Produção Nacional
Revisado por pares
Inactivation of Dnmt3b in Mouse Embryonic Fibroblasts Results in DNA Hypomethylation, Chromosomal Instability, and Spontaneous Immortalization
2005; Elsevier BV; Volume: 280; Issue: 18 Linguagem: Inglês
10.1074/jbc.m413246200
ISSN1083-351X
AutoresJonathan Dodge, Masaki Okano, Frederick A. Dick, Naomi Tsujimoto, Taiping Chen, Shumei Wang, Yoshihide Ueda, Nick Dyson, En Li,
Tópico(s)Genetic Syndromes and Imprinting
ResumoDNA hypomethylation is a hallmark of many types of solid tumors. However, it remains elusive how DNA hypomethylation may contribute to tumorigenesis. In this study, we have investigated how targeted disruption of the DNA methyltransferases Dnmt3a and Dnmt3b affects the growth of mouse embryonic fibroblasts (MEFs). Our studies led to the following observations. 1) Constitutive or conditional deletion of Dnmt3b, but not Dnmt3a, resulted in partial loss of DNA methylation throughout the genome, suggesting that Dnmt3b, in addition to the major maintenance methyltransferase Dnmt1, is required for maintaining DNA methylation in MEF cells. 2) Dnmt3b-deficient MEF cells showed aneuploidy and polyploidy, chromosomal breaks, and fusions. 3) Inactivation of Dnmt3b resulted in either premature senescence or spontaneous immortalization of MEF cells. 4) The G1 to S-phase checkpoint was intact in primary and spontaneously immortalized Dnmt3b-deficient MEFs because the p53 protein was inducible by DNA damage. Interestingly, protein levels of the cyclindependent kinase inhibitor p21 were increased in immortalized Dnmt3b-deficient MEFs even in the absence of p53 induction. These results suggest that DNA hypomethylation may induce genomic instability, which in turn leads to spontaneous immortalization or premature senescence of Dnmt3b-deficient MEFs via a p53-independent mechanism. DNA hypomethylation is a hallmark of many types of solid tumors. However, it remains elusive how DNA hypomethylation may contribute to tumorigenesis. In this study, we have investigated how targeted disruption of the DNA methyltransferases Dnmt3a and Dnmt3b affects the growth of mouse embryonic fibroblasts (MEFs). Our studies led to the following observations. 1) Constitutive or conditional deletion of Dnmt3b, but not Dnmt3a, resulted in partial loss of DNA methylation throughout the genome, suggesting that Dnmt3b, in addition to the major maintenance methyltransferase Dnmt1, is required for maintaining DNA methylation in MEF cells. 2) Dnmt3b-deficient MEF cells showed aneuploidy and polyploidy, chromosomal breaks, and fusions. 3) Inactivation of Dnmt3b resulted in either premature senescence or spontaneous immortalization of MEF cells. 4) The G1 to S-phase checkpoint was intact in primary and spontaneously immortalized Dnmt3b-deficient MEFs because the p53 protein was inducible by DNA damage. Interestingly, protein levels of the cyclindependent kinase inhibitor p21 were increased in immortalized Dnmt3b-deficient MEFs even in the absence of p53 induction. These results suggest that DNA hypomethylation may induce genomic instability, which in turn leads to spontaneous immortalization or premature senescence of Dnmt3b-deficient MEFs via a p53-independent mechanism. Mammalian DNA (cytosine-5) methyltransferases Dnmt1, Dnmt3a, and Dnmt3b catalyze methylation of CpG dinucleotides in genomic DNA. Genetic studies of mice with mutations of these three Dnmt genes have shown that DNA methylation is essential for embryonic development, establishment and maintenance of allele-specific expression of imprinted genes, repression of inactivated X chromosome in female cells, and repression of endogenous viruses and transposable elements (1Chen T. Li E. Curr. Top. Dev. Biol. 2004; 60: 55-89Crossref PubMed Scopus (255) Google Scholar). Complete inactivation of Dnmt1 by gene targeting does not affect ES 1The abbreviations used are: ES, embryonic stem; MEF, mouse embryonic fibroblast; SA, senescence-associated; dpc, days post coitum. 