Aristolochic acid (AA)-DNA adduct as marker of AA exposure and risk factor for AA nephropathy-associated cancer


Aristolochic Acid (AA)-DNA Adduct as Marker of AA Exposure and Risk Factor for AA Nephropathy-Associated Cancer

Dear Sir,

So-called Chinese herbs nephropathy (CHN) has been extensively reviewed.1, 2 Briefly, the disease is a rapidly progressive interstitial fibrosing nephropathy frequently associated with urothelial cancer reported in a series of young Belgian women given slimming pills containing aristolochic acid (AA), a nephrotoxic and carcinogenic plant alkaloid extracted from Aristolochia species. The inclusion of AA resulted from the presumed accidental substitution of Stephania tetrandra (fangji) by Aristolochia fangchi in the preparation of the slimming pills, likely because of the confusion between the Chinese names fangji and fangchi. Exposure of the patients to AA was substantiated by the identification of specific AA-DNA adducts in the kidney and ureteric tissue of CHN patients by 32P-postlabeling. Finally, the more recent demonstration that the sole administration of AA induces in rodents similar lesions to those observed in humans conclusively demonstrated the causal role of this plant alkaloid. This has led to rename so-called CHN as aristolochic acid nephropathy (AAN).

Since the description of the Belgian epidemic in the early 1990s, identical cases have been increasingly reported following the consumption of herbal remedies containing AA outside the context of the Belgian slimming regimen, mainly in China and Japan, and in some cases in Europe.1, 2 In our study, we report the first fully documented cases of AAN in France. Histopathologic and DNA adduct analyses performed on autopsy samples revealed that these cases did indeed represent authentic AAN outside Belgium. It is of interest to note that one of them had developed a transitional cell carcinoma in the right urinary tract with invasive liver metastases. Given the availability of tissue samples from this latter patient's postmortem examination, we were able to evaluate AA-DNA adduct formation in various human tissues outside the urinary tract.

Patient 1 had already been reported in part as a 34-year-old woman who had ingested from January to September 1992 a herbal drug called “Preparation Number 28”.3 This slimming regimen was readily available over-the-counter in French pharmacies but was later shown by HPLC to contain AA. In the absence of renal biopsy, the diagnosis of interstitial nephritis was based on bilateral shrunken kidneys on ultrasound. The likelihood of AAN was based on the short delay between the consumption of the herbal remedy and the development of a rapidly progressive renal failure, and on the absence of the prolonged consumption of other drugs. New follow-up data now include initiation of hemodialysis a few months later (August 1994) and kidney transplantation in 1998. The patient died in 2000. Autopsy revealed extensive renal interstitial fibrosis whose severity, hypocellularity and mainly cortical development, including medullary rays and columns of Bertin, were strongly suggestive of AAN. The grafted kidney displayed moderate interstitial fibrosis consistent with chronic rejection. Finally, necrotic tissue, which was found to be of a high grade, poorly differentiated carcinoma consistent with a transitional cell origin, was observed in the right perirenal space and in the liver. For the detection of AA-DNA adducts, we used the 32P-postlabeling assay as described previously.4 Reference compounds for the identification of AA-DNA adducts were prepared by in vitro incubation.4 As shown in Figure 1, we identified the adenosine adduct of aristolochic acid I (AAI), dA-AAI [7-(deoxyadenosin-N6-yl)-aristolactam I], in kidney, ureter, bladder, liver, lung, stomach, small intestine, spleen, adrenal and brain. This adenosine adduct was the most abundant AA-DNA adduct found in urothelial tissue of the Belgian AAN patients.4, 5 By contrast, DNA isolated from a series of control kidneys from patients with other nephropathies analyzed in parallel were virtually free of DNA adducts in the area where AA-derived adducts were located.4, 5 No dA-AAI adducts were detectable in heart and in the kidney transplant of patient 1. Quantitative analysis showed that levels of the dA-AAI adduct were in the range of 1.0 to 21.9 adducts per 109 nucleotides, being the highest in lung (21.9), spleen (21.2), adrenal (19.5), liver (17.5) and ureter (16.5) (Table I). Only weak DNA binding was observed in bladder (2.7), brain (1.9) and kidney (1.0) (Table I).

