Spontaneous Renal Tubular Hyperplastic and Neoplastic Lesions in Three Sprague-Dawley Rats from a 90-Day Toxicity Study

Abstract

Multiple renal tubular cell adenomas and atypical tubular hyperplasia were diagnosed in 2 high-dose and 1 mid-dose female Sprague–Dawley (Crl:CD®(SD)IGS BR) rats from a 90-day toxicity study of an amino acid found in green tea. The tumors were bilateral multicentric adenomas accompanied by atypical foci of renal tubular hyperplasia in both kidneys of the 3 animals. Toxic tubular changes that typically accompany renal carcinogenesis were not seen in any of the other animals of the study, suggesting rather, an underlying germline mutation of a tumor suppressor gene in these three rats. The histological appearance of these tumors and short latency was reminiscent of the spontaneous lesions reported to arise in Sprague–Dawley rats in the Nihon rat model. Nihon rats develop kidney tumors as a result of a spontaneous mutation in the rat homologue of the Birt-Hogg-Dubé gene (Bhd). Frozen samples of liver from two tumor-bearing rats were assayed for germline alterations in the Bhd gene. The entire coding region (exons 3–13) of the Bhd gene was sequenced, and a guanine (nt106^G) to adenine (nt106^A) polymorphism was detected resulting in a glycine to arginine (G36R) substitution in both tumor-bearing animals. In the study animals, the frequency of the A-allele (adenine) was determined to be 27% (19/70). Interestingly, rats obtained from two other sources (n = 17) only carried the nt106^G-allele, consistent with the published rat sequence for this gene. Genetic fingerprinting of microsatellite loci indicated that the rats had a shared genetic background. Laser capture microdissection (LCM) of the tumor cells demonstrated a loss of heterozygosity in the Bhd gene in neoplastic cells of one of the two animals. Taken together, these data suggest that the tumors observed in these animals arose spontaneously as a result of a shared genetic susceptibility leading to the development of renal tubular neoplasms.

Introduction

Renal tubular cell neoplasms are very rare spontaneous tumors of the S-D rat, having an incidence of approximately 0.1% in a 2-year study for each sex (Giknis and Clifford, 2004). The neoplasms are considered to be slow growing, and even in carcinogenesis studies using high doses of potent renal carcinogens, renal cell neoplasms are seldom seen before 9–12 months following onset of dosing (Hard et al., 1994). The kidney is also a high-frequency target for chemical carcinogens (Gold et al., 2001), with degenerative and proliferative renal tubular changes characteristically observed in most animals administered high doses of a genotoxic chemical, and in all animals receiving a nongenotoxic carcinogen (Khan and Alden, 2002).

In recent years, there have been several reports describing the occurrence of renal tubular cell tumors in just a few animals of a study of S-D or F344 rats unrelated to any corresponding renal cell toxic or proliferative lesions (Hard et al., 1994; Thurman et al., 1995; Savard et al., 2005). In some of the studies, tumors arose as early as 90 days (Hard et al., 1994). In consequence to these observations, an alternative explanation has been put forth for the occurrence of renal cell neoplasia and hyperplasia in just a few rats of a study, in the absence of any toxic/proliferative changes in the animals. In these cases, it appears likely that an underlying susceptibility to spontaneous kidney tumors was present as a result of a predisposing genetic alteration in affected rats. In the case of multiple affected animals in a study, this would presuppose that the affected animals are related, having inherited an otherwise extremely rare spontaneous mutation from a common ancestor, a finding documented in one of the studies (Thurman et al., 1995).

Such an underlying genetic susceptibility would most likely occur in a tumor suppressor gene, and to date, there have been two examples of spontaneously occurring tumor suppressor gene defects in the rat that predispose to spontaneous renal tumors. The tuberous sclerosis complex 2 gene (Tsc-2) originally described in a Long-Evans rat (Eker and Mossige, 1961; Walker et al., 1992; Yeung et al., 1994; Kobayashi et al., 1995), and the Birt-Hogg-Dubé (Bhd) gene described in a Sprague–Dawley rat from Japan (Nihon rat) (Hino et al., 1991; Okimoto et al. 2000, 2004; Kouchi et al. 2006).

