chlorophyll) can be tested for basic PAs by grinding them with
chloroform, heating the filtered extract with a solution (0.1 ml)
of orthochloranil (0.5%) in acetonitrile, and then heating it with
Ehrlich reagent. A magenta colour indicates the presence of an
unsaturated PA. Non-toxic pyrrolizidine alkaloids having a
saturated pyrrolizidine nucleus, and pyrrolizidine alkaloids that
are otonecine esters, such as petasitenine, will not respond to
this test.
2.2.2.6 Indicator dyes
A method generally applicable to tertiary bases has been
adapted for pyrrolizidine alkaloids by Birecka et al. (1981). It
is sensitive, but is not specific for this group of alkaloids, and
it does not distinguish between the saturated and unsaturated
alkaloids. A chloroform solution of the alkaloid is shaken with
acidified aqueous methyl orange. The yellow alkaloid:dye complex
is subsequently released from the chloroform phase, using ethanolic
sulfuric acid, and measured spectrophotometrically.
2.2.2.7 Direct weighing
An insensitive way to determine the alkaloids in, for example,
a plant sample, providing enough is available, is to extract the
alkaloids (section 2.2.1) and weigh them. This will provide a
rough measure of the total bases present in the sample; however,
these may not necessarily be PAs. Nevertheless, the sample can
then be subjected to further tests, e.g., GC-MC, nuclear magnetic
resonance (NMR), or colorimetric analysis. Furthermore,
pyrrolizidine N-oxides are generally too water soluble to be
appreciably extractable from aqueous solution by chloroform. Thus,
if two portions of the sample are extracted, and one of them is
reduced to convert N-oxides to bases, the weight difference between
the two products will represent the alkaloid existing in the form
of N-oxide in the original sample.
2.3 Determination of Metabolites in Animal Tissues
Important metabolites of toxic pyrrolizidine alkaloids in
animals include "pyrrolic" derivatives (dehydro-alkaloids) and
N-oxides. A procedure for measuring pyrrolic metabolites in tissue
samples (such as liver or lung) has been described by Mattocks &
White (1970). The sample (usually 0.5 g) is homogenized in an
ethanolic solution of mercuric chloride; the solids are separated
by centrifugation and heated with Ehrlich reagent to give a soluble
colour that can be measured spectrophotometrically.
The measurement of pyrrolic and N-oxide metabolites, formed by
the action of hepatic microsomal preparations on PAs in vitro, is
an improvement described by Mattocks & Bird (1983).
3. SOURCES AND PATHWAYS OF EXPOSURE
3.1 Hepatotoxic Pyrrolizidine Alkaloids and Their Sources
Plants constitute the only natural source of pyrrolizidine
alkaloids (PAs) that cause toxic reactions in man and animals. PAs
occur in a number of species in the families Boraginaceae,
Compositae, Leguminosae (genus Crotalaria), Ranunculaceae (genus
Caltha), and Scrophulariaceae (genus Castilleja) (Table 1). The
most important genera of PA-containing toxic plants are Crotalaria
(Leguminosae), Senecio (Compositae), Heliotropium, Trichodesma,
Amsinckia, Echium, and Symphytum (Boraginaceae) (Hooper, 1978).
The recorded cases of human toxicity have mainly been caused by at
least 12 different pyrrolizidine alkaloids, mostly derived from
Heliotropium, Senecio, and Crotalaria genera. The Senecio spp.
grow throughout the world; the Crotalaria spp. are mainly found in
the tropics and subtropics (Culvenor, 1980).
