The Japanese Wild-Derived Inbred Mouse Strain, MSM/Ms in Cancer Research

Simple Summary Since human studies on tumor susceptibility alleles require vast numbers of DNA samples from both cancer patients and well-matched controls, their investigation and identification in humans have been complemented using mouse models. However, a number of confounding factors are associated with this type of research, including heterogeneity, weak genetic interactions, and lifestyle habits. Inbred strains relatively recently established from wild mice are often more resistant to carcinogenic stimulation and various pathogens than standard inbred mouse strains. A Japanese wild-derived inbred mouse strain, MSM/Ms has been used to map tumor resistance loci as well as other Quantitative Trait Loci (QTL) in Japan. Furthermore, genetic tools have been developed with MSM/Ms. MSM/Ms genomic sequences are currently available, which have greatly promoted the identification of tumor resistance loci as well as genes controlling quantitative variations and provide a more detailed understanding of gene function. Abstract MSM/Ms is a unique inbred mouse strain derived from the Japanese wild mouse, Mus musculus molossinus, which has been approximately 1 million years genetically distant from standard inbred mouse strains mainly derived from M. m. domesticus. Due to its genetic divergence, MSM/Ms has been broadly used in linkage studies. A bacterial artificial chromosome (BAC) library was constructed for the MSM/Ms genome, and sequence analysis of the MSM/Ms genome showed approximately 1% of nucleotides differed from those in the commonly used inbred mouse strain, C57BL/6J. Therefore, MSM/Ms mice are thought to be useful for functional genome studies. MSM/Ms mice show unique characteristics of phenotypes, including its smaller body size, resistance to high-fat-diet-induced diabetes, high locomotive activity, and resistance to age-onset hearing loss, inflammation, and tumorigenesis, which are distinct from those of common inbred mouse strains. Furthermore, ES (Embryonic Stem) cell lines established from MSM/Ms allow the MSM/Ms genome to be genetically manipulated. Therefore, genomic and phenotypic analyses of MSM/Ms reveal novel insights into gene functions that were previously not obtained from research on common laboratory strains. Tumorigenesis-related MSM/Ms-specific genetic traits have been intensively investigated in Japan. Furthermore, radiation-induced thymic lymphomas and chemically-induced skin tumors have been extensively examined using MSM/Ms.


Introduction
Genome-wide association studies (GWAS) of human cancer have been relatively easily performed recently because next-generation sequencing has been conducted worldwide. However, this research requires a huge number of samples from both cancer patients and well-matched controls. As a number of confounding factors are associated with GWAS, such as heterogeneity, weak genetic interactions, and lifestyle habits, a vast number of cases and controls are needed to reach statistically reliable significance [1,2]. Therefore, mouse size [22], resistance to high-fat-diet-induced diabetes [25], high locomotive activity [26], and resistance to age-onset hearing loss [27,28], inflammation [29], and tumorigenesis [22,[29][30][31][32][33][34][35][36] (Table 1). Furthermore, ES cell lines established from MSM/Ms enable us to manipulate the MSM/Ms genome [37]. Therefore, genomic and phenotypic analyses of MSM/Ms reveal novel insights into gene function that have not been possible using common laboratory strains.  [36] PhIP; 2-Amino-1-methyl-6-phenylimidazo ( We herein focused on the resistance of MSM/Ms to cancer. As we mentioned earlier, inbred mouse strains recently derived from wild mice are often resistant to carcinogenic stimulation. Mus musculus inbred strains have significantly longer telomeres than mouse strains derived from wild mice [38,39]. It is still necessary to map the stable inheritance of telomeres and to assess the expression of telomerase in the MSM/Ms strain. Since it is a hybrid between M. m. musculus and M. m. castaneus, it may have implications for cancer modeling. Radiation-induced thymic lymphomas and chemically-induced skin tumors have been extensively examined using MSM/Ms. We will mainly review these two types of tumor studies using MSM/Ms.

