DNA methylation, through DNMT1, has an essential role in the development of gastrointestinal smooth muscle cells and disease

DNA methylation, through DNMT1, has an essential role in the development of gastrointestinal... DNA methylation is a key epigenetic modification that can regulate gene expression. Genomic DNA hypomethylation is commonly found in many gastrointestinal (GI) diseases. Dysregulated gene expression in GI smooth muscle cells (GI- SMCs) can lead to motility disorders. However, the consequences of genomic DNA hypomethylation within GI-SMCs are still elusive. Utilizing a Cre-lox murine model, we have generated SMC-restricted DNA methyltransferase 1 (Dnmt1) knockout (KO) mice and analyzed the effects of Dnmt1 deficiency. Dnmt1-KO pups are born smaller than their wild- type littermates, have shortened GI tracts, and lose peristaltic movement due to loss of the tunica muscularis in their intestine, causing massive intestinal dilation, and death around postnatal day 21. Within smooth muscle tissue, significant CpG hypomethylation occurs across the genome at promoters, introns, and exons. Additionally, there is a marked loss of differentiated SMC markers (Srf, Myh11, miR-133, miR-143/145), an increase in pro-apoptotic markers (Nr4a1, Gadd45g), loss of cellular connectivity, and an accumulation of coated vesicles within SMC. Interestingly, we observed consistent abnormal expression patterns of enzymes involved in DNA methylation between both Dnmt1-KO mice and diseased human GI tissue. These data demonstrate that DNA hypomethylation in embryonic SMC, via congenital Dnmt1 deficiency, contributes to massive dysregulation of gene expression and is lethal to GI-SMC. These results suggest that Dnmt1 has a necessary role in the embryonic, primary development process of SMC with consistent patterns being found in human GI diseased tissue. Introduction as in vascular SMC (vSMC) found in atherosclerotic The aberrant growth, or loss, of smooth muscle lesions . Recent research has shown that inducing geno- cells (SMCs) is a common symptom found in mic DNA hypomethylation can change SMC phenotypic burdensome gastrointestinal (GI) diseases such as identity, growth patterns, and expression levels of con- 1 4,5 chronic intestinal pseudo-obstruction and megacystis- tractile genes , giving preliminary evidence that altering microcolon-intestinal hypoperistalsis syndrome as well DNA methylation can shift the phenotypic identity of SMCs. Global changes in DNA methylation patterns are characteristic of digestive pathologies such as colorectal Correspondence: Seungil Ro (sro@med.unr.edu) 6 7 cancer and Crohn’s disease . Department of Physiology and Cell Biology, University of Nevada School of The process of methylating DNA at cytosine residues, Medicine, Reno, NV 89557, USA Department of Morphological and Physiological Sciences, University of Fukui, specifically at CpG sites, is a dynamic and reversible epi- Fukui 910-8507, Japan genetic modification . In the mammalian genome, CpG Full list of author information is available at the end of the article These authors contributed equally: Brian G. Jorgensen, Robyn M. Berent, Se Eun Ha. Edited by D. Heery © The Author(s) 2018 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to theCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Official journal of the Cell Death Differentiation Association 1234567890():,; 1234567890():,; Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 2 of 14 Cre-eGFP/+ sites can be found clustered together in regions entitled By crossing smMHC , a mouse line that selec- CpG islands that are often found in close proximity to, or tively expresses Cre recombinase and eGFP under the within, transcription start sites (TSS) as well as the pro- control of a Myh11 promoter beginning at embryonic 9 lox/lox 25 moters of housekeeping and/or tissue-specific genes .The day 12.5, with a Dnmt1 mouse line , we were able to -/-;Cre-eGFP/+ addition of a methyl group to cytosine at its fifth carbon, successfully generate congenital smDnmt1 Cre-eGFP/+ creating 5-methylcytosine (5-mc), by DNA methyl- (Dnmt1-KO) mice. In smMHC (Dnmt1-WT) transferases (DNMT1, DNMT3A, DNMT3B) can cause mice, cells that express Cre recombinase also express repression of associated genes as this mark, at and around eGFP, thus making it possible to identify Cre expression the TSS, has been linked to transcriptionally repressed through enhanced green fluorescent protein (eGFP) chromatin that denies access for transcription factors, such fluorescence. Knockout was confirmed through genotyp- 10,11 as serum response factor (SRF) .Thus, as SRF is a ing with primers specific for Cre, eGFP, known loxP sites, master regulator of contractile genes in SMC ,the tran- and deletion of floxed cassette (Supplementary Figure 1). scription of many contractile smooth muscle genes (Myh11, Only those mice that were positive for Cre, both genomic 13,14 Acta2) and microRNA (miRNAs; mir-133, mir-143/ loxP sites, and the deletion of the floxed cassette (Dnmt1- 15,16 145) is dependent on the binding of SRF to CArG (CC KO) presented the following symptoms: Dnmt1-KO mice 17,18 [A/T] GG) boxes at these loci . CArG boxes are highly were born similar in size and body mass to their wild-type conserved, have high CpG content in the region bound by (WT) littermates, but they have significantly less body SRF, are commonly found within promoter regions/close to mass by postnatal day 15 (P15) continuing through to P21 TSS, and are recognized specifically by SRF . Additionally, (Fig. 1a, b) when death occurs. Gross anatomical images high CpG content sites, such as those around CArG boxes, show intestinal dilation beginning at P10 that continues to are also well-established targets for cytosine methylation by progress until P21 (Fig. 1c) when dilation is severe enough 19,20 DNA methyltransferases . While several studies have to cause intestinal ischemia or perforation of the dete- given valuable insight into the role of DNA methylation in riorated intestinal wall, causing septic shock. Dnmt1-KO SMC, many of them similarly use cytidine analogs mice also have significantly shortened GI tracts in both 4,5 in vitro , which are not selective in their inhibition of any the small intestine (Fig. 1d) and colon (Fig. 1e) by P21. specific DNMT isoform, nor do they present any efficacy of these techniques in vivo in GI-SMC. Dynamic changes of enzymes involved in DNA methylation To further this area of research, we have selectively in Dnmt1-KO mice and diseased human tissue eliminated the methyltransferase, Dnmt1, from the murine DNA methylation is positively regulated by DNMT1/ SMC genome. Dnmt1 is known as the maintenance DNMT3A/DNMT3B and is negatively regulated by the methyltransferase as its main enzymatic function is to Ten-Eleven Translocation (TET) family of proteins faithfully restore methylation marks on newly formed (TET1/TET2/TET3). In the tunica muscularis of Dnmt1- daughter strands of DNA .Deficiency of Dnmt1 caused WT mice, expression levels of DNMT1 and DNMT3A are loss of the intestinal tunica muscularis and halted peristaltic developmentally dynamic while DNMT3B is constantly contractions which induced intestinal dilation and death expressed at all time points (Fig. 2a). DNMT1 is highly around P21. CpG methylation is significantly reduced expressed at embryonic day 18 (E18) and gradually genome-wideatpromoters,exons,and introns. We also decreases from birth to P20, with the opposite pattern found that markers of mature SMC are lost, pro-apoptotic being seen with DNMT3A. Intriguingly, temporal gene expression is increased, and Dnmt1 knockout (KO) expression levels of TET1/TET2/TET3 are reduced in SMCs cytoplasmically accumulate a large number of coated Dnmt1-KO with the reductions of TET2 and TET3 being vesicles. Furthermore, we noted consistent patterns of dys- significant by P15. DNMT1 levels were drastically reduced regulation of DNMT1 and TET3, an enzyme that initiates in Dnmt1-KO after P1 and the protein was not detected at the demethylation of DNA , between KO mice and var- P20 (Fig. 2a). As previously mentioned, in Dnmt1-KO iously diseased human GI tissue, including colon/anal can- smooth muscle tissue, TET2 and TET3 levels were sig- cers and Crohn’s disease. To our knowledge, this is the first nificantly reduced (Fig. 2d), suggesting that expression of in vivo evidence showing the requirement of a DNA these TET isoforms may be dependent on DNMT1. methyltransferase, and the importance of DNA methylation, Furthermore, as Dnmt1-KO smooth muscle tissue in the primary development and differentiation of GI-SMC. developed, potentially compensatory DNMT3A induction occurred (Fig. 2a). Also, as DNMT1 begins to be reduced Results in Dnmt1-KO, so are markers of SMC maturity and Generation, and confirmation, of deletion in SMC- contractibility, including SRF, MYH11, and SM22α restricted Dnmt1 knockout mice (Fig. 2b, d). Since MYH11 and SM22α (Tagln) are SRF -/- 23 12 As global Dnmt1 mice are embryonic lethal , it was targets , the reduction of SRF in Dnmt1-KO could result -/- necessary to generate a SMC-restricted Dnmt1 KO line. in, or exacerbate, the reduced expression of these SRF Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 3 of 14 Fig. 1 SMC-restricted congenital Dnmt1-KO mice are smaller in size/mass with shorter GI tracts that become increasingly dilated over time. a SMC-restricted congenital Dnmt1-KO mice at P20 are smaller when compared to Dnmt1-WT littermates. b No significant differencesinbodymasswere seen between Dnmt1-KO mice and Dnmt1-WT mice until P15 when WT mice continued to gain mass and Dnmt1-KO began to lose mass. Dnmt1-KO slowly began to regain mass by P18 continuing into P21, most likely due to lack of peristaltic movement in the intestine causing buildup of food in the small intestine. c By P10, dilation can be seen in the small intestine of Dnmt1-KO mice with dilation increasing in size until P21 when death occurs. d, e By P21, Dnmt1-KO mice have significantly shortened small intestines and colons. For (b, d, e) n= 10, error bars are SD; *p < 0.005; **p < 0.001, unpaired t-test Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 4 of 14 Fig. 2 Enzymes regulating DNA methylation and mature SMC markers are dynamically altered in the jejunal tunica muscularis of Dnmt1- KO mice and inflamed human colonic tissue. a Western blotting of isolated jejunal tunica muscularis from Dnmt1-WT and Dnmt1-KO mice show temporal changes in expression of enzymes regulating DNA methylation (DNMT1, 3A and 3B) and demethylation (TET1–3). Note a progressive loss of DNMT1, and a massive reduction in expression of TET2 and TET3 in Dnmt1-KO mice. b A loss of DNMT1 correlates with significant reductions in mature SMC markers (SRF, MYH11, and SM22α) that become more pronounced with age indicating that GI-SMCs in Dnmt1-KO mice are indeed losing their mature, contractile status. c Human colonic tunica muscularis shows increased expression of DNMT1 in inflamed tissues along with TET3. Diverticulitis-M (marginal area); Diverticulitis-P (pouch). d, e Quantification of western blotting from Dnmt1-KO and Dnmt1-WT and human healthy and inflamed colonic tissue; n = 2–3 for mouse and human separately, error bars are SEM, *p < 0.005; **p < 0.001, unpaired t-test target genes and their subsequent protein products. Due samples (Fig. 2e). Intriguingly, the correlated expression to aberrant methylation patterns being identified in col- of DNMT1 and TET3 in the inflamed human tissue is orectal cancer , we examined if DNMT1 is differentially consistent with the smooth muscle of Dnmt1-WT and expressed in the inflamed tunica muscularis from human Dnmt1-KO mice (Fig. 2a), suggesting a role for DNMT1 patients with either Crohn’s disease, colon/anal cancer, or in regulating the expression of TET3. diverticulitis. We detected increases in DNMT1 as well as increases in TET3 in the inflamed smooth muscle show- Dnmt1-KO mice gradually lose tunica muscularis and ing that DNMT1 can regulate TET3 across species necessary contractile proteins and accumulate coated (Fig. 2c). Quantitation of protein showed that the vesicles in SMCs increases in DNMT1 and TET3 were significant in the Hematoxylin and eosin staining revealed that inflamed smooth muscle from anal or colon cancer Dnmt1-KO mice begin to lose both muscle layers of the Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 5 of 14 Fig. 3 Electron microscopy (EM) reveals degeneration of smooth muscle and presence of coated vesicles in Dnmt1-KO SMC. a, b EM cross- section images of P21 Dnmt1-WT jejunum. c–f EM cross-section images of P21 Dnmt1-KO jejunum. Both the circular (CM) and longitudinal muscle (LM) layers in Dnmt1-WT jejunum are intact (a, b), complete with organized nuclei in SMC as well as the presence of ICC (ICC), and submucosal cells (SM). Degenerated SMC within the tunica muscularis