Table 1. List of plant genera containing toxic pyrrolizidine alkaloids
(with number of species investigated)
-------------------------------------------------------------------------------------
Family Genera
-------------------------------------------------------------------------------------
Apocynaceae Fernaldia (1), Parsonsia (4),
Boraginaceae Alkanna (1), Amsinckia (4), Anchusa (2), Asperugo (1), Borago (1),
Caccinia (1), Cynoglossum (9), Echium (3), Hackelia (1),
Heliotropium (25), Lappula (2), Lindelofia (7), Lithosperum (1),
Macrotomia (1), Messerschmidtia (1), Myosotis (2), Paracaryum (1),
Paracynoglossum (1), Rindera (5), Solenanthus (4), Symphytum (7),
Tournefortia (2), Trachelanthus (2), Trichodesma (2), Ulugbekia (1)
Compositae Adenostyles (3), Brachyglottis (1), Cacalia (4), Conoclinium (1),
Crassocephalum (1), Doronicum (2), Echinacea (2), Emilia (2),
Erechtites (1), Eupatorium (8), Farfugium (1), Gynura (2),
Ligularia (5), Petasites (4), Senecio (142), Syneilesis (1),
Tussilago (1)
Leguminosae Crotalaria (60)
Ranunculaceae Caltha (2)
Scrophulariaceae Castilleja (1)
-------------------------------------------------------------------------------------
An alphabetical list of pyrrolizidine alkaloids with their
plant sources has been published by Smith & Culvenor (1981) and
Mattocks (1986). An updated version is attached as Appendix I.
The plant genera containing toxic PAs are listed in Table 1
indicating the number of species investigated for PAs. A
comprehensive list of species of plants belonging to each of these
genera, the alkaloids isolated from each, and the part of the plant
containing the alkaloid are presented in Appendix II. Table 1 in
Appendix II includes species known to contain alkaloids of proved
hepatotoxicity, or of a molecular structure that would make them
very probably hepatotoxic. Table 2 in Appendix II includes species
containing pyrrolizidine amino-alcohols or esters, which, while not
having all the features of hepatotoxicity, would need only minor
structural modifications to render them hepatotoxic. Plants of the
same taxonomic groups as the plants of proven hepatotoxicity are
listed in part (a) of the table. There is a possibility that, on
further examination, hepatotoxic alkaloids may be found, as minor
constituents, in strains or parts of these plants not yet
investigated or under specific conditions of growth. It should be
noted that the species that have been investigated and are listed
are only few compared with the total number of species in each
genera. It has been recommended by Smith & Culvenor (1981) that it
would be prudent to regard all species in the family Boraginaceae
and the genera Crotalaria, Senecio, and Eupatorium as potentially
hepatotoxic.
It is pertinent to note that the alkaloid content in different
parts of the plant (e.g., roots, leaves, stalks, flowers, and buds)
varies and is subject to fluctuations according to the climate,
soil conditions, and time of harvesting (Danninger et al., 1983;
Hartmann & Zimmer, 1986). Mattocks (1980) demonstrated that the
alkaloid content of the leaves of Symphytum spp. (Russian
comfrey), which are used as an item of food, varies with their
maturity. The toxic PA content is highest at the beginning of the
vegetative period and declines as the leaves mature. The PA
content of the roots is much higher than that of the leaves, and
dried leaves contain a higher concentration than fresh leaves
(Mattocks, 1986). According to Danninger et al. (1983), in some
species (Symphytum asperum), relatively long storage may lead to a
reduction in the alkaloid content, presumably because enzymes are
released during drying. Candrian et al. (1984b) studied the
stability of PAs in hay and silage containing various amounts of
Senecio alpinus. The PA content of hay remained constant for
several months, but the PAs in silage were mainly degraded.
However, the degradation of PAs was much less complete in the lower
concentration range. A quantitatively significant PA-degradation
product in silage was identified as retronecine. Silage with an
S. alpinus percentage of 3.5 - 23 still contained macrocyclic PAs at
a concentration of about 20 mg/kg wet weight. Such silage was not
considered safe for cattle bearing in mind that a 600-kg calf eats
about 30 kg silage/day, amounting approximately to a daily intake
of about 1 mg PAs/kg body weight. In feeding trials with Senecio
jacobaea, Johnson (1979) found that the minimum lethal dose for
cattle was between 1 and 2 mg PAs/kg body weight per day.
PAs known to have been associated with instances of human toxic
liver disease in different parts of the world are listed in Table
2. Two groups of alkaloids that, according to Culvenor (1983), are
consumed in significant amounts by people in different parts of the
world include:
(a) Echimidine, acetyllycopsamine, and related alkaloids
(many countries)
Leaves of plants of the Symphytum sp. ( Symphytum officinale
(comfrey) and Symphytum x uplandicum) are used traditionally as a
salad and as a medicinal herb in Australia, many countries of
Europe, and the USA. S. officinale has been shown to be
carcinogenic for rats (Hirono et al., 1978). Leaves of Russian
comfrey contain a concentration of alkaloids (mainly echimidine) of
0.1 - 1.5 g/kg. The highest level of daily consumption of the
alkaloids has been estimated to be 5 - 6 mg (Culvenor, 1983).
