Inhibits Senescence Via the INK4A-ARF Pathway
Total Page:16
File Type:pdf, Size:1020Kb
Oncogene (2015) 34, 5807–5820 © 2015 Macmillan Publishers Limited All rights reserved 0950-9232/15 www.nature.com/onc ORIGINAL ARTICLE MOZ (MYST3, KAT6A) inhibits senescence via the INK4A-ARF pathway BN Sheikh1,2, B Phipson3,4, F El-Saafin1,2, HK Vanyai1,2, NL Downer1, MJ Bird4,5, AJ Kueh1, RE May1, GK Smyth3,6, AK Voss1,2,7 and T Thomas1,2,7 Cellular senescence is an important mechanism that restricts tumour growth. The Ink4a-Arf locus (also known as Cdkn2a), which encodes p16INK4A and p19ARF, has a central role in inducing and maintaining senescence. Given the importance of cellular senescence in restraining tumour growth, great emphasis is being placed on the identification of novel factors that can modulate senescence. The MYST-family histone acetyltransferase MOZ (MYST3, KAT6A), first identified in recurrent translocations in acute myeloid leukaemia, has been implicated in both the promotion and inhibition of senescence. In this study, we investigate the role of MOZ in cellular senescence and show that MOZ is a potent inhibitor of senescence via the INK4A-ARF pathway. Primary mouse embryonic fibroblasts (MEFs) isolated from Moz-deficient embryos exhibit premature senescence, which was rescued on the Ink4a- Arf−/− background. Importantly, senescence resulting from the absence of MOZ was not accompanied by DNA damage, suggesting that MOZ acts independently of the DNA damage response. Consistent with the importance of senescence in cancer, expression profiling revealed that genes overexpressed in aggressive and highly proliferative cancers are expressed at low levels in Moz- deficient MEFs. We show that MOZ is required to maintain normal levels of histone 3 lysine 9 (H3K9) and H3K27 acetylation at the transcriptional start sites of at least four genes, Cdc6, Ezh2, E2f2 and Melk, and normal mRNA levels of these genes. CDC6, EZH2 and E2F2 are known inhibitors of the INK4A-ARF pathway. Using chromatin immunoprecipitation, we show that MOZ occupies the Cdc6, Ezh2 and Melk loci, thereby providing a direct link between MOZ, H3K9 and H3K27 acetylation, and normal transcriptional levels at these loci. This work establishes that MOZ is an upstream inhibitor of the INK4A-ARF pathway, and suggests that inhibiting MOZ may be one way to induce senescence in proliferative tumour cells. Oncogene (2015) 34, 5807–5820; doi:10.1038/onc.2015.33; published online 16 March 2015 INTRODUCTION Typically for an oncogene involved in haematological malig- The monocytic leukaemia zinc-finger protein (MOZ) is a MYST- nancy, studies of the normal function of MOZ show that it has an family histone acetyltransferase, which was first identified in a important role in normal haematopoiesis. Studies of loss-of- recurrent translocation, t(8;16)(p11;p13) leading to an aggressive function mutations in MOZ have shown that it is essential for the 10,11 type of acute myeloid leukaemia.1 Patients with a t(8;16) development of hematopoietic stem cells, and the histone acetyltransferase activity of MOZ is required for maintaining the translocation, which generates a fusion transcript between MOZ 12 and CBP, are typically diagnosed with a FAB M4/M5 subtype of self-renewal of HSCs. During embryonic development, MOZ is required for the acetylation of histone 3 at lysine 9 (H3K9ac) at acute myeloid leukaemia, commonly associated with coagulo- 13 14 pathy, erythrophagocytosis and extramedullary dissemination.2,3 Hox, Tbx1 and Tbx5 loci, and for their correct expression. Accordingly, embryos lacking Moz show an extensive anterior The prognosis of patients with a t(8:16) translocation is poor with homeotic transformation of the axial skeleton and neural tube,13 median survival times, after diagnosis, reported between 2 as well as cardiac and craniofacial defects mirroring the human months2 and 4.7 months.4 Gene expression analysis shows that DiGeorge syndrome.14 this forms a distinct subtype of acute myeloid leukaemia that is Recent studies have examined the role of MOZ in cellular typified by upregulation of HOX genes and their co-factor 4,5 senescence. It has been reported that overexpressed MOZ is able MEIS1. Additional chromosomal rearrangements have been to bind to p53, and that MOZ is required to activate p21 fi identi ed that generate chimeric genes in which MOZ is fused to expression in response to DNA damage to induce senescence.15 genes coding for other transcriptional regulators NCOA2 (TIF2), This study suggests that in the absence of Moz, cells are unable to 6–8 NCOA3 and p300. Consistent with the importance of self- senesce and undergo apoptosis instead. However, a conflicting renewal in leukaemia, MOZ-TIF2 is able to induce the property of report has recently been published. Using mice that possess a 9 self-renewal in committed progenitors. catalytically inactive MOZ, Perez-Campo et al.16 suggest that the 1Division of Development and Cancer, The Walter and Eliza Hall Institute of Medical Research, Melbourne, Victoria, Australia; 2Department of Medical Biology, University of Melbourne, Melbourne, Victoria, Australia; 