Page 1 of 40 Diabetes

Myc Is Required for Adaptive ß-Cell Replication in Young Mice but Is not Sufficient in One-Year-Old Mice Fed with a High-Fat Diet.

Carolina Rosselot1,*, Anil Kumar1,*, Jayalakshmi Lakshmipathi1, Pili Zhang1, Geming Lu1, Liora S. Katz1, Edward V. Prochownik3, Andrew F. Stewart1, Luca Lambertini1, Donald K. Scott1,2,**, Adolfo Garcia-Ocaña1,2,**.

1Diabetes, Obesity and Metabolism Institute, Division of Endocrinology, Diabetes and Bone Diseases, The Icahn School of Medicine at Mount Sinai, New York, NY. 2The Mindich Child Health and Development Institute, The Icahn School of Medicine at Mount Sinai, New York, NY. 3Division of Hematology/Oncology, Children’s Hospital of Pittsburgh of UPMC; The Department of Microbiology and Molecular Genetics, The University of Pittsburgh Medical Center; The Hillman Cancer Center and The University of Pittsburgh Liver Research Center, Pittsburgh, PA 15224.

*indicates equal contribution.

**Donald K. Scott and Adolfo Garcia-Ocaña are shared senior authors.

To whom correspondence should be addressed: Adolfo Garcia-Ocaña, email adolfo.garcia- [email protected], phone 212-241 9793, or Donald K. Scott, email [email protected], phone 212-241 2835.

Running title: Myc and adaptive ß-cell replication

Keywords: Myc, Diabetes, aging, metabolic stress, pancreatic ß-cell, DNA methylation, adaptation, regeneration, obesity, proliferation.

The authors have declared that no conflict of interest exists.

1

Diabetes Publish Ahead of Print, published online July 10, 2019 Diabetes Page 2 of 40

Abstract

Failure to expand pancreatic ß-cells in response to metabolic stress leads to excessive work load resulting in ß-cell dysfunction, de-differentiation, death and development of type 2 diabetes. Here we demonstrate that induction of Myc is required for increased pancreatic ß-cell replication and expansion during metabolic stress-induced insulin resistance with short-term high fat diet (HFD) in young mice. ß-cell-specific Myc knockout mice fail to expand adaptively, and show impaired glucose tolerance and ß-cell dysfunction. Mechanistically, PKCζ, ERK1/2, mTOR and PP2A are key regulators of the Myc response in this setting. DNA methylation analysis shows hypomethylation of cell cycle genes that are Myc targets in islets from young mice fed with a short- term HFD. Importantly, DNA hypomethylation of Myc response elements does not occur in islets from one-year-old mice fed with a short-term HFD, impairing both Myc recruitment to cell cycle regulatory genes and ß-cell replication. We conclude that Myc is required for metabolic stress- mediated ß-cell expansion in young mice, but with aging Myc upregulation is not sufficient to induce ß-cell replication by, at least partially, an epigenetically-mediated resistance to Myc action.

2 Page 3 of 40 Diabetes

Introduction

The pancreatic ß-cell adapts to enhanced metabolic demand and insulin resistance by increasing ß-cell mass and function (1-4). This adaptation is orchestrated by signals derived from nutrient metabolism, growth factors and hormone signaling (2,5). However, if adaptive expansion is impaired, ß-cell dysfunction, de-differentiation, death might occur leading to ß-cell failure and type 2 diabetes (T2D) (6,7). Understanding the mechanisms that regulate adequate ß-cell adaptation to increased metabolic demand and insulin resistance is of great importance for the development of potential novel disease modifying treatments.

Myc is a pleiotropic transcription factor that controls multiple cellular functions including proliferation, growth, death, differentiation and genome stability (8,9). Myc is expressed at very low levels, if at all, in quiescent cells. Mild increases (1.5-2-fold) in these normally low levels occur in the course of normal development, growth, and physiology. On the other hand, the expression of Myc is dramatically and irreversibly increased in tumors in which it is involved in regulating cell cycle checkpoints and apoptotic cell death pathways (8-12). Therefore, in order to maintain normal cell function, Myc expression is tightly controlled at the level of transcription, mRNA stability, translation and protein stability (13-16).

