Evidence of a possible epigenetic inactivation mechanism ... - Nature

ã

Oncogene (2001) 20, 2186 ± 2189 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

Evidence of a possible epigenetic inactivation mechanism operating on a region of mouse chromosome 19 in g-radiation-induced thymic lymphomas Javier Santos1,2, Michel Herranz1,2, MoÂnica FernaÂndez1, ConcepcioÂn Vaquero1, Pilar LoÂpez1 and JoseÂ FernaÂndez-Piqueras*,1 1 Departmento de BiologõÂa, Laboratorio de GeneÂtica Molecular Humana, Facultad de Ciencias, Universidad AutoÂnoma de Madrid, 28049-Madrid, Spain

Loss of heterozygosity (LOH) analysis, performed in 68 g-radiation-induced primary thymic lymphomas of F1 hybrid mice, provided evidence of signi®cant LOH on chromosome 19 in a region de®ned by the D19Mit106 (22 cM) and D19Mit100 (27 cM) markers (Thymic Lymphoma Suppressor Region 8, TLSR8). Cd95 and Pten, two genes mapped at this region, were inactivated in a vast majority of these tumors (85.3% for Cd95 and 61.8% for Pten). Moreover, altered expression of Cd95 and Pten occurred concomitantly in 34 of 68 (50%) thymic lymphomas suggesting a coordinated mechanism of inactivation of these genes. Surprisingly, we also found that Jak2, a proto-oncogene located between Cd95 and Pten, was simultaneously inactivated in a signi®cant fraction of the tumors analysed (24 of 34, 70.6%). Taken together these ®ndings and the lack of mutations in the coding sequences of the mentioned genes clearly suggest a possible regional epigenetic inactivation mechanism on mouse chromosome 19 operating during the development of these tumors. Oncogene (2001) 20, 2186 ± 2189. Keywords: Pten; Cd95; Jak2; epigenetic inactivation; girradiation; mouse thymic lymphomas; chromosome 19

We have developed an experimental model based on the use of F1 hybrid mice for detecting potential tumor suppressor gene loci involved in g-radiation-induced lymphomagenesis by revealing loss of heterozygosity (LOH). A genome-wide scan for LOH with a panel of microsatellite markers served us to identify ®ve distinct thymic lymphoma suppressor gene regions on chromosome 4, named as TLSR1-5 (Santos et al., 1996, 1998; MeleÂndez et al., 1999), and the involvement of the p16Ink4a, p15Ink4b and p73 genes in the more advanced stages of thymic lymphomagenesis (Malumbres et al., 1997, 1999; Herranz et al., 1999).

*Correspondence: J FernaÂndez-Piqueras 2 These authors contributed equally to this work. Received 26 September 2000; revised 26 December 2000; accepted 25 January 2001

In this work, an extension of our initial allelotype analysis of tumors from F1 hybrids demonstrated the occurrence of frequent losses of heterozygosity on chromosome 19, de®ning a new critical region which contains the Cd95 receptor (Fas antigen) and the tumor suppressor Pten as candidate genes (Watanabe-Fukunaga et al., 1992; Suzuki et al., 1998; Podsypanina et al., 1999). Surprisingly, we have found that Cd95 and Pten but also Jak2, a proto-oncogene (Yu et al., 1997; Liu et al., 2000) which is located between both genes (Mouse Genome Database, http://www.informatics. jax.org), are simultaneously inactivated in a signi®cant fraction of the tumors. DNAs from 68 frank thymic lymphomas (the more advanced disease stage), induced by g-rays in (C57BL/ 6J6BALB/cJ) F1 mice (48 tumors) and the reciprocal strain (BALB/cJ6C57BL/6J) F1 (20 tumors), were analysed for allelic losses on chromosome 19 with ®ve informative markers (D19Mit19, D19Mit40, D19Mit46, D19Mit100, and D19Mit106). Upon screening, 20 of 68 (29.4%) frank thymic lymphomas showed LOH. No homozygous deletions were found in these tumors. To determine the signi®cance of these allelic losses, we genotyped the same tumor set with other six unlinked polymorphic microsatellites located on chromosomes 1 (Crp), 5 (D5Mit10), 13 (D13Mit36), 15 (D15Mit28), 16 (D16Mit122) and 17 (D17Mit20). All tumors analysed retained heterozygosity for markers located on 5, 13, 15 and 17. LOH was very infrequent at Crp (1/68) and D16Mit122 (4/68). Comparative analyses between LOH occurring on each chromosome 19 marker and those appearing on the other six chromosomes were performed by a w2 contingency test. This analysis revealed that LOH detected for the ®ve markers on chromosome 19 were clearly signi®cant (w2 values ranged from 40.11 ± 68.76, d.f.=1, P50.001). Thymic lymphomas with allelic losses on chromosome 19 (20 tumors) were classi®ed into six dierent categories (I ± VI) (Figure 1). This allelotype analysis revealed the existence of two dierent LOH regions. One encompassed from D19Mit46 (33 cM) to the telomere, which might be coincident with that described previously in X-radiation-induced thymic lymphomas of (BALB/ cHeA6STS/A) F1 hybrid mice (Okumoto et al., 1999). The second one was a new more proximal LOH region of about 5 cM between the D19Mit106

