The Dysregulation of the DLK1-MEG3 Locus in Islets from Patients with Type 2 Diabetes Is Mimicked by Targeted Epimutation of Its Promoter with TALE-DNMT Constructs

Diabetes Volume 67, September 2018 1807

The Dysregulation of the DLK1-MEG3 Locus in Islets From Patients With Type 2 Diabetes Is Mimicked by Targeted Epimutation of Its Promoter With TALE-DNMT Constructs

Vasumathi Kameswaran,1 Maria L. Golson,1 Mireia Ramos-Rodríguez,2,3,4 Kristy Ou,1 Yue J. Wang,1 Jia Zhang,1 Lorenzo Pasquali,2,3,4 and Klaus H. Kaestner1

Diabetes 2018;67:1807–1815 | https://doi.org/10.2337/db17-0682

Type 2 diabetes mellitus (T2DM) is characterized by the insulin to regulate glucose uptake by peripheral tissues. inability of the insulin-producing b-cells to overcome Their autoimmune destruction or functional decline can insulin resistance. We previously identified an imprinted lead to type 1 and type 2 diabetes mellitus (T2DM), re- region on chromosome 14, the DLK1-MEG3 locus, as spectively. Thus, understanding the molecular mecha- being downregulated in islets from humans with T2DM. nisms underlying b-cell physiology is fundamental to In this study, using targeted epigenetic modifiers, we improving current diabetes treatment strategies. prove that increased methylation at the promoter of We previously demonstrated that the imprinted DLK1- STUDIES ISLET b b Meg3 in mouse TC6 -cells results in decreased tran- MEG3 locus is misregulated in islets from donors with scription of the maternal transcripts associated with this T2DM (1). This locus consists of the paternally active DLK1, b locus. As a result, the sensitivity of -cells to cytokine- RTL1,andDIO3 genes, as well as maternally expressed mediated oxidative stress was increased. Additionally, long noncoding RNAs MEG3, RTL1as,andMEG8, a large we demonstrate that an evolutionarily conserved intronic microRNA (miRNA) cluster, and several small nucleolar region at the MEG3 locus can function as an enhancer in RNAs (2,3). The genes in this locus are active in human b- bTC6 b-cells. Using circular chromosome conformation but not a-cells at very high levels (1,4) and repressed in islets capture followed by high-throughput sequencing, we demonstrate that the promoter of MEG3 physically interacts from patients with T2DM (1). This decreased expression with this novel enhancer and other putative regulatory correlates with hypermethylation at the MEG3 promoter. elements in this imprinted region in human islets. Re- Consistent with these human studies, Meg3 expression is markably, this enhancer is bound in an allele-specific decreased in mouse models of type 1 diabetes mellitus and manner by the transcription factors FOXA2, PDX1, and T2DM (5). NKX2.2. Overall, these data suggest that the intronic Little is known about the mechanism by which imprint- MEG3 enhancer plays an important role in the regulation ing at the DLK1-MEG3 locus is regulated, particularly in of allele-specific expression at the imprinted DLK1-MEG3 human islets and b-cells. Monoallelic expression at this locus in human b-cells, which in turn impacts the sensi- locus is established and maintained through specific meth- tivity of b-cells to cytokine-mediated oxidative stress. ylation patterns at two differentially methylated regions (DMRs), the intergenic IG-DMR and MEG3-DMR, over- lapping the promoter of the maternal transcript (6,7). Diabetes refers to a group of metabolic diseases charac- These DMRs are paternally methylated. Although MEG3 terized by an insufficient insulin response to high blood promoter hypermethylation and a concomitant decrease in glucose levels. Pancreatic b-cells are critical regulators of expression have been reported in several human diseases glucose homeostasis as they produce, store, and secrete (8–14), a causal relationship between these observations

1Department of Genetics and Institute for Diabetes, Obesity and Metabolism, Received 14 June 2017 and accepted 18 June 2018. University of Pennsylvania, Philadelphia, PA This article contains Supplementary Data online at http://diabetes 2 Program of Predictive and Personalized Medicine of Cancer, Department of .diabetesjournals.org/lookup/suppl/doi:10.2337/db17-0682/-/DC1. Endocrinology, Germans Trias i Pujol University Hospital and Research Institute, © 2018 by the American Diabetes Association. Readers may use this article as Badalona, Spain long as the work is properly cited, the use is educational and not for profit, and the 3Josep Carreras Leukaemia Research Institute, Badalona, Spain work is not altered. More information is available at http://www.diabetesjournals 4CIBER de Diabetes y Enfermedades Metabólicas Asociadas, Barcelona, Spain .org/content/license. Corresponding author: Klaus H. Kaestner, [email protected]. 