(b) Echimidine and related alkaloids (Australia)
PAs derived from Echium plantagineum, with echimidine as the
major component, have been found in honey secreted by bees feeding
on the plant (Culvenor et al., 1981). The plant is a major source
of honey (section 3.3.4).
3.2 Pneumotoxic and Other Toxic Pyrrolizidine Alkaloids
Not all hepatotoxic alkaloids are pneumotoxic. The commonest
ones used to produce experimental lung injury are fulvine (Barnes
et al., 1964; Kay et al., 1971a; Wagenvoort et al., 1974a,b) and
monocrotaline (Lalich & Ehrhart, 1962; Chesney & Allen, 1973b;
Huxtable et al., 1977). These are also the most active (Mattocks,
1986). The seeds of Crotalaria spectabilis, which contain
monocrotaline, have also been used to study pneumotoxic effects on
experimental animals (Turner & Lalich, 1965; Kay & Heath, 1966; Kay
et al., 1967a) and C. spectabilis has been called the pulmonary
hypertension plant (Kay & Heath, 1969), because of the pulmonary
hypertensionogenic properties of the PAs it contains. Culvenor et
al. (1976a) screened 62 PAs for hepatotoxicity and pneumotoxicity.
Chronic lung lesions were produced by most compounds that induced
chronic liver lesions, though high doses were required in some
instances. It is possible that chronic lung lesions may not occur
in experimental animals because of early death due to acute
toxicity. However, the authors identified a number of PAs that
were particularly prone to produce chronic lung damage in rats
including crispatine, senecionine, seneciphylline, and usaramine
(12-membered macrocyclic, retronecine diesters), anacrotine and
madurensine (crotonecine esters), and the heliotridine esters,
heliosupine, lasiocarpine, and rinderine.
The molecular structure-activity requirements for
pneumotoxicity are the same as those for hepatotoxicity. This is
consistent with their both being caused by the same toxic
metabolites and by the metabolic activation of the alkaloids in the
liver cells to form a reactive pyrrolic dehydro-alkaloid (Culvenor
et al., 1976a).
Trichodesmine and incanine, found in the seeds of Trichodesma
incanum (Yunusov & Plekhanova, 1959), are believed to have been
the causative factors of the "Ozhalangar encephalitis" that was
endemic in Uzbekistan, USSR (1942 - 51), in which the symptoms and
signs were related primarily to the central nervous system
(Shtenberg & Orlova, 1955) (section 7.7).
Table 2. Instances of human toxicity caused by pyrrolizidine alkaloidsa
Principal Plant Country/ Cause of intake Reference
alkaloid Region
Heliotrine and Heliotropium Afghanistan contamination Tandon & Tandon
other alkaloids popovii (1975); Tandon,
similar to B.N. et al.
lasiocarpine (1978); Tandon,
H.D. et al.
(1978);
Mohabbat et al.
(1976)
Senecionine Senecio South contamination Wilmot &
illiciformis; Africa Robertson
Senecio-burchelli (1920)
Senecio spp. South contamination Selzer &
Africa Parker (1951)
Alkaloids of Crotalaria Ecuador medicine Lyford et al.
trichodesmine juncea (1976)
and senecionine
type
Heliotrine and Heliotropium Hong Kong medicine Kumana et al.
lasiocarpine lasiocarpum (1985);
Culvenor et al.
(1986)
Table 2. (cont'd)
Principal Plant Country/ Cause of intake Reference
alkaloid Region
Crotananine and Crotalaria India contamination Tandon, R.K.
cronaburmine nana et al. (1976);
Krishnamachari
et al. (1977);
Siddiqui et al.
(1978a,b)
Heliotrine Heliotropium India medicine Datta et al.
N-oxide eichwaldii (1978a,b)
Monocrotaline Crotalaria West Indies medicine Bras et al.
fulvine retusa; (1954, 1957)
Crotalaria
fulva Stuart & Bras
(1957)
Ilex sp. United medicine McGee et al.
