Rat stomach cancers induced by N-methyl-N′-nitro-N-nitrosoguanidine(MNNG) are widely used as a model of differentiated-type human stomach cancers. ACI/N (ACI) rats are susceptible and BUF/Nac (BUF) rats are resistant to MNNG-induced stomach carcinogenesis, and the presence of an autosomal gene with a dominant BUF allele has been suggested. In this study, we performed a carcinogenicity test by giving MNNG in drinking water to 117 male ACI × (ACI×BUF)F1 backcross rats. Each of 100 effective rats was diagnosed for its “carcinoma development” and when it was bearing stomach carcinoma(s), for histological grade, depth of invasion, and size and number of tumors. Carcinoma development was diagnosed based both on the age of the rat and on the presence of stomach carcinoma(s). Linkage analysis was performed with the genotypes of 161 loci, covering 1637 cM of the rat genome. Contrary to our original expectations, the most influential gene was the one on chromosome(chr.) 15, Gastric cancer susceptibility gene 1(Gcs1), which confers susceptibility to stomach carcinogenesis (LOD, 3.8) with a dominant BUF allele by promoting conversion from adenomas to carcinomas. Two resistance genes on chr. 4 and chr. 3, Gastric cancer resistance gene 1(Gcr1) and Gcr2, were shown to confer dominant resistance (LOD, 2.7 and 2.6, respectively). Gcs1, Gcr1, and Gcr2exerted additive effects on the development of stomach carcinomas. A gene on chr. 16, Gcr3, was indicated to reduce the depth of invasion (LOD, 2.2) and sizes of tumors (LOD, 1.9). No linkage was obtained using the number of tumors. These findings show that the coordinate effect of a susceptibility gene, Gcs1, and two resistance genes, Gcr1 and Gcr2, is responsible for the development of MNNG-induced stomach carcinomas and that Gcr3 is responsible for the growth of a stomach carcinoma, reflected in the depth of invasion and in the tumor size.

Investigations into the genes responsible for cancer predisposition are important. Identification of high-risk populations for a type of cancer can lead to its prevention and its diagnosis and treatment in early stages. For example, the discovery of BRCA1 and BRCA2 made it possible to identify high-risk family members for breast cancers, and genetic tests have been implemented (1). Clarification of the involvement of mismatch repair genes, such as hMSH2 and hMLH1,in hereditary non-polyposis colorectal cancer syndromes has also changed clinical approaches to high-risk patients (2). In the case of stomach cancer, which is the major cause of cancer death in Asian countries, genetic factors have also been implicated (3). Recently, germ-line mutations of the E-cadherin gene were found in families from New Zealand, in which early-onset and highly malignant diffuse-type stomach cancers are clustered (4). However, the gene(s) responsible for predisposition to differentiated-type stomach cancers still remain unknown. Moreover, linkage analysis using human families is facing critical limitations—decreasing numbers of family members and the heterogeneity of responsible genes among families.

The rat stomach carcinoma induced by administration of MNNG3in drinking water provides a good model for human stomach cancer of a differentiated type (5). The histology of stomach carcinomas induced in the model closely resembles that of human differentiated-type stomach carcinomas, and their carcinogenic processes are modulated by various promoting and preventive agents in a way similar to that observed in human populations (6, 7, 8). Differences in the susceptibility to MNNG-induced stomach carcinomas have been known among rat strains (9, 10, 11, 12). Stomach carcinomas are induced in the ACI/N (ACI) strain at a rate of ∼80%,and in the BUF/Nac (BUF) strain at ∼20% (11). The incidences of stomach carcinomas in F1 and F2 rats of ACI and BUF rats indicated involvement of an autosomal gene whose BUF allele is dominant over the ACI allele (11).

As the mechanism of the different susceptibilities, different levels of cell proliferation in response to chronic mucosal damage by MNNG,superimposed with a strong inflammatory reaction, have been implicated (13). Whereas ACI rats show a large increase in the cell proliferation rate after MNNG treatment, BUF rats show only a small increase. It is likely that a high rate of cell proliferation leads to a high rate of mutations, and finally to a high incidence of stomach carcinomas. MNNG is a direct-acting carcinogen, excluding differences in the capacity of metabolic activation as a mechanism for the different cancer susceptibilities between ACI and BUF rats. Moreover,DNA adduct levels in the pylorus of the two strains are the same (5), excluding differences in the local distribution of MNNG as a mechanism. These previous findings suggest that the mechanisms involved in the different susceptibilities between the two strains are not restricted to MNNG-induced stomach carcinogenesis.

To identify the gene(s) involved in the different susceptibility between ACI and BUF rats, where involvement of an autosomal dominant gene of the BUF type has been suggested, we produced 117 male ACI × (ACI×BUF)F1 backcross rats in this study, and performed a carcinogenicity test for 80 weeks.“Carcinoma development” in individual rats and, when a rat was bearing one or more tumors, the histological grade of malignancy, depth of tumor invasion, tumor size, and number of tumors per rat were determined. The results of linkage analysis are reported.

Carcinogenicity Test and Histological Examination.