Regional epigenetic inactivation in g-radiation-induced mouse thymic lymphomas J Santos et al

2187

Figure 1 Allelic loss mapping on mouse chromosome 19 in g-radiation-induced thymic lymphomas from (C57BL/6J6BALB/cJ) F1 mice and the reciprocal strain (BALB/cJ6C57BL/6J) F1. The black bar to the left of the chromosome 19 indicates the critical LOH region (TLSR8). All microsatellite markers are informative in these crosses and positioned according to the genetic linkage map data obtained from the Mouse Genome Database. Thymic lymphomas are classi®ed into six categories (LOH patterns I ± VI). The number of tumors in each LOH class is given in parenthesis. White boxes represent retention of the two alleles and black boxes indicate LOH. Words in the black boxes denote the loss of one allele: BAL, BALB/cJ; B6, C57BL/6J. Allele typings and scoring of LOH were performed as we described previously (Santos et al., 1996)

(22 cM) and D19Mit100 (27 cM) markers. According to the nomenclature established for this type of regions and taking into account those found on mouse chromosomes 12 (TLSR6) and 16 (TLSR7) (Matsumoto et al., 1998), we named this new LOH region as TLSR8. Interestingly, two candidate genes, the pro-apoptotic Cd95 receptor gene (23 cM) and the tumor suppressor Pten gene (25 cM), lie within TLSR8 (Mouse Genome Database). Thus, we examined these genes to determine whether they are inactivated in g-radiationinduced thymic lymphomas. In regard to the analysis of Pten, we ®rst ruled out the possible existence of a non-translated processed cPten in the mouse genome (Figure 2a), similar to that described in humans (Dahia et al., 1998). To examine this possibility, we performed ampli®cations from genomic DNA obtained from three normal thymuses of F1 hybrid mice in the presence of oligonucleotides that amplify an intronless DNA sequence of 240 bp containing exons 1 ± 4, and we did not ®nd PCR product by gel electrophoresis (Figure 2a). PCR ampli®cations of both Pten exon 5 on genomic DNA and Pten exons 1 ± 4 on cDNA from the same three F1 hybrids were used as internal control of the DNA quality and ampli®cation. So doing, the functional status of Pten was studied by RT ± PCR (Figure 2b). Pten expression was clearly patent in the non-treated parental and F1 hybrid individuals. By contrast, we found that 42 of 68 (61.8%) thymic lymphomas had altered expression. Thirty-eight of them showed complete loss of expression and only four tumors exhibited low transcript levels. With

respect to the another candidate gene, Cd95 expression was also detected in all normal samples while lack of Cd95 mRNA was observed in 58 of 68 (85.3%) thymic lymphomas (Figure 2c). These results clearly suggest that the Pten and Cd95 inactivation contribute to the development of g-radiation-induced thymic lymphomas and, therefore, they might be the candidate genes for TLSR8. This assumption is consistent with previous evidence supporting the involvement of these genes in mouse lymphomagenesis. Thus, inactivation of Cd95 has been related to lymphoproliferation in lpr knockout mice (Watanabe-Fukunaga et al., 1992). In addition, Pten +/7 mice are predisposed to develop T-cell lymphomas (Suzuki et al., 1998; Podsypanina et al., 1999). Finally, Pten was found inactivated in some human haematological malignancies (Dahia et al., 1999). Mutational analysis at the coding regions of these two genes, performed on their cDNA sequences (GeneBank accession numbers U92437 and M83649) by SSCP/sequencing, did not reveal any alteration in the tumor samples (data not shown). The absence of mutations in the coding regions of Pten and Cd95 might question the relevance of these genes, but their lack of expression could be explained by some type of epigenetic mechanism related to promoter hypermethylation. Regarding the human model, the inactivation of PTEN in prostate cancer was not associated with promoter methylation (Cairns et al., 1997), despite a putative PTEN reactivation by the demethylating agent 5-azadeoxycytidine has been described (Whang et al., 1998). In human colorectal carcinomas CD95 expresOncogene