Kingdom (1976)
Riddelline Senecio USA medicine Stillman et al.
retrorsine longilobus (1977); Fox et
N-oxide al. (1978);
(with others) Huxtable (1980)
Indicine N-oxide purified USA medicine Letendre et al.
chemical (1984)
Symphytine, Symphytum sp. USA medicine Ridker et al.
symglandine, and (1985);
other symphytum Huxtable
alkaloids et al (1986)
Table 2. (cont'd)
Principal Plant Country/ Cause of intake Reference
alkaloid Region
Lasiocarpine and Heliotropium USSR contamination Dubrovinskii
heliotrine lasiocarpum (1952);
Mnushkin
(1952)
Trichodesmine and Trichodesma USSR contamination Shtenberg &
incanine incanum Orlova (1955);
Yunosov &
Plekhanova
(1959)
a Adapted from: Culvenor (1983) and Mattocks (1986). Refer also to Table 15 for
details and section 7.
3.3 Pathways of Exposure
Naturally-occurring animal disease is caused by the alkaloid-
containing plants growing in fields and pastures or being fed
accidentally as fodder. They are mostly herbaceous or small shrubs
and many thrive in dry and arid climates. One such plant
containing toxic PA alkaloids has been reported to grow in the
western desert of Egypt (Hammouda et al., 1984). The growth of
this group of plants is particularly prolific during, and
following, periods of drought, as has been reported in association
with the outbreaks of human disease in Afghanistan (Tandon &
Tandon, 1975; Mohabbat et al., 1976) and India (Tandon, B.N. et
al., 1976). Alkaloid-containing plants are widespread in the
tropics, especially Crotalaria, of which there are over 300
species in Africa. Ordinarily, the alkaloid-containing plants have
a bitter taste and grazing animals will reject them, unless their
normal fodder is scarce. However, PAs often occur largely as
N-oxides, which are said not to be bitter, and plants containing
PAs are readily eaten by some animal species.
Human intoxication may result from the ingestion of the toxic
substance in either contamined food or herbal infusion.
3.3.1 Contamination of staple food crops
The products of pyrrolizidine alkaloid-containing plants,
generally seeds, may contaminate the staple food and may be eaten
over long periods of time. The fact that these plants may cause
disease is generally not recognized by the people and such
contamination is known to have resulted in large-scale outbreaks of
poisoning (Dubrovinskii, 1952; Mnushkin, 1952; Shtenberg & Orlova,
1955; Tandon & Tandon, 1975; Mohabbat et al., 1976; Tandon, B.N. et
al., 1976, 1977; Tandon, R.K. et al., 1976; Krishnamachari et al.,
1977; Tandon, H.D. et al., 1977) (Table 2, section 3.1).
3.3.2 Herbal infusions
Plants have been used traditionally for medicinal purposes all
over the world. Herbs have been the mainstay of the indigenous
systems of medicine, especially in China, Greece, and India, since
ancient times. Table 3 includes a list of some plants that are
suspected, or known, to contain PAs and have been used as herbal
medicines in different countries (Mattocks, 1986).
Several PA-containing plants are included among the list of
plants used in indigenous systems of medicine in India (Chopra,
1933). As a part of a research study on the etiological factors of
chronic liver disease in Sri Lanka, Arseculeratne et al. (1981)
chemically screened the first 50 plants used in Sri Lanka's
traditional medicine pharmacopoaea, and found that 3 of them
contained PAs. All 3 were hepatotoxic in rats. Of the 3, the
presence of alkaloids in Cassia auriculata and that of PAs in
Hollarhena antidysenterica had not previously been recorded. It
should be noted that the amount of experimental plant material used
in this study was approximately 6.5 g/kg body weight per day, in
contrast to the approximate intake by a human being estimated to be
in the range of 0.3 - 0.6 g/kg body weight per day. Some, but not
all, of the plants reported to be etiological agents in human cases
of veno-occlusive disease can be found in an inventory of medicinal
plants used in different countries (WHO, 1980), which also
indicates the countries that they are used in. The above lists may
not be complete as many such plants may be used in folk medicine
but have not been mentioned in the scientific literature. However,
the lists do indicate the wide and varied use of such toxic herbs
in all parts of the world.
Lately, there has been a growing interest in the developed
countries in organically grown products for food, as well as home
remedies (Table 3), and some of the PA-containing herbs have been
freely available in herbal shops (Schoental, 1968; Burns, 1972).
