The Committee evaluated fumonisins B₁, B₂, and B₃ at the request of the Codex Committee on Food Additives and Contaminants; these toxins had not been evaluated previously by the Committee. In 2000, a monograph on fumonisin B₁ was published (WHO, 2000a), which provided much of the information used in this evaluation.

Fumonisins are fungi produced by fungi of the genus Fusarium. The only species that produce significant quantities of fumonisins are Fusarium verticillioides (Sacc.) Nirenberg (= F. moniliforme (Sheldon)) and the related F. proliferatum (Matsushima) Nirenberg. At least 10 other Fusarium species also produce these toxins. F. verticillioides and F. proliferatum are among the most common fungi associated with maize, the most frequently contaminated food, and can be recovered from both damaged and undamaged maize kernels. These species cause Fusarium kernel rot of maize, an important disease in hot climates. A strong relationship also exists between insect damage and Fusarium kernel rot due to other Fusarium species, such as F. graminearum. Temperature stress may also play a role, especially in cultivars grown outside their area of adaptation. As F. verticillioides and F. proliferatum grow over a wide range of temperatures but only at relatively high water activities (above about 0.9), fumonisins are formed in maize only before harvest or during the early stage of drying. Except under extreme conditions, the concentrations of fumonisins do not increase during grain storage. Formation of fumonisins in the field correlates with the occurrence of F. verticillioides and F. proliferatum, which predominate during late maturity. Fumonisins are widely distributed geographically, and their natural occurrence in maize has been reported in many areas of the world. Of particular concern are the high concentrations found in maize produced and consumed by particular subpopulations, such as subsistence farmers. Considerable annual variations in contamination have been noted. Fumonisins occur infrequently in other foods, such as sorghum, asparagus, rice, beer, and mung beans.

Fumonisins are a group of structurally related compounds. Fumonisin B₁ is the diester of propane-1,2,3-tricarboxylic acid and 2S-amino-12S,16R-dimethyl-3S,5R,10R,14S,15R-pentahydroxyeicosane in which the C-14 and C-15 hydroxy groups are esterified with the terminal carboxy group of propane-1,2,3-tricarboxylic acid. Fumonisin B₂ is the C-10 deoxy analogue of fumonisin B₁ in which the corresponding stereogenic units on the eicosane backbone have the same configuration. The full stereochemistry of fumonisin B₃ and B₄ is unknown, but the amino terminal of fumonisin B₃ has the same absolute configuration as that of fumonisin B₁.

As most biological data were available on fumonisin B₁, and maize is the major source of intake, the Committee focused its evaluation on toxicological studies of fumonisin B₁ and on studies of intake of contaminated maize and maize products. In many studies, culture materials and naturally contaminated maize were used, which can contain several other fumonisins, primarily fumonisins B₂ and B₃. The toxicological profiles of these toxins are very similar to that of fumonisin B₁. Various chemical derivatives of fumonisins have been tested in a number of biological systems to gain insight into structure–activity relationships. Briefly, the fumonisins of the B series that have been investigated are more toxic in vivo than their hydrolysed or N-acetylated counterparts. The free amino group appears to play a specific role in the biological activity of fumonisin B₁.

A thorough review of the biochemical aspects of fumonisin B₁ is contained in Fumonisin B₁ (Environmental Health Criteria 219), published by WHO (2000a), and much of what follows is derived from that review, with relevant new publications. With a few exceptions, only references that were not included in that monograph, recent or critical reviews, or studies with stated doses are cited here.

In many studies on fumonisins in animals in vivo, culture materials or naturally contaminated maize were used in which not only fumonisin B₁ but also other fumonisins (primarily fumonisin B₂ and B₃) can be found. Various chemical derivatives of fumonisins have been studied in several biological systems to gain insight into the structural requirements of fumonisin-induced toxicity and biochemical mechanisms in vivo (see WHO, 2000a, pp. 79–80 for additional details). Briefly, the fumonisins of the B series that have been investigated are more toxic in vivo than their hydrolysed or N-acetylated counterparts. The free amino group has an important role, as the N-acetyl derivatives are less toxic in primary hepatocytes and hydrolysed fumonisin B₁ is more toxic. The lesser toxicity of hydrolysed fumonisin B₁ in rat liver is not due to reduced absorption. The fungus previously known as F. moniliforme Sheldon is referred to in this monograph as F. verticillioides (Sacc.) Nirenberg.

2.1.1 Absorption, distribution, and excretion

In rats and most other animals, the kinetics of absorption of fumonisin B₁ indicates rapid distribution and elimination that is adequately described by a two- or three-compartment model (most recently, Martinez-Larranaga et al., 1999). Little fumonisin B₁ was detected in plasma and tissues after oral administration, indicating that absorption is negligible (­ 4% of the dose). Fumonisin B₂ may be less bioavailable than fumonisin B₁, and proportionally less fumonisin B₂ is excreted in bile (Shephard & Snijman, 1999). In rats treated by gavage, more hydrolysed [¹⁴C]fumonisin B₁ was excreted in urine than [¹⁴C]fumonisin B₁ or the [¹⁴C]fumonisin B₁–fructose adduct. For example, in female Fischer 344NHsd rats, 17% of a dose of hydrolysed [¹⁴C]fumonisin B₁ and 0.7% of a dose of [¹⁴C]fumonisin B₁ was recovered in urine after administration by gavage. The authors concluded that hydrolysed fumonisin B₁ was better absorbed than fumonisin B₁, although the biliary excretion of the two fumonisins was similar (Dantzer et al., 1999). While fumonisin B₁ was distributed to most tissues, the liver and kidney retained most of the absorbed material, the liver retaining more toxin than the kidney in some studies and the inverse in others. In a study in male Wistar rats given pure fumonisin B₁ by gavage, kidney contained > 10 times as much fumonisin B₁ as liver (Martinez-Larranaga et al., 1999). This study confirmed previous reports that fumonisin B₁ persists in rat liver and kidney much longer than in plasma. It was estimated that, after administration to pigs of fumonisin B₁ in the diet at 2–3 mg/kg bw, a withdrawal period of at least 2 weeks would be required to eliminate the toxin from liver and kidney. The material retained in liver and kidney is primarily unmetabolized fumonisin B₁, as shown in several studies of the persistence in kidney of free sphinganine (a biomarker for fumonisin). Thus, while fumonisin B₁ is eliminated rapidly, the concentration of the biomarker in rat kidney (but not liver) is more persistent (most recently, Enongene et al., 2000; Garren et al., 2000). Similar kinetics was seen in the livers of non-human primates given a single dose of pure fumonisin B₁ at 10 mg/kg bw by gavage, which increased the serum concentrations of sphingoid bases, cholesterol, and enzymes indicative of liver function, and these remained elevated for several weeks after dosing (van der Westhuizen et al., 2001).