1The abbreviations used are: ES, embryonic stem; MEF, mouse embryonic fibroblast; SA, senescence-associated; dpc, days post coitum. cell viability. However, Dnmt1-/- embryos die at ∼9.5 days post coitum (dpc), and mouse embryonic fibroblasts (MEFs) that lack Dnmt1 (generated by conditional deletion of Dnmt1) die after a few cell divisions. Inactivation of Dnmt1 in ES cells, embryos, and MEFs results in a genome-wide loss of DNA methylation (2Li E. Bestor T.H. Jaenisch R. Cell. 1992; 69: 915-926Abstract Full Text PDF PubMed Scopus (3201) Google Scholar, 3Lei H. Oh S.P. Okano M. Juttermann R. Goss K.A. Jaenisch R. Li E. Development (Camb.). 1996; 122: 3195-3205Crossref PubMed Google Scholar, 4Jackson-Grusby L. Beard C. Possemato R. Tudor M. Fambrough D. Csankovszki G. Dausman J. Lee P. Wilson C. Lander E. Jaenisch R. Nat. Genet. 2001; 27: 31-39Crossref PubMed Scopus (565) Google Scholar). Inactivation of Dnmt3a results in multiple organ defects and lethality of homozygous mice several weeks after birth, without significant changes in DNA methylation (5Okano M. Bell D.W. Haber D.A. Li E. Cell. 1999; 99: 247-257Abstract Full Text Full Text PDF PubMed Scopus (4449) Google Scholar). Disruption of Dnmt3b results in embryonic lethality at ∼13.5 dpc and hypomethylation of the centromeric minor satellite repeats. In addition, analysis of Dnmt3b-/- 9.5-dpc embryos demonstrates that Dnmt3b plays a major role in de novo methylation of the genome (5Okano M. Bell D.W. Haber D.A. Li E. Cell. 1999; 99: 247-257Abstract Full Text Full Text PDF PubMed Scopus (4449) Google Scholar). Similar to gene targeting of Dnmt1, inactivation of both Dnmt3a and Dnmt3b results in embryonic lethality at ∼9.5 dpc (5Okano M. Bell D.W. Haber D.A. Li E. Cell. 1999; 99: 247-257Abstract Full Text Full Text PDF PubMed Scopus (4449) Google Scholar). ES cells that lack both Dnmt3a and Dnmt3b progressively lose DNA methylation and after extended passage in culture lose almost all DNA methylation at all loci examined while expressing normal levels of Dnmt1 (6Chen T. Ueda Y. Dodge J.E. Wang Z. Li E. Mol. Cell. Biol. 2003; 23: 5594-5605Crossref PubMed Scopus (572) Google Scholar). Genetic studies of human DNA methyltransferases have been carried out in various cancer cell lines. It was shown that targeted disruption of DNMT1 or DNMT3b in the colon cancer cell line HCT116 had little effect on DNA methylation and cell growth, whereas inactivation of both DNMT1 and DNMT3b led to a genome-wide hypomethylation, expression of p16, and reduced cell growth (7Rhee I. Bachman K.E. Park B.H. Jair K.W. Yen R.W. Schuebel K.E. Cui H. Feinberg A.P. Lengauer C. Kinzler K.W. Baylin S.B. Vogelstein B. Nature. 2002; 416: 552-556Crossref PubMed Scopus (1033) Google Scholar, 8Rhee I. Jair K.W. Yen R.W. Lengauer C. Herman J.G. Kinzler K.W. Vogelstein B. Baylin S.B. Schuebel K.E. Nature. 2000; 404: 1003-1007Crossref PubMed Scopus (370) Google Scholar). However, more recent studies showed that inactivation of DNMT1 alone by antisense RNA or RNA interference methods in HCT116 cells and other human cancer cells resulted in loss of DNA methylation and inhibition of cell growth (9Robert M.F. Morin S. Beaulieu N. Gauthier F. Chute I.C. Barsalou A. MacLeod A.R. Nat. Genet. 2003; 33: 61-65Crossref PubMed Scopus (518) Google Scholar). Additional studies are necessary to resolve these discrepancies. Studies of tumor cell lines and primary tumor tissues provide compelling evidence that aberrant change in DNA methylation patterns is a hallmark of cancer. Hypermethylation of CpG-containing gene promoters often targets putative tumor suppressor genes such as p16, RASSF1A, BRCA1, and pRb and is associated with their silencing. Hypomethylation of bulk genomic DNA such as pericentromeric repeats is also frequently observed in tumor cells (10Jones P.A. Baylin S.B. Nat. Rev. Genet. 