Identification of Tumor Resistance Loci
Resistance to radiation-induced thymic lymphomas is dependent on mouse strains [40,41] ( Table 2). BALB/c (Bagg ALBINO/c genotype at the color locus) is a susceptible strain to the radiation-induced mouse thymic lymphomagenesis, whereas MSM/Ms is resistant. The incidence of radiation-induced thymic lymphomas in F1 mice is similar to that in BALB/c, indicating that BALB/c exhibits dominant susceptibility to these lymphomas. Kominami and colleagues at Niigata University (Niigata, Japan) examined an association with the control of lymphoma susceptibility of backcross mice between these two strains to identify the genetic loci involved [33]. Three markers, D2Mit15 (Thyls1 for Thymic lymphoma susceptibility 1), D4Mit12 (Thyls2), and D5Mit5 (Thyls3), showed a strong linkage ( Figure 1). Mice having lymphoma exhibited more heterozygous genotypes at D2Mit15 and D4Mit12 than MSM/Ms homozygosity, indicating that BALB/c has a dominant susceptible allele in these loci, whereas MSM/Ms has resistant alleles. However, the genotype of mice with lymphoma exhibited more MSM/Ms homozygosity than heterozygosity at D5Mit5, indicating MSM/Ms has susceptible alleles in this locus. Cooperative effects on lymphomagenesis were also observed among the three loci.
In an attempt to cut down the candidate interval around D4Mit12, congenic mice carrying a 40-Mb chromosomal region on chromosome 4 from the MSM/Ms genome were generated. These mice were crossed with BALB/c mice to further divide the congenic region. As a result, four subcongenic lines were generated by genotyping with microsatellite markers [42]. Congenic mapping and a haplotype analysis suggested metal-responsive transcription factor-1 (Mtf-1) was a responsible gene for Thyls2 on chromosome 4. Sequence analysis showed lymphoma susceptible strains had a polymorphism in Mtf-1, in which serine exists at position 424 in the proline-rich domain. In contrast, lymphoma-resistant strains were found to have proline in that position. Comparisons of the transcriptional activity of Mtf-1 having both alleles were performed by expressing both constructs in Mtf-1-deficient cells, respectively. As a result, proline at position 424 conferred stronger metal responsiveness. The radiation inducibility of target genes was also stronger in resistant congenic lines having the Mtf-1 allele of the proline type. As the target products induced by irradiation inhibited the cellular stresses, resistant congenic strains to thymic lymphomas having the proline type seemed resistant to radiation effects, which probably resulted in resistance to the development of lymphoma. linkage. In order to confirm the linkage at the three markers to the development of thymic lymphomas, they crossed p53 +/-BALB/c mice with consomic mice having chromosome 19 of MSM/Ms on the BALB/c background. These mice were subjected to radiation and the linkage study was carried out. As a result, D19Mit90 exhibited a significant linkage as well as in the backcross study. In contrast, D19Mit5 and D19Mit123 did not show any linkage. These results suggest two possibilities. Consomic mice were originally generated on the C57BL6/J background and backcrossed to BALB/c mice several times to replace the background strain. The C57BL6/J genome could remain and mask the effect of MSM/Ms alleles of D19Mit5 and D19Mit123. The second is that the linkage at D19Mit5 and D19Mit123 were dependent on epistatic interactions. In other words, the D19Mit90 region could function by itself, whereas the D19Mit5 and D19Mit123 regions could require other regions from MSM/Ms to work in combinations.

Genetic Analysis of the LOH Region
Kominami and colleagues performed a genome-wide LOH analysis of radiation-induced thymic lymphomas obtained from F1 mice between BALB/c and MSM/Ms mice and backcross mice that are heterozygous for several chromosomes. Approximately 50% of those mice were a p53 +/− [45]. The findings obtained revealed that two loci exhibited frequent LOH and both alleles of the two loci were equally lost; one was mapped within a 2.9 cM region (Tlsr12 for Thymic lymphoma suppressor 12) between D12Mit53 and D12Mit279, and the other was mapped near D16Mit122/D16Mit162 (Tlsr16) (Figure 1). D12Mit279 showed 62% of LOH frequency, regardless of the presence of the p53 +/− allele. In contrast, D16Mit122 showed 62% of LOH frequency on the p53 +/− background and 13% on the p53 +/+ background, suggesting the presence of p53 +/− allele increased the frequency of LOH at D16Mit122. These results strongly suggested at least two types of tumor suppressor genes were located around D12Mit279 and D16Mit122, one was independent of p53 and the other was dependent on p53. Kominami and colleagues generated congenic mouse lines covering a 28.4 cM interval of Thyls3 on chromosome 5 by eliminating genomic DNA of susceptible MSM/Ms alleles from F1 mice between BALB/c and MSM/Ms mice to confirm the susceptibility genes in the Thyls3 region of MSM/Ms. MSM/Ms itself is a strain resistant to thymic lymphomagenesis. However, the Thyls3 locus on chromosome 5 of MSM/Ms exceptionally conferred susceptibility to lymphomas [43]. They induced lymphomas in these congenic mice in two ways. Half of the congenic mice were subjected to irradiation, and the other half of the mice were administered N-methyl-N-nitrosourea (MNU), an alkylating agent. As a result, 87.5% of the irradiated congenic mice that are MSM/Ms homozygous at D5Mit5 developed radiogenic lymphomas. On the other hand, 46% of the irradiated congenic mice that are heterozygous at D5Mit5 developed radiogenic lymphomas. MSM/Ms homozygosity at D5Mit5 gave rise to a significantly higher frequency of radiogenic lymphomas. These results strongly suggested MSM/Ms had a strong susceptible allele in Thyls3 on chromosome 5. In contrast, the difference in the frequencies of MNU-induced thymic lymphomas was not detected between MSM/Ms homozygosity and heterozygosity at D5Mit5. These findings indicated a susceptible MSM/Ms allele for thymic lymphomas and Thyls3 conferred susceptibility specifically to radiation carcinogenesis, but not to MNU.