In pregnant rats, low concentrations of fumonisin B₁ were recovered in uteri (0.24–0.44%), individual placentae (0–0.4%), and fetuses (­ 0.015%), indicating the absence of placental transfer of fumonisin B₁. Similar results were reported in rabbits. There is also little evidence of significant transfer during lactation. For example, no fumonisin B₁ was detected in the milk of lactating sows fed diets containing non-lethal concentrations of fumonisin B₁, and there was no evidence of toxicosis in their suckling piglets. In a study in which lactating cows were given fumonisin B₁ intravenously, the carry-over rate into the milk reached a maximum of 0.11%. In other studies, no fumonisins were detected in cows’ milk, and fumonisin B₁ was found in only 1 of 165 samples of milk in the USA at a concentration of 5 ng/ml. The finding that little fumonisin B₁ is retained in tissues, milk, or eggs has led to the conclusion that fumonisin residues in food products derived from animals are insufficient to make them injurious to consumers.

After intraperitoneally or intravenously administered fumonisin B₁ has been distributed, its initial elimination is rapid, with no evidence of metabolism. In vervet monkeys (Cercopithecus aethiops) treated intraperitoneally with fumonisin B₂, elimination was rapid and followed a bi-exponential pattern (half-time, 18 min), similar to that of the elimination of fumonisin B₁ (half-time, 40 min). The elimination kinetics in non-human primates after oral dosing has not been determined, but peak plasma concentrations of fumonisin B₁ and B₂ occurred 1 to several hours after a dose of 7.5 mg/kg bw by gavage, and the plasma fumonisin concentrations ranged from < 20 ng/ml to nearly 210 ng/ml. Thus, the elimination kinetics after oral dosing is not easily described, unlike that of intraperitoneal or intravenous dosing. Furthermore, an oral dose of fumonisin cannot be fully accounted for (Shephard & Snijman, 1999). As the rate of elimination of fumonisin B₁ is a function of body weight, elimination is rapid in mice but would be much longer in humans (Delongchamp & Young, 2001).

2.1.2 Biotransformation

There is little or no evidence that fumonisins are metabolized in vitro or in vivo in animals, even though they are clearly excreted in bile. A study in which primary hepatocytes were exposed to [¹⁴C]fumonisin B₁ showed that the toxin is associated with both the soluble and the insoluble membrane compartments of the cell, and no metabolites were detected after a 44-h incubation. Incubation with rat liver microsomal preparations also showed no metabolism by cytochrome P450 (CYP), microsomal esterase, or any other microsomal enzyme. Incubation with a triglyceride hepatic endothelial lipase or a porcine pancreatic lipase also did not result in hydrolysis of the tricarboxylic acid moieties of fumonisin B₁.

Several studies in which different routes of exposure and different animal species were used showed that fumonisins are excreted primarily in the faeces, either unchanged or with loss of one or both of the tricarboxylic acid side-chains. The material excreted in bile is still biologically active, since fumonisin B₁ given subcutaneously to mice rapidly entered the small intestine, where it inhibited ceramide synthase (Enongene et al., 2000). Loss of the tricarboxylic acid side-chains probably occurs in the gut, since after partial hydrolysis resulting in removal of only one of the two side-chains and full hydrolysis fumonisin B₁ is recovered in the faeces but not in the bile. This finding was confirmed in a study in vervet monkeys (Shephard & Snijman, 1999). Most of the hydrolysed fumonisin B₂ was present as a mixture of the two possible partially hydrolysed forms, while fully hydrolysed fumonisin B₂ was a minor constituent. No hydrolysis products were found in urine, confirming that fumonisin was hydrolysed in the gut, probably by microbial degradation. While there is no evidence that fumonisin is metabolized by CYP enzymes, some studies have shown that fumonisins can alter their activity, and this observation was confirmed in vitro and in vivo (Merrill et al., 1999a; Spotti et al., 2000). In some studies, the effects on CYP activity have been shown to be the result of fumonisin-induced alterations in sphingolipid metabolism (Merrill et al., 1999b). For example, in HepG2 cells, fumonisin B₁ inhibited the induction of CYP 1A1 (which metabolizes aryl hydrocarbons such as methylcholanthrene) (Merrill et al., 1999a). In rats given fumonisin B₁ by gavage at 2 mg/kg bw, there was inhibition of CYP 2C11 and to a lesser extent of CYP 1A2 (Spotti et al., 2000). The inhibition of CYP 2C11 was attributed to suppression of protein kinase activity due to inhibition of sphingolipid biosynthesis. Sphingosine, a sphingolipid that accumulates in animals exposed to fumonisin B₁, was also shown to inhibit CYP 2C11 in rat hepatocytes (Merrill et al., 1999a). Feeding rainbow trout a diet containing fumonisin B₁ at a concentration of 104 mg/kg had no effect on acetylated fumonisin B₁–DNA adduct formation (Carlson et al., 2001).

While there is little evidence that absorbed fumonisins are metabolized in animals, removal of the tricarboxylic acid side-chains (producing hydrolysed fumonisin B₁) converts this inhibitor of ceramide synthase into a substrate for the enzyme. The product of this reaction, N-palmitoyl-hydrolysed fumonisin B₁, also inhibits ceramide synthase in vitro. It is not known whether this product is formed in vivo, but it is more toxic than fumonisin B₁ or hydrolysed fumonisin B₁ for HT-29 cells (Merrill et al., 2001). Since hydrolysed fumonisin B₁ and hydrolysed fumonisin B₂ are major breakdown products in nixtamalized maize products and are also produced in the gut from fumonisin B₁ and B₂, the toxicity of the hydrolysed toxins should be addressed.