In quiescent adult pancreatic islets, Myc expression is rapidly but mildly (~2x) upregulated at the mRNA and protein levels by high glucose both in vitro and in vivo (17,18). Myc expression also is upregulated in islets during pregnancy, where increased metabolic demand and enhanced ß-cell proliferation and mass are present (19-21). Since acute increased metabolic demand leads to a remarkable increase in ß-cell proliferation and a mild increase in Myc expression in vivo, the idea of manipulating Myc expression to favor ß-cell proliferative and regenerative therapies has been pursued over the years (22-24). Transgenic mice expressing very high levels of Myc in ß-cells display increased ß-cell proliferation and apoptosis, downregulation of insulin gene expression and development of diabetes (23). On the other hand, “gentle” induction of Myc expression in rodent and human ß-cells enhances ß-cell replication without induction of cell death or loss of insulin secretion, suggesting that appropriate levels of Myc could have therapeutic potential for ß- cell regeneration (22). Indeed, harmine, a mild (~2x) inducer of Myc expression induces remarkable human ß-cell proliferation in vitro and in vivo with no signs of ß-cell death or de- differentiation (25). Puri et al. have recently shown that Myc is required for postnatal ß-cell proliferation, and that mild, lifelong Myc overexpression in the mouse ß-cell markedly enhances ß-cell mass and leads to sustained mild hypoglycemia, without induction of tumorigenesis (26).

In the current study, we have analyzed the role of Myc in the ß-cell adaptive response to increased metabolic demand. We find that Myc disruption in the rodent ß-cell in vivo and in vitro impairs glucose- and short-term HFD-induced ß-cell proliferation, expansion and function; that the PKCζ- ERK-mTOR-PP2A axis controls the level of phosphorylated/stable Myc in ß-cells; and that gentle, physiological upregulation of Myc expression remarkably increases ß-cell proliferation in islets from both young and old mice. In contrast to young mice, however, Myc action is impaired in the islets of one-year-old mice fed with a short-term HFD. ChIP, DNA methylation analyses and DNA demethylation by 5-azadeoxycytidine treatment suggest that epigenetically-mediated Myc resistance constrains, at least partially, the adaptive proliferation of ß-cells in the context of increased insulin demand in aging.

3 Diabetes Page 4 of 40

Research Design and Methods mRNA Library Preparation, Sequencing and Expression Analysis. RNA preparation, libraries generation and sequencing, and gene expression analysis were performed at the New York Genome Center using standard procedures (27-31). Details are provided in the Supplemental Data. RNAseq data and DNA methylation data (see below) have been deposited in the Gene Expression Omnibus (GEO) data repository (accession number GSE131941).

Genetically Modified Mice. ß-cell–specific inducible Myc knockout mice (βMycKO mice) were generated by combining MIP-creERTAM mice (32) with Myclox/lox mice (33), both in a C57BL/6J mouse background. Cre-mediated recombination and disruption of Myc expression was achieved by intraperitoneal (ip) injection for five consecutive days of 50 μg/g bw of tamoxifen (Tam) (Sigma- Aldrich) dissolved in corn oil (34). All studies were performed with the approval of and in accordance with guidelines established by the Icahn School of Medicine at Mount Sinai Institutional Animal Care and Use Committee.

Short-term High-Fat Diet Feeding. 8- and 52-weeks-old C57Bl6N mice (Charles River, Wilmington, MA) and 14-weeks-old Tam- or corn oil-treated βMycKO mice were fed with the lard- based HFD (41% kcal from fat; TD 96001; Harlan Teklad) or a regular diet (RD) (13.1% kcal from fat; Purina PicoLab 5053; LabDiet) (34). After 7 days, body weights, nonfasting blood glucose and plasma insulin were measured, and pancreata harvested and processed for histological studies or islet isolation.

Glucose Homeostasis. Blood glucose was determined by glucometer and plasma insulin by ELISA (Mercodia). Intraperitoneal glucose tolerance test (IPGTT) was performed in 16–18h– fasted mice injected ip with 2g d-glucose/kg (35).

Immunohistochemistry and Analysis of ß-cell Proliferation and Mass. Paraffin-embedded pancreatic sections were immunostained with DAPI and antibodies for insulin (Dako) and Myc (Y69, LifeSpan). ß-cell proliferation was assessed by insulin and Ki67 (Thermo Fisher Scientific) staining, and at least 2,000 ß-cells were blindly counted per mouse (34). ß-cell mass was measured in three insulin-stained pancreas sections per mouse using ImageJ (National Institutes of Health) (34,35).

Generation of Adenoviruses. Adv.Myc, Adv.KD-PKCζ, Adv.CA-PKCζ, Adv.LacZ, Adv.Cre and Adv.GFP were prepared as previously described (34). Multiplicity of infection (MOI) was determined by optical density at 260nm and by plaque assay.