In order to identify p53 (Trp53)-dependent tumor resistance genes, Kominami and colleagues also attempted to map genetic modifiers of radiation-induced thymic lymphomas and skin tumors in p53 knockout MSM/Ms mice [44]. p53 +/− backcross mice were generated by crossing p53 knockout MSM/Ms mice with FVB/N mice and were subjected to radiation to induce a large number of lymphomas and skin tumors. Genome-wide screening exhibited BALB/c alleles at D19Mit5, D19Mit90, and D19Mit123 on chromosome 19 prolonged the latency of the development of thymic lymphoma and the survival of the mice (Thyls4, Figure 1). These results indicated that BALB/c alleles in the Thyls4 region conferred resistance to radiation-induced thymic lymphomas and MSM alleles in that region conferred susceptibility on p53 +/− background. This is the same situation as shown in Thyls3 on chromosome 5, as mentioned earlier. D19Mit90 and D19Mit123 also showed the linkage to the latency of skin tumors and survival, whereas D19Mit5 did not show the linkage. In order to confirm the linkage at the three markers to the development of thymic lymphomas, they crossed p53 +/− BALB/c mice with consomic mice having chromosome 19 of MSM/Ms on the BALB/c background. These mice were subjected to radiation and the linkage study was carried out. As a result, D19Mit90 exhibited a significant linkage as well as in the backcross study. In contrast, D19Mit5 and D19Mit123 did not show any linkage. These results suggest two possibilities. Consomic mice were originally generated on the C57BL6/J background and backcrossed to BALB/c mice several times to replace the background strain. The C57BL6/J genome could remain and mask the effect of MSM/Ms alleles of D19Mit5 and D19Mit123. The second is that the linkage at D19Mit5 and D19Mit123 were dependent on epistatic interactions. In other words, the D19Mit90 region could function by itself, whereas the D19Mit5 and D19Mit123 regions could require other regions from MSM/Ms to work in combinations.

Genetic Analysis of the LOH Region
Kominami and colleagues performed a genome-wide LOH analysis of radiationinduced thymic lymphomas obtained from F1 mice between BALB/c and MSM/Ms mice and backcross mice that are heterozygous for several chromosomes. Approximately 50% of those mice were a p53 +/− [45]. The findings obtained revealed that two loci exhibited frequent LOH and both alleles of the two loci were equally lost; one was mapped within a 2.9 cM region (Tlsr12 for Thymic lymphoma suppressor 12) between D12Mit53 and D12Mit279, and the other was mapped near D16Mit122/D16Mit162 (Tlsr16) ( Figure 1). D12Mit279 showed 62% of LOH frequency, regardless of the presence of the p53 +/− allele. In contrast, D16Mit122 showed 62% of LOH frequency on the p53 +/− background and 13% on the p53 +/+ background, suggesting the presence of p53 +/− allele increased the frequency of LOH at D16Mit122. These results strongly suggested at least two types of tumor suppressor genes were located around D12Mit279 and D16Mit122, one was independent of p53 and the other was dependent on p53.
They generated a physical map encompassing the Tlsr12 locus with YAC and BAC clones to isolate new probes and to further narrow down the candidate interval having suspected tumor suppressor genes [46]. Allelic loss mapping with polymorphic markers from YAC and BAC clones finally revealed both alleles of two end markers of one BAC clone were retained, and one marker between the two end markers showed LOH in genomic DNA of lymphoma tissues, strongly suggesting the minimal interval of LOH on chromosome 12 was covered by the BAC clone.