2.1.3 Effects on enzymes, biochemical parameters, and cellular regulation

Several biochemical reactions have been proposed to explain all or some of the toxic effects of fumonisins in animals. Two of them invoke disruption of lipid metabolism as the initial site of action, and they are similar in other respects (Gelderblom et al., 2001a; Merrill et al., 2001; Riley et al., 2001). Both of these hypothesized mechanisms are supported by data obtained in in vivo (Table 1) and in vitro (Table 2), in short-term and long-term studies in rodents (National Toxicology Program, 1999; Delongchamp & Young, 2001; Gelderblom et al., 2001a; Riley et al., 2001; Voss et al., 2001), long-term studies of carcinogenicity in trout (Carlson et al., 2001), and short-term studies of toxicity in other animals (WHO, 2000a). The first proposed lipid-based mechanism involves inhibition of ceramide synthase, a key enzyme in the biosynthesis of sphingolipids (reviewed extensively in WHO, 2000a). The second mechanism involves changes in the polyunsaturated fatty acid and phospholipid pools. Both lead ultimately to lipid-mediated alterations in signalling and metabolic pathways crucial to cell growth, death, and differentiation. Several studies in vitro indicate that fumonisin-induced changes in key enzymes involved in cell cycle regulation, differentiation, and/or apoptosis are initial or secondary triggers (most recently, Pinelli et al., 1999; Mobio et al., 2000a; Table 2).

The structural similarity between sphinganine and fumonisin B₁ led to the hypothesis that this mycotoxin acts by disrupting the metabolism or a function of sphingolipids. There is considerable support for the hypothesis that fumonisin-induced disruption of sphingolipid metabolism is an important event in the cascade of events leading to altered cell growth, differentiation, and cell injury observed both in vitro and in vivo (Tables 1–3). A complete review of the literature on this subject is beyond the scope of this monograph, and the interested reader is referred to WHO (2000a) and reviews by Merrill et al. (2001) and Riley et al. (2001)

Fumonisin B₁ strongly inhibited the acylation of sphinganine and sphingosine in all cell lines and all animals, plants, and fungi in which it has been tested. Ceramide synthase recognizes both the amino group (sphingoid-binding domain) and the tricarboxylic acid side-chains (fatty acyl-coenzyme A domain) of fumonisin B₁. While removal of the tricarboxylic acid side-chains reduces the ability of fumonisin B₁ to inhibit ceramide synthase, N-acetylation completely abolishes the inhibitory activity (Norred et al., 2001). Complete inhibition of ceramide synthase by fumonisins causes a rapid increase in the intracellular concentration of sphinganine and sometimes of sphingosine, both in vivo and in vitro. In vivo there is a close relationship between sphinganine accumulation and the expression of toxicity in liver and kidney (Delongchamp & Young, 2001). Once accumulated, free sphingoid bases can persist in tissues (especially kidney) much longer than fumonisin B₁ (Shephard & Snijman, 1999; Enongene et al., 2000). In the urine of rats fed fumonisin B₁, > 95% of the free sphinganine was recovered in dead cells. An oral dose of fumonisin B₁ insufficient to increase the concentration of free sphinganine (1 mg/kg of diet, equivalent to 0.1 mg/kg bw per day) can prolong the half-life of free sphinganine in urine of rats after they have been taken off diets that contained a dose sufficient to cause free sphinganine (10 mg/kg of diet, equivalent to 1 mg/kg bw per day) to accumulate in urine (Wang et al., 1999). This observation that a sub-threshold dose can prolong the increase in free sphinganine caused by a higher dose has been confirmed in mice treated by oral gavage (Enongene et al., 2001). Fumonisin B₁-induced increases in free sphingoid bases and toxicity are both reversible, although elimination of free sphinganine from the liver is faster than from the kidney (Enongene et al., 2000).

A portion of the accumulated sphinganine is metabolized to sphinganine 1-phosphate and then cleaved into a fatty aldehyde and ethanolamine phosphate, both of which can be redirected into other biosynthetic pathways, such as increased biosynthesis of phosphatidylethanolamine. Free sphinganine that is not degraded can be acylated to form C-2 dihydroceramide (Merrill et al., 2001). Disrupted sphingolipid metabolism leads to imbalances in phosphoglycerolipid, fatty acid, and cholesterol metabolism by alterating phosphatidic acid phosphatase and monoacyl-glycerol acyltransferase. Thus, inhibition of ceramide synthase by fumonisin B₁ can cause a wide spectrum of changes in lipid metabolism and associated lipid-dependent signalling pathways (reviewed by Merrill et al., 2001).

In short-term studies with mice, rats, and rabbits, disruption of sphingolipid metabolism, as shown by statistically significant increases in the free sphinganine concentration, occurred at doses at or below those that cause liver or kidney lesions (Table 1; for review, see Riley et al., 2001). In a long-term study in which Fischer 344/N Nctr BR rats were fed diets containing pure fumonisin B₁, the increase in the ratio of sphinganine:sphingosine in kidney and urine correlated with increased incidences of non-neoplastic and neoplastic lesions in the kidney (National Toxicology Program, 1999). In the livers of female B6C3F₁/Nctr BR mice, the concentrations of free sphinganine and the ratio of sphinganine:sphingosine were increased after 3 and 9 weeks on a diet containing fumonisin B₁ at 50 or 80 mg/kg, which doses induced liver adenoma and carcinoma (National Toxicology Program, 1999). In rainbow trout, fumonisin B₁ was not a complete carcinogen, but there was a close correlation between promotion of hepatocarcinogenicity caused by aflatoxin B₁ and the concentration of free sphinganine in liver (Carlson et al., 2001; Riley et al., 2001). A mathematical model based on the data from the study of the National Toxicology Program confirmed that the concentration of sphinganine in mouse liver and rat kidney was a dose-related biomarker of fumonisin B₁-induced cell death (Delongchamp & Young, 2001). In Sprague-Dawley rats fed AIN-76 diets containing pure fumonisin B₁, the LOEL for an increased urinary concentration of free sphinganine was 5 mg/kg of diet (equivalent to 0.5 mg/kg bw per day) (Wang et al., 1999). In rats fed a diet containing a mixture of fumonisins from culture material for 13 days, the NOEL was 1–2 mg/kg of diet (equivalent to 0.1–0.2 mg/kg bw per day) (Solfrizzo et al., 1997).

The potential problems of using increased free sphinganine as a functional biomarker of human exposure to fumonisin B₁ have been reviewed (Turner et al., 1999). In a study in men in China, the ratio of free sphinganine to free sphingosine in urine was significantly greater for men in households where the estimated daily intake of fumonisin B₁ was > 110 µg/kg bw per day (Qui et al., 2001). In several other studies, an increase in free sphinganine in human urine or blood did not appear to be associated with fumonisin intake (e.g. van der Westhuizen et al., 1999).