Islet Isolation and Western Blots. Mouse islets were isolated after collagenase P injection through the pancreatic duct (35). Islet or INS-1 832/13 cell protein extracts were separated on SDS-PAGE, membranes incubated with primary antibodies (see Suppl. Table 1) before followed by peroxidase-conjugated secondary antibodies and chemiluminescence detection (34).

ß-cell Proliferation in Mouse Primary Islet Cell Cultures. After islet trypsinization, cells were plated on 12-mm glass coverslips placed in 24-well plates (34,35). Islet cells were either uninfected or transduced with 100 MOI of the adenoviruses indicated above (34). Thereafter, cells

4 Page 5 of 40 Diabetes

were incubated overnight in fresh medium with 5% FBS containing 2 or 20mM glucose. In some experiments, 5nM endothall (Sigma) was added 30min prior to glucose addition; and, 40µM 10058-F4 (Myc inhibitor,1RH, EMD-Millipore) was added overnight with 11mM glucose or together with daily 30µM 5-azadeoxycytidine for 72h (Sigma). Then, cells were rinsed with PBS, fixed in 4% paraformaldehyde and ß-cell proliferation analyzed as above (34,35).

PP2A Activity. PP2A activity was measured using the PP2A Immunoprecipitation Assay Kit (EMD Millipore). Cells were lysed and protein extracts were mixed with PP2A antibody and protein A slurry for 2h at 4°C. After washes, phosphopeptide and assay buffer were added, tubes incubated for 10min at 30°C before malachite green was added, and absorbance measured at 650nm.

ChIP Assay. INS-1 832/13 cells or islets were exposed to 1% formaldehyde for 10min at room temperature. The ChIP protocols were otherwise as previously described (36). The primer sequences for the PCR reactions can be provided upon request.

DNA Methylation Analysis. Islet DNA samples were barcoded and multiplex-sequenced, and the reads run through a customary DNA methylation pipeline for generating methylation calls at every CpG dinucleotide. Probes were designed to capture different regions of the mouse genome (Suppl. Table 2) spanning a total of 1Mbp. Sequencing was performed at the Epigenomics Core Facility of Weill Cornell Medicine (27-31). Details are provided in the Supplementary Data. We used MethylFlash Methylated DNA 5-mC Quantification Kit (Epigentek) to measure global DNA methylation in mouse islets treated with or without 30µM 5-azadeoxycytidine for 72h.

Statistical Analysis. The data are presented as means ± SE. Statistical analysis was performed using unpaired two-tailed Student t test. P < 0.05 was considered statistically significant.

Results

Transcriptome analysis reveals upregulation of Myc target genes in islets from young mice following short-term HFD feeding. Short-term HFD feeding promotes adaptive ß-cell replication (34,37). We performed RNAseq analysis of islets from 8-week-old mice fed with RD or HFD for one week in order to define, in an unbiased way, the ß-cell transcriptional networks during adaptive ß-cell replication. As shown in Figures 1A and B, mice fed with a HFD for one week displayed the expected increase in ß-cell proliferation, body weight, blood glucose, and plasma insulin. RNAseq analysis revealed that 57 genes were significantly increased by at least 1.5-fold and 10 genes significantly decreased by at least 0.67-fold (adjusted P value <0.05) in the islets of the HFD fed mice (Fig. 1C and Suppl. Fig. 1A). Gene set enrichment analysis identified cell cycle and cell division pathways as the top biological processes in islets of young HFD fed mice (Fig. 1D). A closer examination revealed that 35 of the upregulated genes (Fig. 1E) encode cell cycle regulatory molecules of which 21 (Ccna2, Cdk1, Ccnb1, Ccnb2, Cdc20, Cdca3, Mki67, Cdkn2c, Cdkn1a, Nusap1, Top2a, Ube2c, Cdca2, Rgs2, Hmmr, Kif20a, H2afx, Stmn1, Ect2, Plk1,Ttk) are verified Myc target genes (38). Indeed, gene set enrichment analysis revealed enrichment of Myc target genes in HFD fed mice (Fig. 1F). Analysis of pathways that include genes critical for cell replication were significantly enriched in islets from HFD-fed mice (Fig. 1G).