Sequence analysis of the peak LOH region was then conducted. The findings obtained showed the isolation of the Bcl11b gene from this region [47]. Biallelic changes were more frequently detected in p53-proficient lymphomas than in p53-deficient lymphomas, indicating inactivation of the Bcl11b gene is dependent on the presence of functional p53 Cancers 2021, 13, 1026 6 of 13 in the development of lymphoma. The introduction of Bcl11b into cultured tumor cell lines lacking the expression of Bcl11b exhibited a suppressive effect of Bcl11b on tumor cell growth. Taken together, these findings suggested that biallelic mutations in Bcl11b contributed to mouse lymphomagenesis on a p53 wild-type background.
Bcl11b-deficient mice were then generated by a conventional gene targeting method, in which homologous recombinant ES cells were selected in the presence of neomycin. Bcl11bdeficient mice exhibited neonatal death. A significant block in thymocyte differentiation at the CD4 -CD8the double-negative stage was observed in the neonatal thymus of these knockout mice. In contrast, any impairment was not observed in cells of the B or γδ T cell lineage in thymocyte development [48]. In addition, the unsuccessful recombination of V(β) to D(β) was seen in Bcl11b −/− thymocytes, which also lacked the pre-T cell receptor (TCR) complex due to the lack of Tcrb mRNA. Furthermore, markedly increased apoptosis was observed in the neonatal thymus of Bcl11b -/mice. These findings suggested that Bcl11b is essential for both differentiation and survival in the development of thymocytes.
Bcl11b-deficient mice were mated with p53-deficient mice and subjected to radiation. Bcl11b +/− p53 +/− mice subsequently developed significantly more lymphomas than Bcl11b +/+ p53 +/− mice; however, the wild-type Bcl11b allele was retained and expressed in the majority of lymphomas [49]. These findings suggested that Bcl11b was haploinsufficient for the suppression of thymic lymphomagenesis in p53 +/− mice, namely a condition under which the functional loss of only one allele was enough to confer an advantage for tumorigenesis. Furthermore, haploinsufficiency was supported by the findings that impairment in thymocyte development and survival was seen in Bcl11b +/mouse embryos as well as in Bcl11b −/− mice.
Their genome-wide analysis of allelic loss for radiation-induced thymic lymphomas revealed the LOH region (Tlsr11, Figure 1) on the top of chromosome 11 in addition to chromosomes 12 and 16, as mentioned earlier. As mouse genome information was accumulated at that moment, the Ikaros gene was identified in this region from the database [50]. Fine allelic loss mapping around the Ikaros gene in genomic DNA of lymphomas suggested that the critical region of allelic loss was found in the middle of the Ikaros gene. The Nterminal zinc finger and the activation domains of Ikaros exhibited homozygous deletions and mutations. These findings strongly indicated Ikaros plays a key role in mouse thymic lymphomagenesis. Since the Ikaros gene showed biallelic changes at a high frequency in radiation-induced mouse thymic lymphomas, they investigated two other members of the Ikaros gene family, Helios on chromosome 1 and Aiolos on chromosome 11, [51]. Genetic analysis with adjacent MIT markers to the two genes suggested neither locus showed LOH in genomic DNA of the thymic lymphomas on the p53 wild-type background. In contrast, both Helios and Aiolos loci showed LOH on the p53 heterozygous background. These results suggested tumor-suppressive functions of Helios and Aiolos were dependent on p53 loss.