2002; 3: 415-428Crossref PubMed Google Scholar, 11Baylin S.B. Bestor T.H. Cancer Cell. 2002; 1: 299-305Abstract Full Text Full Text PDF PubMed Scopus (246) Google Scholar). However, the mechanism by which hypomethylation contributes to tumorigenesis remains largely unknown. In this study, we use MEF cells derived from Dnmt3a or Dnmt3b knock-out mice to study how DNA hypomethylation might affect cell growth and chromosome stability. We show that inactivation of Dnmt3b by either constitutive or conditional gene targeting in MEF cells results in global DNA hypomethylation, premature senescence or spontaneous immortalization, and chromosomal instability. In contrast, disruption of Dnmt3a appears to have no effect on these processes in MEF cells. MEF Cell Culture—MEFs were isolated from 12.5- or 13.5-dpc embryos of 129SvJae × C57BL/6 hybrid background and maintained in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum, 50 units/ml penicillin, and 50 μg/ml streptomycin. Each MEF line was expanded to three 10-cm dishes (except those Dnmt3b-/- cultures that grew slowly and were harvested from the initial 10-cm dish) and then frozen in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and 10% Me2SO and labeled as passage 1 (P1). Conditional inactivation of Dnmt3a or Dnmt3b was carried out in vitro by multiple infections of primary MEFs derived from Dnmt3a-/2loxP or Dnmt3b-/2loxP embryos with high-titer Cre adenovirus. Cre-mediated recombination was confirmed by Southern blot and Western blot analyses. Transfection of Dnmt3a or Dnmt3b expression vectors in MEF cells was carried out using Lipofectamine (Invitrogen) and selected with blasticidin as described previously (6Chen T. Ueda Y. Dodge J.E. Wang Z. Li E. Mol. Cell. Biol. 2003; 23: 5594-5605Crossref PubMed Scopus (572) Google Scholar). DNA Methylation Analysis—Genomic DNA was digested with the methylation-sensitive enzyme HpaII and analyzed by Southern hybridization using probes for minor satellite repeats and the endogenous C-type retrovirus. Bisulfite sequencing analysis of the p21 proximal promoter was done as described previously (12Dodge J.E. List A.F. Futscher B.W. Int. J. Cancer. 1998; 78: 561-567Crossref PubMed Scopus (65) Google Scholar), using the following oligonucleotides as nested PCR primers: F1 (5′-GATGTATGTGGTTTTGTTGGTG-3′), F2 (5′-GAGAATAGTTTAGGTGTGGGGG-3′), R1 (5′-CACCCACTAAACTCAACICATTAC-3′), and R2 (5′-CIAAAAAAAACTATTATTCCCTACCAC-3′) (I = inosine). Spontaneous Immortalization and SA-β-Galactosidase Staining Assays—The growth of MEFs was assayed following a standard 3T9 protocol. Briefly, 900,000 cells were plated in a 6-cm dish and grown in a 37 °C incubator with 5% CO2 for 3 days, and the procedure was repeated until wild-type MEFs started to show immortalized growth. All counts were done in duplicate, and the population doublings, which were calculated using the log formula (13Blasko M.A. Lee H.W. Hande M.P. Samper E. Lansdorp P.M. DePinho R.A. Greider C.W. Cell. 1997; 91: 25-34Abstract Full Text Full Text PDF PubMed Scopus (1790) Google Scholar), were plotted on the y axis versus days of culture on the x axis. SA-β-galactosidase staining was performed as described previously (14Dimri G.P. Lee X. Basile G. Acosta M. Scott G. Roskelly C. Medrano E.E. Linskens M. Rubelj I. Pereira-Smith O. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 9363-9367Crossref PubMed Scopus (5655) Google Scholar). Analysis of Metaphase and Anaphase Chromosomes—Metaphase chromosome spreads were prepared following standard procedures. Briefly, after treatment with Colcemid for ∼3 h, cells were harvested, treated twice with 75 mm KCl for 15 min each time, and fixed in methanol:glacial acetic acid (3:1) at room temperature for 1 h. The cell suspension was then dropped onto a microscope slide, and chromosomes were visualized with a light microscope. For anaphase bridge analysis, asynchronous cells grown on chambered slide glass were fixed, stained for mitotic spindle with anti-γ-tubulin antibody (T-6557 monoclonal antibody clone GTU-88; Sigma) followed by Alexa 488-conjugated goat anti-mouse secondary antibody (Molecular Probe), and counterstained for DNA with 4′,6′-diamidino-2-phenylindole. Anaphase cells were visualized by fluorescence microscopy. Cell Cycle, Cell Proliferation, and DNA Damage Checkpoint Assays— For cell cycle and cell proliferation analysis, cells were stained with propidium iodide and anti-bromodeoxyuridine antibodies and analyzed by flow cytometry on a Becton Dickinson FACScan. To induce DNA damage, the cells were dosed with 10 grays of ionizing radiation from a cesium source. For Western blot analysis, proteins were isolated 2 h after irradiation for p53 or 18 h after irradiation for p21. Antibodies against p21 (F-5) and p53 (Pab240) were obtained from Santa Cruz Biotechnology and Chemicon International, Inc., respectively. Dnmt3b Is Required for the Maintenance of DNA Methylation in MEF Cells—We have previously shown that Dnmt3a and Dnmt3b are required for de novo methylation of the mouse genome during early embryogenesis and for both de novo methylation and stable maintenance of DNA methylation patterns in ES cells (5Okano M. Bell D.W. Haber D.A. Li E. Cell. 1999; 99: 247-257Abstract Full Text Full Text PDF PubMed Scopus (4449) Google Scholar, 6Chen T. Ueda Y. Dodge J.E. Wang Z. Li E. Mol. Cell. Biol. 2003; 23: 5594-5605Crossref PubMed Scopus (572) Google Scholar). To examine whether Dnmt3a and Dnmt3b are required for maintaining DNA methylation in somatic cells after the hypermethylated state of the genome is established, we obtained MEF cells for conditional inactivation of these genes. Primary MEF cell cultures were established from embryos that carry one null allele and one conditional allele of Dnmt3a or Dnmt3b. Conditional inactivation of Dnmt3a or Dnmt3b was achieved by infecting the corresponding MEFs with Cre adenovirus. The conditional alleles, which contain two loxP sites flanking the exons that encode the catalytic regions of Dnmt3a or Dnmt3b, had equivalent functions to their wild-type alleles, whereas Cre-mediated deletion of catalytic motifs produced functionally null alleles (Fig. 1). The Dnmt3a conditional allele has been described previously (15Kaneda M. Okano M. Hata K. Sado T. Tsujimoto N. Li E. Sasaki H. Nature. 2004; 429: 900-903Crossref PubMed Scopus (1016) Google Scholar) (Fig. 1A). The Dnmt3b conditional allele with two loxP sites flanking the exons encoding the PC and ENV catalytic motifs is shown in Fig. 1B. The conditional alleles (2P), the Cre-mediated deletion alleles (1P), and the null alleles (-) could be distinguished by Southern blot analysis (Fig. 1, C and D). Both Dnmt3a and Dnmt3b proteins were expressed in wild-type primary MEF cells. Constitutive or conditional inactivation of Dnmt3a or Dnmt3b abolished the corresponding protein product (Fig. 1, E and F). To determine global DNA methylation levels, genomic DNA was digested with the restriction enzyme MspI, which cuts DNA regardless of whether it is methylated or not, or with HpaII, which cuts only unmethylated DNA. The digested DNA was analyzed by agarose gel electrophoresis, followed by ethidium bromide staining (Fig. 2A) and Southern blot hybridization with the endogenous C-type retroviral DNA (Fig. 2B). The ∼100 copies of endogenous virus interspersed throughout the mouse genome were heavily methylated in both wild-type ES and MEF cells (Fig. 2B, lanes 2, 8, and 21). All Dnmt3a mutant (primary or immortalized Dnmt3a-/-, primary Dnmt3a-/2P, and immortalized Dnmt3a-/1P) MEF cells did not show obvious changes in DNA methylation of C-type viral DNA (Fig. 2B, lanes 9–12), as compared with wild-type MEF cells, consistent with the findings reported