1808 DLK1-MEG3 Locus in T2DM Diabetes Volume 67, September 2018 has not been established. In this study, we demonstrate using PyroMark denaturation solution (Qiagen), washed, that hypermethylation of this DMR using targeted DNA and then neutralized using a vacuum preparation station methylation in mouse bTC6 b-cells causes decreased tran- (Qiagen PyroMark Q96 workstation). After annealing the scription of Meg3 and that this repressed expression sequencing primer to the immobilized strand, pyro- exacerbates b-cell death, consistent with our observation sequencing was performed on the PyroMark Q96 MD in islets from human donors with T2DM (1). Furthermore, (Qiagen) using the PyroMark Gold Q96 CDT kit (Qiagen) we identified a putative enhancer within an intron of the according to the manufacturer’s instructions. Data were human MEG3 gene that is bound by transcription factors analyzed using the Pyro Q-CpG software program (Qiagen). that are critical for islet function. We demonstrate that Gene Expression by Quantitative RT-PCR this sequence indeed functions as an active enhancer and Primers used to quantify gene expression of Meg3 and . physically interacts with the MEG3 promoter 16 kb Dlk1 in TALE experiments are listed in Supplementary upstream. Intriguingly, this enhancer is bound by islet Table 3. transcription factors in an allele-specific manner in human islets. Overall, our results suggest an important regulatory Cell Death Assays function for this newly characterized MEG3 enhancer and bTC6 cells were seeded in six-well plates at a density of 3 6 provide insights into the mechanism of imprinting at the 1 10 cells/well. The next day, cells were transfected DLK1-MEG3 locus in b-cells. with 1 mg plasmid and 4 mL Lipofectamine 2000 per well. After72h,cellsweretreatedwith20ng/mLmousetumor necrosis factor-a, 5 ng/mL mouse interleukin-1b, and RESEARCH DESIGN AND METHODS 10 ng/mL mouse interferon-g. Forty-eight hours later, Human Islets cells were treated with CellROX Deep Red (Thermo Fisher Human islets and relevant donor information including Scientific) according to the manufacturer’s instructions age, sex, diabetes status, and BMI were obtained from the and analyzed for GFP and CellROX fluorescence using Islet Cell Resource Center of the University of Pennsylva- an LSRII (BD Biosciences). nia, the National Institute of Diabetes and Digestive and Kidney Diseases–supported Integrated Islet Distribution Dual Luciferase Reporter Assay: Enhancer Activity Program (https://iidp.coh.org), and the National Disease The human MEG3-DMR (hg19, chr14:100,824,307-100, Research Interchange. The donor’s diabetes status was 826,452) and MEG3 enhancer (chr14:101,308,419-101,309, defined by the patient’s medical record and, when avail- 405) were subcloned into pGL3-Basic or pGL4.23[luc2/ minP] luciferase reporter (Promega). A total of 50,000 able, the hemoglobin A (Supplementary Table 1). 1c bTC6 cells was seeded per well of a 24-well plate and Transcription Activator-Like Effector Experiments transfected with 500 ng of plasmid DNA and 10 ng of Transcription activator-like effectors (TALEs) targeting the pRL-SV40 (Promega). Cells were harvested 24 h posttrans- mouse Meg3-DMR (mm10 chr12:109,540,635-109,540,653) fection and processed for luciferase readout. Experiments were designed using an online resource and are as described were performed in triplicates with four technical replicates 6 previously (1). A total of 12 3 10 bTC6 cells in a 10-cm dish per experiment. were transfected with 12 mg of either TALE wild-type (WT) Allele-Specific Chromatin Immunoprecipitation PCR or mutant plasmids using FuGENE HD Transfection Allele-specific chromatin immunoprecipitation (ChIP) was Reagent (Promega). Cells were FACS sorted for GFP+ cells performed according to the schema in Fig. 4A. A total of after 72 h on a Diva 206 (University of Pennsylvania Flow 30 ng of genomic DNA from islet donors was used to identify Cytometry and Cell Sorting Facility). After sorting, the cells donors heterozygous for the single nucleotide polymor- were pooled for total RNA and genomic DNA extraction phism (SNP) rs3783355. SNP genotyping was performed using an AllPrep DNA/RNA mini kit (Qiagen). using TaqMan SNP Genotyping Assay (C_1259770_10, DNA