In cultured cells, the sphingolipid-dependent mechanisms for inducing apoptosis include accumulation of excess ceramide, glucosylceramide (Korkotian et al., 1999), or sphingoid bases and depletion of ceramide, or more complex sphingolipids (Table 2). Conversely, the balance between sphingosine 1-phosphate and ceramide is critical for signalling proliferation or cell survival (Spiegel, 1998). A diversity of alterations in cellular regulation resulting from disruption of sphingolipid metabolism by fumonisin B can also be expected. This is demonstrated in numerous studies of the identify of cell processes mediated by ceramides (Table 3). Many of these processes are relevant to understanding the toxicity and carcinogenicity of fumonisin B₁, in particular, the ability of this toxin to protect oxidant-damaged cells from apoptosis and to alter the proliferative response (Table 3).

There is no doubt that loss of complex sphingolipids also plays a role in the abnormal behaviour, altered morphology, and altered proliferation of fumonisin-treated cells (Tables 1 and 2), in particular the ability of fumonisin B₁ to alter the function of specific glycosphingolipids and lipid rafts (membrane associations of sphingolipids, ceramide-anchored proteins, and other lipids). Examples are functions such as vitamin and toxin transport and cell–cell and cell–substratum contact (Table 3).

Inhibition of ceramide biosynthesis by fumonisins also protects cells from oxidant-induced cell death (Table 3). As the metabolism of ceramides is sensitive to the redox state of the cell, it is of particular interest that mice, pigs, and horses treated with fumonisin have increased amounts of complex sphingolipids containing sphinganine as the long-chain sphingoid-base backbone. The ceramide generated from these complex sphingolipids is dihydroceramide, the form of ceramide that is inactive in ceramide signalling and does not induce apoptosis of oxidant-damaged hepatocytes (Arora et al., 1997). Dihydroceramide also occurs in greater amounts in mouse hepatoma cells, in which 37% of the ceramides contain sphinganine, as compared with 5% in normal rat liver (Rylova et al., 1999).

Essential fatty acids are major constituents of all cell membrane glycerophospholipids, sphingolipids, and triglycerides. Apart from being structural components of all membranes, they are precursors of eicosanoids, prostaglandins, leukotrienes, and other oxygenated derivatives. In addition, the regulated turnover of membrane phospholipids is important in many intracellular signalling systems known to regulate cell growth, death, and differentiation. In rat liver, fumonisin B₁ induced changes in phospholipids and their fatty acid composition that markedly affected the many cell functions that may contribute to its toxicity and carcinogenicity. The following summary is from the review of Gelderblom et al. (2001a).

In primary hepatocytes treated with fumonisin B₁, incorporation of [¹⁴C]palmitic acid into total lipids, neutral lipids, triacylglycerol, and cholesterol esters were decreased, whereas it was increased in phospholipids (Gelderblom et al., 1996a). There was a concomitant increase in the cellular concentrations of phosphatidyl-choline and phosphatidylethanolamine, while the total cholesterol concentration decreased. The concentration and labelling of sphingomyelin was also decreased. Changes in the concentrations of specific polyunsaturated fatty acids were attributed to disruption of the Delta6 desaturase and cyclo-oxygenase metabolic pathways. These could be important in fumonisin B₁-induced toxicity in primary hepatocytes.

Fumonisin B₁ disrupted fatty acid and phospholipid biosynthesis in rat liver in vivo. In contrast to the situation in vitro, the main changes were associated with the phosphatidylethanolamine and phosphatidylcholine fractions, while the concentration of cholesterol was increased in both serum and liver (Tables 1 and 2). A characteristic fatty acid pattern was seen in the livers of rats given diets containing purified fumonisin B₁ at concentrations that caused preneoplastic hepatic lesions in rats treated only with fumonisin B₁ (> 1.6 mg/kg bw per day) or in rats fed fumonisin B₁ after treatment with known liver cancer initiators such as N-nitrosodiethylamine (> 0.8 mg/kg bw per day). The pattern includes:

As glycerophospholipids are important components of many cellular signalling pathways, perturbation of the phospholipid and fatty acid composition of cellular membranes could have a strong effect on processes that control cell growth and cell death (Tables 1 and 2). For example, increased expression of hepatocyte growth factor, transforming growth factor (TGF)-alpha) and especially TGF-beta1 and c-myc was seen in rat liver during short-term feeding with fumonisin B₁ (Lemmer et al., 1999b). Overexpression of TGF-beta1 may contribute to the increased apoptosis seen in the livers of rodents fed fumonisin B₁. The proto-oncogene c-myc is a positive regulator of cell proliferation that is involved in tumour progression (Nagy et al., 1988); it has also been implicated in TGF-beta1 signalling (Alexandrow & Moses, 1995). Increased expression of c-myc and TGF-beta1 may cooperate in the promotion of liver tumours during feeding of fumonisin B₁, possibly by providing an environment that selects for the growth of TGF-beta1-resistant transformed liver cells. Overexpression of c-myc, depletion of growth factors and/or disruption of growth signalling pathways could result in imbalances in cell cycle progression and hence the induction of apoptosis (Steiner et al., 1996). In this regard, fumonisin B₁ overexpressed c-myc in rat liver (Lemmer et al., 1999b), while it disrupted growth-related responses in cell types such as primary hepatocytes and in the liver in vivo (reviewed by Gelderblom et al., 2001a).

In male BDIX rats fed diets containing purified fumonisin B₁ for 2 years, cyclin D1 accumulated in the nuclei of altered hepatocytes in foci, nodules, adenomas, and carcinomas (Ramljak et al., 2000). In male Fischer rats fed diets containing fumonisin B₁ for 21 days, the concentration of cyclin D1 protein in liver was increased up to fivefold in a dose-dependent manner, with no simultaneous increase in mRNA. A fumonisin B₁-induced increase in cyclin-dependent kinase 4 was confirmed by increased phosphorylation of the retinoblastoma protein; the accumulation of cyclin D1 appeared to result from stabilization of the protein associated with activation of protein kinase B (Akt), followed by inhibition of glycogen synthase kinase 3beta activity. Akt is part of the anti-apoptotic phosphatidylinositol 3 kinase cell survival pathway (Dudek et al., 1997) and can be activated by stimuli involving growth factors and cytokines and inhibited by the pro-apototic molecule, ceramide (Franke et al., 1997; Zhou et al., 1998). Therefore, the modulating effects of fumonisin B₁ on both sphingolipids and phospholipids could play a major role in molecular events involving the stability of cyclin D1 protein (Ramljak et al., 2000).