Germline mutations in either of these 2 tumor suppressor genes predispose the rats to multicentric, bilateral renal tubular neoplasia and hyperplasia. In the Eker rat model, tumors have a later onset generally, and are vacuolated and chromophilic, whereas, neoplasms in rats with the Bhd mutation are of early onset (3–4 months with complete penetrance by 8 months; hyperplastic lesions occur within a few weeks) and have a variable clear cell, cystic, papillary, chromophobe, or eosinophilic appearance (Okimoto et al. 2000; Kouchi et al. 2006).

L-theanine, gamma-glutamylethylamide, is an amino acid found in green tea that makes up approximately 1–2% of the dry weight of the green tea leaves. The amino acid is metabolized by the kidney to form glutamic acid and ethylamide (Unno et al., 1999). Worldwide tea consumption is second only to water (Segal, 1996), and there is evidence suggesting that theanine is protective against carcinogenesis. Theanine appears to competitively inhibit glutamate transport into tumor cells, but not normal cells, which results in decreased intracellular glutathione (GSH) levels (Sadzuka et al., 1996, 2001; Sugiyama and Sadzuka, 2003, 2004; Anonymous, 2005). Theanine also reduces the efflux of chemotherapeutic agents, such as doxorubicin, idarubicin, cisplatin, and irinotecan from tumor cells, causing prolonged retention of these compounds in the neoplastic cells (Anonymous, 2005).

Thus, in experimental animals, this component of green tea enhances the effects of various cancer chemotheraputic drugs. It also protects normal cells from damage by these drugs via antioxidant activity, specifically by maintaining cellular GSH levels in normal but not in neoplastic cells (Sugiyama et al., 1999; Sadzuka et al., 2001; Sugiyama and Sadzuka, 2003; Anonymous, 2005). In addition, ascites hepatoma cell line AH109A tumor growth was reduced in rats fed diet containing theanine (Zhang et al., 2002). Finally, an unpublished 78-week chronic toxicity study in B6C3F1 mice fed L-theanine at the maximal tolerated dose of 5% and one other dose level (3%) revealed reduced numbers of lung and liver neoplasms in theanine-fed mice vs. controls (personal communication, Taiyo International, Inc. [Fujii et al., Hiroshima University, 1999]). All other tumors were similar in incidence among the groups. Thus, published and unpublished data document anti-neoplastic effects of theanine.

In this report we describe the occurrence of renal cell tumors in 3 animals of a 90-day toxicity study in S–D rats. The tumors were found to arise in the absence of any other observed toxicity, morphologically resembled those previously observed in the Nihon rat model, and a germline amino acid substitution in the rat Bhd gene was detected. The Tsc2 gene was not examined. Loss of heterozygosity was detected in the Bhd gene of tumor cells from 1 of the 2 rats that were analyzed. Moreover, microsatellite loci analysis of the affected rats found that the animals were genetically related. Taken together, these data suggest that the proliferative lesions observed in these animals arose spontaneously as a result of a shared genetic susceptibility resulting in mutation of the Bhd gene and development of renal tumors.

Materials and Methods

Study Design

In order to assess toxicity of an amino acid found in green tea, male and female Crl:CD®(SD)IGS BR rats were assigned to four toxicity groups (20 animals/sex/group). Each group received diets containing 0 (basal diet) or a targeted dose of 1500, 3000, or 4000 mg/kg of L-theanine/kg of body weight/day (mg/kg/day). Approximately half the animals from each group and sex were sacrificed after 94 days of treatment, and because of equipment failure on ancillary studies unrelated to toxicity, the other half of the animals were sacrificed following 103 days of treatment. Assessment of toxicity was based on mortality, clinical observations, expanded clinical observations, motor activity, clinical pathology, anatomic pathology, and toxicokinetic evaluations. Following completion of the study, the animals were sacrificed by carbon dioxide inhalation and necropsied. Selected organs were weighed and fixed in 10% neutral-buffered formalin. Samples of liver and brain from 10 high-dose and 10 control rats of each sex from both sacrifices were frozen for possible proteomic and genomic studies. Tissues from all high-dose and control animals were processed through paraffin, sectioned at 5 μm, stained with hematoxylin and eosin, and examined microscopically. Only relevant portions of the pathology and ancillary additional studies are reported herein.