in the jejunum of Dnmt1-KO mice (c, d) show cellular fragmentation, as well as a loss of nuclear organization. e A portion of a degenerated SMC with coated vesicles from a Dnmt1-KO mouse found phagocytosed in a macrophage (M). f Degenerated SMCs in Dnmt1-KO mice accumulate coated vesicles (V) trapped in the vicinity of the endoplasmic reticulum (ER) and Golgi (Gol), open vesicles: arrows; closed vesicles: arrowheads tunica muscularis at noticeable levels by P15 but indicating a complete loss of mature SMCs did not have a significant loss of either layer until P21, (Supplementary Figure 2D). The loss of MYH11 with the more drastic tissue loss being found in the cir- in the SMCs of Dnmt1-KO mice results in SMCs cular layer (Supplementary Figure 2). To confirm that are no longer contractile, leading to a that the loss of tissue was indeed due to a loss of weakened tunica muscularis, allowing for intestinal dis- mature SMCs, immunohistochemical analysis was per- tention. Electron microscopy of the tunica muscularis formed. As MYH11 is the most selective marker of from P21 Dnmt1-KO mice revealed a loss of mature and contractile SMCs , antibodies targeting intercellular connections between SMCs, submucosa MYH11 were used to confirm the loss of mature cells, and interstitial cells of Cajal (ICC), the SMCs alongside the endogenous eGFP reporter expres- necessary pacemaking cells in peristaltic contractions sion based on Myh11 expression. As expected, (Fig. 3c). We also noted cellular fragmentation and the MYH11 levels in the tunica muscularis of Dnmt1-WT loss of nuclear/nucleolar organization within Dnmt1-KO mice continued to increase from P10 to P21, SMCs (Fig. 3c, d). Furthermore, we were able to detect while in Dnmt1-KO MYH11 levels waned, starting at P15 macrophages engulfing the deteriorated SMCs in Dnmt1- until there was no detectable MYH11 signal at P21, KO mice (Fig. 3e). Finally, we found that almost all Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 6 of 14 Fig. 4 Transcriptomic analysis reveals dramatic changes of expression in Dnmt1-KO jejunal tunica muscularis. a, b The expression profile of Dnmt1-KO jejunal tunica muscularis shows a higher number of unique mRNA transcripts and more genes that increased their expression than reduced their expression when compared to Dnmt1-WT. Changes in overall miRNA expression were not as dramatic. c Increases in several genes associated with apoptosis or lipid processing were upregulated in Dnmt1-KO mice. Additionally, transcript levels of SMC markers (Srf, Myh11) are reduced along with the SRF transcriptional co-activator, Myocd. d Although overall miRNA changes are not as pronounced as mRNA changes in Dnmt1-KO mice, there are select miRNAs that increase their expression level. e A gene ontology evaluation uncovered the prevalence of five general categories enriched in Dnmt1-KO jejunal tunica muscularis with many genes relating to defense, immunity, lipids, motility, and apoptosis being found Dnmt1-KO SMCs contained small coated vesicles (Fig. 3d, vesicles in their cytosol that congregated in distinct f), which were not associated with Dnmt1-WT SMC regions around the endoplasmic reticulum and Golgi (Fig. 3a, b). Dnmt1-KO SMC appeared to accrue coated (Fig. 3f). Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 7 of 14 Fig. 5 Dnmt1-KO tunica muscularis accumulates lipids, has increased expression in apoptotic genes and characteristic apoptotic DNA fragmentation. a Oil Red O staining of jejunal cross-sections reveals an accumulation of lipids in the muscle layer of Dnmt1-KO mice (white arrows; red stain) that is not seen in Dnmt1-WT mice. b At P21, a clear fragmentation of extracted DNA can be seen in the isolated tunica muscularis of Dnmt1-KO mice. Distinct fragmentation patterns of genomic DNA are generated due to cleavage at specific sites, a characteristic of apoptotic cells. c qPCR of isolated tunica muscularis at P15. As expected, Dnmt1, Srf, and Myh11 all have significantly reduced expression in Dnmt1-KO tunica muscularis as well as significant increases in many genes linked with lipid processes and/or pro-apoptotic pathways (n = 4, error bars are SEM, *p < 0.05, **p < 0.01, unpaired t-test). d In isolated Dnmt1-KO SMC at P15, more dramatic losses of Dnmt1, Srf, and Myh11 were observed as well as increases in the pro-apoptotic markers Gadd45g & Nr4a1 (n = 2, error bars are SEM, *p < 0.05, **p < 0.01, unpaired t-test). Ubb was used as an endogenous control in (c, d) Transcriptomic analysis of tissue/SMC reveals loss of muscularis of both Dnmt1-KO and Dnmt1-WT mice. mature SMC markers and miRNAs and increases in lipid, Summaries and full expression profiles of messenger RNA immune processing, and pro-apoptotic genes (mRNA)/miRNA sequencing results can be found in To gather a more complete understanding of the gene Supplementary Table 1 and Supplementary Table 3, expression changes happening in Dnmt1-KO mice, we respectively. The transcriptome of Dnmt1-KO has 954 performed a transcriptomic analysis on the jejunal tunica unique mRNA gene annotations, compared to only 530 in Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 8 of 14 Fig. 6 Loss of Dnmt1 in GI-SMC causes global genomic CpG hypomethylation. a Reduced representation bisulfite sequencing of Dnmt1-KO mice compared to Dnmt1-WT mice demonstrates that most genes in Dnmt1-KO have at least one CpG site with significant levels of hypomethylation. b CpG methylation is reduced ~20% in Dnmt1-KO mice when compared to Dnmt1-WT mice. c While there is an ~20% reduction of CpG methylation in Dnmt1-KO mice, the patterns of methylation across each chromosome is not altered. d Globally, across the genome, CpG methylation is significantly reduced at promoters, exons and introns (**p < 0.00001, Fisher’s exact test). e Pro-apoptotic marker Nr4a1 shows significant losses of methylation at various genomic locations, including the promoter and exon 1 (*p < 0.01, Fisher’s exact test). f A UCSC genome browser view of the Nr4a1 gene with methylation tracks added (yellow=more methylated, red=less methylated) and the promoter region boxed in purple Dnmt1-WT with approximately equal amounts of unique half their expression, with similar amounts of both miRNA annotations between Dnmt1-KO (184) and -WT upregulated (289) and downregulated (265) miRNA (170) mice (Fig. 4a). When compared to Dnmt1-WT transcripts (Fig. 4b). The differences between the relative mice, Dnmt1-KO mice have 2,829 genes that at least amount and type of mRNA/miRNA transcripts expressed double their expression and only 1,319 that lose at least exposes a remodeled transcriptome in Dnmt1-KO mice. Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 9 of 14 Fig. 7 Only CpG sites show large loss of methylation in Dnmt1-KO. a, b While CHH sites and CHG sites did have individual sites that changed their methylation status in Dnmt1-KO mice, there was no strong genome-wide change in methylation status. c At CpG sites in Dnmt1-KO mice, there was marked hypomethylation at multiple site across the genome which was not observed in either CHH or CHG sites. Each dot represents an individual cytosine that mapped to both Dnmt1-KO and Dnmt1-WT samples; r is Pearson’s correlation coefficient and n is the total number of overlapping sites for each type of cytosine site Dnmt1-KO mice have considerable losses of SMC marker We selected upregulated gene groups associated with transcripts necessary for mature functioning, including lipids and apoptosis for further study. Oil red O staining Acta2 (25.7% loss), Myocd (37.4% loss), Myh11 (53.2% of cross-sections from Dnmt1-KO mice confirmed an loss), Srf(41.7% loss), and Tagln (19.9% loss) (Fig. 4c) as accumulation of lipids in their jejunal tunica muscularis well as reductions in necessary SMC miRNAs miR-143 (Fig. 5a), consistent with our mRNA-sequencing and GO (31.1% loss), miR-145 (40.8% loss), and miR-133 (70.4% term analysis. Next, we examined if Dnmt1-KO SMCs loss) (Fig. 4d). Since the transcription factors SRF and were undergoing apoptosis. DNA fragmentation is a 29 36 MYOCD are required for Myh11 expression , and the hallmark of apoptosis and was observed in Dnmt1-KO previously mentioned miRNAs , the reduction of Srf and tunica muscularis at P21 (Fig. 5b). Our transcriptome Myocd exacerbates the deterioration of SMCs as Srf represents the entirety of the tunica muscularis where the expression in Dnmt1-KO does not become significantly main population of cells are SMCs. However, immune cell reduced until P15 (Supplementary Figure 3) and therefore infiltration into the tunica muscularis may be responsible is likely not the primary insult involved in the loss of GI- for the increases in genes associated with immunity and SMCs. We also noted increases in miRNAs associated other GO categories. With the use of flow cytometry, we with regulating cellular identity and proliferation includ- removed cells from the tunica muscularis that were 31 32 33 + 37 ing, miR-10b , miR-21a , miR-486a , and miR-148a. CD45 , a known marker of immune cells , then selected − + MiR-148a is an established inhibitor of Dnmt1 expres- for cells that were CD45 and eGFP in P15 Dnmt1-KO sion , thus Dnmt1-KO SMCs, through increased miR- and Dnmt1-WT. Indeed, the results of our flow cytometry 148a, likely create even further repression of Dnmt1 identified ~5-fold increase in CD45 cells within Dnmt1- expression. Furthermore, both miR-10b and miR-148a KO tissue when compared to Dnmt1-WT mice while - + expression are known to regulated by genomic CpG CD45 and eGFP SMC were reduced ~3-fold (Supple- 31,35 - + methylation . For mRNA analysis, we selected 392 mentary Figure 5). Sorted CD45 , eGFP cells represented overexpressed genes in the tunica muscularis of Dnmt1- isolated SMCs. We confirmed that isolated SMCs from KO mice and 365 of them had some level of significant Dnmt1-KO mice had almost complete ablation of Dnmt1, demethylation (Supplementary Figure 4). Gene ontology significant reductions in Myh11 and Srf expression, and (GO) examination revealed that the most overrepresented dramatic increases in the pro-apoptotic genes Gadd45g biological process categories included those associated and Nr4a1 (Fig. 5d). However, we found other with lipids, apoptosis, defense responses to stress/stimuli, genes related to apoptosis and lipids upregulated in P15 and immune responses (Fig. 4e and Supplementary tissue that were not upregulated in isolated Dnmt1-KO Table 4). The quantitative increase in expression levels of SMCs, indicating that their upregulation is likely due to many of these genes that were found to be upregulated in immune cells infiltrating into the tunica muscularis Dnmt1-KO varied widely (Fig. 4c). (Fig. 5c, d). Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 10 of 14 Loss of Dnmt1 induces global genomic CpG SMC-restricted congenital Dnmt1-KO mice have shor- hypomethylation tened GI tracts, and lose their tunica muscularis by P21. Employing reduced representation bisulfite conver- Dnmt1-KO GI-SMC have significant genomic CpG sion sequencing, we discovered that ~90% (20,808) of demethylation, a loss of intercellular connectivity, annotated genes in Dnmt1-KO mice had at least one decreases in necessary SMC transcripts (Myh11, Srf, miR- CpG site with significant demethylation in either their 133, miR-143/145), and increases in both pro-apoptotic genes (Gadd45g, Nr4a1) and miRNAs associated with promoter (−1 kb of TSS), exons, or introns and ~17% (4,004) of genes had at least one CpG site with sig- changes in cellular identity (miR-21a, -148a, -186a, -10b). nificant demethylation in all sites, encompassing their In fact, we found aberrant hypermethylation promoter, exons, and introns (Fig. 6a) with genomic flanking a CArG box (CCATATAAGG: antisense) and hypomethylation being significant, on a genome-wide SRF binding site found within the second intron of Srf in scale, across annotated promoter, exon, and intron CpG Dnmt1-KO (Supplementary Figure 6) that may have sites (Fig. 6d). None of these patterns were observed for contributed to the reduction Srf expression, thus genomicCHG or CHH(Hcan be A, C, or T) sites increasing the apoptotic potential of Dnmt1-KO SMC. (Fig. 7a, b). Across the genome of Dnmt1-KO mice, Our results are similar to both congenital and inducible there was a ~20% loss of CpG methylation (Fig. 6c) with intestinal epithelial cell-specific Dnmt1-KO models. Both a majority of genes reducing their average CpG models show genomic hypomethylation, DNA damage, methylation levels from 0.8–0.6 to 0.6–0.4 (Fig. 6b). We and increased apoptosis in affected Dnmt1 knockout 40,41 also observed CpG sites that were strongly hyper- cells . In the congenital intestinal epithelial model, methylated in Dnmt1-KO mice, but these sites were only 35% of Dnmt1 knockout mice live until P20, similar considerably outnumbered by the amount of strongly to our