Genotoxicity of pyrrolizidine alkaloids

Abstract

Pyrrolizidine alkaloids (PAs) are common constituents of many plant species around the world. PA-containing plants are probably the most common poisonous plants affecting livestock and wildlife. They can inflict harm to humans through contaminated food sources, herbal medicines and dietary supplements. Half of the identified PAs are genotoxic and many of them are tumorigenic. The mutagenicity of PAs has been extensively studied in different biological systems. Upon metabolic activation, PAs produce DNA adducts, DNA cross-linking, DNA breaks, sister chromatid exchange, micronuclei, chromosomal aberrations, gene mutations and chromosome mutations in vivo and in vitro. PAs induced mutations in the cII gene of rat liver and in the p53 and K-ras genes of mouse liver tumors. It has been suggested that all PAs produce a set of (±)-6,7-dihydro-7-hydroxy-1-hydroxymethyl-5H-pyrrolizine-derived DNA adducts and similar types of gene mutations. The signature types of mutations are G : C → T : A transversion and tandem base substitutions. Overall, PAs are mutagenic in vivo and in vitro and their mutagenicity appears to be responsible for the carcinogenesis of PAs. Published in 2010 by John Wiley & Sons, Ltd.

INTRODUCTION

Occurrence and Toxicity

Pyrrolizidine alkaloids (PAs) are constitutively formed in many plant species around the world. More than 660 PAs and PA N-oxides have been identified in over 6000 plants mainly contained in the Boraginaceae, Compositae and Leguminosae families (Roeder, 1995, 2000; Stegelmeier et al., 1999). About half of these PAs formed are toxic. Therefore, The PA-containing plants are poisonous to livestock and wildlife and have caused tremendous livestock loss (Arzt and Mount, 1999; de Lanux-Van Gorder, 2000; Fletcher et al., 2009; Fowler, 1968; Knight et al., 1984; Seaman, 1978, 1987; Sharrock, 1969; van der Watt et al., 1972; Wiltjer and Walker, 1974). PAs are also the leading plant toxins associated with disease in humans through contamination of staple foods, honey, milk, herbal teas and herbal medicines (Arseculeratne et al., 1981, 1985; Bach et al., 1989; Bah et al., 1994; Culvenor et al., 1981; Deinzer et al., 1977; Dickinson et al., 1976; Edgar et al., 1992, 2002; Huxtable et al., 1986; Jago, 1969; Kumana et al., 1983: 1985; Mattocks, 1980; Ridker et al., 1985; Roitman, 1981; Steenkamp et al., 2000; White et al., 1984; Zhao et al., 1989).

Chemical Structure

PAs are ester alkaloids composed of a necine (two fused five-membered rings joined by a single nitrogen atom) and a necic acid (one or two carboxylic ester arms) (Fig. 1). Toxic PAs are esters of unsaturated necines having a 1,2 double bond. Structures of several representative carcinogenic PAs are shown in Fig. 2. Riddelliine, retrosine, monocrotaline and symphytine are retronecine-type PAs; lasiocarpine and heliotrine are heliotridine-type PAs; and senkirkine and petasitenine are otonecine-type PAs. In contrast to otonecine-type PAs that contain monocyclic necines, retronecine- and heliotridine-type PAs have bicyclic necine bases and they are enantiomers each other at the C7 position, the retronecine-type PAs possessing an R absolute configuration and the heliotridine-type PAs with an S stereochemistry.

Figure 1.

Schematic structure of pyrrolizidine alkaloid.

Figure 2.

Structures of representative carcinogenic pyrrolizidine alkaloids.

Metabolism

As with many other xenobiotics, PAs require biotransformation to introduce reactive or polar groups so that they can be conjugated to form polar metabolites for excretion out of the body. There are three principal pathways for PA metabolism, including hydrolysis of PAs to release necines and necic acids, N-oxidation to form PA N-oxides and oxidation of PAs to produce dehydropyrrolizidine (pyrrolic ester) derivatives (Fig. 3). The liver is the main organ for PA metabolism, although metabolism in other tissues has also been identified (Lafranconi and Huxtable, 1984).

Figure 3.

Metabolism of pyrrolizidine alkaloids.

When PAs are absorbed into tissues like liver and lung through blood, some of them are cleaved into necines and necic acids by nonspecific esterases. The necines and necic acids are not toxic and the necines may undergo further conjugation to be excreted via the kidneys and urine (Roeder, 1995). This metabolic pathway is an important detoxification route. Rats are very susceptible to toxicity of PAs, partially if not totally, due to lack esterase activity in their livers, whereas guinea pigs possess marked resistance to the toxic effects of PAs because of their particularly high liver esterase activity (Dueker et al., 1992).

Retronecine- and heliotridine-type PAs can become PA N-oxides via N-oxidation of the necine bases while otonecine-type PAs cannot form PA N-oxides because the nitrogen in their necine base is methylated. PA N-oxides are generally regarded as detoxification products because the metabolites can be conjugated for excretion (Williams et al., 1989a, b). However, PA N-oxides can be metabolically converted back to their parent PAs to produce toxic/tumorigenic effects if their parent PAs are toxic/ tumorigenic (Chou et al., 2003b; Mattocks, 1971; Wang et al., 2005c; Yan et al., 2008).

Pyrrolic esters are produced through hydroxylation of the necine base of retronecine- and heliotridine-type PAs at the C3 and C8 position to form 3- or 8-hydroxynecine derivatives followed by spontaneous dehydration. For otonecine-type PA, pyrrolic esters are generated through oxidative N-demethylation of the necine base followed by ring closure and dehydration (Fu et al., 2004). Pyrrolic ester metabolites are very reactive and can bind to one or two molecules of glutathione to form glutathione conjugates for excretion. They also can bind to DNA and proteins to generate DNA adducts, protein adducts and DNA and protein cross-links. Thus, metabolic formation of pyrrolic ester metabolites has been considered as the primary metabolic activation for the genotoxicity and carcinogenicity of PAs. In addition, less reactive longer-lived (±)-6,7-dihydro-7-hydroxy-1-hydroxymethyl-5H-pyrrolizine (DHP) produced by hydrolysis of pyrrolic esters is a considered secondary toxic metabolite due to its antimitotic, mutagenic and carcinogenic effects (Fu et al., 2004; Prakash et al., 1999).