5 Diabetes Page 6 of 40

ChIP in INS-1 832/13 cells confirmed the recruitment of Myc to E-boxes of two selected genes, Ccna2 and Cdk1, which was further stimulated by high glucose (Suppl. Fig. 1B). Upregulation of mRNA from selected genes (Ccna2, Cdk1, Ccnb2) was confirmed in islets from the HFD-fed mice by real-time PCR (Fig. 1H). Notably, in contrast to its target genes, Myc was not increased.

Remarkably, despite the lack of change at the mRNA level, short-term HFD feeding increased Myc protein in islets, assessed by both immunoblots of islet extracts and indirect immunofluorescence in ß-cells (Fig. 1I-J). Thus, the upregulation and activation of Myc protein is a signature event in the adaptive ß-cell expansion in young mice.

Myc is required for adaptive ß-cell replication. Glucose is a well-known inducer of ß-cell proliferation (34,39,40). Myc is upregulated by high glucose in islets and ß-cells in vitro (17,18). Whether Myc is required for glucose-induced ß-cell proliferation is unknown. Therefore, we deleted Myc specifically and conditionally from ß-cells. Initially, islets from Myclox/lox mice (33) (Suppl. Fig. 2A) were treated in vitro with an adenovirus expressing Cre recombinase (Ad.Cre) to achieve DNA recombination (Suppl. Fig. 2B) and downregulation of Myc in islet cells (Suppl. Fig. 2C). Importantly, Myc deletion from cultured Myclox/lox mouse islet cells completely prevented the normal mitogenic response of ß-cells to 20mM glucose (Suppl. Fig. 2D-E).

We next queried whether deletion of Myc from ß-cells in adult mice in vivo would have any impact in glucose homeostasis and compensatory ß-cell proliferation and expansion following short-term HFD feeding. For that purpose, we crossed MIP-creERTAM mice with Myclox/+ mice and generated MIP-creERTAM;Myclox/lox (βMycKO) mice. All the mice used in these studies will have expression of the transgene including the hGH insert (41). βMycKO mice were injected with Tam or corn oil (vehicle) and, following a recovery period, were fed with a HFD or RD for seven days (Fig. 2A). PCR, immunoblot and immunofluorescent labeling confirmed effective blockade of Myc induction in the presence of Tam in βMycKO mice fed a HFD (Fig. 2B-D and Suppl. Fig. 2F). It is important to note that Tam injection did not alter the response of MIP-creERTAM mice to a short-term HFD regarding glucose homeostasis and ß-cell proliferation (34, and Suppl. Fig. 3).

Deletion of Myc in ß-cells did not alter glucose homeostasis in βMycKO mice fed a RD (Fig. 2E- H). However, βMycKO mice fed a short-term HFD displayed increased blood glucose compared with βMycKO mice fed a RD (Fig. 2E). Importantly, the compensatory increase in plasma insulin observed in control mice fed a short-term HFD was not observed in βMycKO mice (Fig. 2F). In addition, βMycKO mice treated with Tam and fed HFD were glucose intolerant compared with Tam-treated βMycKO mice fed the RD or vehicle-treated βMycKO mice fed RD or HFD diets (Fig. 2G-H). Collectively, these results indicate that Myc expression in ß-cells is required for the functional compensatory adaptation induced by acute overnutrition.

To assess the effects of Myc deficiency on compensatory ß-cell proliferation and mass, we analyzed Ki67 labelling in insulin-positive cells in pancreas sections from βMycKO mice fed with RD or HFD for seven days. ß-cell proliferation was significantly increased by HFD feeding in βMycKO mice treated with vehicle (Fig. 2I-J). However, this remarkable increase in ß-cell proliferation was absent in βMycKO mice with Myc deletion in ß-cells (Fig. 2D and 2I-J). The decrease in ß-cell proliferation correlated with a decrease in ß-cell mass in βMycKO mice after

6 Page 7 of 40 Diabetes

HFD feeding (Fig. 2K). Taken together, these results indicate that loss of Myc in ß-cells impairs compensatory ß-cell adaptation to acute overnutrition, hyperglycemia and insulin resistance.

Glucose- and HFD-induced Myc upregulation in ß-cells depends on PKCζ activity. PKCζ activity regulates glucose- and acute HFD-induced ß-cell proliferation (34). Therefore, we wondered whether PKCζ activity might regulate the increase in Myc expression induced by high glucose in vitro and short-term HFD feeding in vivo in mice. Indeed, 20mM glucose remarkably increased Myc expression in INS-1 832/13 cells, and this effect was blocked by overexpressing a kinase-dead (KD) form of PKCζ (Fig. 3A). In addition, one-week of HFD feeding also increased Myc expression in mouse islets and ß-cells which was impaired in transgenic mice expressing KD-PKCζ in ß-cells (Fig. 3B-C).