ACI and BUF rats were purchased from Japan Clea (Tokyo, Japan). F1 rats were produced by mating female ACI rats with male BUF rats, and backcross rats were produced by mating female ACI rats with male F1 rats.

ACI, BUF, and backcross rats were caged in groups of three to four rats and fed Oriental MF diet (Oriental Yeast, Tokyo, Japan). The cages were kept in an air-conditioned room at 25°C and 55% humidity with 12-h light and dark cycles. The rats were administered 83 mg/l MNNG in drinking water ad libitum from the age of 8 weeks through the age of 40 weeks. MNNG was purchased from Sigma-Aldrich (St. Louis,MO), and a fresh 830 mg/l (deionized water) stock solution was prepared once a week. The stock solution was diluted 10-fold with deionized water twice a week. Rats that became moribund or reached 80 weeks of age were sacrificed.

The stomach was opened along the greater curvature, and sections of the glandular stomach were prepared by cutting along the sagittal axis. When a macroscopic tumor was observed, the tumor was cut into two pieces, one of which underwent histological examination and the other was kept frozen. When a macroscopic tumor was absent, slices were made at 2-mm intervals underwent histological examination. Histological examination was performed by an experienced pathologist (M. T.),and when stomach carcinoma was present, its histological grade of malignancy, depth of tumor invasion, maximal diameter, and number of tumors per rat were determined (14). When two or more carcinomas were present in a rat, the most malignant phenotype from one of the carcinomas was adopted as the phenotype of the rat.

The liver and/or tail of each rat was kept frozen, and DNA for genotyping was extracted by serial extraction with phenol and chloroform (15).

Genotyping.

Genotypes of 161 loci, consisting of 146 microsatellite markers, 5 B1-RDA markers (16), and 10 AP-RDA markers (17), were determined for all of the 100 effective rats. Distances between markers summed up to 1637 cM, covering almost the entire genome of the rat. The average of the distances between two markers was 11.6 cM.

Microsatellite primers were synthesized by Sawady Technology(Tokyo, Japan) based on previous reports (18), or purchased from Research Genetics (Huntsville, AZ; Ref. 19). Annealing temperatures of 60°C and 55°C and Mg2+ concentrations of 1.5 and 2.5 mmwere first tested for each pair of primers, and PCR for genotyping was performed in the optimized condition using 25 ng of template DNA. PCR products were run in 3% NuSieve gels or in 6% polyacrylamide gels.

Genotyping with B1-RDA and AP-RDA markers was performed as reported previously (16, 17). “B1-amplicons” and“AP-amplicons” were prepared by amplifying 100 ng of genomic DNA with an appropriate B1- or AP-primer and then dot-blotted onto a nylon membrane after denaturation. The filters were hybridized with B1-RDA or AP-RDA markers, which had been labeled with random hexamers and the Klenow fragment (Multiprime; Amersham-Pharmacia), and purified by gel filtration chromatography (Microspin; Amersham-Pharmacia). Prehybridization, hybridization, wash, and signal detection were performed as reported (16, 17). All of the markers used were selected on the basis that they give signals with BUF but not with ACI.

Linkage Analysis.

A genome map for the backcross rats was calculated with Mapmaker/EXP software (20). A genome-wide survey for QTLs was performed with the Mapmaker/QTL software (20). For the carcinoma development phenotype, a rat was given a digit of 0 or 2, depending on the diagnosis. For the “histological grade” and “depth of tumor invasion” phenotypes, a rat was given scores ranging from 1 to 5. For the “tumor size” phenotype, the maximal diameter of the carcinoma(in mm) in histological sections was used. To assess the effects of the genotypes, χ2 tests and t tests for unequal distribution were performed.

Carcinogenicity in Inbred and Backcross Rats.

The cumulative incidences of the stomach carcinomas in ACI, BUF, and backcross rats are shown in Fig. 1. The first stomach carcinoma in ACI rats was observed at the age of 52 weeks. Therefore, backcross rats that survived 52 weeks or more and whose stomach histology was available were counted as effective animals. Deaths of the ACI and backcross rats before 52 weeks were caused mainly by hydronephrosis and resultant renal failure, which is known to take place in ACI rats at an incidence of 5–20% (21, 22). Deaths of the ACI, BUF, and backcross rats after 52 weeks were due to stomach carcinomas, duodenal sarcomas, pneumonia, and other causes. Incidences of duodenal sarcomas are known to be in the same range in ACI and BUF rats (11). Similar to previous studies (10, 11), stomach carcinomas were observed in 16 of 21 effective ACI rats (76%) and 3 of 22 effective BUF rats (14%). In the backcross rats, 48 of 100 effective rats (48%) developed stomach carcinomas (Table 1).

Linkage with Carcinoma Development.

To diagnose the carcinoma development phenotype of the backcross rats, the rats were classified by the size of stomach carcinomas that they bore and by their age when they were sacrificed (Table 1). The diagnosis on their susceptibility was made by mimicking the sizes and ages observed in inbred strains. Because it was impossible to make a definitive classification before performing linkage analyses, we tested 12 combinations of three criteria for the susceptible rats (S1, S2, and S3) and four criteria for the resistant rats (R1, R2, R3, and R4), as shown in Table 1.