Danninger et al. (1983) listed plants containing PAs that are
commonly used in the Federal Republic of Germany as medicaments
(Table 4). He also listed 9 plants in which alkaloids have only
been identified qualitatively, the toxicity of which has not been,
or has been insufficiently, investigated (Table 5). Similarly,
Roitman (1983) listed 10 plants, in which the presence of PAs is
suspected or has been proved and which are used as herbal teas in
the USA. The lists include 10 plants containing PAs, most of which
have been proved hepatotoxic experimentally, some having highly
carcinogenic promoter activity. Some of these alkaloids have been
associated with human case reports of PA toxicity. The more recent
reports (Table 2) of instances of PA poisoning through the use of
herbal medicines are from developed countries (Lyford et al., 1976;
Stillman et al., 1977; Fox et al., 1978; Kumana et al., 1985;
Ridker et al., 1985). Such use of the herbs is the reason that
veno-occlusive disease is endemic in Jamaica (Bras et al., 1954;
Jellife et al., 1954a,b; Bras & Watler, 1955; Stuart & Bras, 1955,
1957). There are obvious difficulties in exercising any kind of
control to restrict this use only to plants that have been tested
and certified as safe for human use. It is impossible to identify
many such herbs, as they are sold as plants or their amorphous
products in the herbal shops.
Table 3. Some plants containing (or suspected of containing) PAs, which have been used
by people either as herbal medicines (M) or foods (F)
Plant Mode Country Referencea
of use or region
BORAGINACEAE
Anchusa officinalis M Europe Broch-Due & Aasen (1980) B
Borago officinalis M USA Delorme et al. (1977) A
Cynoglossum M East Africa Schoental & Coady (1968) A
geometricum
Cynoglossum M Iran Coady (1973) B
officinale
Heliotropium M India Gandhi et al. (1966a); B
eichwaldii Datta et al. (1978a,b) A
H. europaeum M India, Greece IARC (1976) A
H. lasiocarpum M Hong Kong Kumana et al. (1985); A
Culvenor et al. (1986) A
H. indicum M India, Africa, Schoental (1968a); B
South America, Hoque et al. (1976) B
and elsewhere
H. ramossissimum M Arabia Macksad et al. (1970); B
(ramram) Coady (1973) B
H. supinum M Tanzania Schoental & Coady (1968) A
Table 3 (contd.)
Plant Mode Country Referencea
of use or region
Pulmonaria spp. M USA Delorme et al. (1977) A
Symphytum officinale F, M Japan and Hirono et al. (1978, 1979b) A
M USA Furuya & Hikichi (1971); A
Delorme et al. (1977) A
S. x uplandicum F, M General Hills (1976) B
USA Culvenor et al. (1980a,b) A
S. asperum M USA Pedersen (1975) A
COMPOSITAE
Cacalia decomposita M USA Sullivan (1981) B
(matarique)
C. yatabei F Japan Hikichi & Furuya (1978) B
Farfugium japonicum M Japan Furuya et al. (1971) B
Ligularia dentata F Japan Asada & Furuya (1984) B
Petasites japonicus F, M Japan Hirono et al. (1973, 1979b) A
Senecio abyssinicus M Nigeria Williams & Schoental (1970) B
S. aureus M USA Wade (1977) B
Table 3 (contd.)
Plant Mode Country Referencea
of use or region
S. bupleuroides M Africa Watt & Breyer-Brandwijk (1962) A
S. burchelli F, M South Africa Rose (1972) A
S. coronatus M South Africa Rose (1972) A
S. discolor M Jamaica Asprey & Thornton (1955) B
S. doronicum M Germany Roeder et al. (1980a) B
S. inaequidens F South Africa Rose (1972) B
S. jacobaea M Europe Schoental & Pullinger (1972); B
(ragwort) Wade (1977) B
S. longilobus M USA Stillman et al. (1977); A
(S. douglassi) Huxtable (1979a) B
S. monoensis M USA Huxtable (1980) A
S. nemorensis M Germany Habs et al. (1982) A
spp. fuchsii
S. pierotti F Japan Asada & Furuya (1982) B
S. retrorsus M South Africa Rose (1972) A
(S. latifolius)
Table 3 (contd.)