Step Kidney Sectioning

Following microscopic examination of the tissues, renal tubular adenomas and tubular hyperplasia were diagnosed in 2 high-dose female rats (rat 334 necropsied on day 94 and rat 293 necropsied on day 103), and fixed kidney samples from the remainder of the rats of the toxicity study were processed, stained and examined microscopically. An additional female animal with bilateral renal tubular hyperplasia was detected in the mid-dose group (rat 316 necropsied on day 94).

Except for background lesions attributable to chronic progressive nephropathy of spontaneous origin with similar incidence in all dose groups (Montgomery and Seely 1990; Khan and Alden 2002), renal tubular degenerative and proliferative changes were not seen in any of the other animals of the study. In order to assure that this was the case, the remainder of the renal tissue from all female rats was step-sectioned and examined for tubular cell proliferative and degenerative changes according to recommendations of the National Toxicology Program (NTP) (Eustis et al., 1994).

Analysis of the Bhd Gene

DNA Isolation

Frozen liver tissue from the 2 high-dose female rats with renal adenomas were pulverized using a mortar and pestle under liquid nitrogen, and genomic DNA isolated using a commercially available kit (Wizard Promega). Frozen tissue was not available from the mid-dose rat and it could not be examined; paraffin-embedded liver from this animal had been retained in formalin for several months and DNA was too extensively degraded to be analyzed. Exon 3 of the Bhd gene was initially chosen for analysis, as a hypermutable polypyrimidine tract is located in this exon (Okimoto et al., 2004). This hypermutable region was previously identified as the site of the Bhd mutation in Nihon rats (Okimoto et al., 2004). DNA was isolated from frozen liver samples of 35 additional high-dose and control study animals of both sexes for comparison with the renal tumor-bearing animals. In addition, DNA samples from ear clips of 17 rats from another derivation (NIH x Harlan Sprague–Dawley and Long–Evans from Science Park Research Division, MD Anderson Cancer Center) were compared with the rats of this study.

DNA Amplification, Purification, and Sequence Analysis

PCR primers for genomic sequencing of the rat Bhd gene have been previously reported [see Table 1 (Okimoto et al., 2004)]. PCR products were amplified, confirmed to be of the correct size by gel electrophoresis, and, purified with a commercially available kit (Qiaquick). Sequencing was performed using BigDye Terminator Version 1.1 Cycle Sequencing (ABI). For exon 3 containing the GC-rich hypermutable region, PCR sequencing was initially performed on each DNA sample using 3 different chemistries to obtain accurate sequence results through this region. All 3 sequencing chemistries yielded identical results for any single sample. For all other regions, sequence results were obtained in both directions, and in the case of exon 3, in triplicate for each sample. The Tsc2 gene was not examined.

Laser Capture Microdissection

Sections (6 μm) were taken from the paraffin blocks of the kidney tumors, dewaxed through xylene and graded ethanol to water, stained with eosin, desiccated through ethanol and finally immersed in xylene. Sections were examined by the pathologist (WCH) and tumor cells were removed from the slide for PCR analysis using a PixCell II (Arcturus Bioscience, Inc., Mountain View, CA) laser capture microdissection system. The amplitude and pulse duration of the PixCell II laser were adjusted as necessary to allow complete capture in a 30 μm laser beam. After microdissection of each section, the CapSure LCM caps (Arcturus Bioscience, Inc.) containing the captured tissue were placed in a microtube and DNA isolation was performed from sequential sections of each sample.

Microsatellite Loci Analysis

Known Crl:CD®(SD)IGS BR sibling rats and known outbred unrelated CD rats were tested with the CRL Genetic Testing Services rat background strain characterization panel of 107 markers and compared with the 2 CD rat samples (C99293 and C99334) from the high-dose study animals with renal tumors. The purpose was to determine whether the 2 rats of the study were related genetically. The microsatellite markers are spread across the 20 rat autosomes and the X chromosome at approximately 15 centiMorgan (cM) intervals. Lewis, Brown Norway, and CD controls (siblings and nonsiblings) were also run with the 2 samples.

