Aristolochic acid as a probable human cancer hazard in herbal remedies
By academic.oup.com
Aristolochic acid (AA), the plant extract of Aristolochia spp. (e.g. Aristolochia clematitis, Aristolochia fangchi and Aristolochia manshuriensis), is a mixture of structurally related nitrophenanthrene carboxylic acids, mainly aristolochic acid I (AAI) and aristolochic acid II (AAII) (Figure 1) (Pailer et al., 1955). AA is found primarily in the genus Aristolochia, but may be present in other botanicals. Herbal drugs derived from Aristolochia spp. have been known since antiquity and were used in obstetrics and in the treatment of snake bites (Rosenmund and Reichstein, 1943). Contemporary medicine has used Aristolochia plant extracts for the therapy of arthritis, gout, rheumatism and festering wounds (Rücker and Chung, 1975; Hahn, 1979; Priestap, 1987). The anti-inflammatory properties of AA encouraged the development of pharmaceutical preparations in Germany (Möse, 1966; Möse and Porta, 1974; Kluthe et al., 1982) until Mengs and co-workers observed that AA is a strong carcinogen in rats (Mengs et al., 1982; Mengs, 1983). Subsequently, AA was shown to be a genotoxic mutagen in several short-term tests (Table I). Therefore, all pharmaceutical preparations containing AA have been withdrawn from the market in Germany and in many other countries. However, Aristolochia plants and their extracts have been further used in traditional medicine in some parts of the world (Priestap, 1987; Vishwanath and Gowda, 1987; Houghton and Ogutveren, 1991). Recently the FDA advised consumers to immediately discontinue use of any botanical products containing AA and has published a list of botanical products that have been shown to contain AA (Schwetz, 2001).
So-called Chinese herbs nephropathy (CHN), a unique type of rapidly progressive renal fibrosis associated with the prolonged intake of Chinese herbs during a slimming regimen, was observed for the first time in Belgium in 1991 (Vanherweghem et al., 1993). About 100 CHN cases have been identified so far in Belgium (Table II), half of which needed renal replacement therapy, mostly including renal transplantation (Vanherweghem, 1998). The observed nephrotoxicity has been traced to the ingestion of A.fangchi containing AA inadvertently included in slimming pills (Vanhaelen et al., 1994). So-called CHN has been described in patients in other European and in Asian countries and in the USA (about 170 cases) (Table II), who were exposed to Aristolochia spp. containing AA and had no relationship with the Belgian slimming clinic. Therefore, it has been proposed to designate the interstitial nephropathy in which the unequivocal role of AA has been fully documented as aristolochic acid nephropathy (AAN) (Gillerot et al., 2001; Solez et al., 2001). Recently, a high prevalence of urothelial cancer was found in a large cohort of AAN patients in Belgium (Cosyns et al., 1999; Nortier et al., 2000) and a case with urothelial cancer has also been described in the UK (Lord et al., 2001). This highlights the carcinogenic potential of AA in human beings.
Since the demonstration that AA forms covalent DNA adducts in rodents (Schmeiser et al., 1988; Pfau et al., 1990a; Stiborova et al., 1994) as well as in AAN patients (Table II), AA–DNA adducts have been used as a biomarker of exposure and to investigate the mutagenic and carcinogenic potential of AA. The intention of this article is to summarize data on the genotoxic mechanism for AA carcinogenicity in rodents and to speculate on the mechanism of AA nephrotoxicity and carcinogenicity in humans.