Plant Mode Country Referencea
of use or region
S. vulgaris M Europe Watt & Breyer-Brandwijk (1962) A
(common groundsel)
Netherlands Wade (1977) B
M Iran Coady (1973) B
Syneilesis palmata F Japan Hikichi & Furuya (1976) B
Trichodesma africana M Asia Omar et al. (1983) B
Tussilago farfara M Japan Culvenor et al. (1976a) A
(coltsfoot)
M China Hirono et al. (1976b) A
M Norway Borka & Onshuus (1979) B
M USA Borka & Onshuus (1979); B
Culvenor et al. (1976b); B
LEGUMINOSAE
Crotalaria brevidens F East Africa Coady (1973) B
C. fulva M Jamaica Barnes et al. (1964); A
McLean (1970, 1974) A
Table 3 (contd.)
Plant Mode Country Referencea
of use or region
C. incana M East Africa Schoental & Coady (1968) A
Watt & Breyer-Brandwijk A
(1962)
C. laburnifolia M Tanzania Schoental & Coady (1968) A
F Asia Coady (1973) B
C. mucronata M Tanzania Coady (1973) B
C. recta M, F Tanzania Schoental & Coady (1968); A
Coady (1973) B
C. retusa M, F Africa IARC (1976) A
India Watt & Breyer-Brandwijk (1962) A
C. verrucosa M Sri Lanka Arseculeratne et al. (1981) A
a A = Reference in the reference list of this document.
B = Reference in Mattocks (1986).
Manufactured preparations may also contain PA-containing herbs,
e.g., comfrey-pepsin capsules sold as a digestive aid (Huxtable et
al., 1986).
3.3.3 Use of PA-containing plants as food
Several PA-containing plants are used as food as can be seen in
Table 3 (Mattocks, 1986). Petasites japonicus Maxim, Tussilago
farfara L. (coltsfoot), and Symphytum officinale L. (comfrey or
Russian comfrey) are known as edible plants in Japan, and have been
proved to contain carcinogenic pyrrolizidine alkaloids (Hirono et
al., 1973, 1979a,b). The young flower-stalks of P. japonicus and
the buds of coltsfoot have been used in Japan as human food or
herbal remedies. The leaf and root of comfrey are also used as an
edible vegetable or tonic (Hirono et al., 1978) in Japan and other
countries (Culvenor, 1985). The carcinogenic PAs in these plants
are petasitenine (P. japonicus), senkirkine (coltsfoot), and the
group including symphytine (comfrey). They were also mutagenic in
the Ames system of Salmonella typhimurium and V79 hamster cell line
and induced transformation in cryo-preserved hamster embryonic
cells (Hirono et al., 1979b). Other such PA-containing plants,
used as food in Japan, include young leaves of Syneilesis palmata,
various Cacalia species, and young Senecio pierotti (Mattocks,
1986). According to Culvenor (1985), consumers of comfrey could be
ingesting up to 5 mg PAs per day. Rose (1972) listed a number of
plants of the genus Senecio that are used as spinach in South
Africa. These include S. burchelli, which is known to have caused
an episode of PA poisoning through the ingestion of contaminated
bread (Wilmot & Robertson, 1920).
3.3.4 Contaminated honey
In the USA, Deinzer et al. (1977) reported the presence of all
PAs contained in Senecio jacobaea (ragwort) and proved to be
hepatotoxic, in the honey secreted by bees feeding on the plant.
The total alkaloid content ranged from 0.3 to 3.9 mg/kg. It has
been estimated that an average annual human intake of honey (600 g)
at the highest alkaloid level quoted would contain less than 3 mg
of PAs (Mattocks, 1986). Culvenor et al. (1981) and Culvenor
(1983, 1985) drew attention to the same potential hazard in honey
from Echium plantagineum, a weed that grows widely in Southern
Australia and is a major source of honey, yielding an estimated
2000 - 3000 tonnes per annum for human consumption. Echimidine is
the major component of the alkaloids of Echium, which are present
in concentrations of up to 1 mg/kg. Culvenor (1983) estimated that
individuals may consume up to 80 g honey/day with a corresponding
alkaloid intake of 80 µg/day, if only the Echium honey were used.
No reports of acute human toxicity through this source are
available.