3Division of Bioinformatics, The Walter and Eliza Hall Institute of Medical Research, Melbourne, Victoria, Australia; 4Murdoch Children’s Research Institute, Melbourne, Victoria, Australia; 5Department of Paediatrics, University of Melbourne, Melbourne, Victoria, Australia and 6Department of Mathematics and Statistics, University of Melbourne, Melbourne, Victoria, Australia. Correspondence: Dr T Thomas, Division of Development and Cancer, The Walter and Eliza Hall Institute of Medical Research, 1G Royal Parade, Melbourne, Victoria 3052, Australia or A K Voss, Division of Development and Cancer, The Walter and Eliza Hall Institute of Medical Research, 1G Royal Parade, Melbourne, Victoria 3052, Australia. E-mail: [email protected] or [email protected] 7These authors co-supervised this project and contributed equally. Received 6 March 2014; revised 1 December 2014; accepted 23 January 2015; published online 16 March 2015 MOZ inhibits senescence via the INK4A-ARF pathway BN Sheikh et al 5808 histone acetyltransferase activity of MOZ is important for (P = 0.002) were observed in Moz−/− cultures (Figure 1e). No restraining senescence. It is unclear how the fundamental change in Moz mRNA levels were seen between passages 1 and 6 differences between these studies can be reconciled. (Supplementary Figure 2). Altogether, the increase in (1) In order to resolve the role of MOZ in senescence, we used β-galactosidase activity, (2) increased ROS production and (3) an − − primary mouse embryonic fibroblasts (MEFs) isolated from Moz / increase in Ink4a, Arf and Ink4b mRNA all suggest that Moz- embryos to show that endogenous MOZ is an inhibitor, and not an deficient MEFs fail to proliferate because of premature activator of senescence. Primary MEFs lacking MOZ undergo senescence. premature senescence at passage 3 and express the senescence β − − markers -galactosidase, Ink4a and Arf. Apoptotic cell death was Proliferation in Moz / MEFs is only affected at late passages unaffected in Moz−/− cultures. Using microarray and chromatin As senescence results in cell cycle arrest, we tested DNA synthesis immunoprecipitation (ChIP), we show that MOZ is required for the by examining incorporation of the thymidine analogue BrdU, and maintenance of H3K9ac and H3K27ac at gene loci encoding analysed cell cycle characteristics of Moz-deficient MEFs. Although repressors of the INK4A-ARF pathway including Cdc6, Ezh2 and no differences in BrdU incorporation between wild-type and E2f2. Accordingly, we detect direct binding of MOZ to these gene Moz− / − cultures were apparent at passage 2 (Figures 1f, loci. Consistent with the importance of MOZ in maintaining the P = 0.231), 2.2-fold fewer Moz −/− MEFs stained positive for BrdU transcription of genes required to repress Ink4a-Arf, we show that o fi at passage 4 (P 0.05). These data were consistent with the cell premature senescence in Moz-de cient cells was rescued com- cycle profiles of MEF cultures, showing fewer Moz− / − cells in the pletely by deletion of the Ink4a-Arf locus. This work identifies G2/M-phase compared with controls (Supplementary Figure 3). MOZ as an upstream inhibitor of the INK4A-ARF pathway and Together, these data suggest that proliferation is not a primary premature senescence. defect in early-passage Moz-deficient MEFs. Rather, the reduced proliferation in Moz −/− cells at later passages likely reflects RESULTS increased senescence. Primary Moz-deficient MEFs show premature senescence Apoptosis is not affected in MEFs lacking MOZ To investigate the role of MOZ in cellular senescence, we isolated fi +/ − It has been previously suggested that in response to ultraviolet or and cultured primary broblasts from E12.5 wild-type, Moz and −/− −/− −/− drug-induced DNA damage, Moz MEFs undergo apoptosis at an Moz embryos. Compared with wild-type, Moz cultures failed 15 o increased rate compared with wild type. Therefore, we tested to accumulate cells from passage 3 onwards (Figure 1a; P 0.001). +/ − +/ − whether the failure of cell accumulation in primary Moz and Moz MEFs showed an intermediate phenotype, with a failure to −/− accumulate cells from passage 6 onwards. These data indicated Moz MEF cultures may be due to an increase in cell death. The that Moz-deficient MEFs were either defective in cell proliferation, proportion of apoptotic cells was determined over 72 h after exhibited increased cell death or underwent cellular senescence passages 2, 4 and 6 by annexin V binding of externalized phosphatidylserine. Compared with wild type, there was no prematurely. +/ − A senescent phenotype is characterized by an increase in difference in the proportion of annexin V-binding cells in Moz β 17 cultures at all time points analysed (Figures 1g, P40.10). Similarly, -galactosidase activity, increased reactive oxygen species (ROS) −/− production,18 and an increase in the levels of senescence inducers the proportion of annexin V-binding cells in Moz was similar to p16INK4A and p19ARF. We compared these parameters of cellular wild type, apart from a small increase of approximately 2% in senescence in wild-type and Moz mutant MEFs.