For the rats that survived the full-term of the carcinogenicity test (79–80 weeks of age), rats with stomach carcinomas of a large size could be diagnosed as “susceptible” and those without a carcinoma could be diagnosed as “resistant.” The diameter of the smallest carcinoma in the ACI rats at this period was 5 mm in this study. Therefore, a backcross rat with a carcinoma 3 mm in diameter was not classified as susceptible, and those with carcinomas 4 mm in diameter were classified as susceptible in one criterion (S3) and not in the others (S1, S2). For the rats that were sacrificed at early periods of the carcinogenicity test, only those with a stomach carcinoma could be diagnosed as susceptible, but those without a carcinoma could not be diagnosed as susceptible or resistant because they might have developed a stomach carcinoma if they had survived the full-term. To explore what age was appropriate to mark “early periods,” we tested different criteria for resistant rats (R1–R4).

Linkage analysis was performed using these 12 combinations of the carcinoma development phenotype, and three loci were found to give LOD scores >2.0 with at least one of the 12 combinations (Table 2 and Fig. 2). Because criteria S1, S2, and S3 gave similar LOD score curves,results with S3 and R1-R4 are shown in Fig. 2. The most influential gene, Gcs1, was mapped on chr. 15 with the highest LOD score of 3.8 (criterion R1S3; n = 69), and was found to have a paradoxical effect on the development of stomach carcinomas. Despite the fact that the BUF rat is resistant as an inbred strain, the BUF allele at Gcs1conferred dominant susceptibility to MNNG-induced stomach carcinogenesis. Whereas 13 of 30 rats with the AA genotype(43%) at Gcs1 (represented by D15Rat102)developed stomach carcinomas, 34 of 39 rats with the ABgenotype (87%) developed carcinomas (P = 0.0001; Table 3).

Two genes, Gcr1 and Gcr2, were mapped on chr. 4 and chr. 3 with peak LOD scores of 2.7 and 2.6, respectively,using criterion R3S3. The BUF alleles at these two loci were shown to confer dominant resistance. For Gcr1 (represented by the marker, Ampp; Table 3), 33 of 40 rats with the AA genotype (83%) developed stomach carcinomas, whereas 14 of 33 rats with the AB genotype (42%) developed carcinomas(P = 0.0004). For Gcr2(represented by D3Rat55), 28 of 34 rats with the AA genotype (82%) developed stomach carcinomas, whereas 19 of 39 rats with the AB genotype (49%) developed carcinomas(P = 0.003). When Gcr1 and Gcr2 were combined, 21 of 25 rats (84%) with the AA genotypes at both loci developed stomach carcinomas,whereas 7 of 24 rats (29%) with the AB genotypes at both loci developed carcinomas (P = 0.0002). The effects of the two genes on the development of stomach carcinomas were considered as additive. Gcs1 also showed an additive effect with Gcr1 and Gcr2, but in an opposite direction. All of the 30 rats (100%) with the susceptible genotype at Gcs1 (AB) and at either Gcr1 or at Gcr2 (AA) developed stomach carcinomas.

Linkage with the Histological Grade.

All of the 80 rats with stomach tumors were given a score based on the histology of the tumors they bore (Table 4). Linkage analysis with the score showed a strong linkage with chr. 15,which was considered to be the same with Gcs1 for the rats with tumors (LOD, 3.6; n = 80) but not for those with carcinomas (n = 48; Table 2). This indicated that Gcs1 exerts its effect mainly at the stage of conversion from an adenoma to a carcinoma but that it does not have an effect on the progression of the histological grade of a carcinoma (see “Discussion”). When the backcross rats were classified by the genotypes of Gcs1 (represented by D15Rat102), a major difference was found in the number of rats with an adenoma (Table 4). The LOD score obtained for the rats with tumors (ET group) was 3.6, whereas that obtained for the full-term rats with tumors (FT group) was 3.0.

Linkage with the Depth of Tumor Invasion and Size of Tumor.

Linkage analysis was also performed with the depth of tumor invasion and with the size of tumor (Fig. 3,A). Using the full-term rats, we found that the two phenotypes were linked to the same locus on chr. 16, designated as Gcr3,although the LOD scores were not definitive (Table 2).

The depth of tumor invasion was linked to Gcr3 with a LOD score of 2.2 for full-term rats with a tumor (FT group) and 2.3 for full-term rats with a carcinoma (FC group; Table 2). For the FT group,the average score of depth of tumor was 3.3 in rats with the AA genotype in Gcr3 (represented by D16Rat17), and 2.2 in those with the ABgenotype (Table 5). The BUF allele at Gcr3 was considered to reduce the depth of tumor invasion by a score of 1.1 in Table 5.