Carcinogenicity and Hepatotoxicity

Many PA-containing plants and individual PA compounds have been tested in animal models and shown to be carcinogenic in different tissues. The liver is the main carcinogenic target. However, tumors induced by PAs also were found in lung, kidney, skin, bladder, brain and spinal cord, pancreatic islets and adrenal gland (Allen et al., 1975; Brandange et al., 1970; Chan et al., 1994; Cook et al., 1950; Furuya et al., 1976; Harris and Chen, 1970; Hirono et al., 1976, 1977, 1978, 1979, 1983; Johnson et al., 1978; Kuhara et al., 1980; Mattocks and Cabral, 1982; Mori et al., 1984; Peterson et al., 1983; Rao et al., 1983; Rao and Reddy, 1978; Schoental, 1975; Schoental and Cavanagh, 1972; Schoental et al., 1954, 1970, 1971; Schoental and Head, 1957; Shumaker et al., 1976; Svoboda and Reddy, 1972, 1974; Williams, 1970). Despite no clear evidence that PAs induce tumors in humans, the frequent occurrence of primary liver tumors in the natives of Central Africa and South Africa has been associated with the consumption of traditional medicinal PA-containing plants (Pavlica and Samuel, 1970; Schoental, 1968; Schoental and Coady, 1968; Williams et al., 1967). It has been reported that PA-containing plants induce hepatic veno-occlusive disease in humans. This disease is regarded as specific for PA intoxication. The clinical symptoms usually occur suddenly and include vomiting, enlargement of the liver and bleeding diarrhea. Children are more sensitive to PA intoxications than adults (Bach et al., 1989; Bras et al., 1954; Kumana et al., 1985; Mohabbat et al., 1976; Ortiz Cansado et al., 1995; Ridker and McDermott, 1989; Ridker et al., 1985; Schoental and Coady, 1968; Sperl et al., 1995; Tandon et al., 1976; Weston et al., 1987; Yeong et al., 1990).

In this article, we review and update information on the genotoxicity and mutagenicity of PAs, including primary DNA damage, chromosome damage and mutations induced by PAs or PA-containing plants. For detailed information concerning the chemical and books and other review articles on PAs and PA-containing plants (Cheeke, 1988; Fu et al., 2004; IARC, 1976; IPCS, 1988; Mattocks, 1986; Prakash et al., 1999; Robins, 1984; Roeder, 1995, 2000; Stegelmeier et al., 1999).

Primary DNA Damage Induced by PAs

DNA adducts

Metabolism of PAs in vitro and in vivo generates DHP-derived DNA adducts (Table 1). Candrian et al. (1985) treated rats of both sexes with tritiated senecipylline and senecionine and determined covalent binding of the alkaloids to DNA using HPLC/radioactivity analysis of hydrolyzed DNA 6 h or 4–5 days after the treatment. They found PA-derived DNA adducts in rat livers, lungs and kidneys. The DNA damage was induced 6 h after treatment and persisted during the following 4 days. However, the DNA adducts were not isolated and characterized.

Table 1.

Primary DNA damage induced by pyrrolizidine alkaloids

Agent	Testing system	Result	Reference
DNA adducts Comfrey root extract/comfrey compound oil	In vivo 32P-postlabeling/HPLC analysis	DHP-derived DNA adducts were formed in rat liver	Chou and Fu (2006)
Clivorine	In vitro 32P-postlabeling/HPLC analysis with metabolic activation	DHP-derived DNA adducts were formed in vitro	Xia et al. (2004)
Coltsfoot root extract	In vivo 32P-postlabeling/HPLC analysis	DHP-derived DNA adducts were formed in rat liver	Chou and Fu (2006)
Flos farfara Extract	In vivo 32P-postlabeling/HPLC analysis	DHP-derived DNA adducts were formed in rat liver	Chou and Fu (2006)
Heliotrine	In vitro 32P-postlabeling/HPLC analysis with metabolic activation	DHP-derived DNA adducts were formed	Xia et al. (2008)
Ligularia hodgsonnii extracts	In vitro 32P-postlabeling/HPLC analysis with metabolic activation	DHP-derived DNA adducts were formed	Xia et al. (2004)
Lasiocarpine	In vitro 32P-postlabeling/HPLC analysis with metabolic activation	DHP-derived DNA adducts were formed	Xia et al. (2006)
Monocrotaline	In vivo or in vitro with metabolic activation 32P-postlabeling/HPLC analysis	DHP-derived DNA adducts were formed in vitro and in rat liver	Wang et al. (2005b)
Retrorsine	In vivo or in vitro with metabolic activation 32P-postlabeling/HPLC analysis	DHP-derived DNA adducts were formed in vitro and in rat liver	Wang et al. (2005a)
Riddelliine	In vivo or in vitro with metabolic activation 32P-postlabeling/HPLC analysis	Two enantiomers of DHP-7deoxyguanosin-2N-yl adducts and DHP-modified dinucleotides were formed in vitro and in rat liver	Chou et al. (2003a–c, 2004); Xia et al. (2003); Yang et al. (2001)
Senecionine	In vivo covalent binding analysis in rat using 3H labeling	Uncharacterized DNA adducts were identified in rat liver, lung and kidney	Candrian et al. (1985)
Seneciphylline	In vivo covalent binding analysis in rat using 3H labeling	Uncharacterized DNA adducts were identified in rat liver, lung and kidney	Candrian et al. (1985)
DNA cross-linking Heliosupine	In vitro cells	DNA–DNA crosslinks	Hincks et al. (1991)
Jacobine	Alkaline elution in rat liver	DNA–DNA crosslinks DNA–protein crosslinks	Petry et al. (1986)
Latifoline	In vitro cells	DNA–DNA crosslinks	Hincks et al. (1991)
Monocrotaline (Dehydromonocrotaline)	In vitro cross-linking assay of     PA-exposed cells or pyrrolic     PA-exposed nucleiIn vivo and in vitro alkaline elution assay	DNA–DNA crosslinks DNA–protein crosslinks	Coulombe et al. (1999); Kim et al. (1995); Pereira et al. (1998); Petry et al. (1984); Petry and Sipes (1987); Rieben and Coulombe (2004); Tepe and Williams (1999); Wagner et al. (1993); Weidner et al. (1990)
Retrorsine (dehydroretronecine)	In vitro cross-linking assay of PA-exposed cells or pyrrolic PA-exposed nuclei	DNA–DNA crosslinks	Hincks et al. (1991); Reed et al. (1988)
Riddelliine (dehydroriddelliine)	In vitro cross-linking assay of     PA-exposed cells or pyrrolic     PA-exposed nuclei	DNA–protein crosslinks DNA–DNA crosslinks	Hincks et al. (1991); Kim et al. (1995, 1999)
Senecionine (dehydrosenecionine)	In vitro cross-linking assay of     PA-exposed cells or pyrrolic     PA-exposed nuclei	DNA–protein crosslinks DNA–DNA crosslinks	Coulombe et al. (1999); Hincks et al. (1991); Kim et al. (1995, 1999)
Seneciphylline (dehydroseneciphylline)	In vitro cross-linking assay of     PA-exposed cells or pyrrolic     PA-exposed nuclei	DNA–protein crosslinks DNA–DNA crosslinks	Hincks et al. (1991); Kim et al. (1995)
DNA strand break Heliosupine	In vitro alkaline elution	Negative	Hincks et al. (1991)
Isatidin	In vitro comet assay	Positive	Uhl et al. (2000)
Jacobine	In vivo alkaline elution	Negative	Petry et al. (1986)
Latifoline	In vitro alkaline elution	Negative	Hincks et al. (1991)
Monocrotaline	In vitro alkaline elution	Negative	Hincks et al. (1991); Petry et al. (1984)
Monocrotaline	In vitro comet assay	Positive	Silva-Neto et al. (2010)
Retrorsine	In vitro alkaline elution	Negative	Hincks et al. (1991)
Riddelliine	In vitro alkaline elution	Negative	Hincks et al. (1991)
senecionine	In vitro alkaline elution	Negative	Hincks et al. (1991)
seneciphylline	In vitro alkaline elution	Negative	Hincks et al. (1991)
Unscheduled DNA synthesis Retrorsine	In rat hepatocytes following in vivo treatment	Positive	Griffin and Segall (1986)
Riddelliine	In rat and mouse hepatocytes following in vivo treatment	Positive	Chan (1993); Chan et al. (1994); Mirsalis (1987); Mirsalis et al. (1993); NTP (2003)
Senecionine	In rat hepatocytes following in vivo treatment	Positive	Griffin and Segall (1986)
Seneciphylline	In rat hepatocytes following in vivo treatment	Positive	Griffin and Segall (1986)