DNA was prepared from a tail sample of CRL colony CD rats (siblings and nonsiblings) using the Clontech Nucleospin Tissue Kit. The rats were taken from the same colony and room as those rats of the study. The DNA quantity and quality was verified on a 1% agarose gel. The DNA was diluted to 40 ng/μL with distilled water (dH₂O, Gibco, DNase and RNase free). DNA samples from the liver of the 2 high-dose study rats (C99293 and C99334) had been prepared by MD Anderson Cancer Center in a similar fashion, and sent to CRL for microsatellite analysis. The DNA of each sample was then used as a template in 107 separate polymerase chain reactions (PCR), each containing a primer pair to amplify a particular microsatellite marker. Each reaction was performed in a volume of 20 μL, using 1.1 μL of template DNA, forward and reverse primers at a concentration of 5 μM, 10 μM dNTPs (Bioline), the MasterTaq kit (Eppendorf), and Cresol Red (Rediload) as a loading dye. The reaction conditions were as follows: 1 cycle of 2 minutes at 95°C; 40 cycles of 45 seconds at 95°C, 45 seconds at 56°C, and 1 minute at 72°C; 1 cycle of 5 minutes at 72°C.

A 7 μL aliquot of each reaction was electrophoresed on a 3% SFR Agarose (Amresco)/1x TBE gel containing ethidium bromide along with a 50 base pair DNA ladder as a size standard. Photographs of the reaction products were taken with an AlphaImager (Alpha Innotech), and alleles present in each sample were recorded on an Excel spreadsheet.

Results

Clinical and Clinical Pathology Observations

The complete toxicology report is published separately (Borzelleca et al., 2006). Treatment-related mortality was not observed. One female given 3000 mg/kg/day died on Day 14, and 1 male given 1500 mg/kg/day died on day 35, both from acute inflammation and necrosis of the urinary bladder (cystitis). On day 30, 2 females given 4000 mg/kg/day died immediately following blood collection for clinical pathology. On day 57, a control male was euthanatized due to a severe malocclusion causing respiratory distress. None of these unscheduled deaths were considered a result of test article administration; the remaining animals survived to the scheduled terminal or recovery sacrifice.

Food palatability was decreased and food consumption was reduced in dose-related fashion in dosed males and high-dose females. Weight gains were reduced accordingly, and were not considered a toxic change because of the correlation with food consumption. The kidney-to-body weight percentages were significantly increased in males given 1500, 3000, or 4000 mg/kg/day and sacrificed at day 93 or day 103. The absolute mean kidney weights, the kidney-to-body weight ratios, and the kidney-to-brain weight ratios were significantly increased in females given 4000 mg/kg/day and sacrificed on day 93 or day 103. The kidney-to-body weight ratios were increased in females given 3000 mg/kg/day and sacrificed on day 103, and the absolute kidney weights and the kidney-to-brain weight ratios were significantly increased in females given 1500 mg/kg/day and sacrificed on day 93. Some of these alterations may be related to reduced terminal body weights. There were no histological correlations to these weight changes.

Dose-related clinical pathology changes were not observed. Clinical pathology observations were limited to mildly higher cholesterol for females given 3000 or 4000 mg/kg/day, mildly lower total protein for females given 4000 mg/kg/day, and mildly higher urine pH for males given 3000 or 4000 mg/kg/day. None of these relatively minor findings were considered adverse or indicative of significant target organ toxicity.

Pathology of Lesions That Developed in 90-day Study Animals

Only incidental gross observations were found. Body weights were significantly less in both females and males at the high and mid dose levels. This was attributed to reduced food consumption caused by poor palatability of the diet. Decreased food consumption accompanied by a lower body weight gain was not considered a toxic response.