Genotoxic and carcinogenic mechanism of AA in rodents
Carcinogenic and nephrotoxic effects of AA in rodents
The natural mixture AA is a strong carcinogen in rats (Mengs et al., 1982; Mengs, 1983). In Wistar rats treated orally with 0.1, 1.0 or 10 mg AA/kg body wt/day for 3 months multiple tumours were found after a short induction time (3 months). AA showed mainly a high incidence of tumours in the forestomach at the two high doses, but primary tumours were also found in the renal cortex, renal pelvis and urinary bladder. In a few cases formation of metastases was observed in the regional lymph nodes. At the lowest dose tumours in the forestomach occurred only 12 months after treatment and no urogenital tract tumours were found. However, the observed hyperplasia in the renal pelvis suggests that neoplastic growth might have ensued if the period of observation had been prolonged. Although no carcinogenic activity of AA was initially reported in the liver, a single non-necrogenic dose of AA (10 mg/kg body wt, i.p. injection) given 18 h after two-thirds partial hepatectomy initiated liver cell carcinogenesis (formation of hepatic foci and nodules) (Rossiello et al., 1993). AA is also a potent carcinogen in mice (Mengs, 1988). Oral treatment with 5 mg AA/kg body wt/day for 3 weeks resulted in subsequent tumour formation in the forestomach, lungs, uterus and lymphoid organs. Apart from these carcinogenic effects, acute and subchronic studies in rats and mice showed acute tubular necrosis and renal failure after oral administration of AA (Mengs, 1987; Mengs and Stotzem, 1993). Chronic interstitial fibrosis was observed in rats after i.p. injection of AA (Zheng et al., 2001; Debelle et al., 2002).
Metabolism of AA
The metabolism of AA has been studied in different species including man and has shown that the products of nitroreduction, the corresponding aristolactams (Mix et al., 1982), are the major metabolites found in urine and faeces (Figure 2) (Krumbiegel et al., 1987). The principal metabolite of AAI was aristolactam Ia, produced by two metabolic pathways, one via aristolactam I and the other via AAIa (Figure 2). This interpretation is supported by the results of Schmeiser et al.(1986), which showed that aristolactam I and aristolactam II are also produced in vitro by anaerobic incubation of AAI and AAII with rat liver S9 mix. Under aerobic incubation conditions the major metabolite formed by AAI is AAIa, while AAII remains unaltered. Thus, aristolactam Ia, the major metabolite found in vivo, has not been detected in vitro. The oxygen concentration of tissues in vivo may affect the relative extents of nitroreduction and O-dealkylation for AAI, whereas for AAII only nitroreduction might be influenced by oxygen concentration (Maier et al., 1987). The phase II metabolism of both AAs has not been extensively studied so far, however, large amounts of AA metabolites in the urine and faeces of rodents were present in conjugated form and suggested to be either glucuronides or sulfate esters (Krumbiegel et al., 1987).
Enzymatic activation of AA and DNA adduct formation
Aristolactams represent the final state of reduction of the nitro group of both AAs, but not the DNA-binding species. Aristolactams are not mutagenic themselves and require metabolic activation by an exogenous metabolic system (Table I). Whereas AAI and AAII are direct mutagens in Salmonella strains TA100 and TA1537 (Table I), the mutagenic potency of the corresponding aristolactams in TA100 activated by rat liver S9 mix is about half of that of the parent compounds (Schmeiser et al., 1986). In contrast, both AAs were only weakly mutagenic in strain TA100NR lacking the classical bacterial nitroreductase, indicating that nitroreduction is a crucial step in the pathway of metabolic activation of AA to their ultimate mutagenic species (Schmeiser et al., 1984). Using genetically engineered YG strains, Götzl and Schimmer (1993) confirmed that only the nitro group is important for the mutagenic activity of AA in Salmonella. Nevertheless, both AAs are only weak mutagens in the Ames assay (<1 revertant/nmol) when compared with other nitroaromatic compounds (Purohit and Basu, 2000).
A powerful tool for elucidating the pathway of activation of carcinogens is to characterize and quantify the DNA adducts it forms and to determine what factors either enhance or inhibit adduct formation. The most commonly used method to detect DNA adducts is the highly sensitive 32P-post-labelling assay and detection of DNA adduct formation by AA in vitro and in vivo has been by this assay almost exclusively (Stiborova et al., 1998). Both AAI and AAII form DNA adducts in vitro using rat liver S9 mix, resulting in two major adduct spots for AAI and AAII (Schmeiser et al., 1988). In addition, a minor adduct was formed in incubations with AAI, which is one of the major adducts formed in incubations with AAII. Whereas for AAI the same DNA adducts were observed under aerobic and anaerobic conditions, AAII gave rise to adduct formation only anaerobically. In contrast, no DNA adducts were found for aristolactam I and aristolactam II in the presence of rat liver S9 mix (Schmeiser et al., 1988).