SMC-restricted Dnmt1-KO mice which die around hypomethylated CpG sites (Fig. 7c). We created browser P21 and none have ever survived past P24. When Dnmt1 tracks that visualize and quantify the changes in CpG is inducibly ablated from intestinal epithelial cells at later methylation within the University of California, Santa stages of development, the mouse does not die and sees an Cruz (UCSC) genome browser for any annotated increase in Dnmt3b expression, which also directly com- genomic site. In addition, we added transcriptome level pensated for the loss of DNA methylation due to the RNA-sequencing (RNA-seq) data for both mRNA and knockout of Dnmt1. This pattern of compensatory miRNA from jejunal smooth muscle in both Dnmt1-KO DNMT isoform induction due to loss of a separate and Dnmt1-WT mice (Supplementary Table 3) to the DNMT could potentially be why we see increased browser. The browser tracks were installed at our DNMT3A seen during development in our Dnmt1-KO Smooth Muscle Transcriptome Project webpage: mice, as DNMT3A has been shown to be able to maintain https://med.unr.edu/physio/transcriptome,entitled methylation patterns in mouse embryonic stem cells “UCSC Smooth Muscle Methylome Browser.” This lacking DNMT1 . browser provides a comprehensive reference for tran- While it is clear that GI-SMCs in congenital Dnmt1-KO scriptomic and genomic DNA methylation status at mice are undergoing remodeling and apoptosis, it is not CpG sites in the jejunal smooth muscle of Dnmt1-WT clear what the causative insult of these changes is, and it is and Dnmt1-KO mice. The browser can also interact more likely a combination of several pro-apoptotic with genome-level bioinformatics (e.g., ENCODE) influences including genomic CpG hypomethylation, data publically available in the UCSC Genome epigenome modulation, lack of cellular connectivity, Browser. In Dnmt1-KO mice, Nr4a1 has significant upregulation of pro-apoptotic genes/miRNAs, down- lossesofmethylation at itspromoter aswellasexons 1, regulation of required SMC proteins/miRNAs, exposure 2, 5, and 6 and introns 5 and 6 (Fig. 6e), as shown to lipoproteins, and chromosomal instability. As on the browser (Fig. 6f), suggesting the expression of the it pertains to pro-apoptotic genes upregulated in Dnmt1- pro-apoptotic Nr4a1 could be regulated by CpG KO SMCs, Gadd45g is a marker of DNA damage, acti- methylation levels controlled by DNMT1, agreeing with vator of the p38/c-Jun N-terminal kinase apoptotic 39 43 previous findings . The entirety of our CpG methy- pathway and has been correlated with increased lomic results and analysis for every annotated promoter, expression in a double knockout of DNMT1 and exon, and intron can be found in Supplementary DNMT3B in vitro . Nr4a1 is an orphan nuclear Table 5. receptor that has been shown to be regulated by DNMT1 and the methylation level of its promoter region . Discussion Expression of Nr4a1 is associated with an induction of By eliminating Dnmt1 from the murine genome in a apoptosis through multiple pathways , as well as SMC-restricted model, we showed the in vivo necessity direct inhibition of the growth of SMCs . As for genomic of Dnmt1 in the embryonic development of GI-SMCs. CpG hypomethylation, it is likely a causative feature Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 11 of 14 involved in the transcriptomic remodeling of mRNA and Methods miRNA expression, but probably does not account for all Mice Cre-eGFP/+ 24 changes. Only about 10% of all miRNAs have been shown In a C57BL/6 background, smMHC mice loxllox 25 to change expression based on levels of DNA methyla- were crossed with Dnmt1 mice to make 47 -/-;Cre-eGFP/+ tion , denoting a secondary role of DNMT1 in regulating smDnmt1 mice. PCR confirmed genotypes miRNA expression in GI-SMCs. DNMT1 interacts with (Supplementary Figure 1). Primer sequences are found in up to 58 transcription factors , and many epigenetic Supplementary Table 2. Mice were bred, maintained, and modifiers , and thus it is probable that loss of Dnmt1 has euthanized following guidelines set by the Institutional downstream deleterious consequences outside of Animal Care and Use Committee at the University of genomic hypomethylation. Furthermore, in Dnmt1-null Nevada-Reno Animal Resources. cells, DNA mismatch repair is impeded and microsatellite instability increases ~4-fold , indicating a pivotal, Human tissue methylation-independent role of DNMT1 in DNA Isolated human GI smooth muscle tissue sections were damage repair. While previous in vitro studies of com- donated by Dr. Kent Sasse (Sasse Surgical Associates) and plete Dnmt1 knockout in mouse embryonic stem cells Dr. Laren S. Becker (Stanford University School of Med- have reported more dramatic genomic CpG methylation icine). Use of human tissue was approved by the Institu- reductions from ~75% to ~18% or lower at specific tional Review Board at the University of Nevada, Reno. sites , cell-selective in vivo knockout of Dnmt1 in intestinal epithelial results in CpG methylation reductions Tissue isolation and preparation 40,41 from ~75% to ~45% , similar to our own in vivo The GI tract was extracted into 1× Hank’s Calcium Free findings (Fig. 6c). Buffer. Tissue then underwent one of two processes: (1) Additionally, the observed lack of nuclear/nucleolar smooth muscle isolation and (2) whole tissue organization in electron microscopy images of Dnmt1-KO preparation for cross-sectioning. (1) The muscle layer was could be indicative of catastrophic chromosomal peeled from the mucosa for use in extractions/experi- instability. Of the 25 genes in the CIN25 gene signature of ments. (2) The tissue was fixed in 4% paraformaldehyde chromosomal instability , 23 were found to be upregu- (PFA), a dehydration in 20% sucrose/1× phosphate- lated in Dnmt1-KO smooth muscle tissue (data not buffered saline (PBS), then placed in 1:1 OCT/20% shown). We were also able to show that congenital sucrose and super-cooled by liquid nitrogen. Using a Dnmt1-KO mice have reduced levels of all TET proteins, cryostat microtome (Leica, Wetzlar, Germany), and thus it is likely that Dnmt1-KO mice cannot properly 8 µM thick tissue sections were cut onto slides coated regulate 5-hmc or 5-mc levels, causing dysregulation of with VectaBond (Maravai Biosciences, San Diego, the entire DNA methylation apparatus. We also showed CA, USA), and allowed to dry. Prepared slides were that expression levels of DNMT1 and TET3 are associated used for downstream immunohistochemical/histological and correlated within the inflamed tunica muscularis of staining. colorectal cancer tissue as well as in the jejunal tunica muscularis of both Dnmt1-WT and Dnmt1-KO mice. Isolation of total and small RNAs Further studies are needed to elucidate the associations Isolated muscle tissue was placed in lysis buffer and and interactions between DNMT1 and TET3 in inflamed homogenized by bead beating in an air-cooled Bullet and injured smooth muscle. Blender Storm (Next Advance, Troy, NY, USA). Total Most surprisingly, Dnmt1-KO SMC show an accumu- RNAs and miRNAs were isolated from smooth muscle lation of lipids and lipid associated transcripts within their and sorted SMCs using mirVana miRNA Isolation Kit tunica muscularis. Many of the genes upregulated in our (Ambion, Foster City, CA, USA) as previously described . transcriptomic analysis are commonly associated with Extracted total RNAs and small RNAs were used for vSMC containing atherosclerotic plaques . In vSMC, quantitative PCR (qPCR) and RNA-seq. lipoproteins, based on oxidation status, have been impli- cated in changing the phenotypic identity of SMCs into Quantitative PCR either a proliferative or apoptotic fate . Extracted total RNA was reverse transcribed into In conclusion, the lack of Dnmt1 in GI-SMC resulted in complementary DNA (cDNA) using SuperScript III an inability to maintain de novo methylation patterns and Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA), the consequences of this loss become exacerbated with after DNA removal with DNA-free DNA Removal Kit each round of mitotic division. As can be surmised, fur- (Ambion). cDNA samples were then quantified, diluted, ther research into all observed phenotypic aspects of and tested for quality. Standard qPCR protocol was car- Dnmt1-KO is necessary to understand all cellular con- ried out on a 7900HT Fast Real-Time PCR System sequences of DNMT1 deficiency. (Applied Biosystems, Foster City, CA, USA) as previously Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 12 of 14 described . Primer sequences are found in Supplemen- western blot banding patterns was done using ImageJ and tary Table 2. histograms were produced in SigmaPlot (Systat, San Jose, CA, USA). Immunohistochemistry Slides were exposed to 1× Tris-buffered saline (TBS)/ DNA methylome 0.1% Tween 20 followed by exposure to 4% milk/1× TBS/ Genomic DNA was isolated from jejunal tunica mus- 0.1% Tween 20. Slides were then incubated at 4 °C with 1° cularis at P15 (2 males and 2 females for both Dnmt1-KO antibody. The following day, slides were incubated with 2° and Dnmt1-WT; n = 4) using Qiagen (Hilden, Germany) antibody at room temperature. After washing the slides in AllPrep DNA spin columns. DNA samples were pooled 1× TBS, the slides were mounted and allowed to cure. See and shipped to Zymo Research Corporation (Irvine, CA, Supplementary Table 2 for antibodies/dilutions. Slides USA) where the samples underwent Methyl-MidiSeq were imaged with an Olympus FluoView FV1000 (Tokyo, which involves genomic DNA extraction, fragmentation, Japan) confocal microscope. end modification, bisulfite conversion, limited amplifica- tion, and sequencing. After sequencing, bioinformatic Hematoxylin and eosin staining processing was utilized in comparing amount of methy- Cross-section slides were exposed to hematoxylin, lation at CpG, CHH, and CHG sites between Dnmt1-WT rinsed with water, and then stained in eosin Y. After and Dnmt1-KO samples. Methylation ratios were calcu- several dehydrating baths of increasing ethanol (EtOH) lated (methylated cytosines/total cytosine reads) for concentration (80%, 95%, 100%) and two terminal xylene individual sites and then averaged across regions (exons, baths, the slides were mounted and allowed to cure. Slides introns, promoters = −1 kb of TSS). These averages were were imaged with an Olympus BX43 (Tokyo, Japan) then used in Fisher’s exact test to quantify significant brightfield microscope. changes (p < 0.05) in methylation. The reference genome used was GRCm38/mm10. The data from our bisulfite Oil Red O staining conversion sequencing were submitted to the Gene Cross-section slides were exposed to pure propylene Expression Omnibus (GEO) repository with the following glycol, then moved to 0.5% Oil Red O solution at 60 °C accession numbers: GSM2947783, Dnmt1-WT jejunum; followed by soaking in 85% propylene glycol. slides were GSM2947784, Dnmt1-KO jejunum. then rinsed and counterstained with hematoxylin, rinsed with water, then mounted with aqueous mounting RNA sequencing reagent and allowed to cure. Slides were imaged Total RNAs and small RNAs isolated from Dnmt1-KO with a Keyence BZ-x710 (Osaka, Japan) brightfield and Dnmt1-WT jejunal tunica muscularis at P15 (2 males microscope. and 2 females for both Dnmt1-KO and Dnmt1-WT; n = 4) were shipped to LC Sciences (Houston, TX, USA) Electron microscopy where mRNA-seq and miRNA-seq were performed. Whole jejunum and colon tissue were fixed (3% glu- Sequencing data analysis, alignments, annotations, and taraldehyde and 4% PFA in 0.1 M phosphate buffer) for bioinformatics processing were carried out as previously several days at room temperature. Specimens were then described . The reference genome used was GRCm38/ fixed in 1% OsO at 4 °C, rinsed in distilled water, then mm10. Our RNA-seq data for both mRNA and miRNA block stained with saturated aqueous uranyl acetate were submitted to the GEO repository with the following solution, followed by dehydration in a graded series of accession numbers: GSM2936348, Dnmt1-WT jejunum EtOH and finally embedded in EPON 812. Ultrathin mRNA; GSM2936349, Dnmt1-KO jejunum mRNA; sections were stained with uranyl acetate and lead citrate GSM2936350, Dnmt1-WT jejunum miRNA; and then examined using a Hitachi H-7650 (Tokyo, Japan) GSM2936351 Dnmt1-KO jejunum miRNA. microscope. Cell sorting Western blot Cells were dispersed as previously described . Cells Isolated SM tissue was ground by mortar and pestle were incubated with APC-Cy7 Anti-CD45 (Biolegend, in modified RIPA buffer. Transferred polyvinylidene Clone 30-F11, 1.0 μg/mL; San Diego, CA, USA) followed fluoride membranes were pre-blocked in 5% milk, then by washing with PBS/1% FBS. Resuspended cells had exposed to 1° antibody at 4 °C in 2% milk/0.05% Tween 20 Hoechst 33258 (1 μg/mL) added as a viability marker. at varying dilutions. The membranes were exposed to 2° Cells were sorted and analyzed using the BD Biosciences antibody at room temperature. All antibodies used can be (San Jose, CA, USA) FACSAria II Special Order Research found in Supplementary Table 2. Imaging by UVP EC3 Product with a 130 μm nozzle with sheath pressure at 12 (Upland, CA, USA) Imaging system. Quantification of psi. The 355 nm laser excited Hoechst 33258 with a 450/ Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 13 of 14 50 nm bandpass filter. The eGFP was excited using a 488 3. Liu, R., Leslie,K. L.