for Dnmt3a-/- ES cells and 9.5-dpc embryos (5Okano M. Bell D.W. Haber D.A. Li E. Cell. 1999; 99: 247-257Abstract Full Text Full Text PDF PubMed Scopus (4449) Google Scholar). In contrast, primary Dnmt3b-/- MEFs showed a dramatic decrease of methylation of C-type viral DNA (Fig. 2B, lane 13). Immortalized Dnmt3b-/- and, importantly, Dnmt3b-/1loxP MEFs also showed the same degree of demethylation as primary Dnmt3b-/- MEFs (Fig. 2B, lanes 15 and 16), whereas Dnmt3b-/2P MEFs showed normal levels of DNA methylation in C-type viral DNA (Fig. 2B, lane 14). We also examined the effect of inactivation of both Dnmt3a and Dnmt3b on DNA methylation in MEF cells. Whereas Dnmt3a-/-Dnmt3b-/2P MEFs showed normal levels of DNA methylation (Fig. 2B, lane 17), inactivation of the remaining functional Dnmt3b allele resulted in hypomethylation similar to that observed in Dnmt3b-/1P MEFs (Fig. 2B, lane 18). Notably, Dnmt3a-/-Dnmt3b-/1P MEFs (n = 2) that were passaged in culture for ∼6 months (P40) retained levels of DNA methylation similar to those in early-passage (P1) cells (Fig. 2B, compare lane 19 with lanes 13 and 18), in sharp contrast to the progressive loss of DNA methylation in double mutant ES cells (Fig. 2B, lanes 4–6) (6Chen T. Ueda Y. Dodge J.E. Wang Z. Li E. Mol. Cell. Biol. 2003; 23: 5594-5605Crossref PubMed Scopus (572) Google Scholar). As a control, p53-/- MEFs infected with Cre adenovirus and cultured for the same period of time (∼6 months) retained wild-type levels of DNA methylation (Fig. 2B, lane 20). These results suggest that Dnmt3b is required for the maintenance of global DNA hypermethylation in MEF cells, whereas Dnmt3a is dispensable. Loss of Dnmt3b, but Not Dnmt3a, Alters MEF Proliferation— Growth analysis of seven independent wild-type MEF lines showed that after ∼3 weeks in culture, replicative senescence had occurred, as judged by both a halt in cell proliferation and enlarged cytoplasm (Fig. 3, A and B). A similar growth pattern was observed for Dnmt3a-/-, Dnmt3a-/-Dnmt3b+/-, and Dnmt3a+/-Dnmt3b+/- MEF cells (n = 5 per genotype; data not shown). In contrast, of 10 independent Dnmt3b-/- MEF cultures assayed, none displayed the wild-type pattern of rapid proliferation in early passages followed by a halt in proliferation. Instead, these MEF lines showed either "premature senescence" (n = 3) or spontaneous immortalization (n = 7). As a control, p53-/- MEF lines (n = 2) quickly escaped senescence and became immortalized. Dnmt3b-/-p53-/- MEF lines (n = 2) showed an almost identical growth pattern to that of p53-/- MEF cells, indicating that Dnmt3b is not required for the continuous proliferation of p53-/- cells (Fig. 3A). Expression of Dnmt3b in Immortalized Dnmt3b-/- MEFs Inhibits Cellular Growth—Given that loss of Dnmt3b resulted in alterations in cellular proliferation in MEFs, we asked whether reintroduction of Dnmt3b into Dnmt3b-deficient MEFs would affect the proliferation of these cells. We transfected Dnmt3b or Dnmt3b:PC (a catalytically inactive form) (6Chen T. Ueda Y. Dodge J.E. Wang Z. Li E. Mol. Cell. Biol. 2003; 23: 5594-5605Crossref PubMed Scopus (572) Google Scholar) into an immortalized Dnmt3b1P/1P MEF line and sought to generate stable clones. Whereas clones expressing Dnmt3b:PC were readily obtained and expanded, expression of the wild-type Dnmt3b resulted in severe growth impairment. Expression of Dnmt3b or Dnmt3b:PC in the stable clones was verified by immunoblotting analysis (Fig. 4A). Re-methylation of the known Dnmt3b target sequence, the centromeric minor satellite repeats, occurred in clones expressing Dnmt3b, but not in those expressing Dnmt3b:PC (Fig. 4B). A 3T9 growth assay showed that reduced proliferation correlated with the re-expression of Dnmt3b (Fig. 4C). SA-β-galactosidase activity was detected in 10–20% of cells that expressed Dnmt3b, but not in cells that expressed Dnmt3b:PC (Fig. 4D). In