Note: the pyrrolic PAs in brackets were also measured.

To identify the types of DNA adducts induced by PAs, Fu and his research group developed a ³²P-postlabeling/HPLC method and determined the DNA adduct formation in vitro and in vivo using riddelliine as a model PA (Chou et al., 2003a; Xia et al., 2003; Yang et al., 2001). A set of DHP-derived DNA adducts were identified, among which two were enantiomers of DHP-derived 7′-deoxyguanosin-N2-yl adducts and the other six were DHP-modified dinucleotides. Further studies showed that different types of tumorigenic PAs and PA-containing plants induced the same set of DHP-derived DNA adducts in vivo and in vitro. Those PAs and PA-containing plants included clivorine, heliotrine, lasiocarpine, monocrotaline, retrosine, comfrey root extract, comfrey compound oil, coltsfoot root extract, Flos farfara extract and Ligularia hodgsonnii extracts (Chou and Fu, 2006; Wang et al., 2005a, b; Xia et al., 2004, 2006, 2008). Therefore, it is suggested that the formation of DHP-derived DNA adducts is common to all types of tumorigenic PAs.

The levels of the DHP-derived DNA adducts correlated closely with tumorigenic potency in the rats fed different doses of riddelliine (Chou et al., 2003c; Yang et al., 2001). Riddelliine mainly induced liver hemangiosarcomas in rats and mice (Chan et al., 2003). Liver contains two types of cells involved in tumorigenesis, endothelial and parenchymal cells, and hemangiosarcomas develop from endothelial cells. F344 rats and B6C3F1 mice were treated by gavage 5 days a week for 2 weeks with riddelliine at 1.0 mg/kg for rats and 3.0 mg/kg for mice. The treatment resulted in a significantly greater DHP-derived DNA adduct levels in the endothelial cells than those in the parenchymal cells (Chou et al., 2003c). Thus, the overall results suggest that DHP-derived DNA adducts are, at least partially, responsible for the liver tumor development (Fu et al., 2004). In addition, Fu et al. concluded that this set of DHP-derived DNA adducts has the potential to be utilized as a biomarker of PA exposure and tumorigenicity.

DNA cross-linking

PA metabolites contain two functional groups, at the C7 and the C9 positions of the necine base (Fig. 1), that are capable of binding to two sites in DNA or protein to form DNA or protein cross-linking (Table 1). The antimitotic, toxic and carcinogenic actions of PAs are thought to be caused, at least in part, by these crosslinks (Coulombe et al., 1999; Hincks et al., 1991; Kim et al., 1995, 1999).

Studies on the DNA–protein cross-linking activity of several structurally diverse PAs were conducted using cells or isolated nuclei. The pyrrolic PAs, dehydrosenecionine, dehydromonocrotaline, dehydroseneciphylline and dehydroriddelliine readily induced DNA–protein crosslinks, accounting for approximately 50% of the total cellular DNA crosslinks. The DNA–protein cross-linking potency of PAs coincided with their known toxicity potency in animals. Thus, it was concluded that DNA–protein cross-linking was probably involved in PA-related toxicity (Coulombe et al., 1999; Kim et al., 1995; Petry et al., 1984).

PA-induced DNA–DNA cross-linking was investigated in vivo and in vitro using the model cross-linking PA monocrotaline. Hepatic DNA damage induced by monocrotaline was evaluated following i.p. administration to adult male Sprague–Dawley rats. DNA–DNA crosslinks were characterized by the alkaline elution technique; and a mixture of DNA–DNA interstrand cross-links and DNA–protein cross-links was found (Petry et al., 1984). The active metabolite of monocrotaline, dehydromonocrotaline, was also demonstrated to mediate interstrand DNA cross-link formation (Tepe and Williams, 1999). A study on the types of PA-induced DNA–DNA cross-linking showed that the monocrotaline metabolites produced piperidine- and heat-resistant multiple DNA cross-links that were confirmed by electrophoresis and electron microscopy. It was proposed that the metabolites undergo rapid polymerization to a structure capable of crosslinking several fragments of DNA (Pereira et al., 1998). While one study found the DNA–DNA cross-linking by dehydromonocrotaline lacked base sequence preference (Rieben and Coulombe, 2004), another demonstrated that the metabolites preferentially cross-linked DNA duplexes containing the sequence 5’-CG (Weidner et al., 1990).