The average diameter of tumors was also linked to Gcr3, with LOD scores of 2.1 (all full-term rats; FA group), 1.9 (FT group), and 1.8 (FC group; Table 2). For the FC group, the average diameter was 21 mm in rats with the AA genotype in Gcr3, and 8 mm in those with the AB genotype (Fig. 3 B). The BUF allele at this locus was considered to reduce the diameter of carcinomas by an average of 13 mm. The average tumor diameters were significantly different between the rats with the AAgenotype and those with the AB genotype, for rats in the FC group (P = 0.009) or those in the FT group(P = 0.007).

Linkage with the Number of Tumors.

Among the 100 effective rats, 72 rats developed one tumor, and 8 rats developed two tumors. Although linkage analysis was performed using the number of stomach carcinomas per rat, no linkage was observed.

We mapped one dominant susceptibility gene, Gcs1,and three dominant resistance genes, Gcr1, Gcr2,and Gcr3, involved in MNNG-induced stomach carcinogenesis. The involvement of multiple genes was contrary to our initial expectations of “a single dominant resistance gene” model, which had been proposed simply based on the carcinoma incidences in F1 and F2 populations. Other, more complicated models had not been excluded. Actually, the interaction of the four genes identified in this study explains the carcinoma incidences in the backcross rats in this study. Complex interaction of multiple genes has been reported in the mouse liver cancer model after treatment with urethane (23). Although it is expected that multiple genes are also involved in human cancer susceptibility, each gene with a relatively low penetrance, it is difficult to identify those genes in human families. An efficient approach would be to identify those genes in animal models and then to test the involvement of the genes in a human population.

Rat chr. 15 around Gcs1 is known to have synteny with human chr. 13q (24, 25), where the endothelin receptor type B (Ednrb) is located. Rat chr. 4 around Gcr1has synteny with human 2p13, and it is speculated that the Msh6 gene, encoding the G/T mismatch-binding protein, is located in the region. Although MNNG-induced stomach cancers do not display microsatellite instability (26), the Msh6 gene is one of the candidate genes for Gcr1. The MSH2 gene is also located on 2p in humans, but rat Msh2 was mapped to chr. 6 (27), ruling out Msh2 as a candidate for Gcr1. Rat chr. 3 around Gcr2 has synteny with human 9q34 and harbors the prostaglandin G/H synthase gene (Cox1) and prostaglandin D2 synthase gene. Cox1 is known to be involved in the cytoprotection of gastric mucosa and to be expressed in the stem cells of the pylorus (28). Considering that the suggested mechanism for the different susceptibilities between the two rat strains is the difference in the rates of cell proliferation in response to mucosal damage, Cox1 is a candidate for Gcr2. Rat chr. 16 around Gcr3 has synteny with human chr. 4q or 8p, where loss of heterozygosity is reported in human stomach cancers (29).

Gcs1 is the major gene that controls the development of carcinomas. The histological grade of tumors was also linked to Gcs1 in rats with tumors, but the linkage was much weaker in rats with only carcinomas. This suggested that Gcs1 is mainly involved in the conversion from adenomas to carcinomas. The distribution of the histological grades of the tumors in the rats classified by the genotype of Gcs1 also supports this idea. Linkage of the histological grade with Gcs1 was observed both in the rats of all experimental periods and in the full-term rats. This indicated that the effect of Gcs1 is expressed from early periods to the end of the carcinogenicity test. On the other hand, linkage of the depth of tumor invasion and tumor size with Gcr3 was observed only in the full-term rats, not in the rats of all experimental periods. This was considered because these two parameters change as a tumor grows.

Although LOD scores obtained for Gcr1, Gcr2, and Gcr3 exceeded 1.9, the criterion for“suggested linkage” in backcross rats, they did not exceed 3.3, the criterion for “definitive linkage” (30). After we found that as many as four genes were involved, we considered the number of effective rats, 100, relatively small. However,classification of the backcross rats by the genotypes of one of the four genes clearly showed significant differences in phenotypes. The facts that both the development of carcinoma and histological grade were linked to Gcs1 and that both depth of tumor invasion and sizes of tumors were linked to Gcr3 further indicate that the possibility of false positives is very low.

In summary, we mapped three genes involved in the completion of malignant transformation of a stomach epithelium cell and one gene involved in the growth of the completed cancer cell. The suggested mechanism for the different susceptibilities between ACI and BUF leads us to expect that some of the four genes are also involved in human cancer susceptibility.

Fig. 1.

Cumulative incidence of stomach cancers in ACI, BUF, and backcross rats. Rats that survived 52 weeks or more were counted as effective rats and are shown.

Fig. 1.

Cumulative incidence of stomach cancers in ACI, BUF, and backcross rats. Rats that survived 52 weeks or more were counted as effective rats and are shown.

Close modal
Fig. 2.

Results of linkage analysis with carcinoma development(A) and with tumor size (B). LOD scores using four combinations of classifications (R1S3, R2S3, R3S3, and R4S3;see Table 1) are shown in A.

Fig. 2.

Results of linkage analysis with carcinoma development(A) and with tumor size (B). LOD scores using four combinations of classifications (R1S3, R2S3, R3S3, and R4S3;see Table 1) are shown in A.

Close modal
Fig. 3.

