• carcinogenesis;
  • cancer aetiology;
  • mutagenesis


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results and discussion
  5. Acknowledgements
  6. References

The proposal has been put forward that the primary cause of Balkan endemic nephropathy (BEN) is exposure to food crops contaminated with seeds of Aristolochia spp, which contain high levels of aristolochic acids (AA). Recently, tumour DNA samples from patients with BEN were found to harbour principally A to T mutations in the TP53 tumour suppressor gene (Grollman et al., Proc Natl Acad Sci USA 2007;104:12129–34). Using a novel mutation assay in which we can induce and select mutations in human TP53 sequences in vitro by exposure of cultured cells to a mutagen, we found that A to T mutations were elicited by aristolochic acid at sites in TP53 rarely mutated in human cancers in general, but which were observed in the BEN patients. This concordance of specific mutations in patient tumours and aristolochic acid I-exposed cultures supports the argument that AA has a direct role in the aetiology of BEN-associated cancer. © 2008 Wiley-Liss, Inc.

Aristolochic acids (AAs) extracted from the plant species Aristolochia are nitrophenanthrene carboxylic acids with potent mutagenic activities and are carcinogenic in rodents.1, 2 Herbal remedies containing Aristolochia have been classified as carcinogenic to humans (Group 1) by the International Agency for Research on Cancer (IARC).3 AA exposure elicits an unusual type of renal pathology and greatly increases the risk of urothelial carcinoma in afflicted individuals. A recent cluster of AA nephropathy (AAN) cases with associated urothelial malignancy occurred in a Belgian weight-reduction clinic where patients had been exposed inadvertently to AA.4 AA-DNA adducts, which can persist for years, were detected in the tissues of the AAN patients, unequivocally confirming exposure to the nephrotoxic and carcinogenic substance. This incident provoked renewed attention to the possible role of AA exposure in other patient cohorts displaying the distinctive pathological features of AAN, in particular patients from villages in the Balkans diagnosed with Balkan endemic nephropathy (BEN). The hypothesis that rural populations in the Danube valley at high risk of urothelial cancer have been chronically exposed to AA was greatly strengthened by the recent findings on the unusual spectrum of TP53 mutations displayed by a set of 11 urothelial cancers from BEN patients.5, 6 The predominant DNA sequence change leading to inactivation of the TP53 tumour suppressor gene in these tumours was A:T to T:A. It is plausible that these base changes were caused by exposure to AA because pre-mutagenic AA-specific adducts on adenine, predominantly 7-(deoxyadenosin-N6-yl)aristolactam I, were detected in the Balkan patients' tissues,5 and were reported previously in tissues of AAN patients.4 Furthermore, it is known that these DNA modifications cause predominantly A to T transversions,1 which constitute only a minor component of essentially all spontaneous mutation spectra in eukaryotes, regardless of organism or reporter gene.

Characteristic base changes in tumours can be clues to mutagenic exposures contributing to cancer risk; however, determining the locations of the mutations in a defined sequence, such as the TP53 gene, can provide a greater degree of specificity in mutation profiling than provided by comparing only the frequencies of base substitution types. This information is generally lacking because mutagenesis protocols have been designed with reporter genes that allow rapid and efficient selection of mutants, rather than with human cancer genes, such as TP53, for which no simple strategies to recover mutants were known. To probe further the hypothesis that A to T transversions in BEN associated tumours derive from exposure to AA, we induced and selected mutations in human TP53 sequences of mammalian cells exposed to aristolochic acid I (AAI) in vitro. The procedure we established for such purposes (the HUF assay) scores dysfunctional TP53 mutations in Hupki (human TP53 knock-in) murine fibroblasts (HUFs) that have been immortalised following exposure of the primary cultures to carcinogens.7, 8 The cultures used in the mutation assay are seeded directly from primary embryonic cells taken from the Hupki mouse strain, in which the segment of human TP53 that is targeted in human cancers has replaced the corresponding segment of the p53 murine orthologue.9 This test system allows a direct comparison of mutation profiles found in patients with experimentally generated human TP53 mutations arising in cells exposed to the putative carcinogen.