Methylation Analysis catalog number 4351379; Thermo Fisher Scientific) and Genomic DNA was isolated using an AllPrep DNA/RNA kit TaqMan Genotyping Master Mix (catalog number 4371353; (Qiagen). A total of 325 ng of extracted DNA or unsoni- Thermo Fisher Scientific) on a Stratagene Mx3000P thermo- cated chromatin input was bisulfite treated with the cycler. Chromatin was extracted from islets from donors EpiTect Bisulfite kit (Qiagen) and eluted in 20 mLof without diabetes as previously described (2). The ChIP anti- Buffer EB. PCR and sequencing primers were designed using bodies and conditions used for this experiment were de- PyroMark assay design software version 2.0 (Qiagen) (se- scribed by Charlier et al. (3). The primers used for this quences listed in Supplementary Table 2) to cover CpGs experiment are listed in Supplementary Table 4. Following throughout the Meg3-Dlk1 locus. Bisulfite-converted DNA PCR of the input and ChIP DNA, libraries were prepared was amplified by PCR using the PyroMark PCR kit (Qiagen) using the NuGEN Mondrian SP+ system and sequenced on at 95°C for 15 min followed by 45 cycles at 95°C for 15 s, an Illumina MiSeq to obtain ;200,000 reads per library 57°C for 30 s, and 72°C for 15 s. Biotinylated PCR pro- (sequence read count ranged from 24,568 to 916,096 ducts were immobilized onto streptavidin-coated sephar- reads). For the qualitative assay, 300 ng of PCR-amplified ose beads (GE Healthcare), and DNA strands were separated input or FOXA2 ChIP DNA was digested using the restriction diabetes.diabetesjournals.org Kameswaran and Associates 1809 enzyme BanII. Uncut PCR product (150 bp) and BanII- that had been successfully transfected based on their GFP digested PCR products (113-bp plus 37-bp fragments) were expression. TALE-DNMT–expressing bTC6 cells exhibited run on a 3% gel to visualize differences. an increase in methylation of 20%, along with a 75% de- Circular Chromosome Conformation Capture crease in Meg3 transcript levels, demonstrating directly that Sequencing the activity of the Meg3 locus is controlled by DNA meth- Circular chromosome conformation capture sequencing ylation in b-cells. Cells expressing the catalytically inactive (4C-Seq) was performed on ;10,000 islet equivalents of version of DNMT showed no changes in DNA methylation human islets using the enzymes DpnII (first enzyme) and or Meg3 RNA levels, verifying that the observed changes fi NlaIII (second enzyme) according to Dorrell et al. (4). The were speci c to methyltransferase activity and not an libraries were prepared using the BisPCR2 protocol (1). artifact of transfection and cell sorting (Fig. 1B and D). fi Libraries were sequenced on an Illumina HiSeq on Rapid To con rm that both transfected cell populations had re- Run Mode to obtain 100-bp sequences (single-end reads). ceived comparable amounts of TALE vector, we performed fi Five pooled libraries were sequenced per lane, with 20% quantitative PCR with primers speci c to the TALE con- PhiX supplemented to increase read diversity. Primers struct backbone (Supplementary Fig. 1B). used to determine DpnII digestion efficiency and for library TALE-DNMT constructs can have proximal off-target effects (16). To characterize possible nonspecific targets of preparation are listed in Supplementary Table 5. 4C-Seq fi reads were analyzed as previously described (15). More than TALE binding and methyltransferase activity, we pro led 14 million reads per viewpoint were sorted and aligned to DNA methylation at CpG islands near the Meg3-DMR. The the human genome (hg19). Reads located in fragments IG-DMR of the Dlk1-Meg3 locus is located 16 kb upstream of the Meg3 TSS and also paternally methylated in both human flankedbytworestrictionsitesofthesameenzymeorin and mouse b-cells (1). However, we found this region to fragments ,40 bp were filtered out. Mapped reads were be fully methylated and thus uninformative in bTC6 cells then converted to reads per first enzyme fragment and (Supplementary Fig. 1C). Next, we tested the methylation smoothened using a moving average of 30 fragments per levels of the Dlk1 promoter, located ;87 kb upstream of window. the Meg3 TSS. This region demonstrated a 20% increase In order to identify significant 4C-Seq interactions, we in methylation in cells transfected with WT TALE-DNMT proceed as previously described (15). Briefly, an average compared with cells transfected with the mutant construct, background level was estimated by shuffling the fragments but no change in