Table 4. Medicinal plants containing PAs of known hepatotoxicity, reported as commonly
used in the Federal Republic of Germany, and the PAs contained in thema
Family Genus Species Pyrrolizidine
alkaloids
Compositae Eupatorium E. cannabinum amabilineħ
(hemp agrimony) supinineb
Petasites P. hybirdus senecionineb,c
integerrimineb
senkirkineb
Senecio S. nemorensis fuchsisenecionine
(groundsel) sp. fuchsii senecionineb,c
(Fuch's groundsel)
S. vulgaris senecionineb,c
(groundsel) seneciophyllineb
retrorsineb
riddellineb,c
S. Jacobaea jacobineb
(ragwort) senecionineb,c
seneciphyllineb
jacoline, jaconine
chlorinated PAsd
S. aureus senecionineb,c
(American golden
ragwort)
Tussilago T. farfara senkirkineb
(coltsfoot) (coltsfoot) senecionineb,c
tussilagine
Table 4 (contd.)
Family Genus Species Pyrrolizidine
alkaloids
Alkanna A. tinctoria 7-angelylretronecine
triangularine
dihydroxytriangularine
Anchusa A. officinalis lycopsamine
Boraginaceae Borago B. officinalis lycopsamine/intermedineħ
(borage) acetyllycopsamine/
acetylintermedine
amabiline
supinine
Symphytum S. officinale symphytineb
(comfrey) (comfrey) echimidine(?)
lycopsamine
acetyllycopsamineb
lasiocarpineb,c
heliosupine N-oxide
S. peregrinum lycopsamineb
S. x uplandicum intermedineb
symphytineb
echimidineb
7-acetyllycopsamine
7-acetylintermedine
symlandine
uplandicine
S. asperum asperumine
(prickly comfrey) heliosupine N-oxide
echimidineb
echinatine
Table 4 (contd.)
Family Genus Species Pyrrolizidine
alkaloids
Cynoglossum C. officinale heliosupine N-oxide
(hound's (hound's tongue) echinatine
tongue) acetyl heliosupineb
O-7-angelylhelio-
tridineb
Heliotropium H. europaeum heliotrineb,c,e
(Heliotrope) (common heliotrope) lasiocarpineb,c,e
supinine
heleurine
europine
acetyllasiocarpineb
a Modified from: Danninger et al. (1983).
b Toxic alkaloids.
c Alkaloids known to have caused human toxicity.
d Alkaloids with highly carcinogenic promoter activity.
e Used only in homeopathy.
Table 5. Medicinal plants containing PAs, reported as
commonly used in the Federal Republic of Germany,
the toxicity of which has not been, or has been
insufficiently, investigateda
-----------------------------------------------------------
Family Genus Species
-----------------------------------------------------------
Compositae Eupatorium E. perforatum
Brachyglottis B. repens
Arnica A. montana (mountain arnica)
Boraginaceae Lappula L. intermedia (stickseed)
Pulmonaria P. officinalis (lungwort)
-----------------------------------------------------------
a Modified from: Danninger et al. (1983).
3.3.5 Milk
PAs have been shown to produce toxic effects via transference
into the milk of dams (Schoental, 1959). Retrorsine was
administered orally to 17, and intraperitoneally to 6, lactating
rats weighing 185 - 350 g in 5 - 10 mg doses daily, the first dose
being given during the first 24 h after parturition. The rats
received from 1 to 14 doses, the total intake amounting to
21 - 335 mg/kg body weight. The litters were separated from the
mothers for ´ h following the administration of PA to avoid direct
contamination of the former by licking. Apparently the milk
production was not affected as the stomachs of many of the young,
examined postmortem, were distended with milk. All animals whose
mothers had received a total dose of 138 mg PA or more died within
30 days. Many of the young whose mothers had received smaller
doses survived until they were killed at 6 months. Biopsy of the
liver of the young at various intervals or at autopsy showed marked
changes, even in cases where the mothers did not appear to be
affected. Animals dying at 18 - 30 days showed hydropic or fatty
vacuolation of liver cells. In the liver of animals dying or
killed later, various degrees of haemorrhagic necrosis and increase
in the centrilobular reticulin of the liver, and some thickening of
centrilobular veins were seen. In animals that survived 6 months,
the appearance was less abnormal, but some hyperplastic nodules and
bile-duct proliferation were seen. The lactating rats dosed with
the PAs generally survived longer than the suckling animals and
usually did not show any ill effects, suggesting that the
susceptibility of the suckling rats was greater than that of the
mothers.