Renal tubular cell adenomas and renal tubular cell hyperplasia were seen in two high dose females (4000 mg/kg/day); renal tubular cell hyperplasia was seen in one mid-dose female (3000 mg/kg/day) after standard single sectioning. Following step-sectioning of the remainder of the kidney tissues from these animals (and all other females), the tumors were multiple and bilateral in 1 of the 2 high-dose animals, and multiple and bilateral in the mid-dose female (Table 2). A total of 10 adenomas were seen in all 3 rats, 9 in the cortex and 1 in the outer medulla (Table 3). All tumors of the 2 rats were small, 2 mm or less in diameter. Multiple bilateral foci of tubular cell hyperplasia (N = 62) were seen in the multiple sections of all 3 animals. They were evenly distributed in the cortex and outer stripe of the outer medulla (Table 3). Tubular proliferative changes were not seen in any other animal of the study on the standard or step sections, including the remainder of the high-dose and mid-dose females, and all of the low dose and control females.

Microscopically, the renal tubular neoplasms were expansive, generally cystic, circumscribed growths composed of large, palely eosinophilic, finely granular cytoplasm in polyhedral epithelial cells that formed papillary structures (Figures 1a–1d). The one neoplasm found in the outer medulla had a basophilic appearance; a portion of 1 of the cystic adenomas also had a basophilic appearance. Fewer than 10% of the tumor cells and an occasional hyperplastic focus were vacuolated. The papillae were blunt and supported by a fine fibrovascular stalk infiltrated by inflammatory round cells and neutrophils in some of the tumors. Nuclei in the tumor cells were 1.5–3 times larger than those of the adjacent normal tubular epithelial cells, and generally contained one or two prominent amphophilic nucleoli. Mitotic figures were rare (0–2/high magnification), but observed in greater frequency than normal tubular epithelium.

Other neoplasms had a solid appearance (Figure 1e), and there was occasional mineralization. Renal tubular cell hyperplasia consisted of multiple bilateral foci of slightly basophilic cells with granular cytoplasm and occasionally, weakly acidophilic cells; the latter were seen in the cortex. A brush border could be detected on some of the cells. Many had an atypical appearance (Table 3), appeared to progress from simple tubular hyperplasia to atypical hyperplasia, and were considered precursor lesions of the neoplasms (Figures 2a, 2b). These lesions clearly differed from those of chronic progressive nephropathy (Hard et al., 1995). Clear cells were not observed.

Salivary gland, cardiac, uterine, and vascular lesions were not observed in these rats.

Analysis of the Rat Bhd Gene

Initially, exon 3 containing the hypermutable region of the rat Bhd gene was chosen for analysis, as the polypyrimidine tract located in this exon is a “hot-spot” for mutations. The 2 high-dose tumor-bearing rats were found to have a G to A base pair change at position 106 in the coding region of the Bhd gene. As a result, tumor-bearing rats were heterozygous at codon 36, with the nt106^G allele coding for the predicted glycine at this position and the nt106^A allele coding for an arginine at this position (G36R). In addition to the rats that developed tumors, DNA from a total of 35 nontumor-bearing rats from the high-dose (10 males, 6 females) and control (9 males, 10 females) animals on study of the same derivation was examined.

Of these 35 rats, all of which were from the same breeder source, the frequency of the nt106^A allele was 27% and the distribution of homo- and heterozygotes was G/G = 18, G/A = 15, and A/A = 2. Seventeen additional Sprague–Dawley rats from an alternative source were examined, all of which had only the reported nt106^G at this position, as did an additional Long–Evans rat that was sequenced. Thus, the observed G106A base pair substitution was found only in Sprague–Dawley study animals obtained from a single breeder, and appeared to be polymorphic within this population. No differences in renal histopathology were observed among any of the study rats with a G/A, A/A or G/G genotype.

In addition to the G106A base pair substitution observed in exon 3, two additional deviations from the reported rat Bhd sequence were found in intron 5. Both tumor-bearing rats were homozygous for a T to G base pair substitution at position 2702 from the ATG start site. In addition, at base pair 2790, a G to A substitution was found, with one tumor-bearing rat being heterozygous (G/A) at this position and the other being homozygous (A/A). Of the other 25 study animals sequenced, all were homozygous G/G at 2702 and homozygous A/A at 2790.