Coulombe et al. (1999) characterized the ability of different types of PAs to cross-link cellular DNA in cultured bovine kidney epithelial cells. They found that every PA tested induced DNA crosslinks. The relative potency of PAs in causing DNA cross-linking, however, varied, with an order of seneciphylline > riddelliine > retrorsine > senecionine > heliosupine > monocrotaline > latifoline > retronecine. Their studies suggest that PAs with a macrocyclic necic acid ester and an α,β-unsaturated ester function are more potent cross-linkers. In addition, the stereochemical orientation of the ester linkage was found to have no effect on biological activity (Hincks et al., 1991; Kim et al., 1999). Unfortunately, the structures of the DNA crosslink adducts have never been fully characterized; and the levels of their formation have not been correlated with the tumorigenic potencies of rodents treated with PAs. These warrant further investigation.

DNA strand breakage

Several reports have been published on PA-induced DNA strand breakage (Table 1). Hincks et al. (1991) studied DNA crosslinks and DNA strand breaks induced by several bifunctional PAs, including seneciphylline, riddelliine, retrorsine, senecionine, monocrotaline, heliosupine and latifoline, using the alkaline elution assay. None of the PAs induced detectable amounts of DNA single-strand breaks. Petry et al. (1984, 1986) confirmed that monocrotaline and Jacobine did not induced DNA single-strand breaks using the alkaline elution assay. However, isatidine caused a pronounced effect on human hepatoma cells (HepG2 cells) using the single cell gel electrophoresis assay (Comet assay). The lowest concentration that induced a significant positive effect was 500 μM (Uhl et al., 2000). The authors suggested that HepG2 cells catalyzed the activation reaction, leading to DNA strand breaks. Glial cells from the human glioblastoma cell line GL-15 were treated with 1–5000 μM monocrotaline and the DNA strand breaks were measured using the Comet assay (Silva-Neto et al., 2010). The data showed that the treatment caused significant dose–response increases in cell DNA breaks. The conflicting results from studies using the alkaline elution assay and Comet assay have been explained by the greater sensitivity of the Comet assay for detecting DNA strand breaks (Uhl et al., 2000).

Unscheduled DNA synthesis

The unscheduled DNA synthesis (UDS) test measures the DNA repair synthesis after excision and removal of a stretch of DNA containing the region of damage induced by chemical or physical agents. Therefore, this assay measures the repair of primary DNA damage like DNA adducts. Several UDS studies have been conducted to measure primary DNA damage induced by PAs (Table 1). Riddelliine induced significant elevations in UDS in rat liver (Mirsalis, 1987; Mirsalis et al., 1993). UDS was also detected in hepatocytes cultured from male and female rats and mice following 5 or 30 days of riddelliine treatment by gavage (Chan, 1993; Chan et al., 1994; NTP, 2003). Four PAs or PA metabolites, senecionine, retrorsine, seneciphylline and 19-OH-senecionine, along with four alkenals, trans-4-OH-2-hexenal, trans-4-OH-2-nonenal, nonenal and hexenal, were measured in primary cultures of rat hepatocytes using the UDS test. All eight compounds exhibited positive, dose-related responses as measured by autoradiographic detection. The similar UDS responses by PAs and alkenals suggest that trans-4-OH-2-hexenal is a toxic metabolite of the PAs (Griffin and Segall, 1986).

Chromosomal Damage Induced by PAs

Micronucleus assay

Micronucleus induction by PAs has been widely studied and the results clearly demonstrate that PAs are strong clastogenic agents, producing micronuclei in hepatocytes, bone marrow erythrocytes and peripheral blood cells (Table 2). A PA mixture of crude integerrimine and retrorsine and pure integerrimine was extracted from Senecio brasiliensis, which had been stored for more than 23 years under variable conditions of temperature and humidity and exposed to light. Both the mixture and pure PAs induced significant increases in micronucleus frequency in polychromatic erythrocytes (PCEs) of mouse bone marrow (Santos-Mello et al., 2002). Monocrotaline and Crotalaria seeds containing about 6.84% (dry weight) of monocrotaline were administered to mice in their diet for 6 days; the treatment increased the frequency of micronucleated cells in peripheral blood (MacGregor et al., 1990). Clastogenic damage was evaluated in mice following adult and transplacental exposure to heliotrine and monocrotaline using the micronucleus assays of PCE in mouse adult bone marrow and fetal liver, respectively. Both chemicals significantly increased frequencies of micronucleated PCE in the adult and fetal tissues, but heliotrine resulted in the larger increases in mean value of micronucleated PCE. Also, the induction of micronuclei was significantly higher in fetal than in adult cells. The induction of micronuclei following heliotrine treatment showed a peak expression in PCE at 18 h after injection for adult bone marrow and at 24 h for fetal liver (Sanderson and Clark, 1993). Isatidine, retrorsine and monocrotaline were found to induce micronuclei in cultured rat hepatocytes (Muller-Tegethoff et al., 1995, 1997).

Table 2.