To determine whether Myc is downstream of PKCζ-mediated glucose-induced ß-cell proliferation, we overexpressed Myc in mouse primary ß-cells with or without KD-PKCζ expression. Overexpression of Myc at levels similar to those induced by glucose (Fig. 3D) avoided the downregulation induced by KD-PKCζ and this resulted in increased ß-cell proliferation (Fig. 3E- F). To confirm whether Myc is necessary for PKCζ-mediated ß-cell proliferation, we expressed a constitutively active (CA) form of PKCζ in mouse primary islet cells (42). Expression of CA-PKCζ in mouse islet cells in culture increased Myc expression (Fig. 3G). Interestingly, mouse ß-cell proliferation induced by 11mM glucose or by 11mM glucose plus CA-PKCζ was blocked by the Myc inhibitor, 10058-F4 [1RH, (18), Fig. 3H]. Collectively, these studies indicate that Myc is required for glucose- and PKCζ-induced ß-cell proliferation, and demonstrate that Myc falls downstream of PKCζ.

Glucose induces Myc phosphorylation in ß-cells. High glucose and activation of PKC ζ leads to increased Myc expression in ß-cells (Fig. 3A and 3G). Myc protein stability is controlled by sequential phosphorylation and dephosphorylation events on two highly conserved residues, Thr58 and Ser62 (13,15). Accordingly, we tested if glucose increases Myc phosphorylation through PKCζ. Ser62 phosphorylation of Myc was significantly increased by glucose in ß-cells and this was inhibited by KD-PKCζ (Fig. 4A). High glucose or KD-PKCζ did not significantly alter the expression of phospho-Thr58-Myc (Fig. 4B), PIN1, or the Myc-specific regulatory subunit of PP2A, B56α (Suppl. Fig. 4). Glucose significantly increased the pS62/pT58 ratio, presumably favoring Myc stability, and this ratio was decreased by KD-PKCζ (Fig. 4C). This suggests that PKCζ might regulate the pS62/pT58 ratio by modulating the activities of ERK1/2, GSK3β or the phosphatase PP2A. High glucose increased ERK1/2 and GSK3β phosphorylation in ß-cells but only ERK1/2 activation was inhibited by KD-PKCζ (Fig. 4D), suggesting that PKCζ acts upstream of ERK1/2 activation, a regulatory event observed in other cell types (43). These results suggest that Myc Ser62 phosphorylation, and hence Myc upregulation, could be compromised with PKCζ activity inhibition by affecting ERK1/2 activation. Indeed, the MEK1 inhibitor, PD98059, abolished glucose-induced Ser62 phosphorylation and Myc upregulation in ß-cells (Suppl. Fig. 5A).

Glucose increases Myc expression and ß-cell proliferation by decreasing PP2A activity through PKCζ and mTOR. Studies presented thus far suggest that Myc phosphorylation and upregulation is induced by glucose and regulated by PKCζ-ERK1/2 but do not address whether

7 Diabetes Page 8 of 40

the phosphatase, PP2A, might also contribute to this effect. High glucose decreased PP2A activity in ß-cells, and this decrease was partially reversed by KD-PKCζ (Fig. 4E). Glucose activates mTORC1 in ß-cells and mTORC1 regulates PP2A activity in other cell types (16,34). Rapamycin, an inhibitor of mTORC1 activity, blocked glucose-mediated Ser62 phosphorylation and Myc upregulation (Suppl. Fig. 5B), and blunted the inhibition of PP2A activity induced by glucose (Fig. 4E). Importantly, the selective PP2A inhibitor, endothall, blocked the downregulation induced by KD-PKCζ on both Ser62-Myc phosphorylation and Myc expression (Fig. 4F). Furthermore, PP2A inhibition bypassed the inhibition induced by KD-PKCζ and further enhanced ß-cell proliferation (Fig. 4G). Taken together, these results suggest that mTOR, through modulation of PP2A, regulates glucose-mediated Myc upregulation and stability, and hence ß-cell proliferation.