It has been proposed that glycerophospholipids and the sphingolipid cycle interact to control a variety of cellular processes, including apoptosis. For example, C20:4omega6 generated by cytosolic phospholipase A₂ initiated sphingomyelin hydrolysis, whereas ceramide generated de novo during ceramide synthase stimulated C20:4omega6 release via secretory phospholipase A₂ (Jayadev et al., 1994; Balsinde et al., 1997). Ceramide has also been shown to regulate transcription of cyclooxygenase-2 (Subbaramaiah et al., 1998). A similar interactive pathway is likely to exist for fumonisins in the liver to regulate processes related to cell proliferation and apoptosis. Fumonisin B₁ induces similar changes in phospholipids and in the profile of fatty acids in both the liver and hepatocyte nodules (Abel et al., 2001). However, subsequent effects on sphingolipid and/or prostaglandin production appear to inhibit the growth of normal hepatocytes, which, with the overexpression of TGFbeta-1 and c-myc, could affect apoptosis. Oxidative damage and the resultant lipid peroxidation products may further enhance apoptosis in the liver (Chen et al., 1997). Conversely, the increased C18:1omega9, the decreased long-chain polyunsaturated fatty acid concentration, and a phosphatidyl-ethanolamine-associated increase in C20:4omega6 fatty acids are critical for cell proliferation (Tang et al., 1993; Horribin, 1994), especially in initiated cell populations.

Several studies of fumonisins in vitro showed changes in cellular regulation and function that were attributed to actions independent of altered lipid metabolism (Tables 1 and 2). Many of these effects might be relevant to the toxicity of fumonisins.

Thus, fumonisin-induced disruption of lipid metabolism is observed both in vitro and in vivo. The biochemical consequences of the disruption of sphingolipid metabolism that are most likely to alter cell regulation are increased concentrations of free sphingoid bases and their 1-phosphates, alterations in complex sphingolipids, and decreased ceramide biosynthesis. Because free sphingoid bases and ceramide can induce cell death, inhibition of ceramide synthase can inhibit cell death induced by ceramide but can promote cell death induced by free sphingoid bases. The kinetics of the increases and decreases in the various bioactive sphingolipid pools in liver, kidney, lung, and heart is also important in the toxicity of these toxins. Fumonisins also induced changes in fatty acids and phospholipids in primary rat hepatocytes and rat liver in vivo, which closely reflected those expected from disruption of the Delta6 desaturase enzyme, the rate-limiting enzyme in fatty acid metabolism, and disruption of prostaglandin biosynthesis. The changes in fatty acid and phospholipid metabolism that probably alter cell regulation are changes in the degree of saturation of fatty acids in the phospholipid pools, increases in the ratio of phosphatidylcholine to phosphatidylethanolamine, changes in prostaglandin biosynthesis, and altered ceramide production.

Fumonisins also affect sites of cellular regulation that are apparently independent of the disruption of lipid metabolism. Nevertheless, disruption of lipid metabolism, membrane structure, and signal transduction pathways mediated by lipid second messengers appear to be important in all the proposed mechanisms of action.

2.1.4 Hypothesized cellular mechanisms of toxicity and carcinogenicity

Two cellular modes of action for the toxicity and carcinogenicity of fumonisin B₁ have been proposed that are well supported by results obtained in vivo. In both hypotheses, altered lipid metabolism is the initial biochemical mechanism. In one hypothesis, the initial biochemical lesion is presumed to be inhibition of ceramide synthase (Merrill et al., 2001; Riley et al., 2001), and in the other, the biochemical lesion is attributed to disruption of the Delta6 desaturase and cyclooxygenase metabolic pathways (Gelderblom et al., 2001a). In both hypotheses, it is assumed that other initial sites of action could contribute to the observed cellular responses. The two invoke similar cellular mechanisms, to the extent that fumonisin B₁-induced imbalances in the rates of cell death and proliferation in target tissues are considered to contribute to cancer development (Dragan et al., 2001; Howard et al., 2001a).

Fumonisin-induced disruption of sphingolipid metabolism in target tissues has been demonstrated in many independent studies. Nonetheless, the way in which disrupted sphingolipid metabolism contributes to toxicity in rodents is unclear. Current understanding of the sphingolipid signalling pathways (reviewed by Merrill et al., 2001; Riley et al., 2001) indicates that the balance between the intracellular concentrations of sphingolipid effectors that protect cells from apoptosis (decreased ceramide, increased sphingosine 1-phosphate) and the effectors that induce apoptosis (increased ceramide, increased free sphingoid bases, increased fatty acids) determines the cellular response. Cells sensitive to the proliferative effect of decreased ceramide and increased sphingosine 1-phosphate will be selected to survive and proliferate. Conversely, when the rate of increase in free sphingoid bases exceeds a cell’s ability to convert sphinganine or sphingosine to dihydro-ceramide or ceramide or their sphingoid base 1-phosphate, then free sphingoid bases will accumulate to toxic levels. In this case, cells that are sensitive to sphingoid base-induced growth arrest will cease growing, and insensitive cells will survive. Thus, the kinetics of fumonisin elimination (rapid), the affinity of fumonisin B₁ for ceramide synthase (competitive and reversible), and the kinetics of sphinganine elimination (persistent but reversible) could affect the time course, amplitude, and frequency of peaks in the intracellular concentrations of ceramide, sphingoid base-1 phosphates, and free sphinganine in tissues of animals given diets containing fumonisins. This is important, because the balance between the rates of apoptosis and proliferation is a critical determinant in hepato- and nephrotoxicity and tumorigenesis in animal models (Dragan et al., 2001; Howard et al., 2001a; Voss et al., 2001). At the cellular level, apoptotic necrosis should be considered to be similar to oncotic necrosis (as defined by Levin et al., 1999), in that both lead to a regenerative process involving sustained cell proliferation (Dragan et al., 2001; Hard et al., 2001). Numerous endogenous processes can cause DNA damage; while most are repaired, an unrepaired mutation in DNA can occur. Increased DNA replication can thus increase the risk that damaged DNA is also replicated, resulting in an increased cancer risk. With respect to fumonisin B₁, this hypothesis is best supported by observations in rat kidney, as in liver apoptotic and oncotic necrosis occur concurrently (Dragan et al., 2001). Nonetheless, regeneration after either apoptotic or oncotic necrosis is observed at the same doses that cause cancer development.