Chromosome damage induced by pyrrolizidine alkaloids

Agent	Testing system	Results	References
Micronucleus assay Crotalaria seeds	In vivo peripheral blood of mice	Positive	MacGregor et al. (1990)
Crude mixture of integerrimine, retrorsine and impurities	In vivo polychromatic erythrocytes of mouse bone marrow	Positive	Santos-Mello et al. (2002)
Heliotrine	In vivo polychromatic erythrocytes of mouse adult bone marrow and fetal liver	Positive	Sanderson and Clark (1993)
Integerrimine	In vivo polychromatic erythrocytes of mouse bone marrow	Positive	Santos-Mello et al. (2002)
Isatidine	in vitro rat hepatocytes	Positive	Muller-Tegethoff et al. (1997)
Monocrotaline	In vivo peripheral blood of mice	Positive	MacGregor et al. (1990)
Monocrotaline	In vivo polychromatic erythrocytes of mouse adult bone marrow and fetal liver	Positive	Sanderson and Clark (1993)
Monocrotaline	In vitro rat hepatocyte micronucleus	Positive	Muller-Tegethoff et al. (1995, 1997)
Riddelliine	In vivo peripheral blood and bone marrow of male mice	Week positive	Chan et al. (1994)
Riddelliine	In vivo peripheral blood of rats and mice	Negative	Chan et al. (1994); Mirsalis et al. (1993)
Retrorsine	In vitro rat hepatocyte micronucleus	Positive	Muller-Tegethoff et al. (1995, 1997)
Chromosome aberration Crotalaria retusa	Mouse bone marrow	Positive	Ribeiro et al. (1993)
Fulvine	Human blood cells of children suff ering from veno-occlusive disease	Positive	Martin et al. (1972)
Isatidine	In vitro V79 cells with or without S9	Negative	Muller et al. (1992)
Isatidine	In vitro V79 cells with primary hepatocytes	Positive	Muller et al. (1992)
Heliotrine	In vitro V79 cells with or without S9	Positive	Takanashi et al. (1980)
Heliotropium curassavicum	In vitro CHO cells with or without S9	Positive	Carballo et al. (1992)
Integerrimine	Mouse bone marrow	Positive	Gimmler-Luz et al. (1990)
Lasiocarpine	In vitro V79 cells with or without S9	Positive	Takanashi et al. (1980)
Monocrotaline	In vitro V79 cells with S9 or with primary hepatocyte activation	Positive	Muller et al. (1992)
Petasitenine	In vitro V79 cells with or without S9	Positive	Takanashi et al. (1980)
Riddelliine	In vitro CHO cells with S9.	Positive	Chan (1993); NTP (2003)
Retrorsine	In vitro V79 cells with or without S9; or with primary hepatocytes	Positive	Muller et al. (1992)
Senkirkine	In vitro V79 cells with or without S9	Positive	Takanashi et al. (1980)
Sister chromatid exchange dehydroretronecine	In vitro human lymphocytes	Positive	Ord et al. (1985)
Heliotrine	In vitro V79 cells with primary chick embryo hepatocyte activation	Positive	Bruggeman and van der Hoeven (1985)
Monocrotaline	In vitro V79 cells with primary chick embryo hepatocyte activation	Positive	Bruggeman and van der Hoeven (1985)
Riddelliine	In vitro CHO cells with and without S9	Positive	Chan (1993); NTP (2003)
Seneciphylline	In vitro V79 cells with primary chick embryo hepatocyte activation	Positive	Bruggeman and van der Hoeven (1985)
Senkirkine	In vitro V79 cells with primary chick embryo hepatocyte activation	Positive	Bruggeman and van der Hoeven (1985)

The frequency of micronucleated erythrocytes in mouse and rat peripheral blood samples was not elevated after 4 or 13 weeks of daily gavage treatments of riddelliine at doses up to 10 mg/kg for rats and 25 mg/kg for mice. A weakly positive response was noted in the peripheral blood and bone marrow of male mice administered a single 150 mg/kg riddelliine by gavage (Chan, 1993; Mirsalis et al., 1993).

Chromosomal aberration

Many PAs induce chromosomal aberrations in mammalian cells or in mouse bone marrow when they are appropriately metabolically activated (Table 2). Chromosomal aberrations were induced by riddelliine in Chinese hamster ovary (CHO) cells only in the presence of S9 (Chan, 1993; NTP, 2003). Heliotrine, lasiocarpine, petasitenine and senkirkine induced chromosomal aberrations in V79 cells (Takanashi et al., 1980). While heliotrine and petasitenine induced interchromosomal exchanges, lasiocarpine and senkirkine caused chromatid gaps. Male and female C57Bl/6 mice were treated with integerrimine and their bone marrow cells were collected for measurement of chromosomal aberrations. The chromosomal aberrations were significantly induced in a dose responsive manner. The greatest frequency of chromosomal aberrations was detected 12 h after treatment (Gimmler-Luz et al., 1990).

Muller et al. (1992) investigated chromosomal aberration induction by monocrotaline, retrorsine and isatidine in different activation systems. Two hours of PA treatment of V79 cells with S9 mix led to a strong and concentration-dependent increase for retrorsine, but a negative response for isatidine (retrorsine N-oxide) and a weak positive for monocrotaline. In contrast, an 18 h PA treatment of V79 cells in the presence of primary hepatocytes resulted in clear concentration-dependent positive responses for all three PAs. The authors postulated that primary liver cells could reduce isatidine to retrorsine whereas the S9 mix could not.

Two PA-containing plants were evaluated in the chromosomal aberration test. Ribeiro et al. (1993) found that Crotalaria retusa extracts containing monocrotaline induced chromosomal aberrations in mouse bone marrow cells. Heliotropium curassavicum is a widely employed medicinal plant that contains PAs. In order to analyze its genotoxic effects, Carballo et al. (1992) studied chromosomal aberrations induced by Heliotropium extracts in CHO cells with or without S9 mix. They found that the plant extract induced chromosomal aberrations and the induction was enhanced by the addition of an S9 fraction. The authors suggested that the toxic effects were associated with PAs and their N-oxides.

Genotoxicity studies on PAs in humans are rare. There is one report on chromosomal aberrations in the blood cells of children suffering from veno-occlusive disease, believed to have been caused by fulvine, a cyclic diester of retroneoine (Martin et al., 1972).

Sister chromatid exchanges

Sister chromatid exchange (SCE) is the exchange of genetic material between two identical sister chromatids. SCE has been associated with chromosome damage and tumor induction. Several PAs have been tested and found to be SCE inducers (Table 2). Riddelliine induced SCEs in CHO cells with and without S9 (Chan, 1993; NTP, 2003). Heliotrine, monocrotaline, seneciphylline and senkirkine were studied with the SCE assay in V79 cells. Treatment of the cells with the PAs in the presence of co-cultured primary chick embryo hepatocytes resulted in a strong induction of SCEs. The rank order found was seneciphylline > senkirkine > heliotrine > monocrotaline (Bruggeman and van der Hoeven, 1985). Dehydroretronecine, the common reactive pyrrolic metabolite of retronecine-type PAs, was measured in human lymphocytes for induction of SCE and produced a strong positive result (Ord et al., 1985).

Mutations Induced by PAs

Mutations in bacteria

The Salmonella typhimurium/mammalian microsome test has been extensively used for determining the mutagenicity of PAs (Table 3). Retrorsine was measured using TA98, TA100, TA1535 and TA1537 tester strains in the presence of S9 and the PA was mutagenic for TA1535 and TA1537, indicating that it induced both basepair substitution and frameshift mutations (Wehner et al., 1979). Dehydroretronecine also induced mutations in the S. typhimurium base substitution strain TA92 (Ord et al., 1985). The mutagenicities of clivorine, fukinotoxin, heliotrine, lasiocarpine, ligularidine and senkirkine to S. typhimurium TA100 were demonstrated by pre-incubation of the PAs with S9 mix and bacteria in a liquid medium (Yamanaka et al., 1979). Riddelliine was mutagenic in TA100 with, but not without, S9 activation. Riddelliine was negative in strains TA97, TA98 and TA1535 (Chan, 1993). An acetone extract of tansy ragwort produced positive mutagenic responses in tester strains TA98, TA100, TA1535 and TA1537, with S9 activation (White et al., 1984).

