Detoxification of aristolochic acid I by O-demethylation: Less nephrotoxicity and genotoxicity of aristolochic acid la in rodents

Abstract

Ingestion of aristolochic acids (AA) contained in herbal remedies results in aristolochic acid nephropathy (AAN), which is characterized by chronic renal failure, tubulointerstitial fibrosis and urothelial cancer. AA I and AA II, primary components in AA, have similar genotoxic potential, whereas only AA I shows severe renal toxicity in rodents. AA I is demethylated to form 8-hydroxy-aristolochic acid I (AA Ia) as a major metabolite. However, the nephrotoxicity and genotoxicity of AA Ia has not yet been determined. AA Ia was isolated from urine collected from rats treated with AA I and characterized by NMR and mass spectrometry. The purified AA Ia was administered intraperitoneally to C3H/He male mice for 9 days and its toxicity was compared with AA I. Using (32)P-postlabeling/polyacrylamide gel electrophoresis, the level of AA Ia-derived DNA adducts in renal cortex was approximately 70-110 times lower than that observed with AA I, indicating that AA Ia has only a limited genotoxicity. Supporting this result, when calf thymus DNA was reacted with AA Ia in a buffer containing zinc dust, the formation of AA Ia-DNA adducts was two-orders of magnitude lower than that of AA I. Histopathologic analysis revealed that unlike AA I, no significant changes were detected in the renal cortex of mice treated with AA Ia. Therefore, the contribution of AA Ia to renal toxicity is minimum. We conclude the metabolic pathway of converting AA I to AA Ia functions as the detoxification of AA I.

The ingestion of herbal remedies containing aristolochic acids (AA) has been shown to result in chronic, progressive renal disease designated aristolochic acid nephropathy (AAN) (reviewed by Refs. 1 and 2). The typical clinical syndrome was reported in a group of Belgium women who developed severe renal disease after ingesting slimming pills containing Aristolochia fangchi.^3,4 Nephrotoxicity was manifested by progressive atrophy of renal proximal tubules and development of interstitial fibrosis involving the outer renal cortex and progressing toward the medulla.^5,6 The syndrome was also associated with a high prevalence of renal cancers in the upper urinary tract.^7,8

AA is a mixture of structurally related nitrophenanthrene carboxylic acids including aristolochic acid I (AA I) and its 8-demethoxylated form (aristolochic acid II, AA-II) as major components.⁹ AA I is primarily metabolized along the following two pathways^10–15 (Fig. 1). AA Ia may be demethylated to form 8-hydroxy-aristolochic acid I (AA Ia)^10,16 which, in turn, is subject to phase II conjugation reactions, forming glucuronide or sulfate esters. In fact, using high-resolution mass spectrometry, AA Ia and its O-acetyl-, O-sulfate-, and O-glucuronidated conjugates were detected as the major metabolites in urine of rats treated orally with a mixture of AA I and AA II.¹⁷ Alternatively, the nitro group may be enzymatically reduced to generate the aristolactam I (AL I), which is itself subject to phase II conjugation. AA II, lacking the O-methoxy group, is primarily reduced to aristolactam II (AL II); a part of this metabolite is hydroxylated at C-8 to form 8-hydroxyaristolactam Ia (AL Ia) (Fig. 1).¹

Figure 1.

Scheme outlining the metabolism of aristolochic acids and formation of AA-derived DNA adducts.

Toxicological studies in animals have revealed that AA is associated with renal tubular damage^18,19 and interstitial fibrosis. ^19,20 Acute nephrotoxicity and renal interstitial fibrosis were observed in C3H/He and BALB/c mice treated with a mixture of AA I and AA II.²¹ Using this AAN-animal model, we recently found that AA I is severely nephrotoxic, whereas AA II is not,²² indicating that AA I and/or its metabolites causes the nephrotoxic effect. The formation of AA Ia from AA I was enhanced using liver microsomes prepared from mice treated with 3-methylcholanthrene (3-MC), an inducer of CYP 1A, indicating that CYP 1A is responsible for generating AA Ia.²³ Pretreatment with 3-MC protected mice against renal failure induced by AA I^23,24; therefore, the induction of CYP 1A1/2 may be involved in the AA I detoxification. However, the nephrotoxicity of AA Ia has not been determined.