Fumonisin B₁ alters normal cell proliferation in rat liver in vivo (Gelderblom et al., 1994, 1996a; Li et al., 2000) and in primary rat hepatocytes and many other cell lines in vitro (Gelderblom et al., 1994; Tolleson et al., 1996). Differential inhibition of cell proliferation is a possible mechanism: hepatocytes resistant to fumonisin B₁-induced inhibition of cell growth are selectively stimulated to grow by creating an environment in which the growth of normal cells is impaired. Selective inhibition of normal cell growth could increase the chances of survival of cells initiated by processes such as free-radical damage, leading ultimately to manifestations of cancer initiation in liver, such as glutathione S-transferase, placental form (GST-P)⁺ foci and hepatocyte nodules. Fumonisin B₁ may be not only a complete carcinogen but also promote cancer, and this should be considered in models for risk assessment.

Disruption of the Delta6 desaturase and cyclooxygenase metabolic pathways in the livers of male rats has been well documented. Disruption of lipid-mediated growth stimulation in the liver could be important in establishing a growth differential that results in clonal expansion of certain cell types associated with neoplastic development. For example, disruption of C20:4omega6 metabolism altered the mitogenic response to epidermal growth factor in primary rat hepatocytes, a known property of many hepatocarcinogens (Gelderblom et al., 1999a). Three lines of evidence support the hypothesis that fumonisin B₁-induced alteration of lipid metabolism contributes to establishing a growth differential in rat liver. First, the lipid parameters associated with increased cell proliferation in hepatocyte nodules closely mimic those of normal regeneration in the liver; one major difference is that the changes in the nodules are persistent, whereas they are reversed in regenerating liver (Abel et al., 2001). The increased concentrations of phosphatidylethanolamine and C20:4omega6 are of special interest, as this fatty acid is known to regulate many processes related to cell growth, such as proliferation and apoptosis (Khan et al., 1995). Several studies indicated that fumonisin B₁ interacts with C20:4omega6 metabolism in normal and cancer cell lines in vitro (Gelderblom et al., 1999a; Pinelli et al., 1999; Seegers et al., 2000). Second, alterations in the N-6 fatty acid desaturase pathway and the subsequent decrease in long-chain polyunsaturated fatty acids would result in a more rigid membrane structure. This could provide a survival advantage for hepatocytes under stress, since the membranes will be resistant to oxidative damage. Such membrane changes occur preferentially in hepatocyte nodules, creating an environment for differential growth relative to the surrounding normal tissue. Changes in membrane fluidity could also alter the response of membrane receptors and enzymes by affecting their mobility in the bilayer. Third, lipid metabolites, and in particular glycerophospho-lipids, are important components of many cellular signalling systems that control the balance between cell growth and cell death.

In the classical model of cancer initiation by genotoxic carcinogens, fumonisins did not increase the incidence of hepatocellular foci after single or multiple doses (Gelderblom et al., 1992). Subsequent studies indicated that the ‘effective dose’ for induction of preneoplastic lesions in liver, such as GST-P⁺ foci and hepatocyte nodules, in male Fischer rats depends on the dose and duration of exposure (Gelderblom et al., 1994). The toxicity of fumonisin B₁ in rat liver appears to play an important role in cancer development (Abel & Gelderblom, 1998). Induction of oxidative damage and lipid peroxidation as a consequence of toxicity (Tables 1 and 2) could lead to DNA damage. Changes in the balance of the various cell regulatory molecules, such as those found in the livers of rats fed diets containing fumonisin B₁, are likely to be involved in induction of a growth differential in which the growth of initiated cells is selectively stimulated and cancer develops.

Many of the studies summarized in this section are also described in the IPCS monograph (WHO, 2000a). The pathological findings and doses in pertinent studies are given in Tables 4–6. The diets used in most of the studies summarized in this section differed markedly in nutritional composition. As the proposed mechanisms of action involve alterations in de novo biosynthesis, nutritional factors might be important in toxic end-points. The liver was the target for fumonisin B₁ in all animals in which toxicity was observed, and the kidney was also a target in many animals. In both liver and kidney, fumonisin B-induced toxicity is often characterized initially by increased apoptotic and oncotic necrosis, regeneration, and, in the case of liver, bile-duct hyperplasia (Tables 4 and 5). In rodents, the toxicity of fumonisin B₁ depends on the strain and sex. For example, male BD IX rats appeared to be more resistant to the nephrotoxic effects of fumonisin B₁ than male Fischer 344N, male Sprague-Dawley, and male RIVM:WU rats.

F. verticillioides culture material and naturally contaminated maize can contain various fumonisins and other mycotoxins. However, naturally contaminated maize and culture material of the F. verticillioides isolate known as MRC 826 contain predominantly fumonisins of the B series. Therefore, the results of studies with these materials corroborate those of studies in which pure fumonisin B₁ was used. Although the results were not used for hazard characterization or risk assessment, when possible, the NOEL or LOEL was calculated for comparison with the results of studies with pure fumonisin B₁.

2.2.1 Acute toxicity

No studies have been published on the lethality of single doses of pure fumonisin B₁ in laboratory animals. The few available studies indicate that fumonisins are not acutely toxic. For example, mice given a single dose of 25 mg/kg bw by gavage or subcutaneous injection showed reversible alterations in cytokine expression, serum enzymes activity, and blood cell counts (Bhandari et al., 2001).

2.2.2 Short-term studies of toxicity

Male mice were given a subcutaneous dose of fumonisin B₁ at 0.3, 0.8, 2.3, or 6.8 mg/kg bw per day for 5 days. Apoptosis was detected in the livers of mice at doses > 0.8 mg/kg bw per day and in the kidney at all doses (Sharma et al., 1997; Tsunoda et al., 1998). If it is assumed that 10% of an oral dose would be absorbed in mice, the calculated LOEL for oral administration would be 8 mg/kg bw per day in liver and 3 mg/kg bw per day in kidney. The relative weight of the kidney was decreased at all doses except 0.8 mg/kg bw per day; no effect was observed on the relative weight of the liver. Increased apoptosis in liver was also seen in subsequent studies with various strains of mice and transgenic mouse models. The response to fumonisin differed in some transgenic mouse models in comparison with the wild type (Sharma et al., 2000a,b,c). The results of studies with mice that overexpress or lack tumour necrosis factor (TNF)-alpha suggest that this pathway plays a role in the hepatic toxicity of fumonisin B₁ in mice.

Male and female mice were fed diets containing fumonisin B₁ at a concentration of 1, 3, 9, 27, or 81 mg/kg for 90 days, providing mean intakes of 0.3, 0.9, 2.5, 7.4, and 23 mg/kg bw per day for males and 0.3, 1, 3, 9.7, and 29 mg/kg bw per day for females. The serum concentrations of cholesterol and total bilirubin and the activities of alanine and aspartate aminotransferases, alkaline phosphatase, and lactate dehydrogenase were significantly increased in female mice at the high dose, with no effect in males. The changes were parallelled by histological alterations in the livers of the female mice, mainly restricted to the centrilobular zone. No lesions were reported in the kidney, but some were detected in the adrenal cortex (presumably a normal physiological reaction to stress induced by the treatment) in all females given the highest dose and two females given the next lowest dose (Table 4; Voss et al., 1995a).