expression, suggesting possible chromatin 1,000 times in a 1-Mb window centered on the viewpoint looping of the Dlk1 promoter to the Meg3 promoter DMR and smoothened as described above. This randomized (Fig. 1C and E). We observed no difference in methylation profile was then used to calculate the P value for each levels at two CpG islands close to the Dio3 gene ;650 kb and potential target in the observed 4C-Seq distribution by 737 kb downstream of the Meg3 TSS and found no change in means of Poisson probability function. We considered sig- DNA methylation status at this locus (Supplementary Fig. 2). nificant interactions for each 4C-Seq experiment, regions 2 Several targets of the miRNAs encoded by the Meg3 with a Poisson probability ,1 3 10 10.Inordertotake locus, such as islet amyloid polypeptide p53-induced nu- into account biological variability of the primary tissues, clear protein 1 (TP53INP1), function in controlling the only significant interactions that were consistent in at least sensitivity of b-cells to metabolic stress and proapoptotic two of the three replicates were retained. stimuli (1). Therefore, we hypothesized that the decreased expression of the miRNAs encoded by the Meg3 RESULTS locus would lead to increased susceptibility of b-cells to Promoter Methylation Causes Decreased Expression proapoptotic stimuli. To test this, we measured oxidative of MEG3 stress, as a marker upstream of cell death, in response The MEG3 promoter is hypermethylated in pancreatic to inflammatory cytokine treatment in TALE-DNMT– islets from donors with T2DM compared with donors transfected b-cells (18). We found a significant increase in without diabetes, correlating with a decrease in expression oxidative stress in TALE-DNMT–expressing cells relative of MEG3 and its associated miRNAs, which are all pro- to the control group (Fig. 1E), indicating that decreasing duced from a single primary transcript originating from expression of the Meg3 locus and its associated miRNAs a single promoter (1). To test whether increased methyl- does indeed increase susceptibility to cytokines. This find- ation at this region directly causes a decrease in expression, ing is consistent with the dramatic downregulation of this we used TALEs to target DNA methyltransferases (DNMTs) cluster of miRNAs in islets from individuals with T2DM to the mouse Meg3-DMR sequence 360 bp upstream of the (1), which correlates with the increased sensitivity of Meg3 transcription start site (TSS) (Fig. 1A and Supplemen- T2DM b-cells to metabolic and apoptotic stress. In sum- tary Fig. 1A). These TALE-DNMTs are efficient mediators of mary, using targeted epigenetic modifiers, we have shown targeted DNA methylation (16). As a negative control, we that hypermethylation of the Meg3 promoter causes introduced an inactivating point mutation into the catalytic decreased production of the RNA transcripts produced by domain of DNMT (16,17). We transfected these constructs this locus and a subsequent increase in cytokine-mediated into bTC6 mouse insulinoma cells and FACS-sorted cells b-cell death. 1810 DLK1-MEG3 Locus in T2DM Diabetes Volume 67, September 2018

Figure 1—Targeted methylation of the Meg3-DMR results in decreased expression and increased b-cell death. bTC6 mouse insulinoma cells were transfected with TALE molecules speciﬁc to the mouse Meg3-DMR fused to either a WT or mutant DNMT catalytic domain. Transfected cells were sorted by GFP expression. A: Schema of mouse Meg3-DMR spanning ;2 kb, depicting TALE-DNMT (mm10 chr12:109,540,635- 109,540,653) binding positions relative to the Meg3 TSS. Regions assayed for methylation levels by pyrosequencing are depicted by blue boxes. Percent methylation levels determined by pyrosequencing of CpGs in the Meg3-DMR (mm10 chr12:109,540,014-109,542,041) in untransfected bTC6 (n = 3), WT TALE-DNMT (n = 3), or mutant TALE-DNMT–transfected (n = 2) cells (B) and Dlk1 promoter region (mm10 chr12:109452986-109453847) in untransfected bTC6 cells (n = 2), WT TALE-DNMT GFP+ (n = 3), and mutant TALE-DNMT GFP+ cells (n =1) (C). *P , 0.05; **P , 0.01. D: Relative expression of Meg3 and Dlk1 measured by quantitative RT-PCR between WT (n = 8) and mutant (n =7) TALE-DNMT–transfected bTC6 cells, sorted by GFP levels. Data are represented as mean 6 SEM. **P , 0.01. E: Percent of WT or mutant TALE- DNMT GFP+ bTC6 cells labeled by CellROX ﬂuorescence, a marker of oxidative stress, following cytokine treatment for 48 h (n = 4). Data are represented as mean 6 SEM and P , 0.02.