Upregulation of Myc in ß-cells from one-year-old mice fed short-term HFD is not sufficient to induce ß-cell proliferation. To interrogate whether Myc expression is also regulated in islets from older mice in which adaptive ß-cell proliferation following short-term HFD feeding does not occur (44), we performed RNAseq analysis as well as western blot and indirect immunofluorescence for Myc in islets from one-year-old mice fed with RD or HFD. As shown in Fig. 5A, one-week of HFD feeding significantly increased body weight, blood glucose and plasma insulin in one-year-old mice, similar to events in young mice. In contrast to young mice, however, ß-cell proliferation was low and unaltered by HFD feeding (Fig. 5A). RNAseq analysis of isolated islets revealed that 222 genes were significantly increased by at least 1.5-fold and 61 genes significantly decreased by at least 0.67-fold (Fig. 5B and Suppl. Fig. 1C) (adjusted P value < 0.05) in the one-year-old HFD fed mice. Importantly, gene set enrichment analysis identified the “negative regulation of biological processes” as the main process in islets from these mice (Fig. 5C). Furthermore, of the genes upregulated in young mice fed a HFD (Fig. 1C), only nine (Pbk, Prc1, Cdkn1a, P2ry, Inhbb, Knstrn, Herpud1, mt-Tl1 and Ankrd34b) were also upregulated in islets from one-year-old mice fed a short-term HFD (Fig. 5D). Interestingly, in contrast to young mice, Myc mRNA was significantly upregulated in islets of one-year-old mice fed with short-term HFD (Fig. 5D). However, of the 21 Myc target genes significantly upregulated in islets of young mice fed with HFD (Fig. 1F), only 2 (Prc1, Cdkn1a) were significantly upregulated in one-year-old mice fed with HFD (Fig. 5D). Real time PCR analysis confirmed the increased expression of Myc and the absence of upregulation of selected cell cycle genes in islets from old mice fed a HFD (Fig. 5E). Thus, short-term HFD increases islet Myc expression in one-year-old mice, but Myc target genes fail to respond.

As observed in young HFD-fed mice (Fig. 1I-J), western blot analysis and immunofluorescence showed Myc upregulation in islets and ß-cells in one-year-old mice fed a short-term HFD (Fig. 5F-G). This indicates that Myc is upregulated by acute HFD feeding in both young and one-year- old mice. To address whether aging impairs both Myc action and increased ß-cell proliferation, the effect of mild physiologic Myc overexpression was assessed in young and one-year-old mouse ß-cells in culture (Fig. 5H). ß-cell proliferation was comparably enhanced by mild Myc overexpression in both young and one-year-old ß-cells (Fig. 5I), suggesting that aging per se is not responsible for the impairment of Myc action in one-year-old mice with acute overnutrition.

8 Page 9 of 40 Diabetes

Myc recruitment to cell cycle promoters is enhanced in islets from young, but not one- year-old, mice fed a short-term HFD. To address whether the absence of increased expression of cell cycle genes could correlate with decreased Myc binding to the promoters/enhancers of these genes, we performed ChIP on mouse islets obtained from young and one-year-old mice fed the RD or the HFD (Suppl. Fig. 6). As shown in Fig. 6A, islets from young mice fed a HFD displayed significantly higher Myc binding to the selected genes tested (Ccna2, Cdc20, Cdk1, Cdca2 and Ccnb1) than islets from RD fed mice. On the other hand, Myc binding to these genes was minimal and similar in islets from one-year-old mice fed with HFD or a RD (Fig. 6B). That is, despite a clear increase in Myc expression in islets from both young and one-year-old mice fed a short-term HFD, Myc binding to cell cycle genes is only increased in islets from young HFD mice. Conversely, age per se interferes with the ability of Myc to bind to, and transactivate, key cell cycle genes in response to a short-term HFD.

Increased DNA methylation in E-boxes in promoter/enhancer areas of cell cycle genes in one-year-old mice fed short-term HFD. HFD has been reported to induce methylation changes in gene regulatory regions in multiple tissues, including pancreatic ß-cells (45,46). Therefore, we used a targeted DNA methylation analysis (47) to investigate the DNA methylation pattern of the 21 Myc-targeting cell cycle genes identified in Figure 1 in islets from HFD-fed young and one- year-old mice. Deep methylome sequencing of each potential CpG in the 1Mbp collective genomic regions comprising the 21 genes, along with control regions, revealed that methylation is increased in islet DNA from one-year-old mice fed a HFD as compared to islet DNA from young mice fed the same diet (Fig. 6C, Suppl. Table 1). Analysis of specific genomic subregions revealed hypomethylation of promoters in islets from HFD-fed young mice but not in one-year-old HFD-fed mice. This stood in contrast to relative hypermethylation of gene bodies and intragenic, unmapped regions in one-year-old HFD-fed mice, consistent with previous studies (48,49) (Fig. 6D). Furthermore, analysis of 300bp upstream and downstream of Myc binding sites (E-boxes) in the 1Mbp region indicated clear DNA hypomethylation in islets from HFD fed young mice but not from one-year-old mice (Fig. 6E). Finally, specific analysis of the 250bp upstream and downstream of Myc ChIP peaks in promoter/enhancer regions of the five genes analyzed in the ChIP assays (Fig. 6A-B and Suppl. Fig. 6) showed profound hypomethylation in islets from young mice fed HFD but not in islets of one-year-old mice fed with HFD (Fig. 6F). Together, these results suggest that DNA methylation accrues in islets from one-year-old mice fed short-term HFD specifically at Myc binding sites in promoters/enhancers of cell cycle genes, and thereby impairs Myc binding.