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results and discussion
  5. Acknowledgements
  6. References

Preparation of primary Hupki cell cultures

The humanised p53 locus of the prototype Hupki (human p53 knock-in) strain,9 which harbours human p53 exons 4–9, encodes arginine at the exon 4 codon 72 polymorphic site. To obtain cells from Hupki mice for mutagenicity studies corresponding to both common human alleles, we also generated a variant Hupki strain that encodes the common variant proline at residue 72.10, 11 For this study, the 2 strains were interbred for >3 generations, and 13.5-day-old embryos with the codon 72 genotype arginine/proline and proline/proline were used to prepare primary cultures of embryonic fibroblasts. Each embryo was placed in separate dishes, minced and trypsinised briefly to obtain single cells. After transfer into fresh dishes, cells were incubated for 3–4 days under standard conditions (humidified chambers at 37°C, and at 5% CO2) in Dulbecco's Medium (DMEM) supplemented with 10% fetal calf serum (FCS), and were either used directly or frozen in dimethylsulfoxide (DMSO) until use.

Mutagen assays

Stock solutions of 10 mM aristolochic acid I in sterile water and 1 mM N-methyl-N′-nitro-N-nitrosoguanidine (for positive control experiment, dissolved in DMSO) were prepared and stored frozen until preparation of diluted working solutions, which were discarded after use. Independent cultures of primary embryonic fibroblasts with human TP53 gene codon 72 genotypes codon72arg/pro and codon72pro/pro were cultured in medium containing 50 μM AAI for 2–4 days and then passaged at near confluency in normal medium until immortalised, typically within 15 subsequent passages (10–16 weeks). N-methyl-N′-nitro-N-nitrosoguanidine (MNNG) treatment: Independent primary cell cultures (2 × 105 cells/culture) were exposed to 20 μM MNNG for 2 hr and then serially passaged until immortal.

All cultures were maintained under standard culturing conditions in DMEM supplemented with 10% FCS throughout the experiments. To identify p53 mutations in cultures that had bypassed senescence and evolved into immortalised cell lines, genomic DNA was prepared directly from harvested cells of each immortalised cell line and human TP53 sequences were amplified by the polymerase chain reaction as described in Ref. 8 with intronic primers as follows:

Exon 4



Exons 5 and 6



Exon 7



Exons 8 and 9



We performed direct dideoxy sequencing of PCR products to identify the mutations, which we verified by repeat amplification of genomic DNA and sequencing in both 3′ and 5′ directions. We compared frequencies of base substitution types using Fisher's Exact Test and considered them significantly different if the p-value was <0.005.

Results and discussion

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results and discussion
  5. Acknowledgements
  6. References

In a suite of 4 independent experiments with HUF primary cells treated for 2–4 days at Passage 2 with 50 μM AAI, we established 21 mutant cell lines harbouring a total of 25 TP53 mutations (4 cell lines harboured 2 mutations) (Table I, rows L95–L159). The most common type of base change was A to T transversion (12 of 25 base changes) with an unmistakable strand bias: the pre-mutated adenine residue was positioned on the non-transcribed strand in all instances. Likewise, in tumour DNA of BEN patients, A to T was the most common base substitution detected (14 of 19 mutations), and in all but 1 case, the targeted purine was positioned on the non-transcribed strand.5 These data are in keeping with the observation that bulky DNA adducts such as those formed by AA block transcription, eliciting strand-specific repair activity such that the majority of pre-mutagenic lesions linger on the non-transcribed strand.

Table I. TP53 Mutant Cell Lines Derived from HUF Cells Exposed to AAI
Hupki cell lineGenotype1 (primary cells)Mutated codonSequence changeBase changeAmino acid changep53 functional status2No. entries in IARC database3Source
  • 1

    The primary cells used in these experiments were either heterozygous at codon 72 (R/P: CGC/CCC, or homozygous for the proline allele (P/P: CCC/CCC). HUFs used in previous studies were homozygous at codon 72 for the arginine allele (R/R: CGC/CGC).

  • 2

    The TP53 functional status of cells with a given mutation refers to transcriptional transactivation function in the yeast assay as defined in Petitjean et al.12 and tabulated in the IARC TP53 database ( NF, non-functional; PF, partially functional; F, functional (wild-type activity).

  • 3

    The entry refers to the number of times a given mutation has been observed in human tumour samples as recorded in R12 of the IARC TP53 database. Mutations in bold are identical to mutations reported by Grollman et al.5 that were found in tumours of BEN patients. The mutated base pairs are underlined.