& Martin,K. A.Epigenetic regulationofsmoothmusclecell plasticity. Biochim. Biophys. Acta 1849,448–453 (2015). nm laser with a 530/30 nm bandpass filter. A neutral 4. Hu, B., Gharaee-Kermani, M., Wu, Z. & Phan, S. H. Epigenetic regulation of density filter 2 was used on the forward scatter detector myofibroblast differentiation by DNA methylation. Am.J.Pathol. 177,21–28 due to the high forward scatter properties. Cells that (2010). - + 5. Ning, Y. et al. 5-Aza-2’-deoxycytidine inhibited PDGF-induced rat airway were CD45 and eGFP were sorted into PBS/1% FBS. smooth muscle cell phenotypic switching. Arch. Toxicol. 87,871–881 (2013). Acquisition was performed on BD FACSDiva 8.0 and 6. Hinoue,T.etal. Genome-scale analysis of aberrant DNA methylation in col- TreeStar Flowjo (Ashland, OR, USA) was used to generate orectal cancer. Genome Res. 22,271–282 (2012). 7. Sadler, T. et al. Genome-wide analysis of DNA methylation and gene figures. expression defines molecular characteristics of Crohn’s disease-associated fibrosis. Clin. Epigenetics 8,30(2016). Acknowledgements 8. Ramchandani, S.,Bhattacharya, S. K.,Cervoni,N.&Szyf,M. DNA methylationis We would like to thank Drs. Michael Kotlikoff and Laurie Jackson-Grusby for a reversible biological signal. Proc. Natl. Acad. Sci. USA 96,6107–6112 (1999). Cre-eGFP/+ lox/lox their generous donations of the congenital smMHC and Dnmt1 9. Gardiner-Garden, M. & Frommer, M. CpG Islands in vertebrate genomes. J. Mol. mouse lines respectively as well as Benjamin J. Weigler and Walt F. Mandeville Biol. 196,261–282 (1987). (Laboratory Animal Medicine, University of Nevada, Reno) for the excellent 10. Jones, P. A. Functions of DNA methylation: islands, start sites, gene bodies and animal services provided to the mice. We would also like to thank Dr. Kent M. beyond. Nat. Rev. Genet. 13, 484–492 (2012). Sanders for COBRE Core B and Byoung H. Koh for the cell sorting services as 11. McDonald, O. G. &Owens,G.K.Programming smooth muscle plasticity with well as Dr. Tong Zhou for his assistance in bioinformatic data presentation. B.G. chromatin dynamics. Circ. Res. 100,1428–1441 (2007). J. was supported by an award from the J. M. Hitchcock, Graduate Student 12. Miano,J.M., Long,X.& Fujiwara, K. Serum response factor: master regulatorof Research Fund. the actin cytoskeleton and contractile apparatus. Am.J.Physiol.CellPhysiol. This study was supported by National Institutes of Health (NIH) (DK094886 and 292,C70–C81 (2007). DK103055 to S.R.) and COBRE Phase III FACSAria (P30 GM110767 to Kent M. 13. Miano,J.M., Cserjesi,P., Ligon, K. L.,Periasamy,M.& Olson, E. N. Smooth Sanders). The funders had no role in study design, data collection/analysis, muscle myosin heavy chain exclusively marks the smooth muscle lineage decision to publish, or preparation of the manuscript. during mouse embryogenesis. Circ. Res 75,803–812 (1994). 14. Skalli, O. et al. A monoclonal antibody against α-smooth muscle actin: Author details a new probe for smooth muscle differentiation. J. Cell Biol. 103,2787–2796 Department of Physiology and Cell Biology, University of Nevada School of (1986). Medicine, Reno, NV 89557, USA. Department of Morphological and 15. Torella, D. et al. MicroRNA-133 controls vascular smooth muscle cell pheno- Physiological Sciences, University of Fukui, Fukui 910-8507, Japan. Sasse typic switch in vitro and vascular remodeling in vivo. Circ. Res. 109,880–893 Surgical Associates, Reno, NV 89502, USA. Gastroenterology and Hepatology, (2011). Stanford University School of Medicine, Stanford, CA 94305, USA 16. Cordes, K. R. et al. MiR-145 and miR-143 regulate smooth muscle cell fate and plasticity. Nature 460,705–710 (2009). Authors' contributions 17. Wu, W., Shen, X. & Tao, S. Characteristics of the CArG-SRF binding context in R.M.B., B.G.J., S.E.H., and S.R. designed research; R.M.B., B.G.J., S.E.H., and K.H. mammalian genomes. Mamm. Genome 21,104–113 (2010). performed research; B.G.J., R.M.B., S.E.H., and S.R. analyzed data; K.C.S. and L.S.B. 18. McDonald, O. G.,Wamhoff, B. R.,Hoofnagle,M.H.&Owens, G. K. Controlof contributed materials. B.G.J. wrote the paper and S.R revised it. SRF binding to CArG box chromatin regulates smoothmusclegeneexpres- sion in vivo. J. Clin. Invest. 116,36–48 (2006). Conflict of interest 19. Warnecke, P. M. & Clark,S. J.DNA methylationprofile of the mouse skeletal α- The authors declare that they have no conflict of interest. actin promoter during development and differentiation. Mol. Cell Biol. 19, 164–172 (1999). 20. Yamamura, H. et al. Structure and expression of the human SM22α gene, assignment of the gene to chromosome 11, and repression of the promoter Publisher's note activity by cytosine DNA methylation. J. Biochem. 122,157–167 (1997). Springer Nature remains neutral with regard to jurisdictional claims in 21. Robert, M. F. et al. DNMT1 is required to maintain CpG methylation and published maps and institutional affiliations. aberrant gene silencing in human cancer cells. Nat. Genet 33,61–65 (2003). 22. Pastor, W. A., Aravind, L. & Rao, A. TETonic shift: biological roles of TET proteins Funding in DNA demethylation and transcription. Nat. Rev. Mol. Cell Biol. 14,341–356 Wessex Cancer Trust, CRUK—Southampton Internal Pilot Grant, EU-FP7 Marie (2013). Curie (CANOMICS), Annual Adventures in Research—University of 23. Li, E., Bestor, T. H. & Jaenisch, R. Targeted mutation of the DNA methyl- Southampton, EU-Excellence II—Systems Biology Framework FRA-SYS (Grant transferase gene results in embryonic lethality. Cell 69,915–926 (1992). 4072), Cancer Research UK (Grants C34999/A13719 and RG84119), and MRC 24. Xin,H.-B.,Deng, K.-Y., Rishniw, M., Ji, G. & Kotlikoff, M. I. Smooth muscle Clinician Scientist Fellowship (Grant G1002565). expression of Cre recombinase and eGFP in transgenic mice. Physiol. Genom. 10,211–215 (2002). Supplementary Information accompanies this paper at https://doi.org/ 25. Jackson-Grusby, L. et al. Loss of genomic methylation causes p53-dependent 10.1038/s41419-018-0495-z. apoptosis and epigenetic deregulation. Nat. Genet. 27,31–39 (2001). 26. Coppedè, F. Epigenetic biomarkers of colorectal cancer: focus on DNA Received: 30 October 2017 Revised: 31 January 2018 Accepted: 12 February methylation. Cancer Lett. 342,238–247 (2014). 2018 27. Owens, G. K., Kumar, M. S. & Wamhoff, B. R. Molecular regulation of vascular smooth muscle cell differentiation in development and disease. Physiol. Rev. 84,767–801 (2004). 28. Sanders,K.M., Ördög, T.,Koh,S.D., Torihashi, S. &Ward, S. M. Development and plasticity of interstitial cells of Cajal. Neurogastroenterol. Motil. 11, 311–338 References (1999). 1. Wedel, T. et al. Novel smooth muscle markers reveal abnormalities of the 29. Wang, Z., Wang, D.-Z., Teg Pipes, G. C. & Olson, E. N. Myocardin is a master intestinal musculature in severe colorectal motility disorders. Neurogas- regulatorofsmoothmusclegeneexpression. Proc.Natl. Acad.Sci.USA 100, 7129–7134 (2003). troenterol. Motil. 18,526–538 (2006). 30. Park, C. et al. Serum response factor-dependent microRNAs regulate gastro- 2. Rolle, U., O’Briain, S., Pearl, R. H. & Puri, P. Megacystis-microcolon-intestinal intestinal smooth muscle cell phenotypes. Gastroenterology 141,164–175 hypoperistalsis syndrome: evidence of intestinal myopathy. Pediatr. Surg. Int. (2011). 18,2–5 (2002). Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 14 of 14 31. Li, Z. et al. DNA methylation downregulated mir-10b acts as a tumor sup- 45. Li, Q.-X., Ke, N., Sundaram,R.&Wong-Staal,F.NR4A1, 2, 3 -an orphan nuclear pressor in gastric cancer. Gastric Cancer 18,43–54 (2015). hormone receptor family involved in cell apoptosis and carcinogenesis. Histol. 32. Ji, R. et al. MicroRNA expression signature and antisense-mediated depletion Histopathol. 21,533–540 (2006). reveal an essential role of microRNA in vascular neointimal lesion formation. 46. Mayer, P., Hinze, A. V., Harst, A. & von Kügelgen, I. A receptors mediate the 2B Circ. Res. 100, 1579–1588 (2007). induction of early genes and inhibition of arterial smooth muscle cell pro- 33. Oh, H.-K. et al. Genomic loss of miR-486 regulates tumor progression and the liferation via Epac. Cardiovasc. Res. 90,148–156 (2011). OLFM4 antiapoptotic factor in gastric cancer. Clin. Cancer Res. 17,2657–2667 47. Han,L., Witmer,P.D., Casey, E.,Valle,D. & Sukumar, S. DNAmethylation (2011). regulates microRNA expression. Cancer Biol. Ther. 6,1284–1288 (2007). 34. Nan,W.etal. MicroRNA-21 and microRNA-148a contribute to DNA hypo- 48. Hervouet, E., Vallette, F. M. & Cartron, P.-F. Dnmt1/transcription factor inter- methylation in lupus CD4+T cells by directly and indirectly targeting DNA actions: an alternative mechanism of DNA methylation inheritance. Genes methyltransferase 1. J. Immunol. 184,6773–6781 (2010). Cancer 1,434–443 (2010). 35. Lujambio, A. et al. A microRNA DNA methylation signature for human cancer 49. Qin, W., Leonhardt, H. & Pichler, G. Regulation of DNA methyltransferase 1 by metastasis. Proc. Natl. Acad. Sci. USA 105, 13556–13561 (2008). interactions and modifications. Nucleus 2,392–402 (2011). 36. Bortner,C.D., Oldenburg, N. B. E. &Cidlowski,J.A. The role of DNAfrag- 50. Jin, B. & RobertsonK. D. DNA methyltransferases (DNMTs), DNA damage repair, mentationinapoptosis. Trends Cell Biol. 5,21–26 (1995). and cancer. Adv. Exp. Med. Biol. 754,3–29 (2013). 37. Donovan, J. A. & Koretzky, G. A. CD45 and the immune response. J. Am. Soc. 51. Li, Z. et al. Distinct roles of DNMT1-dependent and DNMT1-independent Nephrol. 4,976–985 (1993). methylation patterns in the genome of mouse embryonic stem cells. Genome 38. Kent, W. J. et al. The Human Genome Browser at UCSC. Genome Res. 12, Biol. 16, 115 (2015). 996–1006 (2002). 52. Martos, S. N. et al. Two approaches reveal a new paradigm of ‘switchable or 39. Chen, Y.-T., Liao, J.-W., Tsai, Y.-C. & Tsai, F.-J. Inhibition of DNA methyltransferase genetics-influenced allele-specificDNA methylation’ with potential in human 1 increases nuclear receptor subfamily 4 group A member 1 expression and disease. Cell Discov. 3, 17038 (2017). decreasesblood glucosein type2diabetes. Oncotarget 7, 39162–39170 53. Carter, S. L., Eklund, A. C., Kohane, I. S., Harris, L. N. & Szallasi, Z. A signature of (2016). chromosomal instability inferred from gene expression profilespredictsclinical 40. Elliott, E. N., Sheaffer, K. L. & Kaestner, K. H. The ‘de novo’ DNA methyl- outcome in multiple human cancers. Nat. Genet. 38,1043–1048 (2006). transferase Dnmt3b compensates the Dnmt1-deficient intestinal epithelium. 54. Lusis, A. J., Fogelman, A. M. & Fonarow, G. C. Genetic basis of atherosclerosis: Elife 5,1–15 (2016). part I. New genes and pathways. Circulation 110,1868–1873 (2004). 41. Elliott, E. N.,Sheaffer, K. L.,Schug,J., Stappenbeck, T. S. &Kaestner, K. H. Dnmt1 55. Johnson, J. L. Emerging regulators of vascular smooth muscle cell function in is essential to maintain progenitors in the perinatal intestinal epithelium. the development and progression of atherosclerosis. Cardiovasc. Res. 103, Development 142,2163–2172 (2015). 452–460 (2014). 42. Liang, G. et al. Cooperativity between DNA methyltransferases in the 56. Ro, S. et al. Cloning and expression profiling of small RNAs expressed in the maintenance methylation of repetitive elements. Mol. Cell Biol. 22,480–491 mouse ovary. RNA 13,2366–2380 (2007). (2002). 57. Lee, M. Y. et al. Smooth muscle cell genome browser: enabling the identifi- 43. Takekawa, M. & Saito, H. A family of stress-inducible GADD45-like proteins cation of novel serum response factor target genes. PLoS One 10, e0133751 mediate activation of the stress-responsive MTK1/MEKK4 MAPKKK. Cell 95, (2015). 521–530 (1998). 58. Ro, S. et al. A model to study the phenotypic changes of interstitial cells of 44. Ying, J. et al. The stress-responsive gene GADD45G is a functional tumor Cajal in gastrointestinal diseases. Gastroenterology 138,1068–1078 (2010). suppressor, with its response to environmental stresses frequently disrupted epigenetically in multiple tumors. Clin. Cancer Res. 11,6442–6449 (2005). Official journal of the Cell Death Differentiation Association http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Cell Death & Disease Springer Journals

DNA methylation, through DNMT1, has an essential role in the development of gastrointestinal smooth muscle cells and disease

Free
14 pages

Loading next page...