AA treatment of rodents is also associated with development of cancer.^25–28 Both AA I and AA II are activated by several mammalian enzymes^10–12,29 to form a cyclic nitrenium ion intermediate which, in turn, reacts with the exocyclic amino groups of deoxyadenosine (dA) and deoxyguanosine (dG) residues in cellular DNA, resulting in the formation of covalent DNA adducts (Fig. 1).^11,30,31 The major DNA adducts detected were 7-(deoxyadenosin-N⁶-yl) aristolactam I (dA-AL-I), 7-(deoxyguanosin-N²-yl) aristolactam I (dG-AL-I), 7-(deoxyadenosin- N⁶-yl) aristolactam II (dA-ALII) and 7-(deoxyguanosin- N^2-yl) aristolactam II (dG-AL-II). Such AL-DNA adducts serve as biomarkers of AA exposure and have been detected in the kidneys of animals^1,22,32 and AAN patients^33–35 using ³²P-postlabeing DNA adduct analysis. The existence of AL-DNA adducts in AAN patients was also established using liquid chromatography tandem mass spectroscopy (LC/MS/MS) analysis.³⁵ AA-initiated carcinogenicity in rodents was associated with activation of H-ras by A:T→T:A transversion mutations that occur exclusively at the first adenine of codon 61.³⁶ Frequent A:T→T:A mutations have also been detected in a urothelial cancer from a patient with AAN.^35,37 Therefore, formation of AA I- and AA II-induced DNA adducts through their active nitrenium intermediate is involved in human renal carcinogenesis.

Like AA I, its major metabolite AA Ia is expected to be activated metabolically, forming its nitrenium intermediate that may react with cellular DNA. As expected, N- and O-glucuronides of aristolactam Ia (AL Ia) were detected in rat urine¹⁶ although hydroxylation of AL II also produces AL Ia.¹⁷ However, AA Ia-derived DNA adducts have not yet been detected; therefore, the contribution of AA Ia to the AA genotoxicity is unclear.

In this study, sufficient amount of AA Ia was isolated by HPLC techniques from urine of rats treated with AA I. We employed an AAN-mouse model to establish the genotoxicity and nephrotoxicity of AA Ia.

Material and Methods

Materials

Potato apyrase was purchased from Sigma-Aldrich (St. Louis, MO), micrococcal nuclease and spleen phosphodiesterase from Worthington Biochemical Corp. (Lakewood, NJ) and 30-phosphatase-free T4 polynucleotide kinase and nuclease P1 from Roche Applied Science (Indianapolis, IN). [γ-³²P]-ATP specific activity, >6,000 Ci/mmol) was obtained from GE Healthcare Bio-sciences (Piscataway, NJ). Pure AA I was separated from a mixture of AA I and AA II (Fisher Scientific, Fairlawn, NJ) using preparative reverse-phase HPLC on a Varian automated ProStar System (Palo Alto, CA), as described previously.³²

Purification of AA Ia from rat urine

Wistar rats (8-weeks-old, male), purchased from Charles River Laboratories, (Wilmington, MA), were used in compliance with guidelines established by the NIH Office of Laboratory Animal Welfare. Animals were acclimated in temperature (22 ± 6 2°C)- and humidity (55 ± 6 5%)-controlled rooms with a 12-hr light-dark cycle for at least 1 week. Regular laboratory chow and tap water were allowed ad lib. Rats were treated daily with AA I (10 mg/kg/day, i.p.) for 35 days. Urine was collected over a 24-hr period before the treatment as the control and each week after the treatment. Approximately 800 ml of urine collected from rats treated with AA I was centrifuged (1,600g) for 5 min and the supernatant was mixed with 4 volume of methanol, centrifuged (1,600g) and then evaporated. The urine samples were dissolved in ~30 ml of DMSO. Using Waters 990 HPLC instrument with photodiodearray monitoring 220–500 nm, ~0.4–0.6 ml of the urine sample was loaded on a μBondapak C18 column (0.78 × 25 cm, 5 μm, Waters), eluted at a flow rate of 2.0 ml/min with a linear gradient of 0.1 M triethyl ammonium acetate, pH 7.4, containing 12 to 50% acetonitrile for 30 min, 50 to 100% acetonitrile for 2 min, followed by an isocratic elution with 100% acetonitrile for 3 min. The same gradient system was used for analytical determinations on a μBondapak C₁₈ column (0.39 × 25 cm, 5 μm, Waters) with a flow rate of 1.0 ml/min. The purified compound was identified as AA Ia using mass- and NMR spectroscopy analyses as follows: AA Ia: ¹H NMR (400 MHz, THF-d₈): 9.64 (br s, 1H), 8.70 (s, 1H), 8.65 (d, 1H), 7.76 (s, 1H), 7.59 (t, 1H), 7.06 (dd, 1H), 6.38 (s, 2H). ¹³C NMR (400 MHz, THF-d₈): 168.06, 156.40, 147.27, 147.10, 147.08, 131.80, 131.57, 125.88, 121.24, 120.20, 119.29, 119.22, 119.02, 113.13, 113.07, 103.81. Nano-electrospray ionization (nano-ESI) was employed on a Quattro LCZ (Micromass, UK) mass spectrometer in the positive ion mode. The sample was suspended in 5 mM ammonium acetate and acetonitrile (50/50) then surveyed in the in MS mode. Two peaks were observed at m/z 345.1 and m/z 366.0 represented the ammonium form of the molecular ion [M+NH⁴]+ and the potassium form of molecular ion [M+K]+, respectively.