The presence of heterozygosity at the Bhd locus made it possible to examine tumors from these animals to determine if additional alterations, such as loss of heterozygosity (LOH) had occurred in this gene, implicating it in tumorigenesis. Laser capture microdissection (Figures 3a and 3b), followed by PCR analysis of the Bhd gene from tumor cells, revealed that LOH had occurred at the Bhd locus in 1 of the 2 animals analyzed (Figure 4). In this animal, which was heterozygous for both the exon 3 (G106A) and intron 5 (G2790A) polymorphisms, reduction to homozygosity was observed at both nt106 and nt2790, strongly suggesting involvement of the Bhd gene in tumor development in these animals.

Genetic Fingerprint Analysis

Inheritance of a common susceptibility to tumorigenesis presupposes that the susceptible rats share a common genetic background. Sibling identity of the rats in the study was not retained, and thus, the genetic relationship of these animals could not be determined from source information. Therefore genetic relatedness of the rats was determined by performing genetic fingerprint analysis utilizing microsatellite loci from the frozen liver samples of the two high dose animals (DNA was not available from the third animal). In this assay, 107 regions of DNA from all chromosomes of each animal were analyzed using published primers; similarities were assessed and compared with tail sample DNA from known siblings of the same litter as well as other colony animals that were not related.

As predicted, little diversity (increased incidence of homozygosity) was observed between the known siblings and wide diversity (low incidence of homozygosity) was observed in the unrelated animals (Table 4). The 2 study animals that developed tumors showed less diversity than nonrelated rats (38% at 103 loci versus 10% at 98 loci respectively), pointing to a shared genetic background. These data indicated that the 2 rats that developed tumors were genetically related, consistent with these animals having a shared genetic susceptibility for renal tumorigenesis.

Discussion

Few clinical and pathological observations were seen in animals exposed to the nongenotoxic amino acid L-theanine found in green tea. The compound is metabolized by the microsomal fraction (Curthoys and Kuhlenschmidt, 1975) of renal tubules (Tsuge et al., 2003), and expectedly, with the high doses administered, relative and absolute increases in kidney weight were observed. However, no changes reflective of the weights were observed microscopically. Renal tubular cell adenomas and atypical tubular cell hyperplasia were seen in 3 of 80 female and 0 of 80 male rats of this study. The neoplasms were bilateral in 2 of the 3 animals; multiple bilateral foci of atypical tubular cell hyperplasia were seen in these same 3 animals, but not in any others. Degenerative or proliferative changes were not seen in renal tubules of any of the other rats of the study, even at the high-dose level. These observations were verified by performing multiple sections through the entire kidney of each of the females, consistent with confirmatory procedures used by the NTP to assess marginal renal carcinogens (Eustis et al., 1994).

The observed lack of proliferative or degenerative changes in dosed animals was not consistent with exposure to a carcinogenic insult by a nephrotoxicant, suggesting other factors, such as an underlying genetic susceptibility may have been involved. Renal tubular degenerative and proliferative changes would be expected with both genotoxic and nongenotoxic carcinogens at the high-dose levels administered to the animals of this study (Bannasch and Zerban, 1999; Khan and Alden, 2002). Moreover, renal tubular tumors induced in the rat by potent renal carcinogens generally have a latent period of 9–12 months before neoplasms are seen, and longer than that with nongenotoxic carcinogens. Even with one of the most potent renal genotoxic carcinogens (dimethylnitrosamine [DMN]), atypical renal tubular hyperplasia is observed at 3 months; rarely, a kidney tumor can be seen at 4 months of dosing, but generally, neoplasms are not seen until after 6 months after dosing (Hard and Butler, 1971). In this study, the animals were dosed for approximately 3 months. Thus, the lack of tubular changes in the dosed rats, except for the 3 animals of concern in this study is highly unusual, and not consistent with either genotoxic or nongenotoxic renal carcinogens.

Other studies performed on L-theanine also support these conclusions: (1) 2 different Ames assays were negative for genotoxicity; (2) no compound-related findings were observed in 2 different 28-day studies in rats at high-dose levels; and (3) there was reduced incidence of neoplasia in a 78-week carcinogenesis assay in B6C3F1 mice dosed at dietary levels of 5% and 2.5% (unpublished studies).