 
/lp/springer_journal/dna-methylation-through-dnmt1-has-an-essential-role-in-the-development-zauuFWuKOK
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s)
Subject
Life Sciences; Life Sciences, general; Biochemistry, general; Cell Biology; Immunology; Cell Culture; Antibodies
eISSN
2041-4889
D.O.I.
10.1038/s41419-018-0495-z
Publisher site
See Article on Publisher Site

Abstract

DNA methylation is a key epigenetic modification that can regulate gene expression. Genomic DNA hypomethylation is commonly found in many gastrointestinal (GI) diseases. Dysregulated gene expression in GI smooth muscle cells (GI- SMCs) can lead to motility disorders. However, the consequences of genomic DNA hypomethylation within GI-SMCs are still elusive. Utilizing a Cre-lox murine model, we have generated SMC-restricted DNA methyltransferase 1 (Dnmt1) knockout (KO) mice and analyzed the effects of Dnmt1 deficiency. Dnmt1-KO pups are born smaller than their wild- type littermates, have shortened GI tracts, and lose peristaltic movement due to loss of the tunica muscularis in their intestine, causing massive intestinal dilation, and death around postnatal day 21. Within smooth muscle tissue, significant CpG hypomethylation occurs across the genome at promoters, introns, and exons. Additionally, there is a marked loss of differentiated SMC markers (Srf, Myh11, miR-133, miR-143/145), an increase in pro-apoptotic markers (Nr4a1, Gadd45g), loss of cellular connectivity, and an accumulation of coated vesicles within SMC. Interestingly, we observed consistent abnormal expression patterns of enzymes involved in DNA methylation between both Dnmt1-KO mice and diseased human GI tissue. These data demonstrate that DNA hypomethylation in embryonic SMC, via congenital Dnmt1 deficiency, contributes to massive dysregulation of gene expression and is lethal to GI-SMC. These results suggest that Dnmt1 has a necessary role in the embryonic, primary development process of SMC with consistent patterns being found in human GI diseased tissue. Introduction as in vascular SMC (vSMC) found in atherosclerotic The aberrant growth, or loss, of smooth muscle lesions . Recent research has shown that inducing geno- cells (SMCs) is a common symptom found in mic DNA hypomethylation can change SMC phenotypic burdensome gastrointestinal (GI) diseases such as identity, growth patterns, and expression levels of con- 1 4,5 chronic intestinal pseudo-obstruction and megacystis- tractile genes , giving preliminary evidence that altering microcolon-intestinal hypoperistalsis syndrome as well DNA methylation can shift the phenotypic identity of SMCs. Global changes in DNA methylation patterns are characteristic of digestive pathologies such as colorectal Correspondence: Seungil Ro (sro@med.unr.edu) 6 7 cancer and Crohn’s disease . Department of Physiology and Cell Biology, University of Nevada School of The process of methylating DNA at cytosine residues, Medicine, Reno, NV 89557, USA Department of Morphological and Physiological Sciences, University of Fukui, specifically at CpG sites, is a dynamic and reversible epi- Fukui 910-8507, Japan genetic modification . In the mammalian genome, CpG Full list of author information is available at the end of the article These authors contributed equally: Brian G. Jorgensen, Robyn M. Berent, Se Eun Ha. Edited by D. Heery © The Author(s) 2018 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to theCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Official journal of the Cell Death Differentiation Association 1234567890():,; 1234567890():,; Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 2 of 14 Cre-eGFP/+ sites can be found clustered together in regions entitled By crossing smMHC , a mouse line that selec- CpG islands that are often found in close proximity to, or tively expresses Cre recombinase and eGFP under the within, transcription start sites (TSS) as well as the pro- control of a Myh11 promoter beginning at embryonic 9 lox/lox 25 moters of housekeeping and/or tissue-specific genes .The day 12.5, with a Dnmt1 mouse line , we were able to -/-;Cre-eGFP/+ addition of a methyl group to cytosine at its fifth carbon, successfully generate congenital smDnmt1 Cre-eGFP/+ creating 5-methylcytosine (5-mc), by DNA methyl- (Dnmt1-KO) mice. In smMHC (Dnmt1-WT) transferases (DNMT1, DNMT3A, DNMT3B) can cause mice, cells that express Cre recombinase also express repression of associated genes as this mark, at and around eGFP, thus making it possible to identify Cre expression the TSS, has been linked to transcriptionally repressed through enhanced green fluorescent protein (eGFP) chromatin that denies access for transcription factors, such fluorescence. Knockout was confirmed through genotyp- 10,11 as serum response factor (SRF) .Thus, as SRF is a ing with primers specific for Cre, eGFP, known loxP sites, master regulator of contractile genes in SMC ,the tran- and deletion of floxed cassette (Supplementary Figure 1). scription of many contractile smooth muscle genes (Myh11, Only those mice that were positive for Cre, both genomic 13,14 Acta2) and microRNA (miRNAs; mir-133, mir-143/ loxP sites, and the deletion of the floxed cassette (Dnmt1- 15,16 145) is dependent on the binding of SRF to CArG (CC KO) presented the following symptoms: Dnmt1-KO mice 17,18 [A/T] GG) boxes at these loci . CArG boxes are highly were born similar in size and body mass to their wild-type conserved, have high CpG content in the region bound by (WT) littermates, but they have significantly less body SRF, are commonly found within promoter regions/close to mass by postnatal day 15 (P15) continuing through to P21 TSS, and are recognized specifically by SRF . Additionally, (Fig. 1a, b) when death occurs. Gross anatomical images high CpG content sites, such as those around CArG boxes, show intestinal dilation beginning at P10 that continues to are also well-established targets for cytosine methylation by progress until P21 (Fig. 1c) when dilation is severe enough 19,20 DNA methyltransferases . While several studies have to cause intestinal ischemia or perforation of the dete- given valuable insight into the role of DNA methylation in riorated intestinal wall, causing septic shock. Dnmt1-KO SMC, many of them similarly use cytidine analogs mice also have significantly shortened GI tracts in both 4,5 in vitro , which are not selective in their inhibition of any the small intestine (Fig. 1d) and colon (Fig. 1e) by P21. specific DNMT isoform, nor do they present any efficacy of these techniques in vivo in GI-SMC. Dynamic changes of enzymes involved in DNA methylation To further this area of research, we have selectively in Dnmt1-KO mice and diseased human tissue eliminated the methyltransferase, Dnmt1, from the murine DNA methylation is positively regulated by DNMT1/ SMC genome. Dnmt1 is known as the maintenance DNMT3A/DNMT3B and is negatively regulated by the methyltransferase as its main enzymatic function is to Ten-Eleven Translocation (TET) family of proteins faithfully restore methylation marks on newly formed (TET1/TET2/TET3). In the tunica muscularis of Dnmt1- daughter strands of DNA .Deficiency of Dnmt1 caused WT mice, expression levels of DNMT1 and DNMT3A are loss of the intestinal tunica muscularis and halted peristaltic developmentally dynamic while DNMT3B is constantly contractions which induced intestinal dilation and death expressed at all time points (Fig. 2a). DNMT1 is highly around P21. CpG methylation is significantly reduced expressed at embryonic day 18 (E18) and gradually genome-wideatpromoters,exons,and introns. We also decreases from birth to P20, with the opposite pattern found that markers of mature SMC are lost, pro-apoptotic being seen with DNMT3A. Intriguingly, temporal gene expression is increased, and Dnmt1 knockout (KO) expression levels of TET1/TET2/TET3 are reduced in SMCs cytoplasmically accumulate a large number of coated Dnmt1-KO with the reductions of TET2 and TET3 being vesicles. Furthermore, we noted consistent patterns of dys- significant by P15. DNMT1 levels were drastically reduced regulation of DNMT1 and TET3, an enzyme that initiates in Dnmt1-KO after P1 and the protein was not detected at the demethylation of DNA , between KO mice and var- P20 (Fig. 2a). As previously mentioned, in Dnmt1-KO iously diseased human GI tissue, including colon/anal can- smooth muscle tissue, TET2 and TET3 levels were sig- cers and Crohn’s disease. To our knowledge, this is the first nificantly reduced (Fig. 2d), suggesting that expression of in vivo evidence showing the requirement of a DNA these TET isoforms may be dependent on DNMT1. methyltransferase, and the importance of DNA methylation, Furthermore, as Dnmt1-KO smooth muscle tissue in the primary development and differentiation of GI-SMC. developed, potentially compensatory DNMT3A induction occurred (Fig. 2a). Also, as DNMT1 begins to be reduced Results in Dnmt1-KO, so are markers of SMC maturity and Generation, and confirmation, of deletion in SMC- contractibility, including SRF, MYH11, and SM22α restricted Dnmt1 knockout mice (Fig. 2b, d). Since MYH11 and SM22α (Tagln) are SRF -/- 23 12 As global Dnmt1 mice are embryonic lethal , it was targets , the reduction of SRF in Dnmt1-KO could result -/- necessary to generate a SMC-restricted Dnmt1 KO line. in, or exacerbate, the reduced expression of these SRF Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 3 of 14 Fig. 1 SMC-restricted congenital Dnmt1-KO mice are smaller in size/mass with shorter GI tracts that become increasingly dilated over time. a SMC-restricted congenital Dnmt1-KO mice at P20 are smaller when compared to Dnmt1-WT littermates. b No significant differencesinbodymasswere seen between Dnmt1-KO mice and Dnmt1-WT mice until P15 when WT mice continued to gain mass and Dnmt1-KO began to lose mass. Dnmt1-KO slowly began to regain mass by P18 continuing into P21, most likely due to lack of peristaltic movement in the intestine causing buildup of food in the small intestine. c By P10, dilation can be seen in the small intestine of Dnmt1-KO mice with dilation increasing in size until P21 when death occurs. d, e By P21, Dnmt1-KO mice have significantly shortened small intestines and colons. For (b, d, e) n= 10, error bars are SD; *p < 0.005; **p < 0.001, unpaired t-test Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 4 of 14 Fig. 2 Enzymes regulating DNA methylation and mature SMC markers are dynamically altered in the jejunal tunica muscularis of Dnmt1- KO mice and inflamed human colonic tissue. a Western blotting of isolated jejunal tunica muscularis from Dnmt1-WT and Dnmt1-KO mice show temporal changes in expression of enzymes regulating DNA methylation (DNMT1, 3A and 3B) and demethylation (TET1–3). Note a progressive loss of DNMT1, and a massive reduction in expression of TET2 and TET3 in Dnmt1-KO mice. b A loss of DNMT1 correlates with significant reductions in mature SMC markers (SRF, MYH11, and SM22α) that become more pronounced with age indicating that GI-SMCs in Dnmt1-KO mice are indeed losing their mature, contractile status. c Human colonic tunica muscularis shows increased expression of DNMT1 in inflamed tissues along with TET3. Diverticulitis-M (marginal area); Diverticulitis-P (pouch). d, e Quantification of western blotting from Dnmt1-KO and Dnmt1-WT and human healthy and inflamed colonic tissue; n = 2–3 for mouse and human separately, error bars are SEM, *p < 0.005; **p < 0.001, unpaired t-test target genes and their subsequent protein products. Due samples (Fig. 2e). Intriguingly, the correlated expression to aberrant methylation patterns being identified in col- of DNMT1 and TET3 in the inflamed human tissue is orectal cancer , we examined if DNMT1 is differentially consistent with the smooth muscle of Dnmt1-WT and expressed in the inflamed tunica muscularis from human Dnmt1-KO mice (Fig. 2a), suggesting a role for DNMT1 patients with either Crohn’s disease, colon/anal cancer, or in regulating the expression of TET3. diverticulitis. We detected increases in DNMT1 as well as increases in TET3 in the inflamed smooth muscle show- Dnmt1-KO mice gradually lose tunica muscularis and ing that DNMT1 can regulate TET3 across species necessary contractile proteins and accumulate coated (Fig. 2c). Quantitation of protein showed that the vesicles in SMCs increases in DNMT1 and TET3 were significant in the Hematoxylin and eosin staining revealed that inflamed smooth muscle from anal or colon cancer Dnmt1-KO mice begin to lose both muscle layers of the Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 5 of 14 Fig. 3 Electron microscopy (EM) reveals degeneration of smooth muscle and presence of coated vesicles in Dnmt1-KO SMC. a, b EM cross- section images of P21 Dnmt1-WT jejunum. c–f EM cross-section images of P21 Dnmt1-KO jejunum. Both the circular (CM) and longitudinal muscle (LM) layers in Dnmt1-WT jejunum are intact (a, b), complete