AAN-mouse model

C3H/He mice (8-weeks-old, male), purchased from Taconic (Germantown, NY), were treated intraperitoneally with 2.5 mg/kg/day of AA Ia or AA I daily for 9 days, as reported earlier.^21,22 Control mice were treated with an identical volume of vehicle (PBS). Body weight was measured daily. The mice were euthanized on day 10 or 24. The left kidney of each mouse was removed and stored in 10% neutral buffered formalin for histopathologic analysis. The contralateral kidney was removed, frozen, and stored at −80°C.

Extraction and digestion of renal DNA

DNA was extracted from frozen tissues using a Qiagen DNeasy Tissue kit (Valencia, CA) according to the manufacturer's protocol. The concentration of DNA was determined by UV spectroscopy as 50 μg/ml = O.D._{260 nm} 1.0. The DNA sample (5.0 μg) was enzymatically digested at 37°C for 16 hr in 100 μl of 17 mM sodium succinate buffer (pH 6.0) containing 8 mM CaCl₂, micrococcal nuclease (30 units) and spleen phosphodiesterase (0.15 unit).³² The reaction mixture was then incubated for another hour with nuclease P1 (1 unit), whereupon 200 μl of water was added. The reaction samples were then extracted twice with 200 μl of butanol; the butanol fractions were combined, back-extracted with 50 μl of distilled water and then evaporated to dryness.

³²P-postlabeling/polyacrylamide-gel electrophoresis (PAGE) analysis

The DNA digestion mixtures were incubated at 37°C for 40 min with 10 μCi of [γ-³²P]-ATP and 30′-phosphatase-free T4 polynucleotide kinase (10 units), followed by incubation with apyrase (50 milliunits) for 30 min, as described previously.³² The ³²P-labeled products were separated by electrophoresis for 4–5 hr on a nondenaturing 30% polyacrylamide gel (35 × 42 × 0.04 cm) with 1,500–1,800 V/20–40 mA. The position of ³²P-labeled adducts was established by Storm 840 (GE Healthcare Bio-sciences, Piscataway, NJ). To quantify the ³²P-labeled products, integrated values were measured using Storm 840 and compared with the standards. Known amounts (0.152–0.000152 pmol) of dA-AL-I- or dG-AL-Imodified oligodeoxynucleotide, prepared by a chemical procedure, were mixed with 5 μg of calf thymus DNA (15,200 pmol) and served as the external standard (characterized as 0.001–1 adduct/10⁵ nucleotides).³² The detection limit for 5 μg DNA was ~5 adducts/10⁹ nucleotides.

Reactivity of AA Ia to calf thymus DNA

Calf thymus DNA (25 μg) was incubated at 37°C overnight (16 hr) with AA Ia or AA I (10, 25 or 100 μM) in 1.0 ml of 50 mM potassium phosphate buffer, pH 7.4 containing zinc dust (1.0 mg). After the reaction, the reaction mixture was centrifuged (1,600g) for 5 min and the supernatant was mixed with 4 volume of methanol and centrifuged (1,600g) to recover the DNA. The DNA was resolved in 1.0 ml distilled water and the concentration of DNA was determined by UV spectroscopy as 50 μg/ml ¼ O.D._{260 nm} 1.0. The DNA sample (1.0 μg) was enzymatically digested and subjected for DNA adduct analysis using ³²P-postlabeling/PAGE as described earlier.