Characterization of a Novel Enhancer in MEG3 Addition of the intronic enhancer, in either orientation, From a previous study that profiled regulatory elements further doubled luciferase activity in mouse bTC6 b-cells based on a comparative analysis of expression data, tran- (Fig. 2B). We additionally validated that the MEG3 en- scription factor binding data, and chromatin marks from hancer increased luciferase activity of an unrelated pro- human islets and FACS-sorted b-cells (15), we identified moter in b-cells by fivefold (Supplementary Fig. 3B). Thus, a putative enhancer within an intron of the MEG3 gene. this intronic region of the MEG3 gene can function as an This region, located ;16 kb downstream of the MEG3 TSS, enhancer in b-cells. is bound by transcription factors critical to islet development and function and marked by histone modifications Long-range Interactions of the MEG3 Promoter and that correlate with active enhancer activity (19) (Fig. 2A). Enhancer Similar H3K27ac enrichment at this region has been ob- We hypothesized that the MEG3 enhancer may physically served in a human lung fibroblast cell line, but not in other interact with the promoter to facilitate transcription of the human cell lines (20) (Supplementary Fig. 3A). We hy- maternal noncoding RNA transcript. To determine long- pothesized that this putative intronic MEG3 enhancer may range interactions between the MEG3 promoter and the be a critical regulator of monoallelic expression at the newly characterized enhancer, we undertook 4C-Seq (21). DLK1-MEG3 locus. This technique provides an unbiased sampling of all inter- First, we sought to validate the activity of the putative acting partners of a selected region of interest (the view- enhancer using luciferase reporter assays. The MEG3-DMR point). We performed 4C-Seq using either the MEG3 promoter sequence itself increased luciferase activity by promoter or enhancer as viewpoints separately, using hu- 14-fold, validating its strong activity in b-cells (1) (Fig. 2B). man islets from three independent donors. This reciprocal diabetes.diabetesjournals.org Kameswaran and Associates 1811

Figure 2—Characteristics of a novel intronic enhancer human islet. A: The chromatin landscape for a putative enhancer in an intron of MEG3 is shown with human islet ChIP sequencing tracks for histone modiﬁcation marks associated with enhancers (H3K4me1 and H3K27ac) and active promoters (H3K4me3). Occupancy of islet transcription factors PDX1, NKX2.2, and FOXA2 at the putative enhancer is also shown. Data retrieved from Pasquali et al. (15). B: Activity of the MEG3 enhancer was validated by luciferase reporter assays. pGL3 vectors with the MEG3-DMR and the enhancer sequence in either its native (For) or reverse (Rev) orientation were transfected into bTC6 cells. Data are represented as mean 6 SEM. n =3.P values calculated by Student t test. **P , 0.1 3 1023; ***P , 0.1 3 1025.

approach allowed us to test the hypothesis that the MEG3 chromatin confirmation at the DLK1-MEG3 locus has been promoter and enhancer interact with each other while also reported in any tissue or species. identifying potential other chromatin interactions of these Consistent with our hypothesis, we found that the MEG3 regulatory elements. To date, no information regarding the promoter displayed frequent interactions with the enhancer, 1812 DLK1-MEG3 Locus in T2DM Diabetes Volume 67, September 2018 and vice versa, in three biological replicates (Fig. 3). Strik- Meg3-DMR is restricted to the maternal chromosome (24). ingly, the promoter and enhancer shared many of their long- To test whether the islet transcription factors that bind the range interactions, including with the paternally expressed MEG3 enhancer are similarly restricted to a single allele, we DLK1 gene. Additionally, these regulatory elements made performed allele-specific ChIP-PCR for FOXA2, NKX2.2, and contacts with other putative enhancers within the imprinted PDX1. In order to differentiate the two alleles, we screened domain, including an enhancer that lies intergenic to DLK1 donor islets for heterozygosity of rs3783355, a common and MEG3, as well as an enhancer cluster downstream of SNP that lies within the enhancer region (Fig. 4). Impor- MEG3 that overlaps with the small nucleolar RNA transcript, tantly, this SNP does not overlap with the consensus binding MEG8. However, some interactions were unique to each motifs for any of the three transcription factors being tested. viewpoint. One such example is the interaction between Using islets from donors heterozygous for rs3783355, we the MEG3 promoter and an enhancer cluster upstream of performed ChIP-PCR and then high-throughput sequencing DLK1. Thus, using 4C-Seq, we have mapped the long-range to determine the relative abundance of the two alleles. We interactions of the MEG3 promoter and enhancer in anticipated that the input material would have an equal human islets and established that the intronic enhancer representation of both alleles, whereas allele-specific tran- in the MEG3 gene 16 kb downstream of the TSS loops back scription factor binding would result in a preferential to the MEG3 promoter. These findings also provide a plau- amplification of a single allele (schema outlined in Fig. sible explanation for the increased DNA methylation we 4A). As predicted, we observed a roughly equal number of saw at the Dlk1 promoter when we performed targeted reads for each allele in the input material for all samples, methylation of the Meg3 promoter DMR (Fig. 1), as the whereas the transcription factor binding was skewed to- two regions are in frequent physical contact in b-cells. ward one allele, with FOXA2 demonstrating the strongest allelic bias (Fig. 4B). Allele-Specific Transcription Factor Binding at the Additionally, for an alternate, qualitative readout of MEG3 Enhancer allele-specific transcription factor binding, we took advan- In mouse midgestation embryos, the open chromatin tage of the fact that rs3783355 lies in the recognition landscape of the Meg3 promoter is restricted to the active site for the restriction enzyme BanII. As expected, the in- maternal allele (22,23). Similarly, in mouse embryonic put samples had both the uncut and digested frag- fibroblasts, binding of the insulator protein CTCF to the ments, corresponding to the A and G alleles, respectively.