Distribution of tumor sizes in full-term rats and the effect of D16Rat17. A, numbers of rats versus the size of tumors. Histological diagnosis of the tumors are shown: □, no tumors; , adenomas; ▪, carcinomas; B, effect of Gcr3, represented by D16rat17, on the size of carcinomas (n = 27). □, AB; ▪, AA.

Fig. 3.

Distribution of tumor sizes in full-term rats and the effect of D16Rat17. A, numbers of rats versus the size of tumors. Histological diagnosis of the tumors are shown: □, no tumors; , adenomas; ▪, carcinomas; B, effect of Gcr3, represented by D16rat17, on the size of carcinomas (n = 27). □, AB; ▪, AA.

Close modal

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

Supported by the Program for Promotion of Fundamental Studies in Health Sciences of the Organization for Pharmaceutical Safety and Research (OPSR); by a Grant-in-Aid from the Ministry of Health and Welfare for the 2nd Term Comprehensive 10-Year Strategy for Cancer Control; and a grant from Princess Takamatsu Cancer Research Fund. M. S. is a recipient of Research Resident Fellowship from the Foundation for Promotion of Cancer Research,Japan.

3

The abbreviations used are: MNNG, N-methyl-N′-nitro-N-nitrosoguanidine;ACI, ACI/N strain; BUF, BUF/Nac strain; QTL, quantitative trait locus; Gcs, gastric cancer susceptibility gene; chr.,chromosome; Gcr, gastric cancer resistance gene.

Table 1

Distribution of the ages of backcross rats and the sizes of stomach cancers

Rats that survived 52 weeks or more were counted as effective rats(n = 100). In addition to histological diagnosis, rats with carcinomas were classified by the maximal diameters of their carcinomas. Sizes of adenomas ranged from 0.5 to 3 mm. Tumors that could not be diagnosed as malignant or benign were classified as “borderline” tumors. Three kinds of criteria for susceptible rats (S1, S2, and S3) and four kinds of criteria for resistant rats (R1, R2, R3, and R4) were set, and 12 kinds of carcinoma development phenotypes were produced by combining S1–S3 and R1–R4. S1, s1 (n = 44); S2, s1–s2(n = 45); S3, s1–s3 (n = 47); R1, r1 (n = 22); R2, r1–r2(n = 25); R3, r1–r3 (n = 26); and R4, r1–r4 (n = 31).

HistologyNo. of rats in each groupTotal
Age (weeks)52–5859–6364–6869–7374–7879–80
Adenocarcinoma         
>5 mm   1s1 5s1 8s1 1s1 23s1 38 
≤5 mm      2s1 1s2 
≤4 mm   1s1 1s1 1s1  2s3 
≤3 mm   1s1    
Borderline    1r3 3r2 
Adenoma   1r1 11r1 23 
No tumor  4r4 1r4 2r1 8r1 20 
Total  15 16 49 100 
HistologyNo. of rats in each groupTotal
Age (weeks)52–5859–6364–6869–7374–7879–80
Adenocarcinoma         
>5 mm   1s1 5s1 8s1 1s1 23s1 38 
≤5 mm      2s1 1s2 
≤4 mm   1s1 1s1 1s1  2s3 
≤3 mm   1s1    
Borderline    1r3 3r2 
Adenoma   1r1 11r1 23 
No tumor  4r4 1r4 2r1 8r1 20 
Total  15 16 49 100 
Table 2

Summary of the results of QTL mapping

Susceptibility was scored as 2 (susceptible) or 0 (resistant). Histological grade and depth of tumor invasion were classified arbitrarily. Tumor size was measured as tumor diameter in mm. LOD scores greater than 2.0 are underlined.