L95R/P239AAC to GACA to GAsn to AspNF43This study
L96R/P131AAC to TACA to TAsn to TyrNF0This study
L96R/P291AAG to TAGA to TLys to Stop(p53null)5This study
L97R/P305AAG to TAGA to TLys to Stop(p53null)14This study
L98R/P273CGT to TGTC to TArg to CysNF611This study
L99R/P104CAG to CTGA to TGln to LeuF1This study
L104R/P135TGC to TGGC to GCys to TrpPF24This study
L104R/P159GCC to CCCG to CAla to ProNF26This study
L112R/P131AAC to ATCA to TAsn to IleNF6This study
L113R/P168CAC to CAGC to GHis to GlnPF1This study
L113R/P169ATG to TTGA to TMet to LeuPF0This study
L114R/P258GAA to GACA to CGlu to AspNF4This study
L115R/P179CAT to CGTA to GHis to ArgNF139This study
L128P/P209AGA to TGAA to TArg to Stop(p53null)12This study
L129P/P249AGG to TGGA to TArg to TrpNF34This study
L130P/P135TGC to TGGC to GCys to TrpPF24This study
L152P/P55ACT to TCTA to TThr to SerPF0This study
L153P/P135TGC to TGGC to GCys to TrpPF24This study
L154P/P194CTT to TTTC to TLeu to PheNF25This study
L155P/P249AGG to TGGA to TArg to TrpNF34This study
L156P/P209AGA to TGAA to TArg to Stop(p53null)12This study
L157P/P164AAG to TAGA to TLys to Stop(p53null)14This study
L157P/P213CGA to CCAG to CArg to ProNF6This study
L158P/P154GGC to GTCG to TGly to ValNF59This study
L159P/P245GGC to GCCG to CCys to AlaNF12This study
L6R/R209AGA to TGAA to TArg to Stop(p53null)12Ref. 7
L7R/R176TGC to TGGC to GCys to TrpNF18Ref. 7
L8R/R280AGA to AGTA to TArg to SerNF16Ref. 7
L9R/R158CGC to GGCC to GArg to GlyNF20Ref. 7
L9R/Rint8AG to TGA to T Splice site Ref. 7
L10R/R281GAC to GTCA to TAsp to ValNF4Ref. 7
L11R/R286GAA to GTAA to TGlu to ValNF5Ref. 7
L12R/R193CAT to CTTA to THis to LeuNF50Ref. 13
L13R/R130CTC to CACT to ALeu to HisNF2Ref. 13
L14R/R313AGC to TGCA to TSer to CysF1Ref. 13
L15R/R139AAG to TAGA to TLys to Stop(p53null)3Ref. 13
L16R/R203GTG to GTCC to GVal to Valsilent1Ref. 13

A to T transversions were rare in control HUF cell lines from cultures not treated with mutagen (1 A to T substitution among 52 p53 mutant cell lines from untreated cultures: Refs. 7, 8 and14 and unpublished data). To verify that the HUF assay generates different mutation patterns with different classes of mutagens, we also performed a positive control experiment, using the alkylating agent MNNG, which is known to target guanine residues. As shown in Table IIa, the hallmark mutation in HUF cell lines we derived from exposure to MNNG is G to A transition (10/13 mutations, 77%) in keeping with prediction from mutagen studies in other test systems with this compound. The difference between AAI and MNNG experiments in HUF cells with respect to the predominant base change in the p53 gene is highly significant (Table IIb), showing clearly the impact of treatment on p53 mutation patterns detected in the assay. Our previous work8, 10 has shown that benzo[a]pyrene, another class of mutagen that targets guanine as does MNNG but induces principally G to T transversions, induces the anticipated G to T transversions in Hupki cells, thus yields a profile distinct from both AAI (A to T transversions) and MNNG (G to A transitions) in the HUF assay.

Table II. TP53 Mutations in Cell Lines from MNNG-Treated HUFs and TP53 Base Substitutions in Mutagen-Treated Cultures
  • Treatment: MNNG 20μM 2 hr. Primary cells: HUFs Codon 72: CGC/CCC Arg/Pro.

  • 1

    Fisher's Exact test.

(a) TP53 mutations in cell lines from MNNG-treated HUFs
Cell lineCodonWt codonMutated codonBase changeAmino acid changeHomo/hetero
MNNG 77241TCCTTCC to TSer to PheHetero
MNNG 77272GTGATGG to AVal to MetHetero
MNNG 80282CGGTGGC to TArg to TrpHomo
MNNG 84321AAATAAA to TK to StopHetero
MNNG 88152CCGTCGC to TPro to SerHetero
MNNG 88159GCCGTCC to TAla to ValHetero
MNNG 9699TCCTTCC to TSer to PheHomo
MNNG 98165CAGTAGC to TQ to StopHetero
MNNG 100245GGCGACG to AGly to AspHetero
MNNG 105241TCCTTCC to TSer to PheHomo
MNNG 113281GACGAGC to GAsp to GluHetero
MNNG 114113TTCGTCT to GPhe to ValHetero
MNNG 115141TGCTACG to ACys to TyrHetero
(b) TP53 base substitutions in mutagen-treated cultures
G to A102<0.0001
A to T120<0.005