Identification of Tumor Resistance Loci
Resistance to chemically-induced skin tumors is highly dependent on mouse strains [52] ( Table 2). We previously reported that MSM/Ms is a dominant resistant strain to chemicallyinduced skin tumors when they are crossed with a highly susceptible strain, FVB/N (Friend Virus B/NIH), and treated with DMBA/TPA according to the standard two-stage skin carcinogenesis protocol [36]. In order to carry out genome-wide screening of genetic modifiers for DMBA/TPA-induced skin tumors, we generated p53 +/− or p53 +/+ backcross mice between FVB/N and MSM/Ms mice. Approximately 50% of these backcross mice were p53 +/− , which allowed us to screen p53-dependent skin tumor modifier loci as well as p53-non-dependent modifier loci. Genome-wide genetic modifier screening showed a significant linkage to the number of papillomas on chromosomes 6 (Stmm4 for Skin tumor modifier of MSM/Ms) and 7 (Stmm1, 2) and a possible linkage on chromosomes 1 (Stmm12), 3 (Stmm6, 7), 5 (Stmm5), 11 (Stmm11), 12 (Stmm8), 13 (Stmm9), and 17 (Stmm10) (Figure 1). We then classified tumors into three size categories (<2 (hereafter "micro"), 2-6 (hereafter Cancers 2021, 13, 1026 7 of 13 "middle"), and >6 mm (hereafter "large")) and performed linkage analysis to identify stage-dependent linkage loci. The Skts1 locus on chromosome 7 showed a strong linkage in mice that developed micro or middle-sized papillomas. However, we did not see any linkage on chromosome 7 in mice that developed large papillomas, whereas a linkage to large papillomas and carcinomas was seen at a different locus on the chromosome (Stmm3, Figure 1). Stmm3 locus, which was detected around the Cdkn2a/p19 Arf gene showed p53dependency because this locus was detected only in p53 +/+ backcross mice, not detected in p53 +/− backcross mice. Furthermore, a suggestive linkage conferring "susceptibility" to carcinoma was also found on chromosome 5 (Stmm5). This is the only "susceptibility" locus MSM/Ms conferred in our study, although MSM/Ms itself is very resistant. These findings strongly suggested that multiple loci regulate each stage of tumorigenesis, some of them showed p53 dependency.

Tumor Resistance Loci on Chromosome 7
To confirm the presence of tumor resistance loci on chromosome 7, we selected resistant backcross mice and eliminated MSM/Ms genomic DNA by backcrossing the mice to FVB/N mice in order to generate congenic mouse lines spanning the linkage region on chromosome 7 [53]. We first generated congenic mouse lines covering the Stmm1 and Stmm2 regions, respectively. Successive rounds of crossing and repeated DMBA/TPA chemical carcinogenesis experiments on each line allowed us to eliminate the whole Stmm2 region from the candidate interval and to cut down the Stmm1 region to approximately 3 cM regions.
We then checked allelic imbalances on chromosome 7 to investigate the specific location of somatic changes in the Stmm1 region on chromosome 7. Two allelic imbalance peaks were observed within the 3 cM region identified with the multiple congenic lines covering the Stmm1 region. The combination of these two peak regions reduced the total physical size of the Stmm1 region to approximately 5.4 Mb [53].
We continued congenic mapping and DMBA/TPA chemical carcinogenesis and narrowed down the candidate interval of Stmm1 to a region of 3.4 Mb on chromosome 7. Pth (parathyroid hormone) was detected among the genes mapped within the Stmm1 region [54]. PTH is well-known to function cooperatively with vitamin D to regulate calcium and phosphate homeostasis in the blood. However, the role of PTH in skin tumorigenesis has not yet been elucidated in detail. Previous studies reported that PTH slowed down the proliferation and differentiation of cells in the epidermis, suggesting its role in skin tumorigenesis [55]. We first measured the concentration of intact PTH (iPTH) in sera of cancer-resistant MSM/MS and susceptible FVB mice. As a result, significantly higher iPTH level was detected in sera from MSM/Ms than from FVB/NJ mice. Therefore, skin carcinogenesis experiments were carried out with MSM-BAC transgenic (Pth MSM -Tg) and Pth knockout heterozygous mice (Pth +/− ). Pth MSM -Tg mice developed a significantly lower number of tumors compared to the wild-type mice. In contrast, Pth +/− mice developed a significantly higher number of tumors. These results strongly suggested iPTH in sera regulated skin tumorigenesis in a dose-dependent manner. Moreover, differentiation markers, such as Loricrine and Keratine 10 were highly expressed in the epidermis of Pth MSM -Tg, suggesting PTH promoted differentiation of cells in the epidermis. Furthermore, the in vitro experiments showed the coding SNP (rs51104087, Val28Met) in the mouse Pro-PTH encoding