Dickinson et al. (1976) demonstrated the presence of PAs in the
milk of dairy cattle fed or dosed with ragwort (Senecio jacobaea).
When 4 cows were administered the dried plant material at levels of
up to 10 g/kg body weight per day through rumen cannula, PA levels
of up to 0.84 mg/kg were observed in the milk. However, only one
(jacoline) of the several PAs contained in the plant was secreted.
Calves, bucket fed on the milk did not show any signs of PA
toxicity.
Dickinson (1980) repeated the study on goats. Four milk goats
were freshly prepared with rumen cannulae. The kids were separated
from their dams and were fed milk twice a day. Dried tansy ragwort
plant material with a PA content of 0.16% (dry weight) was
administered through the cannulae to each goat at a dosage rate of
10 g/kg body weight per day over 125 days. During this period,
each of the 4 kids received milk from their dams at approximately
125 ml/kg per day in addition to ad lib feeding on alfalfa grass
hay. Six PAs were isolated from the plant material: jacobine,
jaconine, jaconline, jacozine, senecionine, and seneciphylline.
Milk samples collected twice daily showed PA contents of
225 - 530 µg/litre with a mean of 381 µg/litre. No apparent health
effects were noted in the kids, and only mild hepatic damage was
suspected in the dams, on the basis of liver function tests. Fifty
percent of the kids were killed after 10 weeks. No lesions of PA
toxicity were seen. The dams were rebred and appeared normal
throughout the gestation period. However, three dams aborted at
almost full term, and the fetuses were born dead. One of the dams
died shortly after parturition and showed evidence of severe liver
damage characteristic of PA toxicity. Another, which delivered
normally, also showed a lesser degree of liver damage at biopsy.
Data relating to PA secretion were compared with similar
earlier data on cows. Mean secretion of PAs in cows appeared much
higher, e.g., 684 µg/litre. The authors concluded that the amount
of PAs secreted in the goat's milk did not cause any serious
deleterious effects in the kids.
Johnson (1976) fed long-term lethal doses of Senecio jacobaea,
by stomach tube, to 6 cows. Feeding started at term or within 30
days post-partum, and continued until what was considered to be a
lethal dose had been fed. The daily dose of the plant ranged from
1 to 4.4 g/kg body weight, the total amount fed representing 5 -
15% of body weight over a period of 54 - 126 days. Five cows died
within 98 days; one, in a moribund state, was killed on day 126.
The calves suckled for 30 - 126 days. Suckling started immediately
after birth in the case of 4 calves and 10 and 30 days later,
respectively, in the 2 remaining calves. Three calves were killed
with their dams or soon after, and 3 were retained for 1 year for
observation. Milk samples from 2 cows were collected and pooled in
14- to 16-day lots during 64 days of feeding of the Senecio plant.
Each pooled sample was administered intragastrically to a group of
rats in daily doses of 12 ml for 15 - 30 days. A control group of
rats were fed raw milk from cows not fed Senecio. Blood samples of
the dams and the calves were analysed for glutamic oxaloacetic
transaminase (GOT), lactic dehydrogenase (LDH), and gamma-glutamyl
transpeptidase (GGTP). Serum-enzyme levels in all cows indicated
statistically significant deviations suggesting liver dysfunction,
and the livers at autopsy had characteristic features of PA
toxicosis. The LDH and GOT levels in calves were generally
abnormal after 20 - 45 days of suckling. The abnormalities ranged
from mild to a 15- to 170-fold increase. One calf was autopsied at
the peak increase of serum-enzymes and was found to have mild focal
hepatitis. No significant pathological features were seen in the
livers of other animals, nor of the rats, some of which were
retained for up to 150 days.
Goeger et al. (1982) fed dried Senecio jacobaea (tansy ragwort)
to lactating goats in a proportion of 25% of the feed. The milk
contained 7.5 µg PA/kg dry weight. The milk produced by the goats
was pooled and then bottle fed to appetite to 2 Jersey bull calves
(1 day old) that also had access to tansy ragwort-free hay for 109
and 124 days, respectively. They were then weaned and given normal
feed and observed for 6 months, after which they were killed and
autopsied. In another study, rats were fed a diet con