A to T transversions are a comparatively rare class of TP53 base substitution in human tumours (IARC TP53 Database R12, A:T to T:A substitutions: 5.1% of >24,000 mutations15) and also rare in urinary tract tumours not associated with AA exposures (4% of mutations). Curiously, there is no strand bias among the (non-BEN-associated) A to T transversions reported in human tumours (655 premutated adenines on the transcribed strand; 608 premutated adenines on the nontranscribed strand), suggesting that the origins of these mutations, whether from replication errors, DNA damage due to endogenous reactive oxygen species, or other causes, involve mutagenic mechanisms that are probably distinct from the bulky adduct-base mispairing pathway known for various exogenous mutagens including AA. Also noteworthy is the fact that most of the A to T mutations we observed in cell lines from AAI treated HUF cultures have never been reported in non-AA-associated human cancers of the urinary tract system (UTS) (IARC TP53 database, 1,468 tumours: topography classes combined: kidney, ureter, bladder, renal pelvis, other urinary organs), and the remaining AA-HUF mutations were observed in UTS tumours at most only twice.

Of the 2,314 possible TP53 mutants that can be derived theoretically from single nucleotide substitutions in the human TP53 coding sequence, more than 1,200 have been reported at least once in human tumours.12, 16 Whereas some specific human tumour mutations have been detected in >500 tumours (>2%), for example, codon 248 C to T transitions (Arg to Trp), others are infrequent. As noted earlier, human tumour TP53 mutations that correspond exactly in base change and gene location to the A to T mutations we identified in AAI-HUF cell lines are rare (Table I, penultimate column; <20 occurrences among >24,000 human TP53 mutations [<0.1%]). Strikingly, however, 6 of the AAI-HUF mutations uncommon in the database were identical to mutations found in urothelial tumours from BEN patients (Table I). In all, there are 5 codons harbouring A to T transversions that correspond in BEN tumours and HUF cell lines (of a total of 11 codons in tumours and 16 codons in HUF cells found targeted for this base change) (Fig. 1). The concordant A to T substitutions found in BEN patients and in the HUF assay included transversions at codon 131 and codon 209, which arose twice in independent mutagenesis experiments with AAI treated cells, pointing to candidate AAI signature (hotspot) mutations. Previous HUF assays with AAI using HUF cells of the Hupkic72Arg/arg genotype also yielded mutations (at codons 209 and 280) matching the BEN patient dataset that was reported subsequently.7, 13 Furthermore, another of the experimentally induced A to T mutations (at codon 139) matches the TP53 mutation reported in an urothelial carcinoma of a patient in the UK exposed to aristolochic acid from an herbal preparation to treat eczema.17 Thus, although these specific mutations are rare in human tumours overall and rare in non AA-associated urothelial cancers as well (few-to-zero occurrences in the TP53 database), they are reiterated mutations in the 2 datasets (AAI-HUF and BEN patients), underscoring the possibility that they contribute to distinctive features of an AA mutation signature in TP53.

thumbnail image

Figure 1. TP53 A:T to T:A mutations in tumours of BEN patients and in AAI-treated HUFs. BEN patients: Data are from Ref. 5. HUFs: Hupki cell lines derived from primary cultures treated with AAI. Codons harbouring A to T transversions in 2 or more samples are labelled.

Download figure to PowerPoint

The mycotoxin ochratoxin A (OA) has also been considered in conjunction with the aetiology of BEN, but is less plausible as a candidate mutagen inducing A to T transversions in the p53 gene of BEN patients. OA is a weak mutagen at best, and a recent definitive study has shown that HPRT gene base substitutions of mammalian cells exposed to OA did not include A to T transversions and were similar to the spontaneous mutation pattern.18

Given that TP53 mutations in tumours of BEN patients correlate remarkably well with AAI-HUF experimental mutations, yet are of a rare mutation type in other human tumours, it is increasingly plausible that AA has a causative role in the aetiology of BEN-associated tumourigenesis.


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results and discussion
  5. Acknowledgements
  6. References

We thank Mrs. A. Weninger and Mr. K.-R. Muehlbauer for excellent technical assistance.