addition, polynucleated cells were observed in some Dnmt3b-transfected clones, but not in any Dnmt3b:PC-transfected clones (Fig. 4E). Consistent with our observation that mutation of Dnmt3a had no effect on cellular proliferation, re-introduction of Dnmt3a into Dnmt3a-/-Dnmt3b-/1loxP MEFs did not alter cellular growth (data not shown). Loss of Dnmt3b, but Not Dnmt3a, Is Associated with Chromosomal Abnormalities—We performed metaphase analysis of primary wild-type, Dnmt3a-/-, Dnmt3b-/-, and p53-/- MEFs. p53-/- MEFs were previously described to be polyploid and served as a positive control. We found that Dnmt3b-/- MEFs contained more polyploid cells (∼40%) than wild-type (∼20%) or Dnmt3a-/- MEFs (∼20%) (Fig. 5A; data not shown). Consistent with this observation, FACScan analysis revealed that Dnmt3b-/- MEFs grown in asynchronous cultures had higher percentages of tetraploid cells than wild-type MEFs (data not shown). Dnmt3b-/- MEFs also showed increased aneuploidy. Whereas ∼70% of wild-type MEFs contained 40 chromosomes, <50% of Dnmt3b-/- MEFs contained 40 chromosomes (Fig. 5B). Remarkably, we identified one Dnmt3b1P/1P MEF line that contained numerous fused and broken chromosomes, and this correlated with the presence of an anaphase bridge in ∼20% of the nuclei (Fig. 5, D and E). Chromosome fusion was also observed in one of the two Dnmt3a-/-Dnmt3b-/1P MEF lines examined. In contrast, none of the four Dnmt3a mutant MEFs examined contained obvious chromosomal abnormalities (Fig. 5C). Interestingly, anaphase bridge formation was reversed in the Dnmt3b1P/1P clones that had Dnmt3b re-expressed, but it was retained in the Dnmt3b:PC clones (Fig. 5F). Inactivation of Dnmt3b Results in Increased Levels of p21 Protein—To determine the possible mechanisms by which Dnmt3b may regulate cellular proliferation, we examined the function of the G1 to S-phase checkpoint that is regulated largely by the p53 pathway. MEF cells were treated with or without 10 grays of γ-irradiation, and 25 h later, bromodeoxyuridine incorporation was quantitated by FACScan analysis (bromodeoxyuridine was added 1 h prior to FACScan). Similar to wild-type MEFs, both primary and immortalized Dnmt3b-/- MEFs showed reduced cellular proliferation in response to DNA damage by γ-irradiation, whereas p53-/- MEFs, as a control, continued to enter S phase after DNA damage (Fig. 6A). Furthermore, immunoblotting analysis indicated that immortalized Dnmt3b-/- and Dnmt3a-/-Dnmt3b-/1P MEF lines had increased levels of p53 protein 2 h after DNA damage (Fig. 6B). These results suggested that the G1 to S-phase checkpoint is intact in Dnmt3b-deficient MEF cells. We then examined the protein levels of p21, a major downstream mediator of p53-mediated G1 arrest, in Dnmt3b-deficient MEF cells 18 h after γ-irradiation. Whereas all three spontaneously immortalized Dnmt3b-/- MEF lines showed slightly increased levels of p21 in response to DNA damage, untreated cells, surprisingly, had elevated baseline levels of p21 protein (Fig. 6C), despite their lack of p53 induction (Fig. 6B). High levels of p21 protein were observed in one Dnmt3b-/- line (3b-/-#2), one Dnmt3b1P/1P line, and one Dnmt3a-/-Dnmt3b-/1P line (Fig. 6C). Given that hypermethylation of gene promoters is associated with transcriptional silencing, we performed bisulfite genomic sequencing of the p21 proximal promoter in wild-type, Dnmt3b-/-, and Dnmt3a-/-Dnmt3b-/1loxP MEFs. Similar to the majority of previous reports that have examined the methylation status of the p21 proximal promoter, we did not find any DNA methylation in the ∼400-bp sequence in any of these samples (data not shown). This suggests that demethylation of p21 promoter CpG island is not the underlying mechanism of the p21 induction observed in Dnmt3b-deficient MEFs (16Milutinovic S. Knox J.D. Szyf M. J. Biol. Chem. 2000; 275: 6353-6359Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar). Maintenance