TraitPopulation usedChr. 3 D3Rat55Chr. 4 AmppChr. 15 D15Rat102Chr. 16 D16Rat17
LOD scoreEffectLOD scoreEffectLOD scoreEffectLOD scoreEffect
Carcinoma development R1S3 (n = 69) 1.88 −0.64 2.38 −0.74 3.84 0.95 0.10 −0.16 
 R2S3 (n = 72) 2.41 −0.72 2.48 −0.74 3.84 0.95 0.11 −0.16 
 R3S3 (n = 73) 2.59 −0.75 2.70 −0.76 3.41 0.87 0.05 −0.11 
 R4S3 (n = 78) 1.83 −0.63 1.67 −0.61 2.67 0.77 0.00 0.00 
Histological grade EAa (n = 100) 0.90 −0.60 0.89 −0.58 2.22 0.96 0.02 −0.09 
 ET (n = 80) 0.48 −0.39 0.26 −0.29 3.55 1.13 0.01 −0.05 
 EC (n = 48) 0.10 0.11 0.64 0.30 0.01 0.04 1.11 −0.37 
 FA (n = 49) 0.76 −0.80 0.87 −0.85 2.11 1.42 1.02 −0.95 
 FT (n = 41) 0.44 −0.55 0.48 −0.57 2.98 1.44 1.08 −0.87 
 FC (n = 27) 0.07 0.14 0.54 0.39 0.34 −0.36 1.80 −0.65 
Depth of tumor invasion EA (n = 100) 0.55 −0.21 0.74 −0.31 0.36 0.37 0.22 −0.29 
 ET (n = 80) 0.20 −0.23 0.08 −0.14 1.20 0.56 0.33 −0.30 
 EC (n = 48) 0.16 0.25 0.13 0.24 0.00 0.03 1.27 −0.68 
 FA (n = 49) 0.33 −0.50 0.21 −0.40 1.06 0.88 1.69 −1.13 
 FT (n = 41) 0.07 −0.20 0.00 −0.05 1.13 0.79 2.17 −1.13 
 FC (n = 27) 0.27 0.46 0.56 0.66 0.03 −0.19 2.33 −1.31 
Tumor size EA (n = 100) 0.18 −1.65 0.05 −0.86 1.42 4.55 0.60 −3.04 
 ET (n = 80) 0.05 −1.02 0.01 0.34 1.30 5.06 0.61 −3.55 
 EC (n = 48) 0.17 2.66 0.38 4.13 0.10 2.13 1.50 −7.62 
 FA (n = 49) 0.11 −2.30 0.07 −1.78 1.32 7.59 2.07 −9.63 
 FT (n = 41) 0.03 −1.34 0.00 −0.45 1.09 7.87 1.92 −10.30 
 FC (n = 27) 0.19 4.36 0.26 5.30 0.07 3.17 1.77 −12.80 
TraitPopulation usedChr. 3 D3Rat55Chr. 4 AmppChr. 15 D15Rat102Chr. 16 D16Rat17
LOD scoreEffectLOD scoreEffectLOD scoreEffectLOD scoreEffect
Carcinoma development R1S3 (n = 69) 1.88 −0.64 2.38 −0.74 3.84 0.95 0.10 −0.16 
 R2S3 (n = 72) 2.41 −0.72 2.48 −0.74 3.84 0.95 0.11 −0.16 
 R3S3 (n = 73) 2.59 −0.75 2.70 −0.76 3.41 0.87 0.05 −0.11 
 R4S3 (n = 78) 1.83 −0.63 1.67 −0.61 2.67 0.77 0.00 0.00 
Histological grade EAa (n = 100) 0.90 −0.60 0.89 −0.58 2.22 0.96 0.02 −0.09 
 ET (n = 80) 0.48 −0.39 0.26 −0.29 3.55 1.13 0.01 −0.05 
 EC (n = 48) 0.10 0.11 0.64 0.30 0.01 0.04 1.11 −0.37 
 FA (n = 49) 0.76 −0.80 0.87 −0.85 2.11 1.42 1.02 −0.95 
 FT (n = 41) 0.44 −0.55 0.48 −0.57 2.98 1.44 1.08 −0.87 
 FC (n = 27) 0.07 0.14 0.54 0.39 0.34 −0.36 1.80 −0.65 
Depth of tumor invasion EA (n = 100) 0.55 −0.21 0.74 −0.31 0.36 0.37 0.22 −0.29 
 ET (n = 80) 0.20 −0.23 0.08 −0.14 1.20 0.56 0.33 −0.30 
 EC (n = 48) 0.16 0.25 0.13 0.24 0.00 0.03 1.27 −0.68 
 FA (n = 49) 0.33 −0.50 0.21 −0.40 1.06 0.88 1.69 −1.13 
 FT (n = 41) 0.07 −0.20 0.00 −0.05 1.13 0.79 2.17 −1.13 
 FC (n = 27) 0.27 0.46 0.56 0.66 0.03 −0.19 2.33 −1.31 
Tumor size EA (n = 100) 0.18 −1.65 0.05 −0.86 1.42 4.55 0.60 −3.04 
 ET (n = 80) 0.05 −1.02 0.01 0.34 1.30 5.06 0.61 −3.55 
 EC (n = 48) 0.17 2.66 0.38 4.13 0.10 2.13 1.50 −7.62 
 FA (n = 49) 0.11 −2.30 0.07 −1.78 1.32 7.59 2.07 −9.63 
 FT (n = 41) 0.03 −1.34 0.00 −0.45 1.09 7.87 1.92 −10.30 
 FC (n = 27) 0.19 4.36 0.26 5.30 0.07 3.17 1.77 −12.80 
a

EA, all effective rats; ET, rats with tumors; EC, rats with carcinomas; FA, full-term rats; FT, full-term rats with tumors;FC, full-term rats with carcinomas.

Table 3

Coordinate effect of Gcs1, Gcr1, and Gcr2 on carcinoma development

Incidences of stomach cancers are shown in each group of rats classified by the genotypes of Gcs1 (D15Rat102), Gcr1 (Ampp), and Gcr2(D3Rat55). Criterion R3S3 was used to diagnose susceptible and resistant rats in this Table.