Figure 1.

Autoradiographic pattern of DNA adducts found in kidney tissue of a Belgian AAN patient (a) and in various human tissues of patient 1: kidney (b), ureter (c), bladder (d), liver (e), lung (f), stomach (g), small intestine (h), heart (i), adrenal (j), spleen (k) and brain (l). The nuclease P1 enrichment version of the 32P-postlabeling method was used as previously described.4 Polyethyleneimine-cellulose (PEI-cellulose) thin layer plates from Machery and Nagel (Düren, Germany) have been used. Origins in the bottom left corner were cut off before exposure. Screen enhanced autoradiography was at −80°C for 12–24 hr. Chromatographic conditions used: D1: 1 M sodium phosphate, pH 6.8; D3: 3.5 M lithium formate, 8.5 M urea, pH 4.0; D4: 0.8 M LiCl, 0.5 M Tris-HCl, 8.5 M urea, pH 9.0; D5: 1.7 M NaH2PO4, pH 6.0. dA-AAI, 7-(deoxyadenosin-N6-yl)-aristolactam I. dG-AAI, 7-(deoxyguanosin-N2-yl)-aristolactam I. dA-AAII, 7-(deoxyadenosin-N6-yl)-aristolactam II. All other spots on the thin layer plates are of unknown origin (systemic).

Table I. Quantitative Analysis of AA-DNA Adducts in Various Human Tissue Samples of French Patients with AAN
CaseGender/age (years)Duration of herbal intake (months)TissueRAL1 in tissue samples (mean/109 nucleotides)2 dA-AAI3
  • 1

    Relative adduct labeling.

  • 2

    Mean of duplicate determinations in separate experiments.

  • 3

    dA-AAI, 7-(deoxyadenosin-N6-yl)-aristolactam I.

  • 4

    “Preparation Number 28”.

  • 5

    “Preparation Number 23”.

  • 6

    ND, not detected.

1F/34124 (1992)Left kidney1.0
   Kidney transplantND6
   Left ureter16.5
   Small intestine10.0
2F/3445 (1990)/1.54 (1992)Right kidney3.7
   Left kidney54.0

Patient 2 was a 34-year-old woman who had consumed a herbal drug mixture called “Preparation Number 23” for 4 months in 1990. In 1992, she absorbed the same “Preparation Number 28” as the previous patient for 1.5 months. AA was identified in both herbal mixtures.3 Hemodialysis was initiated in 1999. The patient died in 2001. Autopsy revealed severe, diffuse, hypocellular mainly cortical renal interstitial fibrosis including medullary rays and columns of Bertin as well as atypia of the collecting ducts that were strongly suggestive of AAN. Multivisceral calcifications, probably the result of secondary hyperparathyroidism, were also noted. Moderately extensive interstitial myocardial fibrosis, possibly ischaemic in origin, was also found. DNA was only isolated from both kidney specimens and assayed for the presence of AA-DNA adducts by 32P-postlabeling.4 Specific dA-AAI adducts were detected in both kidneys, with adduct levels of 3.7 and 54 adducts per 109 nucleotides in the right and left kidney, respectively (Table I).