2188

Figure 2 Expression analyses of Pten, Cd95 and Jak2. (a). Exclusion of a processed cPten in the mouse genome. Lanes 1, 4 and 7 show an intronless DNA fragment of 248 bp covering exons 1 ± 4 ampli®ed from normal cDNA samples. Lanes 2, 5 and 8 indicate the lack of ampli®cation of this fragment when normal genomic DNA was used as template. Pten exon 5 ampli®cations with internal exonic primers were used as control of genomic DNA (lanes 3, 6 and 9). (b) RT ± PCR analysis of Pten on representative frank lymphomas. (c) Representative analysis of Cd95 by RT ± PCR in frank lymphomas. (d) RT ± PCR analysis of Jak2. Cdk4 was used as a control of RT ± PCR and the loading levels. Tumors BC6A and BC13B, showed complete loss expression of Pten, Cd95 and Jak2. RT ± PCR reactions were performed by using the Enhanced Avian RT ± PCR Kit (Sigma, Saint Louis, MO) with the following primers: for Pten, 5'TGACAGCCATCATCAAAGAGA-3' (forward), and 5'CTGCAGTTAAATTTGGCGGT-3' (reverse); for Cd95, 5'TGCAGTTGCTGAGATGAACC-3' (forward) and 5'GGAAGGTCTTCAATTAACTGCG-3' (reverse); and for Jak2, 5'-TCTGGCAACAGACAAGTGGA-3' (forward) and 5'GCTCTCCGTCAAGGATTGAG-3' (reverse). The co-ampli®cation of the mouse Cdk4 gene was carried out by using the primers 5'-TGGCTGCCACTCGATATGAAC-3' (forward) and 5'CCTCAGGTCCTGGTCTATATG-3' (reverse). The primers used for Cd95 were those published in the Mouse Genome Database (MGI: 1199209). Detection of the PCR products was performed by direct ethidium bromide staining of 1.5% agarose gels. Quanti®cation of PCR products was done as we described before (Malumbres et al., 1997)

sion may be lost but not due to methylation (Butler et al., 2000). Interestingly, we have previously found that g-radiation-induced thymic lymphomas exhibit both no mutations and promoter hypermethylation associated with loss of expression of the p15INK4b and p16INK4a genes (Malumbres et al., 1997). Moreover, emerging data suggest that exposure to ionizing radiation triggers changes in DNA methylation patterns (Dubrova et al., 2000). In this context, it is worth noting that aberrant epigenetic modi®cations have been widely described in human cancers (Feinberg, 1993; Feinberg et al., 1995), and that a growing list of tumor suppressor genes can be inactivated in association with abnormal methylation of gene promoter regions (Baylin and Herman, 2000; Dammann et al., 2000). Oncogene

On the other hand, it is particularly intriguing that alterations in Pten and Cd95 occurred concomitantly in a signi®cant fraction of frank thymic lymphomas (34 of 68 (50%), w2=17.46, d.f.=1, P50.001). Since these genes are located in an interval of 2 cM (Mouse Genome Database), a possible explanation for this ®nding is that Pten and Cd95 may be coordinately regulated as consequence of an abnormal regional epigenetic mechanism operating in the development of these tumors. If our hypothesis is true any gene located within the 2 cM interval de®ned by Cd95 and Pten should be inactivated in those thymic lymphomas exhibiting simultaneous lack of expression of both genes. According to the Mouse Genome Database the only gene mapping at this region is Jak2, a protooncogene involved in the prevention of g-radiationinduced apoptosis (Liu et al., 2000), which has been associated with the development of some mouse T-cell lymphomas (Yu et al., 1997). Contrary to what expected for a proto-oncogene, we have found that most of thymic lymphomas exhibited simultaneous inactivation of Cd95, Pten and Jak2 (24 of 34 (70.6%), w2=8.01, d.f.=1, P=0.004) (Figure 2d). Thus, these data strongly suggest that a mouse chromosome 19 region extending from Cd95 to Pten is epigenetically silenced in a signi®cant fraction of g-radiation-induced thymic lymphomas. In this context, a minority of tumors showing no simultaneous inactivation of Cd95, Jak2 and Pten might be explained by the disruption of speci®c regulatory regions produced by the exposure to g-irradiation (Sachs et al., 2000). De®nitive con®rmation of such a mechanism would require the molecular analysis of the promoter regions of the mouse Cd95, Jak2 and Pten genes in order to correlate loss of gene expression with CpG island hypermethylation. Unfortunately, these sequences have not yet been described. In summary, data presented here represent, to our knowledge, the ®rst report on the existence of speci®c allelic losses de®ning the critical region TLSR8 in mouse tumors, and the involvement of a possible regional epigenetic inactivation on mouse chromosome 19 operating during the development of g-radiationinduced thymic lymphomas.