Figure 3—The MEG3 promoter and enhancer interact with other putative enhancers within the DLK1-MEG3–imprinted region. Genome browser image of the selective interactions of the MEG3 promoter and enhancer within 350 kb of the MEG3 promoter. 4C-Seq was performed using the MEG3 promoter and enhancer, respectively, as viewpoints (indicated with red triangles) using three human islet donors (represented by different shaded tracks) per viewpoint. Bars under each track represent the significant interaction sites (P , 1 3 1028). The MEG3 promoter and enhancer make frequent contact with putative enhancers within the imprinted locus. ChIP sequencing for the histone modification marks and the putative enhancer identification was performed by Pasquali et al. (15). diabetes.diabetesjournals.org Kameswaran and Associates 1813

Figure 4—Allele-specific transcription factor occupancy at the MEG3 enhancer. A: Schematic representation of the allele-specific ChIP experimental design. Transcription factor ChIP is performed on islets from donors heterozygous for the SNP rs3783355. Following amplification using primers surrounding the SNP, the PCR products are sequenced to quantitatively determine the relative representation of the two alleles. B: Relative amplification (percent sequencing reads) of the rs3783355 alleles as determined by high-throughput sequencing of input, NKX2.2, FOXA2, and PDX1 ChIP DNA from islet donors heterozygous for rs3783355 (G/A). C: The minor allele of rs3783355 alters the recognition sequence of a restriction enzyme, BanII. Following ChIP-PCR to determine FOXA2 occupancy at the MEG3 enhancer, the PCR products were digested with BanII to qualitatively assess the allelic representation of rs3783355. A representative gel of BanII-digested input and FOXA2 ChIP-PCR products from islets from two donors heterozygous for rs3783355 is shown. Data are represented as mean 6 SEM. P values calculated using Student t test. ***P , 0.001.

Conversely, the FOXA2 transcription factor ChIP samples of the maternal RNAs including a cluster of 54 miRNAs in primarily contained the digested (G allele) fragments with islets from donors with T2DM (1), but had not established minimal uncut (A allele) fragments (Fig. 4C), confirming a causal link between these observations. To circumvent allele-specific occupancy by FOXA2. any confounding nonspecific effects of the use of global These two methods independently verify that the islet- DNA demethylating agents like 5-aza-2-deoxycytidine, we active transcription factors FOXA2, NKX2.2, and PDX1 used TALE molecules fused to a DNMT to specifically direct that bind to the MEG3 enhancer do so in a monoallelic methylation to the Meg3-DMR in mouse b-cells. Using this fashion. However, as we cannot obtain genetic information targeted approach, we demonstrate that increased methyl- regarding the parents of the organ donors who provided ation of the Meg3-DMR in b-cells does in fact cause decreased islets for our study, we cannot determine which allele is Meg3 expression. This approach also enabled us to expand inherited maternally or paternally. Nevertheless, our re- upon our previous observation that targets of the miRNAs in sults suggest that the enhancer is regulated in an allele- this locus are related to b-cell death and apoptosis, providing specific manner in human islets, with the most plausible a functional validation of our previous results. explanation that it is bound by the transcription factors on We also observe an increase in methylation at the the active, maternally inherited allele. promoter of the paternally expressed Dlk1. This change in Dlk1 methylation levels is likely the result of the three- DISCUSSION dimensional chromatin architecture of this region, as we We have previously reported hypermethylation of the also demonstrate that the DLK1 and MEG3 promoters MEG3 promoter and a concomitant decrease in expression physically interact in human islets using our chromatin 1814 DLK1-MEG3 Locus in T2DM Diabetes Volume 67, September 2018 conformation capture analysis. The fact that the MEG3 Research Center Functional Genomics Core was supported by the National regulatory elements make physical contact with DLK1 sug- Institutes of Health (P30-DK-19525). gests that these interactions occur in trans between the two Duality of Interest. No potential conflicts of interest relevant to this article parental chromosomes or, alternatively, that the cis inter- were reported. Author Contributions. V.K. wrote the manuscript and researched data. action is constitutive on both alleles, but gene activation M.L.G., M.R.