with organized nuclei in SMC as well as the presence of ICC (ICC), and submucosal cells (SM). Degenerated SMC within the tunica muscularis in the jejunum of Dnmt1-KO mice (c, d) show cellular fragmentation, as well as a loss of nuclear organization. e A portion of a degenerated SMC with coated vesicles from a Dnmt1-KO mouse found phagocytosed in a macrophage (M). f Degenerated SMCs in Dnmt1-KO mice accumulate coated vesicles (V) trapped in the vicinity of the endoplasmic reticulum (ER) and Golgi (Gol), open vesicles: arrows; closed vesicles: arrowheads tunica muscularis at noticeable levels by P15 but indicating a complete loss of mature SMCs did not have a significant loss of either layer until P21, (Supplementary Figure 2D). The loss of MYH11 with the more drastic tissue loss being found in the cir- in the SMCs of Dnmt1-KO mice results in SMCs cular layer (Supplementary Figure 2). To confirm that are no longer contractile, leading to a that the loss of tissue was indeed due to a loss of weakened tunica muscularis, allowing for intestinal dis- mature SMCs, immunohistochemical analysis was per- tention. Electron microscopy of the tunica muscularis formed. As MYH11 is the most selective marker of from P21 Dnmt1-KO mice revealed a loss of mature and contractile SMCs , antibodies targeting intercellular connections between SMCs, submucosa MYH11 were used to confirm the loss of mature cells, and interstitial cells of Cajal (ICC), the SMCs alongside the endogenous eGFP reporter expres- necessary pacemaking cells in peristaltic contractions sion based on Myh11 expression. As expected, (Fig. 3c). We also noted cellular fragmentation and the MYH11 levels in the tunica muscularis of Dnmt1-WT loss of nuclear/nucleolar organization within Dnmt1-KO mice continued to increase from P10 to P21, SMCs (Fig. 3c, d). Furthermore, we were able to detect while in Dnmt1-KO MYH11 levels waned, starting at P15 macrophages engulfing the deteriorated SMCs in Dnmt1- until there was no detectable MYH11 signal at P21, KO mice (Fig. 3e). Finally, we found that almost all Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 6 of 14 Fig. 4 Transcriptomic analysis reveals dramatic changes of expression in Dnmt1-KO jejunal tunica muscularis. a, b The expression profile of Dnmt1-KO jejunal tunica muscularis shows a higher number of unique mRNA transcripts and more genes that increased their expression than reduced their expression when compared to Dnmt1-WT. Changes in overall miRNA expression were not as dramatic. c Increases in several genes associated with apoptosis or lipid processing were upregulated in Dnmt1-KO mice. Additionally, transcript levels of SMC markers (Srf, Myh11) are reduced along with the SRF transcriptional co-activator, Myocd. d Although overall miRNA changes are not as pronounced as mRNA changes in Dnmt1-KO mice, there are select miRNAs that increase their expression level. e A gene ontology evaluation uncovered the prevalence of five general categories enriched in Dnmt1-KO jejunal tunica muscularis with many genes relating to defense, immunity, lipids, motility, and apoptosis being found Dnmt1-KO SMCs contained small coated vesicles (Fig. 3d, vesicles in their cytosol that congregated in distinct f), which were not associated with Dnmt1-WT SMC regions around the endoplasmic reticulum and Golgi (Fig. 3a, b). Dnmt1-KO SMC appeared to accrue coated (Fig. 3f). Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 7 of 14 Fig. 5 Dnmt1-KO tunica muscularis accumulates lipids, has increased expression in apoptotic genes and characteristic apoptotic DNA fragmentation. a Oil Red O staining of jejunal cross-sections reveals an accumulation of lipids in the muscle layer of Dnmt1-KO mice (white arrows; red stain) that is not seen in Dnmt1-WT mice. b At P21, a clear fragmentation of extracted DNA can be seen in the isolated tunica muscularis of Dnmt1-KO mice. Distinct fragmentation patterns of genomic DNA are generated due to cleavage at specific sites, a characteristic of apoptotic cells. c qPCR of isolated tunica muscularis at P15. As expected, Dnmt1, Srf, and Myh11 all have significantly reduced expression in Dnmt1-KO tunica muscularis as well as significant increases in many genes linked with lipid processes and/or pro-apoptotic pathways (n = 4, error bars are SEM, *p < 0.05, **p < 0.01, unpaired t-test). d In isolated Dnmt1-KO SMC at P15, more dramatic losses of Dnmt1, Srf, and Myh11 were observed as well as increases in the pro-apoptotic markers Gadd45g & Nr4a1 (n = 2, error bars are SEM, *p < 0.05, **p < 0.01, unpaired t-test). Ubb was used as an endogenous control in (c, d) Transcriptomic analysis of tissue/SMC reveals loss of muscularis of both Dnmt1-KO and Dnmt1-WT mice. mature SMC markers and miRNAs and increases in lipid, Summaries and full expression profiles of messenger RNA immune processing, and pro-apoptotic genes (mRNA)/miRNA sequencing results can be found in To gather a more complete understanding of the gene Supplementary Table 1 and Supplementary Table 3, expression changes happening in Dnmt1-KO mice, we respectively. The transcriptome of Dnmt1-KO has 954 performed a transcriptomic analysis on the jejunal tunica unique mRNA gene annotations, compared to only 530 in Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 8 of 14 Fig. 6 Loss of Dnmt1 in GI-SMC causes global genomic CpG hypomethylation. a Reduced representation bisulfite sequencing of Dnmt1-KO mice compared to Dnmt1-WT mice demonstrates that most genes in Dnmt1-KO have at least one CpG site with significant levels of hypomethylation. b CpG methylation is reduced ~20% in Dnmt1-KO mice when compared to Dnmt1-WT mice. c While there is an ~20% reduction of CpG methylation in Dnmt1-KO mice, the patterns of methylation across each chromosome is not altered. d Globally, across the genome, CpG methylation is significantly reduced at promoters, exons and introns (**p < 0.00001, Fisher’s exact test). e Pro-apoptotic marker Nr4a1 shows significant losses of methylation at various genomic locations, including the promoter and exon 1 (*p < 0.01, Fisher’s exact test). f A UCSC genome browser view of the Nr4a1 gene with methylation tracks added (yellow=more methylated, red=less methylated) and the promoter region boxed in purple Dnmt1-WT with approximately equal amounts of unique half their expression, with similar amounts of both miRNA annotations between Dnmt1-KO (184) and -WT upregulated (289) and downregulated (265) miRNA (170) mice (Fig. 4a). When compared to Dnmt1-WT transcripts (Fig. 4b). The differences between the relative mice, Dnmt1-KO mice have 2,829 genes that at least amount and type of mRNA/miRNA transcripts expressed double their expression and only 1,319 that lose at least exposes a remodeled transcriptome in Dnmt1-KO mice. Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 9 of 14 Fig. 7 Only CpG sites show large loss of methylation in Dnmt1-KO. a, b While CHH sites and CHG sites did have individual sites that changed their methylation status in Dnmt1-KO mice, there was no strong genome-wide change in methylation status. c At CpG sites in Dnmt1-KO mice, there was marked hypomethylation at multiple site across the genome which was not observed in either CHH or CHG sites. Each dot represents an individual cytosine that mapped to both Dnmt1-KO and Dnmt1-WT samples; r is Pearson’s correlation coefficient and n is the total number of overlapping sites for each type of cytosine site Dnmt1-KO mice have considerable losses of SMC marker We selected upregulated gene groups associated with transcripts necessary for mature functioning, including lipids and apoptosis for further study. Oil red O staining Acta2 (25.7% loss), Myocd (37.4% loss), Myh11 (53.2% of cross-sections from Dnmt1-KO mice confirmed an loss), Srf(41.7% loss), and Tagln (19.9% loss) (Fig. 4c) as accumulation of lipids in their jejunal tunica muscularis well as reductions in necessary SMC miRNAs miR-143 (Fig. 5a), consistent with our mRNA-sequencing and GO (31.1% loss), miR-145 (40.8% loss), and miR-133 (70.4% term analysis. Next, we examined if Dnmt1-KO SMCs loss) (Fig. 4d). Since the transcription factors SRF and were undergoing apoptosis. DNA fragmentation is a 29 36 MYOCD are required for Myh11 expression , and the hallmark of apoptosis and was observed in Dnmt1-KO previously mentioned miRNAs , the reduction of Srf and tunica muscularis at P21 (Fig. 5b). Our transcriptome Myocd exacerbates the deterioration of SMCs as Srf represents the entirety of the tunica muscularis where the expression in Dnmt1-KO does not become significantly main population of cells are SMCs. However, immune cell reduced until P15 (Supplementary Figure 3) and therefore infiltration into the tunica muscularis may be responsible is likely not the primary insult involved in the loss of GI- for the increases in genes associated with immunity and SMCs. We also noted increases in miRNAs associated other GO categories. With the use of flow cytometry, we with regulating cellular identity and proliferation includ- removed cells from the tunica muscularis that were 31 32 33 + 37 ing, miR-10b , miR-21a , miR-486a , and miR-148a. CD45 , a known marker of immune cells , then selected − + MiR-148a is an established inhibitor of Dnmt1 expres- for cells that were CD45 and eGFP in P15 Dnmt1-KO sion , thus Dnmt1-KO SMCs, through increased miR- and Dnmt1-WT. Indeed, the results of our flow cytometry 148a, likely create even further repression of Dnmt1 identified ~5-fold increase in CD45 cells within Dnmt1- expression. Furthermore, both miR-10b and miR-148a KO tissue when compared to Dnmt1-WT mice while - + expression are known to regulated by genomic CpG CD45 and eGFP SMC were reduced ~3-fold (Supple- 31,35 - + methylation . For mRNA analysis, we selected 392 mentary Figure 5). Sorted CD45 , eGFP cells represented overexpressed genes in the tunica muscularis of Dnmt1- isolated SMCs. We confirmed that isolated SMCs from KO mice and 365 of them had some level of significant Dnmt1-KO mice had almost complete ablation of Dnmt1, demethylation (Supplementary Figure 4). Gene ontology significant reductions in Myh11 and Srf expression, and (GO) examination revealed that the most overrepresented dramatic increases in the pro-apoptotic genes Gadd45g biological process categories included those associated and Nr4a1 (Fig. 5d). However, we found other with lipids, apoptosis, defense responses to stress/stimuli, genes related to apoptosis and lipids upregulated in P15 and immune responses (Fig. 4e and Supplementary tissue that were not upregulated in isolated Dnmt1-KO Table 4). The quantitative increase in expression levels of SMCs, indicating that their upregulation is likely due to many of these genes that were found to be upregulated in immune cells infiltrating into the tunica muscularis Dnmt1-KO varied widely (Fig. 4c). (Fig. 5c, d). Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 10 of 14 Loss of Dnmt1 induces global genomic CpG SMC-restricted congenital Dnmt1-KO mice have shor- hypomethylation tened GI tracts, and lose their tunica muscularis by P21. Employing reduced representation bisulfite conver- Dnmt1-KO GI-SMC have significant genomic CpG sion sequencing, we discovered that ~90% (20,808) of demethylation, a loss of intercellular connectivity, annotated genes in Dnmt1-KO mice had at least one decreases in necessary SMC transcripts (Myh11, Srf, miR- CpG site with significant demethylation in either their 133, miR-143/145), and increases in both pro-apoptotic genes (Gadd45g, Nr4a1) and miRNAs associated with promoter (−1 kb of TSS), exons, or introns and ~17% (4,004) of genes had at least one CpG site with sig- changes in cellular identity (miR-21a, -148a, -186a, -10b). nificant demethylation in all sites, encompassing their In fact, we found aberrant hypermethylation promoter, exons, and introns (Fig. 6a) with genomic flanking a CArG box (CCATATAAGG: antisense) and hypomethylation being significant, on a genome-wide SRF binding site found within the second intron of Srf in scale, across annotated promoter, exon, and intron CpG Dnmt1-KO (Supplementary Figure 6) that may have sites (Fig. 6d). None of these patterns were observed for contributed to the reduction Srf expression, thus genomicCHG or CHH(Hcan be A, C, or T) sites increasing the apoptotic potential of Dnmt1-KO SMC. (Fig. 7a, b). Across the genome of Dnmt1-KO mice, Our results are similar to both congenital and inducible there was a ~20% loss of CpG methylation (Fig. 6c) with intestinal epithelial cell-specific Dnmt1-KO models. Both a majority of genes reducing their average CpG models show genomic hypomethylation, DNA damage, methylation levels from 0.8–0.6 to 0.6–0.4 (Fig. 6b). We and increased apoptosis in affected Dnmt1 knockout 40,41 also observed CpG sites that were strongly hyper- cells . In the congenital intestinal epithelial model, methylated in Dnmt1-KO mice, but these sites were