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results and discussion
  5. Acknowledgements
  6. References
  • 1
    Arlt VM,Stiborova M,vom Brocke J,Simoes ML,Lord GM,Nortier JL,Hollstein M,Phillips DH,Schmeiser HH. Aristolochic acid mutagenesis: molecular clues to the aetiology of Balkan endemic nephropathy-associated urothelial cancer. Carcinogenesis 2007; 28: 225361.
  • 2
    Stiborova M,Frei E,Arlt VM,Schmeiser HH. Metabolic activation of carcinogenic aristolochic acid, a risk factor for Balkan endemic nephropathy. Mutat Res 2008; 658: 5567.
  • 3
    International Agency for Research on Cancer.Some traditional herbal medicines, some mycotoxins, naphthalene and styrene. IARC Monographs Evaluation of Carcinogenic Risks in Humans, Vol. 82. Lyon, France: IARC, 2002.
  • 4
    Nortier JL,Martinez MC,Schmeiser HH,Arlt VM,Bieler CA,Petein M,Depierreux MF,De Pauw L,Abramowicz D,Vereerstraeten P,Vanherweghem JL. Urothelial carcinoma associated with the use of a Chinese herb (Aristolochia fangchi). N Engl J Med 2000; 342: 168692.
  • 5
    Grollman AP,Shibutani S,Moriya M,Miller F,Wu L,Moll U,Suzuki N,Fernandes A,Rosenquist T,Medverec Z,Jakovina K,Brdar B, et al. Aristolochic acid and the etiology of endemic (Balkan) nephropathy. Proc Natl Acad Sci USA 2007; 104: 1212934.
  • 6
    Hranjec T,Kovac A,Kos J,Mao W,Chen JJ,Grollman AP,Jelakovic B. Endemic nephropathy: the case for chronic poisoning by aristolochia. Croat Med J 2005; 46: 11625.
  • 7
    Liu Z,Hergenhahn M,Schmeiser HH,Wogan GN,Hong A,Hollstein M. Human tumor TP53 mutations are selected for in mouse embryonic fibroblasts harboring a humanized TP53 gene. Proc Natl Acad Sci USA 2004; 101: 29638.
  • 8
    Liu Z,Muehlbaer K-R,Schmeiser HH,Hergenhahn M,Belharazem D,Hollstein MC. P53 mutations in benzo(a)pyrene-exposed human p53 knock-in murine fibroblasts correlate with p53 mutations in human lung tumors. Cancer Res 2005; 65: 25837.
  • 9
    Luo JL,Yang Q,Tong WM,Hergenhahn M,Wang ZQ,Hollstein M. Knock-in mice with a chimeric human/murine TP53 gene develop normally and show wild-type TP53 responses to DNA damaging agents: a new biomedical research tool. Oncogene 2001; 20: 3208.
  • 10
    Reinbold M,Luo JL,Nedelko T,Jerchow B,Murphy ME,Whibley C,Wei Q,Hollstein M. Common tumour TP53 mutations in immortalized cells from Hupki mice heterozygous at codon 72. Oncogene 2008; 27: 278894.
  • 11
    Pietsch EC,Humbey O,Murphy ME. Polymorphisms in the TP53 pathway. Oncogene 2006; 25: 160211.
  • 12
    Petitjean A,Mathe E,Kato S,Ishioka C,Tavtigian SV,Hainaut P,Olivier M. Impact of mutant TP53 functional properties on TP53 mutation patterns and tumor phenotype: lessions from recent developments in the IARC TP53 database. Hum Mutat 2007; 28: 6229.
  • 13
    Feldmeyer N,Schmeiser HH,Muehlbauer K,Belharazem D,Knyazev Y,Nedelko T,Hollstein M. Further studies with a cell immortalization assay to investigate the mutation signature of aristolochic acid in human TP53 sequences. Mutation Res 2006; 608: 1638.
  • 14
    Vom Brocke J,Schmeiser HH,Reinbold M,Hollstein M. MEF immortalization to investigate the ins and outs of mutagenesis. Carcinogenesis 2006; 27: 21417.
  • 15
    Olivier M,Eeles R,Hollstein M,Khan MA,Harris CC,Hainaut P. The IARC TP53 database: new online mutation analysis and recommendations to users. Hum Mutat 2002; 19: 60714.
  • 16
    Kato S,Han S-Y,Liu w,Otsuka K,Shibata H,Kanamaru R,Ishioka C. Understanding the function-structure and function-mutation relationships of p53 tumor suppressor protein by high-resolution missense mutation analysis. Proc Natl Acad Sci USA 2003; 100: 84249.
  • 17
    Lord GM,Hollstein M,Arlt VM,Roufosse C,Pusey CD,Cook T,Schmeiser HH. DNA adducts and p53 mutations in a patient with aristolochic acid-associated nephropathy. Am J Kidney Dis 2004; 43: e11e17.
  • 18
    Palma N,Cinelli S,Sapora O,Wilson SH,Dogliotti E. Ochratoxin A-induced mutagenesis in mammalian cells is consistent with the production of oxidative stress. Chem Res Toxicol 2007; 20: 10317.