Acknowledgments This work was supported in part by grants PM99/003 (Ministerio EducacioÂn y Cultura, Spain) and 08/0043/1998 (Comunidad AutoÂnoma de Madrid, Spain) to J FernaÂndezPiqueras, and BMH4-98-3426 (BIOMED 2 program, European Union) and 08/0008/1999 (Comunidad AutoÂnoma de Madrid, Spain) to J Santos.


2189

References Baylin SB and Herman JG. (2000). Trends Genet., 16, 168 ± 174. Butler LM, Dobrovic A and Cowled PA. (2000). Br. J. Cancer, 82, 131 ± 135. Cairns P, Okami K, Halachmi S, Halachmi N, Esteller M, Herman JG, Jen J, Isaacs WB, Bova GS and Sidransky D. (1997). Cancer Res., 57, 4997 ± 5000. Dahia PL, Fitzgerald MG, Zhang X, Marsh DJ, Zheng Z, Pietsh T, von Deimling A, Haluska FG, Haber DA and Eng C. (1998). Oncogene, 16, 2403 ± 2406. Dahia PLM, Aguiar RC, Alberta J, Kum JB, Caron S, Sill H, Marsh DJ, Ritz J, Freedman A, Stiles C and Eng C. (1999). Hum. Mol. Genet., 8, 185 ± 193. Dammann R, Li C, Yoon J-H, Chin PL, Bates S and Pfeifer GP. (2000). Nature Genet., 25, 315 ± 319. Dubrova YE, Plumb M, Gutierrez B, Boulton E and Jereys AJ. (2000). Nature, 405, 37. Feinberg AP. (1993). Nature Genet., 4, 110 ± 113. Feinberg AP, Rainier S and DeBaunn MR. (1995). J. Natl. Cancer Inst. Monogr., 17, 21 ± 26. Herranz M, Santos J, Salido E, FernaÂndez-Piqueras J and Serrano M. (1999). Cancer Res., 59, 2068 ± 2071. Liu R, Liu C-B, Golam Mohi M, Arai K and Watanabe S. (2000). Oncogene, 19, 571 ± 579. Malumbres M, PeÂrez de Castro I, Santos J, MeleÂndez B, Mangues R, Serrano M, Pellicer A and FernaÂndezPiqueras J. (1997). Oncogene, 14, 1361 ± 1370. Malumbres M, PeÂrez de Castro I, Santos J, FernaÂndezPiqueras J and Pellicer A. (1999). Oncogene, 18, 385 ± 396.

Matsumoto Y, Kosugi S, Shinbo T, Chou D, Ohashi M, Wakabayashi Y, Sakai K, Okumoto M, Mori N, Aizawa S, Niwa O and Kominami R. (1998). Oncogene, 16, 2747 ± 2754. MeleÂndez B, Santos J and FernaÂndez-Piqueras J. (1999). Oncogene, 18, 4166 ± 4169. Okumoto M, Park Y-G, Song C-W and Mori N. (1999). Cancer Lett., 135, 223 ± 228. Podsypanina K, Hedrick-Ellenson L, Nemes A, Gu J, Tamura M, Yamada KM, CordoÂn-Cardo C, Catoretti G and Fisher P. (1999). Proc. Natl. Acad. Sci. USA, 96, 1563 ± 1568. Sachs RK, Hiatky LR and Trask BJ. (2000). Trends Genet., 16, 143 ± 146. Santos J, PeÂrez de Castro I, Herranz M, Pellicer A and FernaÂndez-Piqueras J. (1996). Oncogene, 12, 669 ± 676. Santos J, Herranz M, PeÂrez de Castro I, Pellicer A and FernaÂndez-Piqueras J. (1998). Oncogene, 17, 925 ± 929. Suzuki A, De la Pompa JL, Stambolic V, Elia AJ, Sasaki T, Del Barco-Barrantes I, Ho A, Wakeham A, Itie A, Khoo W, Fukumoto M and Mak TW. (1998). Curr. Biol., 8, 1169 ± 1178. Watanabe-Fukunaga R, Brannan CI, Copeland NG, Jenkins NA and Nagata S. (1992). Nature, 356, 314 ± 317. Whang YE, Wu X, Suzuki H, Reiter RE, Tran C, Vessella RL, Said JW, Isaacs WB and Sawyers CL. (1998). Proc. Natl. Acad. Sci. USA, 95, 5246 ± 5250. Yu C-L, Jove R and Burako SJ. (1997). J. Immunol., 159, 5206 ± 5210.

Oncogene