Histologic analysis of AA-treated mouse kidney

Kidneys were fixed in 10% buffered formalin for histopathologic analysis. Sections, cut at 2–3 μm, stained with H&E and Mallory's trichrome were visualized at a magnification of ×200; the cortical tissues were then quantitatively scored for injury. The scorer was blinded as to treatment. A total of 12 pathologic changes were used to quantity renal toxicity: (a) alterations in vasa recti; (b) tubular necrosis; (c) tubular regeneration; (d) tubular mitoses; (e) increase in Bowman's space; (f) tubularization of Bowman's space; (g) calcification or calcium salt deposition; (h) interstitial inflammation; (i) casts; (j) apoptosis; (k) edema and (l) tubular brush border loss.^22,38 The scores obtained were defined as follows: 0, no histologic toxicity; 0.5–1.0, mild histologic toxicity; 1.0–2.0, moderate histologic toxicity; above 2.5, severe histologic toxicity. The histologic score for each mouse was determined using a weighted formula that included all of the 12 pathologic changes, such that the total score ¼ a + 2b + c + d + e + f + 2g + 2h + 2i + j + k + l.

Results

Isolation and characterization of AA Ia

Several rats were treated subcutaneously with AA I (10 mg/ kg/day) for 35 days. Urine was collected before and after the treatment using metabolic cages. As shown in Figure 2a, no AA I-related metabolites were observed, using HPLC/photodiode array system, in urine obtained before the AA I treatment. However, four AA I metabolites were detected in urine from AA I-treated rats (Fig. 2b).

Figure 2.

HPLC isolation of AA Ia from urine of rats treated with AA I. Rat urine was collected before AA I treatment as the control (a) and once every week after treating subcutaneously with AA I (10 mg/kg/day) for 35 days (b). Urine (0.25 ml) from the untreated and treated rats was mixed with 4 volume of methanol, centrifuged (1,600g) and then evaporated to dryness. The urine sample was dissolved in DMSO (50 μl) and then subjected HPLC system. HPLC condition was described in the “Material and Methods.” The purified AA Ia (c) and a mixture of AA I and AA II (d) were also subjected to the same HPLC system. UV spectrum of AA Ia, AA I and AA II are also shown (e).

The retention time (t_R = 10.3) of a major metabolite was different from that of AA I (t_R = 15.0) and AA II (t_R = 12.8) (Fig. 2d). The UV maxima of this metabolite are at 315 and 398 nm and differ from that of AA I (320 and 394nm) or AA II (302 and 372 nm) (Fig. 2e). From 800 ml of urine collected from AAI-treated rats ~30 mg of this metabolite was isolated using a semipreparative HPLC column. Based on the result using an analytical HPLC column, the purity of this compound obtained was >99% (Fig. 2c); significant impurities including AA I and AA II ere not observed with a detection limit of 5 ng.

Nano-ESI on Quattro LCZ mass spectrometer detected a peak at m/z 345.2 representing as the ammonium form of the molecular ion [M+NH⁴]+ and a peak at m/z 366.0 representing the potassium form of the molecular ion [M+K]+ (Fig. 3b). The ¹H NMR analysis also identified the urine metabolite as AA Ia (Fig. 3a). The methylenedioxy group is characterized by a singlet peak located at 6.38 ppm for two hydrogen. The peaks in the aromatic region are also consistent with the structure as they relate to AA I itself. The singlet at 7.76 ppm is assigned to the hydrogen at the C2-position. The two doublets of doublets and the triplet at 7.06 ppm at C7-position, 8.65 ppm at C5-position and 7.59 ppm at C6 position show an ortho and meta coupling pattern that is characteristic of three contiguous hydrogen on an aromatic where the ring is asymmetrically substituted. The remaining proton absorption at 8.70 ppm represents the hydrogen atem at C9-position. The broad absorption at 9.64 ppm is indicative of a phenolic hydrogen and the extremely broad absorption centered at 11.68 ppm is due to the carboxylic acid proton The absorption of the latter proton is visible only after the spectrum is sufficiently amplified (see insert in Fig. 3a). These data in conjunction with the mass spectra are highly consistent with the proposed structure and is further confirmed by the absence of a singlet (4.04 ppm), characteristic of the hydrogen of the methoxy group in AA I.³⁹

Figure 3.

Determination of AA Ia. The purified AA Ia was characterized by ¹H NMR (a) and mass (b) spectroscopy.

Three other products eluted at 5.3, 8.1 and 8.6 min (Fig. 2b) had UV spectra similar to AA I or AA Ia, indicating to be AA-related metabolites. As the urine collected was not treated with β-glucuronidase/sulfatase, they may be phase II conjugates of AA I and/or AA Ia.