-R., K.O., Y.J.W., J.Z., and L.P. researched data. K.H.K. contributed is determined by other trans-acting determinants, such as fi to discussion and reviewed and edited the manuscript. K.H.K. is the guarantor transcription factor binding. Allele-speci c occupancy of of this work and, as such, had full access to all of the data in the study and transcription factors and histone marks has been demon- takes responsibility for the integrity of the data and the accuracy of the data strated at cis-regulatory elements of several imprinted re- analysis. gions, including the Meg3-DMR (22,23,25). Our data suggest that the novel MEG3 enhancer is also bound by islet tran- References scription factors in an allele-specific manner in human islets 1. Kameswaran V, Bramswig NC, McKenna LB, et al. Epigenetic regulation of and thereby contributes to monoallelic gene expression. the DLK1-MEG3 microRNA cluster in human type 2 diabetic islets. Cell Metab Our characterization of this novel enhancer is particularly 2014;19:135–145 noteworthy in light of the observation that a SNP located 2. da Rocha ST, Edwards CA, Ito M, Ogata T, Ferguson-Smith AC. Genomic ;2.5 kb upstream of this MEG3 enhancer in the human imprinting at the mammalian Dlk1-Dio3 domain. Trends Genet 2008;24:306– locus is correlated with risk for type 1 diabetes mellitus. The 316 SNP identified as the risk variant lies within the same intron 3. Charlier C, Segers K, Wagenaar D, et al. Human-ovine comparative sequencing of a 250-kb imprinted domain encompassing the callipyge (clpg) locus of MEG3 as the enhancer characterized in our study (26) and and identification of six imprinted transcripts: DLK1, DAT, GTL2, PEG11, anti- is part of the same linkage disequilibrium block. It remains to PEG11, and MEG8. Genome Res 2001;11:850–862 be determined whether this SNP and other variants in 4. Dorrell C, Schug J, Lin CF, et al. Transcriptomes of the major human linkage disequilibrium with it confer risk for type 1 dia- pancreatic cell types. Diabetologia 2011;54:2832–2844 betes mellitus by altering regulation of the genes in the 5. You L, Wang N, Yin D, et al. Downregulation of long noncoding RNA Meg3 imprinted locus by affecting the enhancer region character- affects insulin synthesis and secretion in mouse pancreatic beta cells. J Cell ized in this study. This is indeed supported by evidence that Physiol 2016;231:852–862 in general, disease variants identified by genome-wide as- 6. Schmidt JV, Matteson PG, Jones BK, Guan XJ, Tilghman SM. The Dlk1 and sociation studies are frequently found in cis-regulatory Gtl2 genes are linked and reciprocally imprinted. Genes Dev 2000;14:1997– elements (27,28). 2002 7. Takada S, Paulsen M, Tevendale M, et al. Epigenetic analysis of the Dlk1-Gtl2 It is likely that other unidentified regulators contribute imprinted domain on mouse chromosome 12: implications for imprinting control to the control of imprinting at this locus. Recent studies from comparison with Igf2-H19. Hum Mol Genet 2002;11:77–86 have suggested that similar to other nuclear long non- 8. Benetatos L, Dasoula A, Hatzimichael E, Georgiou I, Syrrou M, Bourantas KL. coding RNAs, MEG3 directly interacts with the polycomb Promoter hypermethylation of the MEG3 (DLK1/MEG3) imprinted gene in multiple repressive complex (PRC2) in embryonic stem cells to guide myeloma. Clin Lymphoma Myeloma 2008;8:171–175 the repressive histone modification mark H3K27me3 to 9. Zhao J, Dahle D, Zhou Y, Zhang X, Klibanski A. Hypermethylation of the its target sites (29,30). A careful characterization of MEG3- promoter region is associated with the loss of MEG3 gene expression in human PRC2 complex targets in adult pancreatic islets will provide pituitary tumors. J Clin Endocrinol Metab 2005;90:2179–2186 better insights into the role of this silencing complex in 10. Guo W, Dong Z, Liu S, et al. Promoter hypermethylation-mediated down- mediating allelic expression at this locus. Additionally, regulation of miR-770 and its host gene MEG3, a long non-coding RNA, in the – although islet transcription factors bound by both the pro- development of gastric cardia adenocarcinoma. Mol Carcinog 2017;56:1924 1934 moter and the enhancer, such as NKX2.2, are likely medi- 11. Sun M, Xia R, Jin F, et al. Downregulated long noncoding RNA MEG3 is ators of the looping interaction between these regions, it is associated with poor prognosis and promotes cell proliferation in gastric cancer. possible that architectural elements unexplored in this Tumour Biol 2014;35:1065–1073 current study are also involved in stabilizing this interaction. 