Both patients fulfilled the well-established diagnostic criteria of AAN:1, 2 1) herbal drugs proven to contain AA have been ingested, 2) the typical renal mainly cortical hypocellular interstitial fibrosis was observed and 3) the most abundant AA-DNA adduct, dA-AAI, was identified in the urinary tract. In addition, the development of a malignancy consistent with a transitional cell carcinoma of the right urinary tract in 1 of the patients is in line with the near 50% incidence of upper urinary tract urothelial malignancy found in Belgian AAN patients.5, 6 An invasive transitional cell carcinoma of the urinary tract associated with the presence of AA-DNA adducts was also reported in an English patient with AAN.7 This report therefore highlights once again the carcinogenic potential of AA in humans beings. Of note, herbal remedies containing species of the genus Aristolochia were recently classified as carcinogenic to humans by the International Agency for Research on Cancer (IARC).8 Several AAN cases unrelated to the Belgian epidemic have been reported worldwide.1, 2 Most of them live in Asia where herbal drugs including the Aristolochia species have been used for centuries. The diagnosis of AAN in France was suggested on the basis of the intake of herbal drugs containing AA in 4 patients.3 The present observations fulfill all the diagnostic criteria of AAN described above and bring the number of reported AAN patients outside the Belgian epidemic in Europe up to 11.1, 2 Several countries including France, Germany, the UK, Canada and Australia have already banned the use of herbs containing AA.9 Moreover, the FDA advised consumers already back in 2000 to immediately discontinue the use of any botanical products containing aristolochic acid and published a list of botanical products that had been shown to contain AA.10 However, despite the actions of the FDA, a recent report published in 2003 demonstrated that herbal products containing AA, sold for gastrointestinal symptoms, weight loss, cough and immune stimulation, were still available for sale on U.S. websites.11 Therefore, our report points once more to the urgent need to submit the so-called “natural” harmless drugs of herbal medicine to a critical evaluation of benefits and side effects prior to their release for medical use just as it is mandatory for drugs in the Western world.9

In rodents, the formation of DNA adducts by AA has been extensively studied in several tissues.2, 8, 12 In humans, by contrast, little is known concerning AA-related DNA adduct formation and their persistence in other tissues than the urinary tract. A variety of human enzymes including cytosolic nitroreductases (e.g., DT-diaphorase and xanthine oxidase), cytochrome P450 (CYP) enzymes (e.g., CYP1A1, CYP1A2 and NADPH:CYP reductase) and peroxidases (e.g., prostaglandin H synthase) capable of bioactivating AA have been identified during the last years.2, 13 The detection of the dA-AAI adduct in the kidney, ureter, bladder, liver, lung, stomach, small intestine, spleen, adrenal and brain removed from patient 1 suggests the presence of AA activating enzymes in the cells of those tissues and that AA and/or its metabolites are distributed via the blood stream to other organs. Significant levels of the dA-AAI adduct were also detected postmortem in tissue samples of a Belgian AAN patient not only in the kidney but also in liver, pancreas, lymph nodes, stomach and lung.14 It is of interest to note that AA-DNA adducts were found in the central nervous system, indicating that AA and/or its metabolites are capable to pass through the blood brain barrier. This finding is also in line with the demonstration of CYP and NADPH:CYP reductase enzyme activity in the human brain.15 The absence of AA-DNA adducts in the myocardial tissue of this patient should be added to that previously reported in muscle and skin samples from AAN patients.4 No AA-DNA adducts were found in the kidney transplant of patient 1, demonstrating that a current exposure to AA can be excluded, at least in patient 1.