5-azadeoxycytidine treatment decreases global DNA methylation, increases Myc binding to cell cycle promoters and induces ß-cell replication in old HFD-fed mouse islets in a Myc-dependent way. We next tested whether global DNA demethylation induced by 5- azadeoxycytidine (50) treatment could enhance Myc binding and induce ß-cell proliferation in isolated islets from one-week HFD-fed one-year-old mice. Indeed, treatment with 5- azadeoxycytidine mildly but significantly decreased global DNA methylation, enhanced Myc binding to cell cycle promoters and increased ß-cell proliferation (Fig. 6G-I). Importantly, simultaneous treatment with the Myc inhibitor 10058-F4 completely abolished ß-cell proliferation induced by 5-azadeoxycytidine treatment, indicating that DNA methylation in islets from short- term HFD-fed one-year-old mice impairs Myc action to activate cell cycle target genes and ß-cell replication.

9 Diabetes Page 10 of 40

Discussion

Insulin resistance in young animals is a well-known maneuver for enhancing pancreatic ß-cell replication and mass (2, 34, 37, 40, 51). Unfortunately, the intracellular signals and networks that control ß-cell compensatory growth remain incompletely understood. Using short-term HFD as a challenge that rapidly increases body weight and insulin demand, we have unraveled the intracellular mechanisms that regulate HFD-mediated adaptive ß-cell replication and expansion in mice; this information can perhaps be leveraged to increase proliferation in human ß-cells. We have found that: (i) Myc protein abundance, but not Myc gene expression, is increased in ß-cells from young mice fed acutely with a HFD; (ii) Myc upregulation is required for compensatory ß-cell proliferation and expansion in this setting; (iii) Myc upregulation is mediated by a previously unrecognized PKCζ/ERK/mTOR/PP2A pathway; (iv) Myc gene and protein expression are also rapidly upregulated in ß-cells from one-year-old mice on a short-term HFD, but this is not sufficient to activate ß-cell proliferation; (v) in contrast, mild Myc overexpression in islets from one-year-old mice on a RD increases ß-cell proliferation, indicating that ß-cells from one-year-old mice are responsive to the mitogenic action of Myc; (vi) binding of Myc to cell cycle gene promoters/enhancer regions is increased in response to a short-term HFD in islets from young, but not one-year-old, mice; and, (vii) methylation near Myc binding sites in regulatory regions of cell cycle genes is decreased in islets from young mice but not in islets from one-year-old mice fed a HFD. Collectively, these data suggest that normally, in young mice, short-term HFD might selectively induce the demethylation of Myc binding sites near cell cycle genes, and that does not occur in one-year-old mice potentially impairing the mitogenic action of Myc (Fig. 7).

Adaptive ß-cell expansion in young animals is a well-documented process that occurs in response to an increase in insulin demand (2,34,37,40,51,52). Here, as expected, one week of HFD feeding in young mice led to increased body weight, insulin resistance, glucose intolerance, increased ß- cell function and enhanced ß-cell proliferation and mass (34,37). The extracellular and intracellular effectors implicated in this adaptive ß-cell growth include SerpinB1, insulin, IRS2, PKCζ, mTOR, FoxM1, cyclin D2, PLK1 and CENP-A (34,53-57). In the current study, we sought to interrogate, in an unbiased way, the transcriptional profile of islets from mice fed a HFD for one week, with the idea of unraveling networks of genes controlling this process. Not surprisingly, gene ontology enrichment analysis showed that most of the genes upregulated in the islets of young mice fed a HFD could be ascribed to the cell cycle process. Importantly, however, an unexpectedly high percentage of these genes are Myc targets. This was a particularly surprising observation because it occurred in the absence of a change in the expression of Myc gene itself, raising two obvious questions: Might Myc protein be increased in the absence of Myc mRNA in islets from these mice? And, if so, is Myc centrally involved in the adaptive ß-cell proliferation and expansion confronted with increased insulin demand? The common answer to both questions is affirmative: we show that Myc protein is an upstream master regulator of cell cycle genes in ß- cells in the context of compensatory growth to acute overnutrition and insulin resistance.