of a hypermethylated state of the genome is essential for mammalian development. Whereas biochemical and genetic studies have demonstrated that Dnmt1 is the major maintenance methyltransferase, we have recently shown that Dnmt3a and Dnmt3b also play a role in maintaining DNA methylation patterns in ES cells (6Chen T. Ueda Y. Dodge J.E. Wang Z. Li E. Mol. Cell. Biol. 2003; 23: 5594-5605Crossref PubMed Scopus (572) Google Scholar). In this study, we show that conditional inactivation of Dnmt3b results in reduction of DNA methylation in MEF cells, indicating that the Dnmt3 family of DNA methyltransferases is required for maintaining normal DNA methylation levels in somatic cells as well. In contrast to the largely redundant function of Dnmt3a and Dnmt3b in ES cells, Dnmt3a does not appear to compensate for the loss of Dnmt3b in MEFs. Given that both Dnmt3a and Dnmt3b are expressed in MEFs, it is possible that chromatin remodeling associated with cellular differentiation makes most genomic loci inaccessible to Dnmt3a but does not affect their accessibility to Dnmt3b. However, we cannot exclude the possibility that Dnmt3a is involved in the maintenance of DNA methylation in other cell types or during later stages of development. Another interesting observation is that Dnmt3a-/-Dnmt3b-/- MEF cells, unlike Dnmt3a-/-Dnmt3b-/- ES cells, do not show progressive loss of methylation in culture, suggesting that in the absence of Dnmt3a and Dnmt3b, Dnmt1 is capable of maintaining a higher level of DNA methylation in MEFs than in ES cells. Collectively, these results suggest that the relative contributions of the Dnmt1 and Dnmt3 families of DNA methyltransferases to the maintenance of global methylation vary in a cell type- and developmental stage-specific manner. Inactivation of Dnmt1 has been shown to cause severe global demethylation and p53-dependent apoptosis in MEF cells (4Jackson-Grusby L. Beard C. Possemato R. Tudor M. Fambrough D. Csankovszki G. Dausman J. Lee P. Wilson C. Lander E. Jaenisch R. Nat. Genet. 2001; 27: 31-39Crossref PubMed Scopus (565) Google Scholar). In this study, we demonstrate that Dnmt3b-deficient MEF cells can tolerate moderate global demethylation and survive, but they show abnormal cellular proliferation. Unlike wild-type MEF cells, which undergo replicative senescence after ∼3 weeks in culture, the majority of Dnmt3b-/- MEF lines (7 of 10) escape senescence and become spontaneously immortalized because they continue to proliferate during the course of our study. Meanwhile, some Dnmt3b-/- MEF lines (3 of 10) show decreased proliferation and become senescent prematurely. We analyzed components of the p53 pathway that are known to regulate cell proliferation. We showed that the p53 protein could be activated by radiation-induced DNA damage in primary as well as immortalized Dnmt3b-deficient MEFs. Interestingly, the expression of the cyclin-dependent kinase inhibitor p21 was elevated in a p53-independent manner. It remains to be determined whether up-regulation of p21 is a contributing factor for the premature senescence observed in the small number of Dnmt3b-deficient MEF lines. Clearly, high levels of p21 are incapable of halting G1 progression in most Dnmt3b-deficient cells, suggesting that the pRB pathway may be defective. Additional studies are necessary to determine the molecular mechanisms by which Dnmt3b regulates cell proliferation. It is unknown why some Dnmt3b-deficient cells undergo spontaneous immortalization, whereas others undergo premature senescence. The genetic background of the MEF lines does not appear to account for the variation because all the MEF cells were isolated from embryos of 129SvJae × C57BL/6 hybrid background, and male and female cells showed no obvious differences in proliferation. One possibility is that genomic instability causes different