The structures of the major AA–DNA adducts were elucidated spectroscopically as 7-(deoxyadenosin-N6-yl)aristolactam I (dA–AAI), 7-(deoxyguanosin-N2-yl)aristolactam I (dG–AAI) and 7-(deoxyadenosin-N6-yl)aristolactam II (dA–AAII) (Figure 3) (Pfau et al., 1990b, 1991). It was also shown that the dA–AAII adduct is formed from AAI through a demethoxylation reaction of AAI (Stiborova et al., 1994). A second major guanosine adduct formed by reaction of AAII with deoxyguanosine 3′-monophosphate and DNA was tentatively determined as 7-(deoxyguanosin-N2-yl)aristolactam II (dG–AAII) (Stiborova et al., 1994). These chemical structures indicate that a cyclic N-acylnitrenium ion with a delocalized positive charge, as the ultimate carcinogenic species, binds preferentially to the exocyclic amino groups of purine nucleotides in DNA or is hydrolysed to the corresponding 7-hydroxyaristolactam (Figure 3). This preference for reaction with the exocyclic amino group is unusual for nitroaromatic compounds since their major target site in DNA is the C-8 atom of deoxyguanosine. However, this fits in with the concept introduced by Dipple (1995) that polycyclic arylaminating and polycyclic aralkylating agents that delocalize charge and are substantially distorted from planarity react extensively at the amino groups of both deoxyguanosine and deoxyadenosine. It is known that in the activation of carcinogenic nitroaromatics and aromatic amines acetylation of the amino or hydroxyamino group plays a key role. Therefore, the activation of AA is a unique example of intramolecular acylation, which leads to the ultimate carcinogen.
Enzymatic activation of both AAs by buttermilk xanthine oxidase and rat DT-diaphorase, cytosolic nitroreductases, produced a similar adduct pattern to that obtained by rat liver S9 mix-mediated metabolism (Schmeiser et al., 1988; Stiborova et al., 2001a, 2002), confirming that nitroreduction is the crucial step in the pathway of metabolic activation of AAs to their ultimate DNA binding species. It was also demonstrated that both AAs could be activated by rat liver microsomes via simple nitroreduction (Schmeiser et al., 1997). This hepatic microsomal activation of AA was attributed to cytochrome P450 (CYP) 1A1 and CYP1A2 and, although to a minor extent, to NADPH:CYP reductase using specific CYP/NADPH:CYP reductase inhibitors and purified enzymes (Stiborova et al., 2001b,c).
All four purine AA–DNA adducts were identified by 32P-post-labelling in vivo in different organs of rats treated orally with five daily doses (10 mg/kg body wt) of AAI and AAII (Pfau et al., 1990a). The adduct patterns in DNA from forestomach and kidney, target tissues of AA-mediated carcinogenesis, and from non-target tissues like stomach, liver and lung were similar, indicating that adduct formation is not directly correlated with initiation of the carcinogenic process and subsequent tumour formation in target tissues in rats. In this in vivo study DNA binding by AAI was in general 10 times higher compared with AAII. For AAI total relative adduct labelling was highest in forestomach DNA, with ~3 adducts/106 nt. In the bladder (also a target tissue) DNA binding by AAII was much greater than for AAI (relative adduct labelling was ~0.4 adducts/106 nt for AAI, compared with 0.8 adducts/106 nt for AAII) (Pfau et al., 1990a). This difference in organotropic activity could be related to different phase II metabolic pathways for AAI and AAII. Whereas AAIa may be excreted as an O-glucuronide, AAII, unlike AAI, is metabolized to the corresponding lactam, which can only form an N-glucuronide, which could be hydrolysed in the bladder due to the acidic nature of urine and form DNA-reactive species. These combined data indicate that AAI may be responsible for the induction of carcinoma in the gastrointestinal tract while AAII could give rise to neoplastic changes and to toxic effects in the urinary tract. This suggestion is further supported by the fact that in rats treated with pure AAI a high incidence of tumours of the forestomach was observed but no neoplastic changes were found in the urinary tract (Schmeiser et al., 1990).