In vitro reactivity of AA Ia with DNA

When calf thymus DNA was incubated at 37°C overnight with AA Ia or AA I in the presence of zinc dust (Table 1), AA I (100 μM) promoted a high level of dA-AL-I (1.41 adducts/10³ nucleotides) and dG-AL-I (1.66 adducts/10³ nucleotides) adducts. In contrast, formation of AA Ia-derived dA- and dG-adducts (3.26 and 1.20 adducts/10⁶ nucleotides, respectively) was at least three order of magnitudes lower, indicating that the in vitro reactivity of AA Ia is much lower than that of the parent AA I.

Table 1.

Formation of AA-DNA adducts in calf thymus DNA exposed to AA I or AA Ia

	AA I adducts/103 dN		AA Ia adducts/106 dN
μM	dA-AL-I	dG-AL-I	dA-AL-Ia	dG-AL-Ia
0	N.D.	N.D.	N.D.	N.D.
10	0.20 ± 0.011	0.36 ± 0.02	1.11 ± 0.09	0.87 ± 0.08
25	0.92 ± 0.05	1.20 ± 0.08	1.39 ± 0.10	0.91 ± 0.07
100	1.41 ± 0.10	1.66 ± 0.12	3.26 ± 0.14	1.20 ± 0.10

¹Data are expressed as mean values ± S.D. from analyses of three samples.

Mouse model of AA nephropathy

To determine the nephrotoxicity of AA Ia, C3H/He mice were treated intraperitoneally with AA Ia or AA I (2.5 mg/ kg/day for 9 days) and were euthanized on the 1st day (day 10) and 15th day (day 24) after ending the treatment. Control mice were treated only with vehicle. One kidney of each mouse was used for histologic examination and the contralateral kidney was used for analysis of DNA adducts. The kidneys of all AA I-treated mice appeared pale at day 10, compared with the controls. No differences in appearance were observed between the kidneys of AA Ia-treated and control mice. To establish the renal toxicity of AA Ia, kidneys collected at day 10 or 24 from mice treated with AA Ia or AA I were subjected to pathologic examination. While AA I was found to cause acute tubular necrosis, extensive cortical interstitial fibrosis, mild interstitial inflammation and occasional tubular apoptosis (Table 2) no significant histological differences between the control and AA Ia-treated kidneys were observed. Thus, AA Ia administration is not significantly associated with the nephrotoxic effects observed after AA I administration.

Table 2.

Pathologic score of mice treated with AA I or AA Ia

Compound	Treatment	Total score1
Control		1.8 ± 1.1
AA I	Day 10	15.0 ± 2.1
	Day 24	18.8 ± 6.0
AA Ia	Day 10	0 ± 0
	Day 24	0 ± 0

¹Data are expressed as mean values ± S.D. based on analyses of three mice. The grading system used for classifying changes in renal pathology is described in “Material and Methods.”

Formation of AA-derived DNA adducts in mice treated with AA Ia

To investigate DNA damage induced by AA Ia, kidneys were collected at day 10 and 24 and DNA adduct formation in the cortex was determined using ³²P-postlabeling/PAGE analysis and compared with the results obtained from AA I-treated mice. dA-AL-I (10.8 ± 2.6 adducts/10⁶ nucleotides for day 10 and 6.6 ± 0.6 adducts/10⁶ nucleotides for day 24) and dG-AL-I (3.4 ± 1.0 adducts/10⁶ nucleotides for day 10 and 1.2 ± 0.1 adducts/10⁶ nucleotides for day 24) were detected in the cortex of AA I-treated mice (Fig. 4). However, the level of AA Ia-derived dA (0.15 ± 0.01 adducts/10⁶ nucleotides for day 10 and 0.06 ± 0.02 adducts/10⁶ nucleotides for day 24) and AA Ia-derived dG adducts (0.04 ± 0.01 adducts/ 10⁶ nucleotides for day 10) were at least 70 times lower than that observed with AA I-treated mice. Especially, no dG adduct was detected in AA Ia-treated mice at day 24. As the migration of AA II (a desmethoxy form of AA I)-derived DNA adducts on the gel was not significantly different from that of AA I-DNA adducts,^22,32 AA Ia-derived adducts were expected to migrate similar to those of AA I- or AA II-DNA adducts. Thus, unlike AA I, AA Ia induced only a limited amount of DNA adduct in kidneys.

Figure 4.