Table 3.

Mutagenicity of pyrrolizidine alkaloids

Agent	Testing system	Results	References
In bacteria Clivorine	Salmonella typhimurium TA100 with S9	Positive	Yamanaka et al. (1979)
Dyhydroretrosine	S. typhimurium TA100 and T97	Positive	Ord et al. (1985)
Fukinotoxin	S. typhimurium TA100 with S9	Positive	Yamanaka et al. (1979)
Heliotrine	S. typhimurium TA100 with S9	Positive	Yamanaka et al. (1979)
Heliotrine	Escherichia coli WP2 and WP2 uvrA	Negative	Green and Muriel (1975)
Isatidine	S. typhimurium TA100 with S9	Negative	Rubiolo et al. (1992)
Lasiocarpine	S. typhimurium TA100 with S9	Positive	Yamanaka et al. (1979)
Ligularidine	S. typhimurium TA100 with S9	Positive	Yamanaka et al. (1979)
LX201	S. typhimurium TA100 with S9	Positive	Yamanaka et al. (1979)
Monocrotaline	S. typhimurium TA100 with S9	Negative	Rubiolo et al. (1992)
Monocrotaline	E. coli WP2 and WP2 uvrA	Negative	Green and Muriel (1975)
Retrorsine	S. typhimurium TA1535 and TA1537 with S9	Positive	Wehner et al. (1979)
Retrorsine	S. typhimurium TA100 with S9	Weakly positive	Rubiolo et al. (1992)
Riddelliine	S. typhimurium TA97, TA98 or TA1535 with or without S9	Negative	Chan (1993)
Riddelliine	S. typhimurium TA100 with S9	Positive	Chan (1993)
Senecivernine	S. typhimurium TA100 with S9	Weakly positive	Rubiolo et al. (1992)
Senecionine	S. typhimurium TA100 with S9	Negative	Rubiolo et al. (1992)
Seneciphylline	S. typhimurium TA100 with S9	Weakly positive	Rubiolo et al. (1992)
Senecio inaequidens and	S. typhimurium TA100 with S9	Weakly positive	Rubiolo et al. (1992)
S. fuchsia	S. typhimurium TA100 with S9	Weakly positive	Rubiolo et al. (1992)
S. cacliastershowed	S. typhimurium TA100 with S9	Weakly positive	Rubiolo et al. (1992)
Senkirkine	S. typhimurium TA100 with S9	Positive	Yamanaka et al. (1979)
Tansy ragwort	S. typhimurium TA1535, TA1537, TA98 and TA100 with S9	Positive	White et al. (1984)
In Drosophila 7-Acetyl-intermedine	Wing spot test	Positive	Frei et al. (1992)
7-Acetyl-lycopsamine	Wing spot test	Positive	Frei et al. (1992)
Heliotrine	Wing spot test	Positive	Brink (1969); Frei et al. (1992); Sivlingham and Brink (1988).
Heliotrine	Sex-linked recessive lethal test	Positive	Clark (1959)
Indicine	Wing spot test	Positive	Frei et al. (1992)
indicine-N-oxide	Wing spot test	Positive	Frei et al. (1992)
Integerrimine	Wing spot test	Positive	Campesato et al. (1997)
Intermedine	Wing spot test	Positive	Frei et al. (1992)
Jacoline	Wing spot test	Positive	Frei et al. (1992)
Lycopsamine	Wing spot test	Positive	Frei et al. (1992)
Monocrotaline	Wing spot test	Positive	Frei et al. (1992)
Monocrotaline	Sex-linked recessive lethal test	Positive	Clark (1959)
Retrorsine	Wing spot test	Positive	Frei et al. (1992)
Senecionine	Wing spot test	Positive	Frei et al. (1992)
Seneciphylline	Wing spot test	Positive	Frei et al. (1992)
Seneciphylline	Sex-linked recessive lethal test	Positive	Candrian et al. (1984)
Senkirkine	Wing spot test	Positive	Frei et al. (1992)
Senkirkine	Sex-linked recessive lethal test	Positive	Candrian et al. (1984)
Supinine	Wing spot test	Negative	Frei et al. (1992)
Symlandine	Wing spot test	Positive	Frei et al. (1992)
Symphytine	Wing spot test	Positive	Frei et al. (1992)
In rodent Comfrey	Transgenic rat cII assay	Positive in liver and lung; The major types of induced-mutations are G : C → T : A and tandem base substitutions	Mei et al. (2005); Mei and Chen (2007)
Riddelliine	Transgenic rat cII assay	Positive in liver and showed the liver endothelial specificity; The major types of induced-mutations are G : C → T : A and tandem base substitutions	Mei et al. (2004a, b)
Riddelliine	Mutations in the K-ras gene in mouse liver tumor induced by riddelliine	K-ras codon 12 G : C → T : A transversion; 58% incidence vs 0% in spontaneous tumors	Hong et al. (2003)
	Mutations in the p53 gene in mouse liver tumor induced by riddelliine	75% incidence vs 0% in spontaneous tumors	Hong et al. (2003)

In spite of the results described above, there were contradictory reports on the system for detection of mutagenicity of PAs. Clark (1976) reported that PAs did not produce mutagenicity in the S. typhimurium test system, even in the presence of a liver microsome preparation. Rubiolo et al. (1992) found that several PAs gave negative results in the test system. Retrorsine, senecivernine, seneciphylline and extracts from PA-containing plants Senecio inaequidens, S. fuchsia and S. cacliastershowed only showed weak mutagenic activity. The authors concluded that the S. typhimurium/mammalian microsome system was not a sensitive assay for detection of Pas’ mutagenicity unless a suitable activation enzyme system was applied.

Mutations in Drosophila

Drosophila test systems are particularly suitable for testing PAs due to the fly’s versatile metabolic bioactivation system and its excellent sensitivity to cross-linking genotoxins (Candrian et al., 1984). Heliotrine was a powerful mutagen in D. melanogaster (Brink, 1969; Clark, 1959; Sivlingham and Brink, 1988). Seneciphylline and senkirkine were found to produce sex-linked recessive lethal in males of D. melanogaster using the 3-day feeding method (Candrian et al., 1984). In addition, monocrotaline was positive in Drosophila sex-linked recessive lethal assay (Clark, 1976). Drosophila flies fed with milk from lactating rats given an oral dose of 25 mg seneciphylline/kg showed 1.2% sex-linked recessive lethal compared with 0.3% in controls (Candrian et al., 1984).