region enhanced its processing and secretion of PTH as well as intracellular calcium levels. Collectively, these findings demonstrated that PTH promotes terminal differentiation in keratinocytes by increasing intracellular calcium levels in keratinocytes, which results in resistance to tumors (Figure 2). as Loricrine and Keratine 10 were highly expressed in the epidermis of Pth -Tg, sug-gesting PTH promoted differentiation of cells in the epidermis. Furthermore, the in vitro experiments showed the coding SNP (rs51104087, Val28Met) in the mouse Pro-PTH encoding region enhanced its processing and secretion of PTH as well as intracellular calcium levels. Collectively, these findings demonstrated that PTH promotes terminal differentiation in keratinocytes by increasing intracellular calcium levels in keratinocytes, which results in resistance to tumors (Figure 2). and Stmm1b (about 4.7 Mb). Both congenic lines including Stmm1a and Stmm1b exhibited a strong suppressive effect on papilloma development, respectively. Pth was identified in the Stmm1b region. Pak1 (serine/threonine p21-activated kinases 1) was identified in Stmm1a region [56]. It is well-known to exhibit oncogenic activity in several cancers. Therefore, Pak1 knockout mice were generated using ES cells from MSM/Ms with the CRISPR/Cas9 system, and DMBA/TPA skin carcinogenesis experiments were performed using Pak1 +/− F1 (FVB/N × MSM/Ms) mice. As a result, Pak1 +/− mice developed almost no tumor. Immunohistochemistry revealed that PAK1 was strongly expressed in Langerhans cells (LCs) as well as in keratinocytes. Furthermore, Pak1 homozygous knockout mice on MSM/Ms background (Pak1 −/− MSM/Ms − ) showed a significant decrease in the number of LCs. F1-Pak1 +/− mice exhibited a decrease in the number of epidermal stem cells in the skin bulge and an increase in the number of Th17 cells in the skin. We generated Pak1 knockdown cells using LC-derived XS52 cells (XS52-Pak1KD) and carried out co-culture experiments with keratinocyte-derived C5N cells. As a result, the proliferation of C5N cells was and Stmm1b (about 4.7 Mb). Both congenic lines including Stmm1a and Stmm1b exhibited a strong suppressive effect on papilloma development, respectively. Pth was identified in the Stmm1b region. Pak1 (serine/threonine p21-activated kinases 1) was identified in Stmm1a region [56]. It is well-known to exhibit oncogenic activity in several cancers. Therefore, Pak1 knockout mice were generated using ES cells from MSM/Ms with the CRISPR/Cas9 system, and DMBA/TPA skin carcinogenesis experiments were performed using Pak1 +/− F 1 (FVB/N × MSM/Ms) mice. As a result, Pak1 +/− mice developed almost no tumor. Immunohistochemistry revealed that PAK1 was strongly expressed in Langerhans cells (LCs) as well as in keratinocytes. Furthermore, Pak1 homozygous knockout mice on MSM/Ms background (Pak1 −/− MSM/Ms − ) showed a significant decrease in the number of LCs. F 1 -Pak1 +/− mice exhibited a decrease in the number of epidermal stem cells in the skin bulge and an increase in the number of Th17 cells in the skin. We generated Pak1 knockdown cells using LC-derived XS52 cells (XS52-Pak1KD) and carried out co-culture experiments with keratinocyte-derived C5N cells. As a result, the proliferation of C5N cells was significantly enhanced in the presence of supernatants of XS52-Pak1KD cells. Taken together, these findings indicated that Pak1 was required for the maintenance of epidermal stem cells, which showed an abnormal growth in the absence of Pak1 in LCs and were not maintained correctly, resulting in the resistance to tumors (Figure 3).

Tumor Resistance Loci on Chromosome 4
In our previous genome-wide linkage study to map genetic modifiers conferring resistance to DMBA/TPA-induced skin tumors, Stmm3 was found to possess loci that conferred strong resistance to larger papillomas on chromosome 4. To confirm the presence of the tumor resistance loci identified in this region on chromosome 4, we generated congenic mice harboring this region of the MSM genome on the FVB/N background [57]. DMBA-TPA carcinogenesis experiments on each line cut down physical interval to less than 34 Mb on proximal chromosome 4. We also examined somatic changes in the tumors of congenic mice to further cut down the interval. As a result, allelic imbalances, high frequencies of MSM allele loss, or FVB allele gain were detected, suggesting that a physical distance was cut down to approximately 25 Mb. significantly enhanced in the presence of supernatants of XS52-Pak1KD cells. T gether, these findings indicated that Pak1 was required for the maintenance of ep stem cells, which showed an abnormal growth in the absence of Pak1 in LCs and w maintained correctly, resulting in the resistance to tumors (Figure 3).