only 35% of Dnmt1 knockout mice live until P20, similar considerably outnumbered by the amount of strongly to our SMC-restricted Dnmt1-KO mice which die around hypomethylated CpG sites (Fig. 7c). We created browser P21 and none have ever survived past P24. When Dnmt1 tracks that visualize and quantify the changes in CpG is inducibly ablated from intestinal epithelial cells at later methylation within the University of California, Santa stages of development, the mouse does not die and sees an Cruz (UCSC) genome browser for any annotated increase in Dnmt3b expression, which also directly com- genomic site. In addition, we added transcriptome level pensated for the loss of DNA methylation due to the RNA-sequencing (RNA-seq) data for both mRNA and knockout of Dnmt1. This pattern of compensatory miRNA from jejunal smooth muscle in both Dnmt1-KO DNMT isoform induction due to loss of a separate and Dnmt1-WT mice (Supplementary Table 3) to the DNMT could potentially be why we see increased browser. The browser tracks were installed at our DNMT3A seen during development in our Dnmt1-KO Smooth Muscle Transcriptome Project webpage: mice, as DNMT3A has been shown to be able to maintain https://med.unr.edu/physio/transcriptome,entitled methylation patterns in mouse embryonic stem cells “UCSC Smooth Muscle Methylome Browser.” This lacking DNMT1 . browser provides a comprehensive reference for tran- While it is clear that GI-SMCs in congenital Dnmt1-KO scriptomic and genomic DNA methylation status at mice are undergoing remodeling and apoptosis, it is not CpG sites in the jejunal smooth muscle of Dnmt1-WT clear what the causative insult of these changes is, and it is and Dnmt1-KO mice. The browser can also interact more likely a combination of several pro-apoptotic with genome-level bioinformatics (e.g., ENCODE) influences including genomic CpG hypomethylation, data publically available in the UCSC Genome epigenome modulation, lack of cellular connectivity, Browser. In Dnmt1-KO mice, Nr4a1 has significant upregulation of pro-apoptotic genes/miRNAs, down- lossesofmethylation at itspromoter aswellasexons 1, regulation of required SMC proteins/miRNAs, exposure 2, 5, and 6 and introns 5 and 6 (Fig. 6e), as shown to lipoproteins, and chromosomal instability. As on the browser (Fig. 6f), suggesting the expression of the it pertains to pro-apoptotic genes upregulated in Dnmt1- pro-apoptotic Nr4a1 could be regulated by CpG KO SMCs, Gadd45g is a marker of DNA damage, acti- methylation levels controlled by DNMT1, agreeing with vator of the p38/c-Jun N-terminal kinase apoptotic 39 43 previous findings . The entirety of our CpG methy- pathway and has been correlated with increased lomic results and analysis for every annotated promoter, expression in a double knockout of DNMT1 and exon, and intron can be found in Supplementary DNMT3B in vitro . Nr4a1 is an orphan nuclear Table 5. receptor that has been shown to be regulated by DNMT1 and the methylation level of its promoter region . Discussion Expression of Nr4a1 is associated with an induction of By eliminating Dnmt1 from the murine genome in a apoptosis through multiple pathways , as well as SMC-restricted model, we showed the in vivo necessity direct inhibition of the growth of SMCs . As for genomic of Dnmt1 in the embryonic development of GI-SMCs. CpG hypomethylation, it is likely a causative feature Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 11 of 14 involved in the transcriptomic remodeling of mRNA and Methods miRNA expression, but probably does not account for all Mice Cre-eGFP/+ 24 changes. Only about 10% of all miRNAs have been shown In a C57BL/6 background, smMHC mice loxllox 25 to change expression based on levels of DNA methyla- were crossed with Dnmt1 mice to make 47 -/-;Cre-eGFP/+ tion , denoting a secondary role of DNMT1 in regulating smDnmt1 mice. PCR confirmed genotypes miRNA expression in GI-SMCs. DNMT1 interacts with (Supplementary Figure 1). Primer sequences are found in up to 58 transcription factors , and many epigenetic Supplementary Table 2. Mice were bred, maintained, and modifiers , and thus it is probable that loss of Dnmt1 has euthanized following guidelines set by the Institutional downstream deleterious consequences outside of Animal Care and Use Committee at the University of genomic hypomethylation. Furthermore, in Dnmt1-null Nevada-Reno Animal Resources. cells, DNA mismatch repair is impeded and microsatellite instability increases ~4-fold , indicating a pivotal, Human tissue methylation-independent role of DNMT1 in DNA Isolated human GI smooth muscle tissue sections were damage repair. While previous in vitro studies of com- donated by Dr. Kent Sasse (Sasse Surgical Associates) and plete Dnmt1 knockout in mouse embryonic stem cells Dr. Laren S. Becker (Stanford University School of Med- have reported more dramatic genomic CpG methylation icine). Use of human tissue was approved by the Institu- reductions from ~75% to ~18% or lower at specific tional Review Board at the University of Nevada, Reno. sites , cell-selective in vivo knockout of Dnmt1 in intestinal epithelial results in CpG methylation reductions Tissue isolation and preparation 40,41 from ~75% to ~45% , similar to our own in vivo The GI tract was extracted into 1× Hank’s Calcium Free findings (Fig. 6c). Buffer. Tissue then underwent one of two processes: (1) Additionally, the observed lack of nuclear/nucleolar smooth muscle isolation and (2) whole tissue organization in electron microscopy images of Dnmt1-KO preparation for cross-sectioning. (1) The muscle layer was could be indicative of catastrophic chromosomal peeled from the mucosa for use in extractions/experi- instability. Of the 25 genes in the CIN25 gene signature of ments. (2) The tissue was fixed in 4% paraformaldehyde chromosomal instability , 23 were found to be upregu- (PFA), a dehydration in 20% sucrose/1× phosphate- lated in Dnmt1-KO smooth muscle tissue (data not buffered saline (PBS), then placed in 1:1 OCT/20% shown). We were also able to show that congenital sucrose and super-cooled by liquid nitrogen. Using a Dnmt1-KO mice have reduced levels of all TET proteins, cryostat microtome (Leica, Wetzlar, Germany), and thus it is likely that Dnmt1-KO mice cannot properly 8 µM thick tissue sections were cut onto slides coated regulate 5-hmc or 5-mc levels, causing dysregulation of with VectaBond (Maravai Biosciences, San Diego, the entire DNA methylation apparatus. We also showed CA, USA), and allowed to dry. Prepared slides were that expression levels of DNMT1 and TET3 are associated used for downstream immunohistochemical/histological and correlated within the inflamed tunica muscularis of staining. colorectal cancer tissue as well as in the jejunal tunica muscularis of both Dnmt1-WT and Dnmt1-KO mice. Isolation of total and small RNAs Further studies are needed to elucidate the associations Isolated muscle tissue was placed in lysis buffer and and interactions between DNMT1 and TET3 in inflamed homogenized by bead beating in an air-cooled Bullet and injured smooth muscle. Blender Storm (Next Advance, Troy, NY, USA). Total Most surprisingly, Dnmt1-KO SMC show an accumu- RNAs and miRNAs were isolated from smooth muscle lation of lipids and lipid associated transcripts within their and sorted SMCs using mirVana miRNA Isolation Kit tunica muscularis. Many of the genes upregulated in our (Ambion, Foster City, CA, USA) as previously described . transcriptomic analysis are commonly associated with Extracted total RNAs and small RNAs were used for vSMC containing atherosclerotic plaques . In vSMC, quantitative PCR (qPCR) and RNA-seq. lipoproteins, based on oxidation status, have been impli- cated in changing the phenotypic identity of SMCs into Quantitative PCR either a proliferative or apoptotic fate . Extracted total RNA was reverse transcribed into In conclusion, the lack of Dnmt1 in GI-SMC resulted in complementary DNA (cDNA) using SuperScript III an inability to maintain de novo methylation patterns and Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA), the consequences of this loss become exacerbated with after DNA removal with DNA-free DNA Removal Kit each round of mitotic division. As can be surmised, fur- (Ambion). cDNA samples were then quantified, diluted, ther research into all observed phenotypic aspects of and tested for quality. Standard qPCR protocol was car- Dnmt1-KO is necessary to understand all cellular con- ried out on a 7900HT Fast Real-Time PCR System sequences of DNMT1 deficiency. (Applied Biosystems, Foster City, CA, USA) as previously Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 12 of 14 described . Primer sequences are found in Supplemen- western blot banding patterns was done using ImageJ and tary Table 2. histograms were produced in SigmaPlot (Systat, San Jose, CA, USA). Immunohistochemistry Slides were exposed to 1× Tris-buffered saline (TBS)/ DNA methylome 0.1% Tween 20 followed by exposure to 4% milk/1× TBS/ Genomic DNA was isolated from jejunal tunica mus- 0.1% Tween 20. Slides were then incubated at 4 °C with 1° cularis at P15 (2 males and 2 females for both Dnmt1-KO antibody. The following day, slides were incubated with 2° and Dnmt1-WT; n = 4) using Qiagen (Hilden, Germany) antibody at room temperature. After washing the slides in AllPrep DNA spin columns. DNA samples were pooled 1× TBS, the slides were mounted and allowed to cure. See and shipped to Zymo Research Corporation (Irvine, CA, Supplementary Table 2 for antibodies/dilutions. Slides USA) where the samples underwent Methyl-MidiSeq were imaged with an Olympus FluoView FV1000 (Tokyo, which involves genomic DNA extraction, fragmentation, Japan) confocal microscope. end modification, bisulfite conversion, limited amplifica- tion, and sequencing. After sequencing, bioinformatic Hematoxylin and eosin staining processing was utilized in comparing amount of methy- Cross-section slides were exposed to hematoxylin, lation at CpG, CHH, and CHG sites between Dnmt1-WT rinsed with water, and then stained in eosin Y. After and Dnmt1-KO samples. Methylation ratios were calcu- several dehydrating baths of increasing ethanol (EtOH) lated (methylated cytosines/total cytosine reads) for concentration (80%, 95%, 100%) and two terminal xylene individual sites and then averaged across regions (exons, baths, the slides were mounted and allowed to cure. Slides introns, promoters = −1 kb of TSS). These averages were were imaged with an Olympus BX43 (Tokyo, Japan) then used in Fisher’s exact test to quantify significant brightfield microscope. changes (p < 0.05) in methylation. The reference genome used was GRCm38/mm10. The data from our bisulfite Oil Red O staining conversion sequencing were submitted to the Gene Cross-section slides were exposed to pure propylene Expression Omnibus (GEO) repository with the following glycol, then moved to 0.5% Oil Red O solution at 60 °C accession numbers: GSM2947783, Dnmt1-WT jejunum; followed by soaking in 85% propylene glycol. slides were GSM2947784, Dnmt1-KO jejunum. then rinsed and counterstained with hematoxylin, rinsed with water, then mounted with aqueous mounting RNA sequencing reagent and allowed to cure. Slides were imaged Total RNAs and small RNAs isolated from Dnmt1-KO with a Keyence BZ-x710 (Osaka, Japan) brightfield and Dnmt1-WT jejunal tunica muscularis at P15 (2 males microscope. and 2 females for both Dnmt1-KO and Dnmt1-WT; n = 4) were shipped to LC Sciences (Houston, TX, USA) Electron microscopy where mRNA-seq and miRNA-seq were performed. Whole jejunum and colon tissue were fixed (3% glu- Sequencing data analysis, alignments, annotations, and taraldehyde and 4% PFA in 0.1 M phosphate buffer) for bioinformatics processing were carried out as previously several days at room temperature. Specimens were then described . The reference genome used was GRCm38/ fixed in 1% OsO at 4 °C, rinsed in distilled water, then mm10. Our RNA-seq data for both mRNA and miRNA block stained with saturated aqueous uranyl acetate were submitted to the GEO repository with the following solution, followed by dehydration in a graded series of accession numbers: GSM2936348, Dnmt1-WT jejunum EtOH and finally embedded in EPON 812. Ultrathin mRNA; GSM2936349, Dnmt1-KO jejunum mRNA; sections were stained with uranyl acetate and lead citrate GSM2936350, Dnmt1-WT jejunum miRNA; and then examined using a Hitachi H-7650 (Tokyo, Japan) GSM2936351 Dnmt1-KO jejunum miRNA. microscope. Cell sorting Western blot Cells were dispersed as previously described . Cells Isolated SM tissue was ground by mortar and pestle were incubated with APC-Cy7 Anti-CD45 (Biolegend, in modified RIPA buffer. Transferred polyvinylidene Clone 30-F11, 1.0 μg/mL; San Diego, CA, USA) followed fluoride membranes were pre-blocked in 5% milk, then by washing with PBS/1% FBS. Resuspended cells had exposed to 1° antibody at 4 °C in 2% milk/0.05% Tween 20 Hoechst 33258 (1 μg/mL) added as a viability marker. at varying dilutions. The membranes were exposed to 2° Cells were sorted and analyzed using the BD Biosciences antibody at room temperature. All antibodies used can be (San Jose, CA, USA) FACSAria II Special Order Research found in Supplementary Table 2. Imaging by UVP EC3 Product with a 130 μm nozzle with sheath pressure at 12 (Upland, CA, USA) Imaging system. Quantification of psi. The 355 nm laser excited Hoechst 33258 with a 450/ Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 13 of 14 50 nm bandpass filter. The eGFP was excited using a 488 3. Liu, R., Leslie,K. L.