Detection of AA-DNA adducts in C3H/He mice using ³²P-postlabeling/PAGE. C3H/He mice were treated i.p. with AA I or AA II at a dose of 2.5 mg/kg/day for 9 days. Tissues were collected 1 day (indicated as day 10) or 15 days (indicated as day 24) after the final treatment, and the renal cortex DNA was extracted. DNA adduct level was determined using ³²P-postlabeling/PAGE. Standards represent 5 lg of calf thymus DNA (1.52 × 10⁷ fmol dNs) containing a known amount (1.52 × 10² fmol dNs) of dA-AL-I- or dG-AL-I-modified oligomer, representing 1 adduct/10⁵ dNs.

Discussion

To determine genotoxic and nephrotoxic potential of AA Ia, a major metabolite of AA I, several milligrams of AA Ia was required for animal studies. Several synthetic approaches have been attempted in our laboratory to demethylate the Omethoxy group of AA I without success. As AA Ia is a major metabolite of AA I and excreted in urine of rodents,¹⁶ we have therefore attempted to recover AA Ia from a large volume of urine from rats treated with AA I. As expected, ~30 mg of pure AA Ia was isolated using an HPLC system and characterized by mass and NMR spectroscopy. Such an amount was sufficient for determining the nephrotoxic and genotoxic potential of AA Ia using both an in vitro experimental system and an AAN-mouse model used previously in our laboratory.²²

As AA Ia is simply O-demethylated form of AA I, AA Ia is expected to undergo a phase 1 metabolic reductive pathway similar to that observed with AA I, producing the lactam form (AL Ia) through the active nitrenium intermediate that reacts with DNA,¹⁶ as proposed previously by other researchers.39 To determine the chemical reactivity of AA Ia to DNA, AA Ia was incubated with calf thymus DNA in the presence of zinc dust that can catalyze the production of the reactive nitrenium ion. Surprisingly, the reactivity of AA Ia with DNA was at least three orders of magnitude lower than that observed with the parent AA I (Table 1). AA Ia may be expected to expel the proton on the 8-hydroxy group before any reaction could occur with an external nucleophile. Although the nitrenium ion intermediate produced by AA Ia would still be electrophilic in nature, it would be much less reactive than that of AA I. Thus, the 8-hydroxyl moiety formed by demethylation of AA I may weaken the relative electrophobic nature of the nitrenium ion produced from AA Ia, resulting in the lower reactivity with DNA. Therefore, such lower reactivity of AA Ia was expected to appear in animal studies.

To explore the relative genotoxic and nephrotoxic potentials of AA Ia, we employed an AAN-mouse model.²¹ AA-sensitive C3H/He mice were treated with AA Ia or AA I. Unlike AA I-treated mice,²² the body weight of AA Ia-treated mice were not significantly decreased during the experiment; the growth was similar to that observed with the untreated controls (data not shown). We used a quantitative ³²P-postlabeling/ PAGE assay and external standards for analysis of DNA adducts.³² AA I promoted a high level of AA-derived DNA adducts in renal cortex as reported earlier.^22,32 In contrast, the formation of AA Ia-DNA adducts was at least two orders of magnitude lower than that observed with AA I, indicating that contribution of AA Ia to AA I-induced genotoxicity is minimal. The mechanism proposed for the in vitro experiment may be involved in reducing the formation of AA Ia-DNA adducts in vivo. In addition, since AL I formed through the AA I nitrenium intermediate has shown a limited genotoxic potential in rats,³² AA I, not its metabolites, may be solely responsible for forming renal DNA adducts.

AA I also caused acute tubular necrosis and extensive cortical interstitial fibrosis in addition to mild interstitial inflammation and occasional tubular apoptosis as observed earlier.²² No differences in appearance were observed between the kidneys of AA Ia-treated and control mice, indicating that AA Ia is less nephrotoxic. AA Ia, a major metabolite of AA I, may not be involved in nephrotoxicity and genotoxicity associated with AA I. The O-demethylation of AA I to AA Ia may be a major metabolic pathway for detoxifying AA I. We conclude that AA I, acting through its reductively activated nitrenium intermediate, could react with a specific protein to irreversibly inhibit a pathway critical to the function of renal proximal tubule cells, resulting in cell death and fibrogenesis. In conclusion, AA Ia, a major metabolite of AA I, has less genotoxic potential and less nephrotoxicity in mice, indicating that 8-demethylation of AA I is a major detoxification metabolic pathway. AA I is the chemical species responsible for the profound nephrotoxic effects of AA I.