Integerrimine was positive in the somatic mutation and recombination wing spot test (SMART) in D. melanogaster. Analysis of the dose–response data indicated that 85–90% of the genotoxic events were due to mitotic recombination activity while 10–15% of them resulted from somatic mutations (Campesato et al., 1997).

Frei et al. (1992) examined 15 PAs and one PA N-oxide (indicine N-oxide) for their genotoxic potency with respect to the structure/activity relationship in the wing spot test of D. melanogaster following oral administration (Table 3). All PAs tested except the C9-monoester supinine were clearly genotoxic. The results demonstrated that the macrocyclic diester-type PAs were the most genotoxic; the open diesters PAs were intermediate; and 7-hydroxy C9-monoester types of PAs were the least active. Stereoisomeric PAs mostly showed similar activity. An increasing number of hydroxy groups in the PA molecule seemed to reduce genotoxic potency. A rank order with decreasing genotoxic potency (senkirkine as 100%) follows: senkirkine (100.0), monocrotaline (90.0), seneciphylline (54.5), senecionine (39.1), 7-acetyl-intermedine (22.5), heliotrine (13.4), retrorsine (8.3), 7-acetyl-lycopsamine (7.9), symphytine (3.8), Jacoline (1.8), symlandine (1.7), intermedine (0.49), indicine (0.27), lycopsamine (0.19), indicine-N-oxide (0.07) and supinine (0.002).

Mutations in rodents

The mutagenicity of riddelliine in rat liver was investigated using Big Blue transgenic rats (Mei et al., 2004a, b). Transgenic mutation assays provide a unique opportunity for studying the induction of tissue-specific mutation, and the assays also permit quantitative measurements of mutant frequencies (MFs) in the PA-target tissues and molecular analysis of the PA-specific mutational spectra. Groups of six female transgenic Big Blue rats were gavaged with 0.1, 0.3 and 1.0 mg riddelliine per kg body weight 5 days a week for 12 weeks and sacrificed 1 day after the last treatment. The middle and high doses resulted in liver tumors in an NTP bioassay (NTP, 2003). A significant dose-dependent increase in MF was found (Fig. 4).

Figure 4.

Mutant frequencies induced by riddelliine in liver of female transgenic Big Blue rats. The upper panel shows mutant frequencies induced by riddelliine treatment in the cII genes in rat liver as a function of dose. The lower panel compares mutant frequencies with other biological consequences after riddelliine treatments in rat liver hepatocytes (solid bars) and endothelial cells (open bars). Data are from the literature (Chou et al., 2003c; Hirono et al., 1978; Mei et al., 2004a).

Riddelliine mainly induces liver hemangiosarcomas that develop from endothelial cells in rat or mouse liver (Chan et al., 2003). To identify the cell type-specific response to mutagenicity of PA, we separated the endothelial cells from the parenchymal cells in the livers of Big Blue rats treated with riddelliine (Mei et al., 2004a). While there was no difference in the cII MFs of liver parenchymal cells in control and riddelliine-treated rats, the cII MF of liver endothelial cells from treated rats was significantly greater than the cII MF of endothelial cells from control rats. The induction of mutation also correlated with induction of DNA adducts and tumors by riddelliine in rat liver (Fig. 4). These results suggest that the relatively high mutagenicity of riddelliine in rat liver endothelial cells may be partially responsible for the tumorigenic specificity of this agent.

A statistically significant difference was found between the mutational spectra from the riddelliine-treated and the control rats. The major types of mutations induced by riddelliine were G : C → T : A transversion and tandem base substitutions of GG → TT and GG → AT (Table 4). The types of mutations induced by riddelliine are consistent with riddelliine adducts involving G : C base pairs (Chou et al., 2003a). Riddelliine reacts with guanine, adenine and thymine, but not with cytosine, and the relative reactivity of guanine with DHP is greater than other bases. The GG → TT and GG → AT tandem base substitutions were believed to result from intra-strand crosslinks in adjacent guanine bases forming DHP-modified dimers (Chou et al., 2003a; Mei et al., 2004b). This unique spectrum may serve as a signature for genetic damage produced by riddelliine and other PAs. The types of mutations induced by riddelliine suggest that both mononucleotide and dinucleotide DNA adducts involving G : C base pairs are responsible for its mutagenicity (Mei et al., 2004a, b).

Table 4.

Summary of the types of mutations in cII gene from riddelliine- or comfrey-treated and control Big Blue rats

Type of mutation	Control (%)		Treatment (%)
	In liver	In lung	Riddelliine in liver	Comfrey in liver	Comfrey in lung
G : C → C : G	11	9	5	6	8
G : C → A : T	43	63	26	12	28
G : C → T : A	20	9	35	42	29
A : T → T : A	2	0	5	2	16
A : T → C : G	7	2	6	3	4
A : T → G : C	2	2	5	4	6
Frameshift	15	13	10	13	5
Complex mutation	0	2	0	1	0
Tandem–base substitution	0	0	8	17	4

Note: the data are from the literature (Mei et al., 2004b, 2005, 2007)

PA-containing comfrey is a rat liver toxin and carcinogen (Hirono et al., 1978). In order to evaluate the mechanisms underlying its carcinogenicity, we examined the mutagenicity of comfrey in the transgenic Big Blue rat model. Groups of six 6-week-old male Big Blue rats were fed either a basal diet or the comfrey diet for 12 weeks. MFs were determined for the liver and lung cII gene of the rats treated with comfrey. The MFs in both liver and lung were increased by the comfrey treatment with a much higher MF in liver compared to that in lung (Mei et al., 2005; Mei and Chen, 2007). These results correlated with the previous report that tumors were induced by comfrey in liver and liver was the major target tissue (Hirono et al., 1978). Sequencing analysis of the comfrey-induced cII mutant DNA showed the PA mutational signature, with a high induction of G : C → T : A transversions and tandem base substitutions (Table 4). Therefore, these mutational data support the hypothesis that the mutations induced by comfrey in rat liver and lung were due to the PAs in comfrey (Mei et al., 2005; Mei and Chen, 2007).

Hong et al. (2003) examined the mutations occurring in the K-ras protooncongene and p53 tumor suppressor gene in riddelliine-induced liver hemangiosarcomas. They found that 58% riddelliine-induced tumors contained K-ras codon 12 G → T mutation, a PA signature mutation, and 75% riddelliine-induced tumors contained p53 mutations, the types of which were not identified. In contrast, spontaneous hemangiosarcomas from control mice lacked mutations in both the K-ras and p53 genes. It was concluded that p53 and K-ras mutations in riddelliine-induced hemangiosarcomas most likely occurred as a result of the mutagenic effects of riddelliine (Hong et al., 2003).