Adult male and female mice were given fumonisin B₁ at a daily dose of 1, 5, 15, 35, or 75 mg/kg bw per day by gavage for 14 days. Hepatotoxicity was observed in animals of each sex, but renal toxicity was seen only in females. Females were more sensitive than males to these effects. Single-cell necrosis was found in the liver at a dose of 35 mg/kg bw per day in males and 15 mg/kg bw per day in females. Increased hepatocyte mitosis was seen at 75 mg/kg bw per day in males and > 5 mg/kg bw per day in females. Mild single-cell necrosis was detected in the cortical and medullary tubules only in female mice at 15–75 mg/kg bw per day. Males (at doses > 35 mg/kg bw per day) and females (at doses > 15 mg/kg bw per day) had moderate diffuse vacuolization of adrenal cortical-cell cytoplasm. Mild thymic cortical lymphocytolysis was noted in a few female mice that received doses > 35 mg/kg bw per day (Table 4; Bondy et al., 1997).

Male and female mice were fed diets containing fumonisin B₁ at a concentration of 99, 160, 230, or 480 mg/kg for 28 days in order to establish the doses for a 2-year bioassay. Males at the highest concentration developed liver lesions. Changes in clinical chemical end-points and cell cycle progression indicative of increased proliferation paralleled the pathological changes. Similar changes were seen in females but at all doses. Thus, female mice were more sensitive than males to the hepatic toxicity of fumonisin B₁. No NOEL could be identified, as liver lesions were seen in females at all doses (Table 4; National Toxicology Program, 1999).

Male rats were fed diets containing fumonisin B₁ at a concentration of 0.1% for 33 days. Major changes were reported in the liver and mild changes in kidney (Table 5). After 3 days of treatment of male rats with fumonisin B₁ at 240 mg/kg bw per day by gavage, major pathological lesions were observed in the liver and only minor changes in kidney. Severe disseminated acute myocardial necrosis and severe pulmonary oedema were observed in two rats. At lower doses but longer treatment (9–12 days), pathological changes were observed only in liver. Early signs of bile-duct proliferation and fibrosis radiating from the portal areas were noted, and the nuclei of a few hepatocytes were enlarged (Gelderblom et al., 1988).

Culture material from F. verticillioides, strain MRC 826, was fed to male rats for 77 days. All developed cirrhosis, intraventricular cardiac thrombosis and nephrosis (Kriek et al., 1981a).

When male rats were fed diets containing fumonisin B₁ at a concentration of 0.1% for 55 days, the pathological lesions reported in the livers were similar to those described in male BD IX rats. Hepatocyte nodules were induced when the fumonisin B₁-fed rats were subjected to promotion by 2-acetylaminofluorene and partial hepatectomy (Gelderblom et al., 1992). In subsequent studies, in which rats were given diets containing 25–750 mg/kg, histopathological lesions were observed in the livers of animals at 500 and 750 mg/kg of diet after 21 days. Similar changes, although less pronounced, were observed in the rats that received 100 and 250 mg/kg of diet. A few necrotic cells were detected in the livers of rats at 50 mg/kg of diet, but none were observed at 25 mg/kg of diet. Measurements indicated that the mean daily intake of fumonisin B₁ at these six doses was 1.7, 3.5, 7.2, 15, 25, and 31 mg/kg bw per day. Twenty-one days after the promoting treatment, GST-P⁺ foci and nodules were induced at 250 mg/kg of diet or a daily intake > 15 mg/kg bw. The dose-dependency for inducing the putative preneoplastic lesions in liver was also seen after daily treatment by gavage for 14 days. After the promoting treatment, hepatocyte nodules and toxic effects in the liver were seen only in rats that received a daily dose > 8.5 mg/kg bw. These studies showed that fumonisin B₁-induced hepatoxicity was a prerequisite for development of the preneoplastic lesions in rat liver (Table 5; Gelderblom et al., 1994).

In a study of the cancer promoting potential of fumonisin B₁, similar dietary concentrations were fed to male rats over 21 days. Pathological changes were reported in the livers of rats receiving 50 mg/kg of diet (3.5 mg/kg bw per day) but not in those that received 10 mg/kg of diet (0.7 mg/kg bw per day), and cancer promotion was seen in N-nitrosodiethylamine-initiated rats at the higher concentration (Table 5; Gelderblom et al., 1996b). Microscopic lesions were seen in the kidneys of rats that received the low dose (W.C.A. Gelderblom, unpublished data) and in the inner cortex and outer medulla of rats at the high dose. The changes were common in animals at the high dose and only mild at 10 mg/kg of diet (0.7 mg/kg bw per day). These results indicate that similar lesions are induced by comparable doses of fumonisin B₁ in the livers of male BD IX and Fischer rats, but in Fischer rats lesions were detected in the livers at a dietary concentration as low as 50 mg/kg and in the kidneys at a concentration as low as low as 10 mg/kg.

The kinetics of hepatocyte injury in liver was investigated in male rats fed a diet containing fumonisin B₁ at 250 mg/kg for 5 weeks. Hepatocyte necrosis and apoptosis were found mainly in zone 3 of the liver lobule. Hepatocyte injury and death were mirrored by desmin-positive hepatic stellate-cell proliferation and marked fibrosis, with progressive disturbance of the architecture and formation of regenerative nodules. Oval-cell proliferation was seen in the presence of hepatocyte mitotic activity. Oval-cell (OV-6 positive) proliferation was noted from week 2; glutathione S-transferase-positive hepatic foci and nodules developed; and, later, oval cells were seen inside some of the ‘atypical’ nodules (Table 5; Lemmer et al., 1999b).

In a subsequent experiment, groups of 12 male Fischer rats were fed diets containing fumonisin B₁ at a concentration of 250 mg/kg for 5 weeks followed by 100 mg/kg of diet for the remainder of the experiment, up to 6 months, when exposure was discontinued (‘stop model’). Liver biopsy samples were collected from a subpopulation of the rats for another 6 months. The body-weight gain of treated animals was significantly lower after 6 months. A variety of preneoplastic and neoplastic hepatic lesions were observed in the samples collected. The intake of fumonisin B₁ was estimated to be 15 mg/kg bw per day (Table 5; Vessey et al., 1999; Lemmer, 2000).