Oncogene activation by AA
Protooncogenes have been identified as genetic targets that are involved in chemical carcinogenesis (Balmain and Brown, 1988). In rodents many chemical carcinogens activate the ras protooncogene by a single point mutation, resulting in the alteration of amino acid residue 12, 13 or 61. Likewise, AA-initiated carcinogenesis in rodents is associated with a distinct molecular characteristic, activation of H-ras by a specific AT→TA transversion mutation in codon 61 (CAA). This mutation occurs exclusively at the first adenine of codon 61 in all forestomach and ear duct tumours of rats treated with AAI (Schmeiser et al., 1990) and was confirmed in tumours of the forestomach and lung of mice treated with the plant extract AA (Schmeiser et al., 1991). The mutagenic activity of AA was also investigated in different organs of the λ/lacZ transgenic mouse (Muta™Mouse) after intragastric treatment with 15 mg AA/kg body wt once a week for 4 weeks (Kohara et al., 2002). Increased mutation frequencies in the lacZ and cII genes were observed in the target organs (forestomach, kidney and bladder) compared with non-target organs (e.g. glandular stomach and liver). Moreover, mainly AT→TA transversion mutations were found by sequence analysis of cII mutants in the target organs. This selectivity of AAI for mutations at adenine residues is consistent with the extensive formation of dA–AAI adducts in the target organs in rats (Pfau et al., 1990a; Stiborova et al., 1994). Moreover, an apparently life-long persistence of dA–AAI adducts in forestomach DNA was found, whereas dG–AAI adducts were continuously removed from the same DNA over a 36 week period in rats treated with a single dose of AAI (Fernando et al., 1993). As suggested by others, it is possible that persistent DNA adducts may occupy specific genomic sites that are not amenable to repair and that these DNA adducts may be converted into the mutations found in target genes of carcinogenesis, e.g. cellular oncogenes (Randerath et al., 1985).
Mutagenic activity of AA–DNA adducts and DNA binding specificity of AA
Oligonucleotides containing defined DNA adducts placed at specific sites are useful tools for investigating how individual chemical lesions formed in DNA by carcinogens are converted into mutations (Singer and Essigmann, 1991). To examine the mutagenic activity of AA–DNA adducts, mono-adducted oligonucleotides containing the major AA–DNA adducts located at a defined site have been used in primed DNA replication reactions with phage T7 DNA polymerase (Broschard et al., 1994) and human DNA polymerase α (Broschard et al., 1995). It was found that dAMP and dTMP were incorporated equally well opposite the adenine adducts (dA–AAI and dA–AAII), whereas the guanine adducts (dG–AAI and dG–AAII) led to preferential incorporation of dCMP. The translesional bypass past adenine adducts of AA indicates a mutagenic potential resulting from dAMP incorporation by DNA polymerase, suggesting that an AT→TA transversion mutation would be the mutagenic consequence. Incorporation of dTMP opposite the adenine adducts or dCMP opposite the guanine adducts results in a non-mutagenic event. Therefore, the adenine adducts have a higher mutagenic potential compared with the guanine adducts, which may explain the apparent selectivity for mutations found at adenine residues in codon 61 of the H-ras gene in AA-induced rodent tumours (Schmeiser et al., 1990, 1991) and the preferential induction of AT→TA transversion mutations in the cII gene in target organs of the AA-treated Muta™Mouse (Kohara et al., 2002).
Moreover, this assay showed that, regardless of the type of AA–DNA adduct examined, DNA synthesis was blocked predominantly (80–90%) at the nucleotide 3′ of each adduct (Broschard et al., 1994). Thus, DNA polymerase arrest due to the presence of bulky AA–DNA adducts can be used to examine sequence-specific DNA binding by AA in genes involved in the carcinogenic process. To a certain degree, it is possible to relate the DNA binding specificity of a carcinogen to specific mutations found in a target gene for tumour formation (Denissenko et al., 1996). Using an adduct-specific polymerase arrest assay it was demonstrated that both adenines in codon 61 of the H-ras gene in a plasmid are AA–DNA bindings sites (Arlt et al., 2000), indicating that the mutations observed in AA-treated rodents may originate from adduct formation in this codon, thereby triggering tumorigenesis.
Source: https://academic.oup.com/mutage/article/doi/10.1093/mutage/17.4.265/1065881/Aristolochic-acid-as-a-probable-human-cancer