Tumor Resistance Loci on Chromosome 4
In our previous genome-wide linkage study to map genetic modifiers confe sistance to DMBA/TPA-induced skin tumors, Stmm3 was found to possess loci t ferred strong resistance to larger papillomas on chromosome 4. To confirm the p of the tumor resistance loci identified in this region on chromosome 4, we genera genic mice harboring this region of the MSM genome on the FVB/N backgrou DMBA-TPA carcinogenesis experiments on each line cut down physical interva than 34 Mb on proximal chromosome 4. We also examined somatic changes in the of congenic mice to further cut down the interval. As a result, allelic imbalanc frequencies of MSM allele loss, or FVB allele gain were detected, suggesting that a distance was cut down to approximately 25 Mb.
A congenic line with the minimal interval was crossed with p53 +/− FVB mice a mice were treated with DMBA/TPA. The findings obtained revealed strong inhib fects on the development of papilloma in p53 +/+ congenic mice. In contrast, papill velopment was not markedly suppressed in p53 +/− congenic mice. Therefore, the ca gene was expected to be functionally dependent on p53, which was consistent w previous findings by the initial linkage analysis of backcross mice [36]. Cdkn2a, dependent kinase inhibitor gene, was detected in the minimal interval. Two differ teins, p16 Ink4a and p19 Arf are encoded by this locus. These two proteins share exons produced from alternative splicing. These two well-known tumor suppressors, p1 p19 Arf were previously shown to promote the growth-inhibitory effect of pRb and tein, respectively. In addition, the stimulation of separate promoters in the upst exon1α (encoding p16 Ink4a ) and exon1β (encoding p19 Arf ) initiated RB-and p53-de programs, respectively [58,59]. As our initial linkage and the congenic study su the gene responsible Stmm3 was functionally dependent on p53, p19 Arf seemed mo A congenic line with the minimal interval was crossed with p53 +/− FVB mice and these mice were treated with DMBA/TPA. The findings obtained revealed strong inhibitory effects on the development of papilloma in p53 +/+ congenic mice. In contrast, papilloma development was not markedly suppressed in p53 +/− congenic mice. Therefore, the candidate gene was expected to be functionally dependent on p53, which was consistent with our previous findings by the initial linkage analysis of backcross mice [36]. Cdkn2a, a cyclin-dependent kinase inhibitor gene, was detected in the minimal interval. Two different proteins, p16 Ink4a and p19 Arf are encoded by this locus. These two proteins share exons and are produced from alternative splicing. These two well-known tumor suppressors, p16 Ink4a and p19 Arf were previously shown to promote the growth-inhibitory effect of pRb and p53 protein, respectively. In addition, the stimulation of separate promoters in the upstream of exon1α (encoding p16 Ink4a ) and exon1β (encoding p19 Arf ) initiated RBand p53-dependent programs, respectively [58,59]. As our initial linkage and the congenic study suggested the gene responsible Stmm3 was functionally dependent on p53, p19 Arf seemed more likely to be responsible for Stmm3 than p16 Ink4a . To prove p19 Arf is responsible for Stmm3 and exclude the possibility that p19 Arf is responsible, we eliminated MSM/Ms alleles of p16 Ink4a and p19 Arf respectively in MSM/Ms ES cell lines with CRISPR/Cas9 system and generated p16 Ink4a and p19 Arf knockout mice on the MSM/Ms background. These mice were crossed with FVB/N mice to generate p16 Ink4aFVB/-(p16 Ink4aMSM allele knockout F 1 ) and p19 ArfFVB/-(p19 ArfMSM allele knockout F 1 ) mice. They were treated with DMBA/TPA. As a result, a significantly higher number of papillomas developed in p19 ArfFVB/− mice. In contrast, the number of papillomas that developed in p16 Ink4aFVB/− mice did not markedly increase. Therefore, the p19 ArfMSM allele appeared to confer greater resistance to the development of papilloma than the p16 Ink4aMSM allele [60], which indicated that p19 Arf was more likely to be the responsible gene for Stmm3 than p16 Ink4a . We then separately knocked out the MSM/M allele of p19 Arf (p19 ArfMSM ) and FVB/N allele of p19 Arf (p19 ArfFVB ) on (FVB/N × MSM/) F 1 background and subjected these mice to DMBA/TPA skin carcinogenesis. The numbers of total papillomas and larger papillomas were significantly higher in MSM/Ms allele knockout mice (p19 Arf FVB/-). These results indicated MSM/Ms allele of p19 Arf more strongly inhibited papilloma genesis and growth than the FVB/N allele, confirming MSM/Ms allele was a resistance gene. We also showed that the p53 pathway was more efficiently activated by the p19 ArfMSM allele than by the p19 ArfFVB allele in vitro (Figure 4). Furthermore, novel polymorphisms in human CDKN2A that were near the SNP in mouse Cdkn2a were shown to be associated with the risk of human breast cancers. These results strongly suggested that linkage study started with MSM/Ms mice allowed us to identify the responsible gene for Stmm3 locus, and finally proved its human orthologue functioned as a tumor resistance/susceptibility gene also in humans.