12. Peng W, Si S, Zhang Q, et al. Long non-coding RNA MEG3 functions as Overall, our results demonstrate that targeted methyl- a competing endogenous RNA to regulate gastric cancer progression. J Exp Clin ation of the MEG3-DMR is sufficient to repress the locus Cancer Res 2015;34:79 and increase b-cell susceptibility to cytokines and provide 13. Yin DD, Liu ZJ, Zhang E, Kong R, Zhang ZH, Guo RH. Decreased ex- evidence for the role of a novel enhancer in the regulation pression of long noncoding RNA MEG3 affects cell proliferation and predicts of imprinting at the DLK1-MEG3 locus in human cells. a poor prognosis in patients with colorectal cancer. Tumour Biol 2015;36: These results extend our understanding of allelic expres- 4851–4859 sion of genes in this important locus and thus its mis- 14. Zhang J, Yao T, Wang Y, Yu J, Liu Y, Lin Z. Long noncoding RNA MEG3 is downregulated in cervical cancer and affects cell proliferation and apoptosis by regulation in T2DM and other diseases. regulating miR-21. Cancer Biol Ther 2016;17:104–113 15. Pasquali L, Gaulton KJ, Rodríguez-Seguí SA, et al. Pancreatic islet enhancer clusters enriched in type 2 diabetes risk-associated variants. Nat Genet 2014;46: Acknowledgments. The authors thank the University of Pennsylvania 136–143 Diabetes Research Center for the use of the Functional Genomics Core. 16. Bernstein DL, Le Lay JE, Ruano EG, Kaestner KH. TALE-mediated epigenetic Funding. This work was supported by the National Institutes of Health (R01- suppression of CDKN2A increases replication in human fibroblasts. J Clin Invest DK-088383 and UC4-DK-104119). The University of Pennsylvania Diabetes 2015;125:1998–2006 diabetes.diabetesjournals.org Kameswaran and Associates 1815

17. Li JY, Pu MT, Hirasawa R, et al. Synergistic function of DNA methyl- 24. Lin S, Ferguson-Smith AC, Schultz RM, Bartolomei MS. Nonallelic tran- transferases Dnmt3a and Dnmt3b in the methylation of Oct4 and Nanog. Mol Cell scriptional roles of CTCF and cohesins at imprinted loci. Mol Cell Biol 2011;31: Biol 2007;27:8748–8759 3094–3104 18. Tanaka Y, Tran PO, Harmon J, Robertson RP. A role for glutathione per- 25. Verona RI, Thorvaldsen JL, Reese KJ, Bartolomei MS. The transcrip- oxidase in protecting pancreatic beta cells against oxidative stress in a model of tional status but not the imprinting control region determines allele-specific glucose toxicity. Proc Natl Acad Sci U S A 2002;99:12363–12368 histone modifications at the imprinted H19 locus. Mol Cell Biol 2008;28: 19. Creyghton MP, Cheng AW, Welstead GG, et al. Histone H3K27ac separates 71–82 active from poised enhancers and predicts developmental state. Proc Natl Acad Sci 26. Wallace C, Smyth DJ, Maisuria-Armer M, Walker NM, Todd JA, Clayton DG. U S A 2010;107:21931–21936 The imprinted DLK1-MEG3 gene region on chromosome 14q32.2 alters sus- 20. ENCODE Project Consortium. An integrated encyclopedia of DNA elements in ceptibility to type 1 diabetes. Nat Genet 2010;42:68–71 the human genome. Nature 2012;489:57–74 27. Sakabe NJ, Savic D, Nobrega MA. Transcriptional enhancers in development 21. Zhao Z, Tavoosidana G, Sjölinder M, et al. Circular chromosome conformation and disease. Genome Biol 2012;13:238 capture (4C) uncovers extensive networks of epigenetically regulated intra- and 28. Visel A, Rubin EM, Pennacchio LA. Genomic views of distant-acting en- interchromosomal interactions. Nat Genet 2006;38:1341–1347 hancers. Nature 2009;461:199–205 22. Carr MS, Yevtodiyenko A, Schmidt CL, Schmidt JV. Allele-specific histone 29. Zhao J, Ohsumi TK, Kung JT, et al. Genome-wide identification of polycomb- modifications regulate expression of the Dlk1-Gtl2 imprinted domain. Genomics associated RNAs by RIP-seq. Mol Cell 2010;40:939–953 2007;89:280–290 30. Kaneko S, Bonasio R, Saldaña-Meyer R, et al. Interactions between JARID2 23. McMurray EN, Schmidt JV. Identification of imprinting regulators at the and noncoding RNAs regulate PRC2 recruitment to chromatin. Mol Cell 2014;53: Meg3 differentially methylated region. Genomics 2012;100:184–194 290–300