& Martin,K. A.Epigenetic regulationofsmoothmusclecell plasticity. Biochim. Biophys. Acta 1849,448–453 (2015). nm laser with a 530/30 nm bandpass filter. A neutral 4. Hu, B., Gharaee-Kermani, M., Wu, Z. & Phan, S. H. Epigenetic regulation of density filter 2 was used on the forward scatter detector myofibroblast differentiation by DNA methylation. Am.J.Pathol. 177,21–28 due to the high forward scatter properties. Cells that (2010). - + 5. Ning, Y. et al. 5-Aza-2’-deoxycytidine inhibited PDGF-induced rat airway were CD45 and eGFP were sorted into PBS/1% FBS. smooth muscle cell phenotypic switching. Arch. Toxicol. 87,871–881 (2013). Acquisition was performed on BD FACSDiva 8.0 and 6. Hinoue,T.etal. Genome-scale analysis of aberrant DNA methylation in col- TreeStar Flowjo (Ashland, OR, USA) was used to generate orectal cancer. Genome Res. 22,271–282 (2012). 7. Sadler, T. et al. Genome-wide analysis of DNA methylation and gene figures. expression defines molecular characteristics of Crohn’s disease-associated fibrosis. Clin. Epigenetics 8,30(2016). Acknowledgements 8. Ramchandani, S.,Bhattacharya, S. K.,Cervoni,N.&Szyf,M. DNA methylationis We would like to thank Drs. Michael Kotlikoff and Laurie Jackson-Grusby for a reversible biological signal. Proc. Natl. Acad. Sci. USA 96,6107–6112 (1999). Cre-eGFP/+ lox/lox their generous donations of the congenital smMHC and Dnmt1 9. Gardiner-Garden, M. & Frommer, M. CpG Islands in vertebrate genomes. J. Mol. mouse lines respectively as well as Benjamin J. Weigler and Walt F. Mandeville Biol. 196,261–282 (1987). (Laboratory Animal Medicine, University of Nevada, Reno) for the excellent 10. Jones, P. A. Functions of DNA methylation: islands, start sites, gene bodies and animal services provided to the mice. We would also like to thank Dr. Kent M. beyond. Nat. Rev. Genet. 13, 484–492 (2012). Sanders for COBRE Core B and Byoung H. Koh for the cell sorting services as 11. McDonald, O. G. &Owens,G.K.Programming smooth muscle plasticity with well as Dr. Tong Zhou for his assistance in bioinformatic data presentation. B.G. chromatin dynamics. Circ. Res. 100,1428–1441 (2007). J. was supported by an award from the J. M. Hitchcock, Graduate Student 12. Miano,J.M., Long,X.& Fujiwara, K. Serum response factor: master regulatorof Research Fund. the actin cytoskeleton and contractile apparatus. Am.J.Physiol.CellPhysiol. This study was supported by National Institutes of Health (NIH) (DK094886 and 292,C70–C81 (2007). DK103055 to S.R.) and COBRE Phase III FACSAria (P30 GM110767 to Kent M. 13. Miano,J.M., Cserjesi,P., Ligon, K. L.,Periasamy,M.& Olson, E. N. Smooth Sanders). The funders had no role in study design, data collection/analysis, muscle myosin heavy chain exclusively marks the smooth muscle lineage decision to publish, or preparation of the manuscript. during mouse embryogenesis. Circ. Res 75,803–812 (1994). 14. Skalli, O. et al. A monoclonal antibody against α-smooth muscle actin: Author details a new probe for smooth muscle differentiation. J. Cell Biol. 103,2787–2796 Department of Physiology and Cell Biology, University of Nevada School of (1986). Medicine, Reno, NV 89557, USA. Department of Morphological and 15. Torella, D. et al. MicroRNA-133 controls vascular smooth muscle cell pheno- Physiological Sciences, University of Fukui, Fukui 910-8507, Japan. Sasse typic switch in vitro and vascular remodeling in vivo. Circ. Res. 109,880–893 Surgical Associates, Reno, NV 89502, USA. Gastroenterology and Hepatology, (2011). Stanford University School of Medicine, Stanford, CA 94305, USA 16. Cordes, K. R. et al. MiR-145 and miR-143 regulate smooth muscle cell fate and plasticity. Nature 460,705–710 (2009). Authors' contributions 17. Wu, W., Shen, X. & Tao, S. Characteristics of the CArG-SRF binding context in R.M.B., B.G.J., S.E.H., and S.R. designed research; R.M.B., B.G.J., S.E.H., and K.H. mammalian genomes. Mamm. Genome 21,104–113 (2010). performed research; B.G.J., R.M.B., S.E.H., and S.R. analyzed data; K.C.S. and L.S.B. 18. McDonald, O. G.,Wamhoff, B. R.,Hoofnagle,M.H.&Owens, G. K. Controlof contributed materials. B.G.J. wrote the paper and S.R revised it. SRF binding to CArG box chromatin regulates smoothmusclegeneexpres- sion in vivo. J. Clin. Invest. 116,36–48 (2006). Conflict of interest 19. Warnecke, P. M. & Clark,S. J.DNA methylationprofile of the mouse skeletal α- The authors declare that they have no conflict of interest. actin promoter during development and differentiation. Mol. Cell Biol. 19, 164–172 (1999). 20. Yamamura, H. et al. Structure and expression of the human SM22α gene, assignment of the gene to chromosome 11, and repression of the promoter Publisher's note activity by cytosine DNA methylation. J. Biochem. 122,157–167 (1997). Springer Nature remains neutral with regard to jurisdictional claims in 21. Robert, M. F. et al. DNMT1 is required to maintain CpG methylation and published maps and institutional affiliations. aberrant gene silencing in human cancer cells. Nat. Genet 33,61–65 (2003). 22. Pastor, W. A., Aravind, L. & Rao, A. TETonic shift: biological roles of TET proteins Funding in DNA demethylation and transcription. Nat. Rev. Mol. Cell Biol. 14,341–356 Wessex Cancer Trust, CRUK—Southampton Internal Pilot Grant, EU-FP7 Marie (2013). Curie (CANOMICS), Annual Adventures in Research—University of 23. Li, E., Bestor, T. H. & Jaenisch, R. Targeted mutation of the DNA methyl- Southampton, EU-Excellence II—Systems Biology Framework FRA-SYS (Grant transferase gene results in embryonic lethality. Cell 69,915–926 (1992). 4072), Cancer Research UK (Grants C34999/A13719 and RG84119), and MRC 24. Xin,H.-B.,Deng, K.-Y., Rishniw, M., Ji, G. & Kotlikoff, M. I. Smooth muscle Clinician Scientist Fellowship (Grant G1002565). expression of Cre recombinase and eGFP in transgenic mice. Physiol. Genom. 10,211–215 (2002). Supplementary Information accompanies this paper at https://doi.org/ 25. Jackson-Grusby, L. et al. Loss of genomic methylation causes p53-dependent 10.1038/s41419-018-0495-z. apoptosis and epigenetic deregulation. Nat. Genet. 27,31–39 (2001). 26. Coppedè, F. Epigenetic biomarkers of colorectal cancer: focus on DNA Received: 30 October 2017 Revised: 31 January 2018 Accepted: 12 February methylation. Cancer Lett. 342,238–247 (2014). 2018 27. Owens, G. K., Kumar, M. S. & Wamhoff, B. R. Molecular regulation of vascular smooth muscle cell differentiation in development and disease. Physiol. Rev. 84,767–801 (2004). 28. Sanders,K.M., Ördög, T.,Koh,S.D., Torihashi, S. &Ward, S. M. Development and plasticity of interstitial cells of Cajal. Neurogastroenterol. Motil. 11, 311–338 References (1999). 1. Wedel, T. et al. Novel smooth muscle markers reveal abnormalities of the 29. Wang, Z., Wang, D.-Z., Teg Pipes, G. C. & Olson, E. N. Myocardin is a master intestinal musculature in severe colorectal motility disorders. Neurogas- regulatorofsmoothmusclegeneexpression. Proc.Natl. Acad.Sci.USA 100, 7129–7134 (2003). troenterol. Motil. 18,526–538 (2006). 30. Park, C. et al. Serum response factor-dependent microRNAs regulate gastro- 2. Rolle, U., O’Briain, S., Pearl, R. H. & Puri, P. Megacystis-microcolon-intestinal intestinal smooth muscle cell phenotypes. Gastroenterology 141,164–175 hypoperistalsis syndrome: evidence of intestinal myopathy. Pediatr. Surg. Int. (2011). 18,2–5 (2002). Official journal of the Cell Death Differentiation Association Jorgensen et al. Cell Death and Disease (2018) 9:474 Page 14 of 14 31. Li, Z. et al. DNA methylation downregulated mir-10b acts as a tumor sup- 45. Li, Q.-X., Ke, N., Sundaram,R.&Wong-Staal,F.NR4A1, 2, 3 -an orphan nuclear pressor in gastric cancer. Gastric Cancer 18,43–54 (2015). hormone receptor family involved in cell apoptosis and carcinogenesis. Histol. 32. Ji, R. et al. MicroRNA expression signature and antisense-mediated depletion Histopathol. 21,533–540 (2006). reveal an essential role of microRNA in vascular neointimal lesion formation. 46. Mayer, P., Hinze, A. V., Harst, A. & von Kügelgen, I. A receptors mediate the 2B Circ. Res. 100, 1579–1588 (2007). induction of early genes and inhibition of arterial smooth muscle cell pro- 33. Oh, H.-K. et al. Genomic loss of miR-486 regulates tumor progression and the liferation via Epac. Cardiovasc. Res. 90,148–156 (2011). OLFM4 antiapoptotic factor in gastric cancer. Clin. Cancer Res. 17,2657–2667 47. Han,L., Witmer,P.D., Casey, E.,Valle,D. & Sukumar, S. DNAmethylation (2011). regulates microRNA expression. Cancer Biol. Ther. 6,1284–1288 (2007). 34. Nan,W.etal. MicroRNA-21 and microRNA-148a contribute to DNA hypo- 48. Hervouet, E., Vallette, F. M. & Cartron, P.-F. Dnmt1/transcription factor inter- methylation in lupus CD4+T cells by directly and indirectly targeting DNA actions: an alternative mechanism of DNA methylation inheritance. Genes methyltransferase 1. J. Immunol. 184,6773–6781 (2010). Cancer 1,434–443 (2010). 35. Lujambio, A. et al. A microRNA DNA methylation signature for human cancer 49. Qin, W., Leonhardt, H. & Pichler, G. Regulation of DNA methyltransferase 1 by metastasis. Proc. Natl. Acad. Sci. USA 105, 13556–13561 (2008). interactions and modifications. Nucleus 2,392–402 (2011). 36. Bortner,C.D., Oldenburg, N. B. E. &Cidlowski,J.A. The role of DNAfrag- 50. Jin, B. & RobertsonK. D. DNA methyltransferases (DNMTs), DNA damage repair, mentationinapoptosis. Trends Cell Biol. 5,21–26 (1995). and cancer. Adv. Exp. Med. Biol. 754,3–29 (2013). 37. Donovan, J. A. & Koretzky, G. A. CD45 and the immune response. J. Am. Soc. 51. Li, Z. et al. Distinct roles of DNMT1-dependent and DNMT1-independent Nephrol. 4,976–985 (1993). methylation patterns in the genome of mouse embryonic stem cells. Genome 38. Kent, W. J. et al. The Human Genome Browser at UCSC. Genome Res. 12, Biol. 16, 115 (2015). 996–1006 (2002). 52. Martos, S. N. et al. Two approaches reveal a new paradigm of ‘switchable or 39. Chen, Y.-T., Liao, J.-W., Tsai, Y.-C. & Tsai, F.-J. Inhibition of DNA methyltransferase genetics-influenced allele-specificDNA methylation’ with potential in human 1 increases nuclear receptor subfamily 4 group A member 1 expression and disease. Cell Discov. 3, 17038 (2017). decreasesblood glucosein type2diabetes. Oncotarget 7, 39162–39170 53. Carter, S. L., Eklund, A. C., Kohane, I. S., Harris, L. N. & Szallasi, Z. A signature of (2016). chromosomal instability inferred from gene expression profilespredictsclinical 40. Elliott, E. N., Sheaffer, K. L. & Kaestner, K. H. The ‘de novo’ DNA methyl- outcome in multiple human cancers. Nat. Genet. 38,1043–1048 (2006). transferase Dnmt3b compensates the Dnmt1-deficient intestinal epithelium. 54. Lusis, A. J., Fogelman, A. M. & Fonarow, G. C. Genetic basis of atherosclerosis: Elife 5,1–15 (2016). part I. New genes and pathways. Circulation 110,1868–1873 (2004). 41. Elliott, E. N.,Sheaffer, K. L.,Schug,J., Stappenbeck, T. S. &Kaestner, K. H. Dnmt1 55. Johnson, J. L. Emerging regulators of vascular smooth muscle cell function in is essential to maintain progenitors in the perinatal intestinal epithelium. the development and progression of atherosclerosis. Cardiovasc. Res. 103, Development 142,2163–2172 (2015). 452–460 (2014). 42. Liang, G. et al. Cooperativity between DNA methyltransferases in the 56. Ro, S. et al. Cloning and expression profiling of small RNAs expressed in the maintenance methylation of repetitive elements. Mol. Cell Biol. 22,480–491 mouse ovary. RNA 13,2366–2380 (2007). (2002). 57. Lee, M. Y. et al. Smooth muscle cell genome browser: enabling the identifi- 43. Takekawa, M. & Saito, H. A family of stress-inducible GADD45-like proteins cation of novel serum response factor target genes. PLoS One 10, e0133751 mediate activation of the stress-responsive MTK1/MEKK4 MAPKKK. Cell 95, (2015). 521–530 (1998). 58. Ro, S. et al. A model to study the phenotypic changes of interstitial cells of 44. Ying, J. et al. The stress-responsive gene GADD45G is a functional tumor Cajal in gastrointestinal diseases. Gastroenterology 138,1068–1078 (2010). suppressor, with its response to environmental stresses frequently disrupted epigenetically in multiple tumors. Clin. Cancer Res. 11,6442–6449 (2005). Official journal of the Cell Death Differentiation Association

Journal

Cell Death & DiseaseSpringer Journals

Published: Apr 27, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off