sible gene for Stmm3 than p16 . We then separately knocked out the MSM/M allele of p19 Arf (p19 ArfMSM ) and FVB/N allele of p19 Arf (p19 ArfFVB ) on (FVB/N × MSM/) F1 background and subjected these mice to DMBA/TPA skin carcinogenesis. The numbers of total papillomas and larger papillomas were significantly higher in MSM/Ms allele knockout mice (p19 Arf FVB/-). These results indicated MSM/Ms allele of p19 Arf more strongly inhibited papilloma genesis and growth than the FVB/N allele, confirming MSM/Ms allele was a resistance gene. We also showed that the p53 pathway was more efficiently activated by the p19 ArfMSM allele than by the p19 ArfFVB allele in vitro ( Figure 4). Furthermore, novel polymorphisms in human CDKN2A that were near the SNP in mouse Cdkn2a were shown to be associated with the risk of human breast cancers. These results strongly suggested that linkage study started with MSM/Ms mice allowed us to identify the responsible gene for Stmm3 locus, and finally proved its human orthologue functioned as a tumor resistance/susceptibility gene also in humans.

Closing Remarks
MSM/Ms mice were originally captured in Mishima, Japan in the 1970s. After longterm repeated sister-brother mating, the mice were established as a pure inbred mouse strain and have been utilized as a valuable genetic tool for mouse genetics for the past 40 years, mainly in Japan. MSM/Ms mice were originally used mainly for genetic mapping by SSLP markers because of their extensive polymorphism with standard inbred mouse strains. In the field of cancer biology, tumor cell lines were often established from F1 mice between MSM/Ms and inbred mice and used for LOH analysis. As SSLP markers were established, they were used for positional cloning of single trait mutants as well as multigenic traits, including cancer. In this review, we focused on radiation-induced lymphomas

Closing Remarks
MSM/Ms mice were originally captured in Mishima, Japan in the 1970s. After longterm repeated sister-brother mating, the mice were established as a pure inbred mouse strain and have been utilized as a valuable genetic tool for mouse genetics for the past 40 years, mainly in Japan. MSM/Ms mice were originally used mainly for genetic mapping by SSLP markers because of their extensive polymorphism with standard inbred mouse strains. In the field of cancer biology, tumor cell lines were often established from F1 mice between MSM/Ms and inbred mice and used for LOH analysis. As SSLP markers were established, they were used for positional cloning of single trait mutants as well as multigenic traits, including cancer. In this review, we focused on radiation-induced lymphomas and chemically-induced skin tumors. A lot of other cancers and multigenic diseases were examined with MSM/Ms. However, a very limited number of studies reached particular responsible genes. In the majority of studies, genetic loci were mapped on chromosomes, but particular genes were not identified, because of the lack of genome information previously. In other words, a large number of interesting genes are still yet to be identified in MSM/Ms. Genetic resources are already established with MSM/Ms, such as the complete genome sequence information [21], a full set of mouse consomic strains [61], the BAC clone library [24], Microsatellite database [23], and ES cells [37]. Using currently available genetic tools developed from MSM/Ms will allow us to identify responsible genes and polymorphisms for complex traits including cancer, which were previously difficult to handle.

Conclusions
Recently, new carcinogenesis models, such as organoids, were developed and nextgeneration sequencing of human samples was widely carried out. However, the benefit of mouse models is that all the experiments can be carried out in a situation where the immune system and the tumor microenvironment are maintained in the same way as in humans. The important thing is that several carcinogenesis models complement one another.
Finally, the integration of this information from carcinogenesis models and human GWAS information will facilitate the identification of human cancer susceptibility genes and polymorphisms, which will enable us to predict cancer risk and develop strategies for targeted therapy Author Contributions: K.O. wrote the manuscript. M.S. and E.I. collected data from the literature. Y.W. revised the manuscript. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement:
This study was conducted according to the guidelines of the Care and Use of Laboratory Animals of the Ministry of Education, Culture, Sports, Science, and Technology of Japan, and approved by the Ethics Committee of Chiba Cancer Center (protocol code: 19-13, approved on 24 July 2019).
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.