Male and female rats were fed diets containing fumonisin B₁ at a concentration of 99, 160, 230, or 480 mg/kg for 28 days, in order to set the doses for a 2-year bioassay, providing average daily doses of the toxin of 12, 20, 28, and 56 mg/kg bw per day. The body weights of both males and females were at the highest dose were decreased. Similar pathological lesions and effects were seen in the livers and kidneys of males and females. The male kidney was the most sensitive target for fumonisin B₁, and the liver was more severely affected in females than in males. The earliest cellular response in both liver and kidney was increased apoptosis, which was accompanied by increased cell proliferation. Structural degeneration as a result of apoptosis was seen in both liver and kidney. In females, the NOEL for bile-duct hyperplasia and decreased liver weight was 28 mg/kg bw per day, and that for liver degeneration and increased hepatocellular mitosis was 20 mg/kg bw per day. The NOEL for increased hepatocellular apoptosis and for decreased kidney weight, increased mitosis, increased structural degeneration, and increased mitosis in males was < 12 mg/kg bw per day (Table 5; Tolleson et al., 1996; National Toxicology Program, 1999; Howard et al., 2001b).

Male and female rats were fed diets containing fumonisin B₁ at a concentration of 1, 3, 9, 27, or 81 mg/kg for 90 days. No hepatotoxic effects were seen in male or female rats, whereas renal toxicity was found in male rats fed diets containing > 9 mg/kg. Female rats were more resistant, as renal changes were observed only at 81 mg/kg of diet (5.7 mg/kg bw per day); however, when kidney weight was used as a marker, a significant difference was found between control females and females fed 9 mg/kg of diet (0.7 mg/kg bw per day). The NOEL for renal toxicity in male rats was 0.2 mg/kg bw per day (Table 5; Voss et al., 1995a).

A maize sample naturally contaminated with F. verticillioides that had been involved in an epizootic of equine leukoencephalomalacia in the USA was fed to male rats for 5–6 months. The liver was the major organ affected (Table 6; Wilson et al., 1985). Retrospective analyses indicated the the total concentration of fumonisin B₁ and B₂ was 33 mg/kg, and only trace amounts of aflatoxin B₁ and B₂ were present. Dietary deficiencies, especially those related to methionine and choline, were suggested to have accelerated and/or promoted the liver lesions. The mean fumonisin B intake was calculated to be equivalent to 2.3–3.2 mg/kg bw per day (Gelderblom et al., 2001a).

Male and female rats were fed diets containing fumonisin B₁ at a concentration of 15, 50, or 150 mg/kg for 4 weeks, providing estimated daily intakes of 1.4, 4.7, and 14 mg/kg bw per day for males and 1.4, 4.1, and 13 mg/kg bw per day for females. Only mild changes were observed by light microscopy in the livers of rats fed the high dose. The NOEL for effects on the liver was thus 4.1–13 mg/kg bw per day. The nephrotoxic changes were localized in the proximal convoluted tubules of males fed diets containing > 15 mg/kg and of females at > 50 mg/kg. The NOEL for effects on the kidney was < 1.4 mg/kg bw per day in males and 1.4 mg/kg bw per day in females (Table 5; Voss et al., 1993, 1995b). Ultrastructural changes were observed in the livers in males and females at doses > 15 and > 50 mg/kg of diet, respectively. The activity of serum enzymes and the concentrations of cholesterol and triglycerides were increased at 150 mg/kg of diet (Riley et al., 1994). In a subsequent study in male rats dosed intraperitoneally with fumonisin B₁ at 2 mg/kg bw per day for 4 days, alanine aminotransferase activity, serum cholesterol concentration, and apoptosis (on the basis of morphology and the terminal UTP end-labelling assay) were increased in liver on day 3. Although increased proliferation was also seen on day 3, it was not statistically significant until day 5 (only liver was examined on days 1, 3, and 5) (Li et al., 2000).

In male and female rats given fumonisin B₁ at 1, 5, 15, 35, or 75 mg/kg bw for 11 days by gavage, the histopathological changes in the kidneys were similar to those described in other studies, and males were more sensitive to the renal effects (< 1.0 mg/kg bw per day for males and 5 mg/kg bw per day for females). Primary changes associated with bone-marrow toxicity were also found in both sexes. Hepatotoxicity, reduced liver weight, and increased vacuolization of adrenal cortical cells were found in female rats treated with doses > 15 mg/kg bw per day. Elevated cholesterol concentrations were observed in female rats from 5 mg/kg bw per day. Changes in serum enzyme activity and cholesterol concentration in males were markedly increased from 15 mg/kg bw per day, while single-cell necrosis and mitosis were seen at 15–75 mg/kg bw per day. Mild lymphocytosis in the thymic cortex was seen in rats treated at 5 mg/kg bw per day and persisted in rats at the highest dose. It was concluded that male rats are marginally more sensitive than female rats to fumonisin B₁ (Table 5; Bondy et al., 1996, 1998).

The number of GST-P⁺ hepatocytes was increased in males and females given 35 and 75 mg/kg bw per day by gavage, respectively, for 11 days. In the same study, the area occupied by GST-P⁺ minifoci was increased significantly in both sexes at 75 mg/kg bw per day. The hepatocyte proliferation rate, measured as the expression of proliferating cell nuclear antigen (PCNA), was significantly enhanced in the livers of rats in which the number of GST-P⁺ cells was increased ((Mehta et al., 1998). Pathological changes occurred at the same doses in males and females, although they appeared to be more severe in males (Bondy et al., 1996, 1998). After intraperitoneal administration of a single dose of 7.5 or 10 mg/kg bw to male rats on four consecutive days, only the higher dose significantly enhanced the induction of GST-P⁺ minifoci. The authors suggested that, like known genotoxic carcinogens, fumonisin B₁ can induce GST-P⁺ hepatocytes and their subsequent development into minifoci in the presence of enhanced hepatocyte proliferation, presumably in response to toxicity.

Maize samples associated with outbreaks of equine leukoencephalomalacia fed to male rats for 28 days caused degenerative changes in the liver and kidney similar to those seen with purified fumonisin B₁ (Table 6; Voss et al., 1989). The toxicity was probably exacerbated by the nutritionally imbalanced diet. On the basis of the estimated feed intake and the estimated fumonisin B content of the two samples (150 and 20 mg/kg of feed), the mean daily intake of fumonisin B₁ and B₂ was equivalent to 13 and 1.7 mg/kg bw, respectively (Plattner et al., 1990).