The mutagenic and carcinogenic significance of AA-DNA adducts and their life-long persistence have been demonstrated in rats, the dA-AAI adduct being associated with a characteristic A→T transversion mutation at the first adenine of codon 61 (CAA) found in the H-ras protooncogene of tumors induced by AAI.16 The long persistence of this adduct in various organs in rats2, 4 is in line with its detection in several tissues of those two AAN patients many years after they had stopped taking the herbal medicine containing AA. In a series of Belgian AAN patients, the dA-AAI adduct was detectable almost 8 years after the patients stopped taking the herbal slimming regimen.5 As discussed previously,4 the possible explanation for the apparent life-long persistence of specific DNA adducts in particular organs is that these lesions reside in the genome of an essentially dormant subpopulation of tissue cells. This suggests that AA exposure itself may have caused mitosis inhibition and/or growth arrest of this subpopulation. The tissular adduct levels showed a wide variation ranging from 1.0 to 21.9 and from 3.7 to 54.0 adducts per 109 nucleotides for patient 1 and 2, respectively (Table I). In patient 1, adduct levels in kidney and bladder, the target tissues for AAN-associated urothelial malignancy, were remarkably lower than in the ureter (Table I). Although the levels of dA-AAI adducts in the ureter from an English patient with AAN were 10-fold higher than in the kidney,7 it is of interest to note that in urothelial tissues from Belgian AAN patients the levels of dA-AAI adducts were higher in renal than in ureteric tissue.5 It is therefore tempting to speculate that other components given during the slimming regimen in Belgium may have influenced the metabolic activation and/or persistence of AA-DNA adducts in renal tissue. Indeed, in rats treated with AA and dexfenfluramine, an anorexigen given with the slimming regimen, levels of AA-DNA adducts in liver and renal tissue samples were significantly increased compared to rats treated with AA alone.17 On the other hand, in patient 2 the adduct levels in the left kidney were approximately 15 times higher than in the right kidney (Table I). Large differences in DNA adduct formation between right and left kidney of the same patient were also observed in some Belgian patients with AAN (Arlt and Schmeiser, unpublished observation). It is possible that differences in renal damage (e.g. necrotic, apoptotic or neoplastic processes) may be responsible for differences in DNA adduct formation and their persistence. More importantly, in patient 1, high adduct levels were found both in nontarget organs such as the liver, lung, spleen, adrenal, stomach and small intestine and in target organs such as the ureter (Table I). By contrast, the highest levels of DNA adducts in AAI-treated rats were measured in the DNA of the target organ of carcinogenesis, the forestomach, with lower levels in most nontarget tissues (e.g., liver, kidney and bladder).12 In rats, AA showed mainly a high incidence of tumors in the forestomach, but primary tumors were also found in the renal cortex, renal pelvis and urinary bladder.2, 8 Moreover, AA-treatment of mice resulted in subsequent tumor formation in the forestomach, lungs, uterus and lymphoid organs.2, 8 The multisystemic formation and persistence of AA-DNA adducts in AAN patients therefore strongly suggests the possibility of the future development of multisystemic tumours in AAN cases similarly to what has previously been reported in AA-exposed rodents.

DNA adduct-based risk assessment is very likely to correlate with dose and reflect modulations by DNA repair rates but, by necessity, will overlook other factors that impart susceptibility and may increase cancer risk, such as genetic polymorphisms or mutations in oncogenes and tumor suppressor genes.18 This study shows that other factors than DNA adduct formation by AA may also be critical for the high incidence of urothelial tumors observed in AAN patients. Urothelial tumors of AAN patients in Belgium were associated with alterations in the p53 tumor suppressor gene.6 Therefore, p53 mutations in AAN-related urothelial malignancy may provide a link to the molecular mechanisms leading to tumors by AA in human beings.19 Interestingly, a characteristic A→T transversion mutation was found recently at the first adenine of codon 139 (AAG) of the p53 gene in urothelial tumor cells from an English AAN patient.20 Moreover, specific A→T transversion mutations in the human p53 gene associated with the presence of AA-DNA adducts were detected in primary embryonic cells from a human p53 knock-in (Hupki) mouse strain after exposure to AA.21 On the other hand, specific DNA damage due to AA in urothelial cells and cell-specific alterations at the transcription level of proteins might impair physiological processes.22 This may not only be of primary importance in explaining the rapidly progressive nature of AAN but also a potential mechanism on the seeming tissue specificity of the AAN-associated oncogenesis compared to the widespread occurrence of AA-DNA adducts.

In summary, those French cases of AAN highlight the carcinogenic potential of AA in humans and the need for vigilance as all products known to or suspected of containing AA should be banned from the market worldwide.

Yours sincerely,

Volker M. Arlt, Véronique Alunni-Perret, Gérald Quatrehomme, Patrick Ohayon, Laetitia Albano, Hacène Gaïd, Jean-Francois Michiels, Alain Meyrier, Elisabeth Cassuto, Manfred Wiessler, Heinz H. Schmeiser, Jean-Pierre Cosyns