SUMMARY

There have been a number of outbreaks of human poisoning as a result of ingestion of contaminated grain as well as case reports of poisoning caused by intentional ingestion of herbal medicines containing PAs. In recent years, there has been an increasing use of herbal medicines and dietary supplements to treat various chronic diseases and to promote health. Thus, humans may be exposed to PAs via these herbs or dietary supplements, such as Sympthytum spp. (comfrey), which have been deliberately ingested. Although there are no epidemiological data regarding the carcinogenicity of PAs in humans, a number of studies demonstrated that various PAs are carcinogens in experimental animals. Several PAs were classified as possibly carcinogenic to humans (group 2B) (IARC, 1976). Comparison of the total intake resulting in human toxicity with the total doses to death observed in the chronic toxicity studies in rats indicates that human beings are more susceptible and suggests that humans may survive for sufficient time to develop cancer after only a brief exposure at toxic levels or a longer exposure at a markedly lower levels (IPCS, 1988). Thus, a potential cancer risk for human beings should be seriously considered. To assess the risk to humans, it is necessary to determine whether tumorigenic PAs are mutagenic carcinogens because the selection of an appropriate model for conducting quantitative cancer risk assessment is based upon an understanding of the chemical’s mode-of-action.

It seems that the carcinogenic activity of individual alkaloids parallels their mutagenic behavior, but not their relative hepatotoxicities (IPCS, 1988). Our review of the available information demonstrated that the tumorigenic PAs are mutagenic. The mutagenic activity is mediated by the formation of PA metabolites binding to DNA, thus resulting in DNA damage, gene mutations and chromosomal mutations. Since the mutagenicities have been found in the tumorigenic target tissues and in oncogenes of PA-induced tumors, it is reasonable to conclude that the PAs induce tumors via a mutagenic mode of action. A postulated mechanism of pyrrolizidine alkaloid carcinogenesis is shown in Fig. 5.

Figure 5.

Postulated mechanism of pyrrolizidine alkaloid carcinogenesis.

PA metabolic activation plays a major role in their mutagenicity. This is evidenced by the fact that the main genotoxic and tumorigenic target organ is the liver where the most PAs’ metabolic activation occurs. The contradictory reports between in vitro and in vivo mutagenicities of PAs also suggest the importance of PAs’ metabolism for their mutagenicities. PAs were not mutagenic or showed weak mutagenic activity in S. typhimurium, especially when suitable activation enzymes were not employed. Surprisingly, no reports were found on mutagenicity of PAs using in vitro mammalian cell assays, perhaps due to insensitivities of these in vitro systems to PAs’ mutagenic insults. In contrast to the in vitro systems, the in vivo test systems like Drosophila and transgenic rodents are particularly suitable for testing PAs due to the in vivo system’s versatile metabolic bioactivation systems. The mutagenic activities of some PAs that were identified as negative or weakly positive in bacteria were shown to be strong positives in Drosophila or rodent test systems.

A large body of evidence shows that PAs induced DNA adducts and DNA crosslinks while the results for PA-induced DNA breaks were less clear. Pyrrolic esters or DHP can bind to DNA and generate DHP-derived DNA adducts. Two of the DNA adducts were identified as enantiomers of DHP-derived 7′-deoxyguanosin-N2-yl adducts and the others were characterized as DHP-modified dinucleotides. The PA-induced DNA adducts could be formed in common by all PAs. The observation that common types of mutations, G : C → T : A and tandem base pair substitutions, were induced by different PAs strengthen this theory (Table 4). PAs have proved to be strong cross-linking agents. The bifunctional PAs are capable of forming DNA–DNA or DNA–protein crosslinks. DNA–protein crosslinks comprise about half of the total cellular DNA crosslinks. It also has been suggested that PA metabolites can undergo polymerization to form a structure capable of producing multiple DNA crosslinks. Different PAs have different potencies in causing DNA cross-linking. PAs with a macrocyclic necic acid ester and anα,β-unsaturated ester function demonstrated the most potent cross-linking activity. While DNA adducts may induce more gene mutations, DNA crosslinks tend to induce chromosome mutations.

PAs are both gene mutagens and chromosomal mutagens, perhaps more potent as chromosomal mutagens than as gene mutagens. Many PAs were positive in micronucleus, chromosomal aberration and sister chromatid exchange assays which detect chromosomal damage, and indicate the likely induction of chromosomal mutations. Although many of the micronucleus tests were conducted in surrogate tissues, bone marrow and peripheral blood cells, PAs mainly were positive most of the time, albeit weak. Chromosomal aberrations have been found in the blood cells of children suffering from veno-occlusive disease, believed to be caused by the PA fulvine. PAs were positive in many different S. typhimurium strains, although the assay was relatively insensitive to PAs’ mutagenicity. PAs induced mutations in Drosophila, primarily measured with the wing spot or sex-linked recessive lethal assays. The systems appeared to be particularly suitable for testing PA mutagenicity. The mutagenic potency for a number of PAs has been ranked in these systems. PAs induced mutations in the transgenic cII gene in rat liver and lung. Riddelliine induced higher MFs in the endothelial cells than the parenchymal cells in rat livers, which correlated with induction of hemangiosarcomas developed from rat endothelial cells. Signature mutations of PAs have been identified as G : C → T : A transversion and tandem base substitutions. This signature marker may be used as a fingerprint of PA exposure. Mutations have been detected in the p53 tumor suppressor gene and K-ras oncogene of liver tumors induced by riddelliine in mice. The signature type of mutation, G : C → T : A, was found in the oncogene.

The consistency between the induction of mutations and tumors in liver suggests that gene and chromosomal mutations are major factors in the induction of tumors by PAs. Gene and chromosomal mutations in oncogenes and tumor suppressor genes can initiate tumorigenesis. The initiated cells can be promoted and progressed into tumors under the effects of mutations induced by PAs in different genes. Gene expression studies have confirmed that exposure of PAs significantly changes genes that are involved in cancer, cell death, tissue development, cellular movement, tissue morphology, cell-to-cell signaling and interaction, and cellular growth and proliferation. (Guo et al., 2007; Mei et al., 2006, 2007). Although there is little evidence for mutagenicity of PAs in humans, mutagenicity data from studies in vitro, in vivo and in oncogenes provide strong evidence that PAs are mutagenic carcinogens.