Gcs1 (D15Rat102)AAAB
Gcr1 (Ampp)AAABAAAB
Gcr2 (D3Rat55)AAABAAABAAABAAAB
Effective rats (n12 13 13 12 11 
Rats with stomach cancer (n13 12 
Incidence (%) 67 50 23 100 100 100 36 
Gcs1 (D15Rat102)AAAB
Gcr1 (Ampp)AAABAAAB
Gcr2 (D3Rat55)AAABAAABAAABAAAB
Effective rats (n12 13 13 12 11 
Rats with stomach cancer (n13 12 
Incidence (%) 67 50 23 100 100 100 36 
Table 4

Classification of the backcross rats by the histological grade and effect of Gcs1

BUF allele in Gcs1 (D15Rat102) increased the incidence of stomach cancer, mainly by promoting conversion from adenomas to carcinomas.

HistologyScoreAll effective ratsRats surviving full term
TotalD15Rat102TotalD15Rat102
AAABAAAB
No tumor  20 11 
Adenoma 23 17 11 
Borderline 
Adenocarcinoma  48 14 34 27 21 
Well-differentiated 28 19 14 12 
Moderately differentiated 18 14 11 
Poorly differentiated or signet-ring cell 
Total  100 47 53 49 22 27 
HistologyScoreAll effective ratsRats surviving full term
TotalD15Rat102TotalD15Rat102
AAABAAAB
No tumor  20 11 
Adenoma 23 17 11 
Borderline 
Adenocarcinoma  48 14 34 27 21 
Well-differentiated 28 19 14 12 
Moderately differentiated 18 14 11 
Poorly differentiated or signet-ring cell 
Total  100 47 53 49 22 27 
Table 5

Classification of the backcross rats by the depth of tumor invasion and effect of Gcr3

BUF allele in Gcr3 (D16Rat17) was shown to reduce the depth of tumor invasion in the full-term rats.

DepthScoreAll effective ratsRats surviving full term
TotalD16Rat17TotalD16Rat17
AAABAAAB
No tumor  20 11 
Mucosa 11 
Submucosa 35 14 21 18 14 
Muscularis propria 18 11 
Subserosa 10 
Serosa exposed 
Total  100 43 57 49 17 32 
DepthScoreAll effective ratsRats surviving full term
TotalD16Rat17TotalD16Rat17
AAABAAAB
No tumor  20 11 
Mucosa 11 
Submucosa 35 14 21 18 14 
Muscularis propria 18 11 
Subserosa 10 
Serosa exposed 
Total  100 43 57 49 17 32 

We thank Y. Hosoya for excellent technical assistance.