The early onset of the renal cell neoplasms without corresponding degenerative and proliferative tubular changes in non-tumor-bearing animals is most consistent with what is observed for spontaneous tumors that arise in genetically susceptible animals (Everitt et al., 1992, 1995; Okimoto et al., 2000; Hino et al., 2003). The histopathology of the majority of the neoplasms in the rats of this study (papillary cystadenomas, eosinophilic adenomas) is similar to features of the neoplasms described for both the Eker rat (with a germline mutation in the Tsc-2 gene) (Wolf et al., 1995), and the Nihon rat (with germline mutation in the Bhd gene) (Everitt et al., 1992, 1995; Okimoto et al., 2000; Hino et al., 2003; Kouchi et al., 2006), although the vacuolated appearance as described in both the Eker and Nihon rat tumors was seen only in approximately 10% of the cells, and clear cells were not seen. All the renal neoplasms seen in the 3 rats of this study were 2 mm or less in diameter. However, an N of 3 is a small number for morphologic comparison with either model.

Other lesions as reported in the Nihon rat such as cardiac rhabdomyomatosis, clear cell appearance of the mandibular salivary gland ducts, clear cells in the uterine endometrium, and hypertrophic arterial changes (Kouchi et al., 2006), were not seen in the 3 rats of this study.

In this study, liver samples from the high-dose and control animals were frozen for possible gene or protein analysis. This was fortuitous, as it provided unfixed samples for genetic analysis, which permitted the determination both of genetic fingerprinting and Bhd gene analysis. Liver DNA was analyzed by PCR for germline mutations in the Bhd gene, with an initial focus on exon 3, which contains a polypyrimidine hypermutable region of this gene. Although no mutations were observed in this polypyrimidine tract, a polymorphism at nt106 in the Bhd gene was identified. Both heterozygosity and homozygosity was observed in the 37 study animals examined (2 tumor-bearing rats and 35 additional study animals), the distribution being G/G = 18, G/A = 17 and A/A = 2.

In the Nihon rat, renal cell neoplasms begin as preneoplastic lesions (focal tubular hyperplasia) as early as 3 weeks, progress to adenomas at 8 weeks, with complete penetrance of renal cell carcinoma by 6 months (Hino et al., 1991). In this 90-day study, neoplasms were observed in animals that were approximately 20–21 weeks old at study termination, consistent with the short tumor latency of Nihon rats carrying a defect in the Bhd gene. In line with Knudson’s “2-hit hypothesis” the finding of a loss of heterozygosity of the Bhd gene in the tumor from 1 of the rats (the other remained heterozygous) strongly suggests that this tumor suppressor gene was involved in the development of this lesion (Knudson, 1985). After identifying this polymorphism in both tumor-bearing and non-tumor-bearing animals in the study population, multiple kidney sections from non-tumor-bearing rats carrying the nt106^A allele were re-examined; neither renal neoplasms or nor tubular hyperplasias were observed in any of these animals.

Given their shared genetic background, why were renal tubular cell tumors seen in just three animals? One possibility is that G106A polymorphism simply serves as a marker of the shared genetic background of these animals, and that another genetic factor(s) present in the affected animals was responsible for the susceptibility for development of renal tumors. Sprague–Dawley rats are an outbred line, and other genes, which could act as modifiers of tumor susceptibility may also be segregating within this population, alone or in combination with the G106A polymorphism. Even in genetically susceptible animals, tumor development is stochastic, with multiple genetic events, both inherited and acquired, required for tumor development. Not all inherited tumor susceptibility genes are high penetrance, and in many cases, such as for metabolizing enzyme polymorphisms, these predisposing mutations have a high prevalence but a low penetrance, and result in only a small increase in risk.

The G106A substitution in exon 3 of the Bhd gene was present in high dose animals with tumors, and also in dosed and control animals with no neoplasia and hyperplasia. Although the G106A polymorphism present in the germline of the study animals results in a nonconservative amino acid change (glycine to arginine), it is unclear at this time if this polymorphism has a functional consequence. As the function of the Bhd gene product folliculin is unknown, functional assays to determine if the G106A polymorphism identified in this report has a functional impact on this tumor suppressor gene are not possible at this time. However, in the future, it should be possible to determine whether the G106A substitution has affected the function of this tumor suppressor gene and has contributed to an increased tumor susceptibility in animals carrying this allele.