genetic aberrations in the Dnmt3b-deficient MEF lines, resulting in different phenotypes. Previous studies provide a basis for further speculation. Targeted disruption of Lsh, which encodes a member of the SNF2 family of chromatin remodeling proteins, has been shown to cause global hypomethylation and decreased proliferation of MEF cells (17Dennis K. Fan T. Geiman T. Yan Q. Muegge K. Genes Dev. 2001; 15: 2940-2944Crossref PubMed Scopus (314) Google Scholar, 18Fan T. Yan Q. Huang J. Austin S. Cho E. Ferris D. Muegge K. Cancer Res. 2003; 63: 4677-4683PubMed Google Scholar). Like Dnmt3b, Lsh has been implicated in maintaining heterochromatin organization (19Yan Q. Cho E. Lockett S. Muegge K. Mol. Cell. Biol. 2003; 23: 8416-8428Crossref PubMed Scopus (61) Google Scholar, 20Bachman K.E. Rountree M.R. Baylin S.B. J. Biol. Chem. 2001; 276: 32282-32287Abstract Full Text Full Text PDF PubMed Scopus (373) Google Scholar, 21Chen T. Tsujimoto N. Li E. Mol. Cell. Biol. 2004; 24: 9048-9058Crossref PubMed Scopus (201) Google Scholar). This raises the possibility that altered heterochromatin structure and function, as a consequence of global hypomethylation, might be the underlying cause of premature senescence observed in some Dnmt3b-deficient MEF cells. Another example of altered cellular proliferation was recently reported for uniparental MEFs, which, similar to Dnmt3b-deficient MEFs, exhibit dramatically contrasting patterns of growth, depending on parent of origin. "Androgenetic" MEFs, which contain two copies of the paternal genome, are prone to spontaneous immortalization and transformation, whereas "parthenogenetic" MEFs, which contain two copies of the maternal genome, undergo premature senescence and apoptosis. These effects have been attributed to loss of imprinting (22Hernandez L. Kozlov S. Piras G. Stewart C.L. Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 13344-13349Crossref PubMed Scopus (73) Google Scholar). Whereas a clear role for Dnmt3b in the maintenance of genomic imprinting has yet to be determined, it is possible that abnormal expression of imprinted genes, due to DNA demethylation, might determine whether Dnmt3b-deficient MEF cells would undergo spontaneous immortalization or premature senescence. Numerous studies have established a link between alterations in DNA methylation and cancer. Whereas hypermethylation of promoter CpG islands of tumor suppressor genes provides a plausible explanation that methylation-mediated inactivation of tumor suppressor genes promotes tumorigenesis, how DNA hypomethylation contributes to tumorigenesis remains largely unknown (10Jones P.A. Baylin S.B. Nat. Rev. Genet. 2002; 3: 415-428Crossref PubMed Google Scholar, 23Chen R.Z. Pettersson U. Beard C. Jackson-Grusby L. Jaenisch R. Nature. 1998; 395: 89-93Crossref PubMed Scopus (789) Google Scholar, 24Eden A. Gaudet F. Waghmare A. Jaenisch R. Science. 2003; 300: 455Crossref PubMed Scopus (1084) Google Scholar, 25Gaudet F. Hodgson J.G. Eden A. Jackson-Grusby L. Dausman J. Gray J.W. Leonhardt H. Jaenisch R. Science. 2003; 300: 489-492Crossref PubMed Scopus (1224) Google Scholar). Our results show that inactivation of Dnmt3b in MEFs results in hypomethylation and chromosomal instability, which in turn may lead to spontaneous immortalization. It remains to be determined whether immortalized Dnmt3b-/- MEFs are prone to oncogenic transformation. Because DNMT3b is expressed at very low levels in most human somatic tissues, it is conceivable that any reduction of DNMT3b below a certain threshold would result in genome-wide hypomethylation and chromosomal instability. It would also be interesting to examine whether oncogenic transformation of normal cells would lead to reduction of DNMT3b expression and DNA hypomethylation. Additional studies are necessary to elucidate the role of DNMT3b in oncogenic transformation and cancer progression.
Referência(s)