1
Warmuth M. A., Sutton L. M., Winer E. P. A review of hereditary breast cancer: from screening to risk factor modification.
Am. J. Med.
,
102
:
407
-415,  
1997
.
2
Lynch H. T., Smyrk T., Lynch J. An update of HNPCC (Lynch syndrome).
Cancer Genet. Cytogenet.
,
93
:
84
-99,  
1997
.
3
Peddanna N., Holt S., Verma R. S. Genetics of gastric cancer.
Anticancer Res.
,
15
:
2055
-2064,  
1995
.
4
Guilford P., Hopkins J., Harraway J., McLeod M., McLeod N., Harawira P., Taite H., Scoular R., Miller A., Reeve A. E. E-Cadherin germline mutations in familial gastric cancer.
Nature (Lond.)
,
392
:
402
-405,  
1998
.
5
Ohgaki H., Sugimura T. Experimental gastric cancer Sugimura T. Sasako M. eds. .
Gastric Cancer
,
:
73
-86, Oxford University Press New York  
1997
.
6
Tatematsu M., Takahashi M., Fukushima S., Hananouchi M., Shirai T. Effects in rats of sodium chloride on experimental gastric cancers induced by N-methyl-N′-nitro-N-nitrosoguanidine or 4-nitroquinoline-1-oxide.
J. Natl. Cancer Inst.
,
55
:
101
-106,  
1975
.
7
Yamane T., Takahashi T., Kuwata K., Oya K., Inagake M., Kitao Y., Suganuma M., Fujiki H. Inhibition of N-methyl-N′-nitro-N-nitrosoguanidine-induced carcinogenesis by (−)-epigallocatechin gallate in the rat glandular stomach.
Cancer Res.
,
55
:
2081
-2084,  
1995
.
8
Furukawa H., Iwanaga T., Koyama H., Taniguchi H. Effect of sex hormones on carcinogenesis in the stomachs of rats.
Cancer Res.
,
42
:
5181
-5182,  
1982
.
9
Sugimura T., Fujimura S. Tumour production in glandular stomach of rat by N-methyl-N′-nitro-N-nitrosoguanidine.
Nature (Lond.)
,
216
:
943
-944,  
1967
.
10
Bralow S. P., Gruenstein M., Meranze D. R. Host resistance to gastric adenocarcinomatosis in three strains of rats ingesting N-methyl-N′-nitro-N-nitrosoguanidine.
Oncology
,
27
:
168
-180,  
1973
.
11
Ohgaki H., Kawachi T., Matsukura N., Morino K., Miyamoto M., Sugimura T. Genetic control of susceptibility of rats to gastric carcinoma.
Cancer Res.
,
43
:
3663
-3667,  
1983
.
12
Tatematsu M., Aoki T., Inoue T., Mutai M., Furihata C., Ito N. Coefficient induction of pepsinogen 1-decreased pyloric glands and gastric cancers in five different strains of rats treated with N-methyl-N′-nitro-N-nitrosoguanidine.
Carcinogenesis (Lond.)
,
9
:
495
-498,  
1988
.
13
Ohgaki H., Tomihari M., Sato S., Kleihues P., Sugimura T. Differential proliferative response of gastric mucosa during carcinogenesis induced by N-methyl-N′-nitro-N-nitrosoguanidine in susceptible ACI rats, resistant Buffalo rats, and their hybrid F1 cross.
Cancer Res.
,
48
:
5275
-5279,  
1988
.
14
Japanese Gastric Cancer Association. Japanese classification of gastric carcinoma, 2nd English ed. Gastric Cancer, 1: 10–24, 1998.
15
Sambrook, J., Fritsch, E. F., and Maniatis, T. Isolation of high-molecular-weight DNA from mammalian cells. In: J. Sambrook, E. F. Fritsch, and T. Maniatis (eds.), Molecular Cloning, 2nd ed., Vol. 2, pp. 9.14–9.23. New York: Cold Spring Harbor Laboratory, 1989.
16
Ushijima T., Nomoto T., Sugimura T., Housman D. E., Nagao M. Isolation of 48 genetic markers appropriate for high throughput genotyping of inbred rat strains by B1 repetitive sequence-representational difference analysis.
Mamm. Genome
,
9
:
1008
-1012,  
1998
.
17
Yoshida Y., Ushijima T., Yamashita S., Imai K., Sugimura T., Nagao M. Development of the arbitrarily primed-representational difference analysis method and chromosomal mapping of isolated high throughput rat genetic markers.
Proc. Natl. Acad. Sci. USA
,
96
:
610
-615,  
1999
.
18
Yamada J., Kuramoto T., Serikawa T. A rat genetic linkage map and comparative maps for mouse or human homologous rat genes.
Mamm. Genome
,
5
:
63
-83,  
1994
.
19
Jacob H. J., Brown D. M., Bunker R. K., Daly M. J., Dzau V. J., Goodman A., Koike G., Kren V., Kurtz T., Lernmark A., Levan G., Mao Y-P., Petteson A., Pravence M., Simon J. S., Szpirer C., Szpirer J., Trolliet M. R., Winer E. S., Lander E. S. A genetic linkage map of the laboratory rat, Rattus norvegicus.
Nat. Genet.
,
9
:
63
-69,  
1995
.
20
Lander E. S., Green P., Abrahamson J., Barlow A., Daly M. J., Lincoln S. E., Newburg L. MAPMAKER: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations.
Genomics
,
1
:
174
-181,  
1987
.
21
Solleveld H. A., Boorman G. A. Spontaneous renal lesions in five rat strains.
Toxicol. Pathol.
,
14
:
168
-174,  
1986
.
22
Marshall F. F., Ewing L. L., Zirkin B. R., Cochran R. C. Testicular atrophy associated with agenesis of the epididymis in the ACI rat.
J. Urol.
,
127
:
155
-158,  
1982
.
23
Dragani T. A., Canzian F., Manenti G., Pierotti M. A. Hepatocarcinogenesis: a polygenic model of inherited predisposition to cancer.
Tumori
,
82
:
1
-5,  
1996
.
24
Serikawa T., Cui Z., Yokoi N., Kuramoto T., Kondo Y., Kitada K., Guenet J. L. A comparative genetic map of rat, mouse, and human genomes.
Exp. Anim.
,
47
:
1
-9,  
1998
.
25
Watanabe T. K., Bihoreau M-T., McCarthy L. C., Kiguwa S. L., Hishigaki H., Tsuji A., Browne J., Yamasaki Y., Mizoguchi-Miyakita A., Oga K. A radiation hybrid map of the rat genome containing 5,255 markers, Nat.
Genet.
,
22
:
27
-36,  
1999
.
26
Hirayama Y., Wakazono K., Yamamoto M., Kitano M., Tatematsu M., Nagao M., Sugimura T., Ushijima T. Rare mutations of p53, K-ras and β-catenin genes and absence of K-sam and c-erbB-2 amplifications in N-methyl-N′-nitro-N-nitrosoguanidine-induced rat stomach cancers.
Mol. Carcinog.
,
25
:
42
-47,  
1999
.
27
Hirayama Y., Ushijima T., Kuramoto T., Kitano M., Sugimura T., Nagao M. Linkage mapping of the rat Msh2 DNA mismatch repair gene on chromosome 6.
Exp. Anim.
,
48
:
63
-64,  
1999
.
28
Iseki S. Immunocytochemical localization of cyclooxygenase-1 and cyclooxygenase-2 in the rat stomach.
Histochem. J.
,
27
:
323
-328,  
1995
.
29
Yustein A. S., Harper J. C., Petroni G. R., Cummings O. W., Moskaluk C. A., Powell S. M. Allelotype of gastric adenocarcinoma.
Cancer Res.
,
59
:
1437
-1441,  
1999
.
30
Lander E. S., Kruglyak L. Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results.
Nat. Genet.
,
11
:
241
-247,  
1995
.