Glucose is a mitogenic signal for the ß-cell and a well-documented stimulator of Myc expression (17,18) suggesting a potential role for Myc in glucose-mediated ß-cell proliferation. Indeed, transgenic mice with marked (~50-150x) overexpression of Myc in ß-cells display transient ß-cell

10 Page 11 of 40 Diabetes

proliferation, followed rapidly by high rates of ß-cell apoptosis and ß-cell dysfunction, and consequent diabetes (23). Several subsequent studies have associated high level, non- physiologic Myc overexpression with impaired insulin secretion, decreased insulin gene expression and ß-cell apoptosis independently of hyperglycemia (58,59). Interestingly, suppression of Myc-induced apoptosis in ß-cells favors the mitogenic properties of Myc and triggers carcinogenic progression (24). Taken together, these studies suggest that both the level of expression of Myc and/or the expression of apoptosis inhibitory pathways in ß-cells could ultimately determine ß-cell mass, insulin secretion and glucose homeostasis in physiological and pathological situations.

Recent studies from our group have indicated that modest (~2-7x) Myc upregulation increases rat and human ß-cell proliferation without induction of apoptosis, alteration of insulin secretion, or progression to malignancy (22, 25). In addition, recent studies have predicted Myc as an upstream regulator of increased expression of proliferative genes in mouse islets during pregnancy (19,21). Collectively, these studies suggest that Myc is a physiological regulator of normal ß-cell proliferation, and that mild upregulation of Myc is beneficial for ß-cell proliferation. Indeed, these findings appear to apply to the entire pancreas since pancreatic Myc inactivation in mice does not alter endocrine progenitors but decreases proliferation and alters differentiation of exocrine progenitors leading to decreased acinar mass and transdifferentiation of acinar cells into adipocytes (60). Interestingly, pancreatic contents of insulin and glucagon and islet function are similar in pancreas deficient Myc mice and control mice both at birth and 2 months of age (60). This suggests that Myc may not be required for ß-cell development, growth or function under basal conditions. However, recent studies indicate that deletion of endogenous Myc in ß-cells using Ins-Cre and Myclox/lox mice reduces postnatal ß-cell proliferation and decreases ß-cell mass in adulthood (26). On the other hand, chronic modest overexpression of Myc in ß-cells of transgenic mice increases ß-cell replication and mass, with no evidence of malignant transformation for the life span of the mice (26). Eventually, after more than a year of overexpression, ß-cells reverted towards an immature phenotype (26). Taken together, these studies suggest that Myc plays an important role in ß-cell homeostasis in physiological conditions. However, none of these studies addressed whether Myc is necessary for the adaptive expansion of ß-cells in response to increased insulin demand. The current study is the first to address this point and to show that Myc is unequivocally required for adaptive ß-cell proliferation, function and expansion in response to overnutrition in mice. This places Myc as a key transcription factor required for the ß-cell response to obesity and insulin resistance.

Having demonstrated that Myc is required for the adaptive response of ß-cells, it became important to determine how Myc expression is upregulated with increased insulin demand. Phosphorylation of Myc at Ser62 by ERK1/2 transiently increases Myc stability while phosphorylation of Thr58 by GSK3β causes dephosphorylation of Ser62 by PP2A, triggering ubiquitination and proteosomal degradation (13,15). Therefore, ERK1/2, GSK3ß and PP2A could play a role in the upregulation of Myc expression with increased high glucose and overnutrition. We have recently demonstrated that PKCζ is an upstream regulator of adaptive ß-cell expansion by controlling the activity of mTOR (34,42). Since mTOR is a known repressor of PP2A activity (16), we wondered whether PKCζ activity could control Myc levels via mTOR and PP2A. We report here that inhibition of PKCζ activity diminishes the ratio of pS62/pTh58-Myc in ß-cells

11 Diabetes Page 12 of 40

incubated with high glucose concentrations, and this effect involves ERK1/2 activation and inhibition of PP2A via mTOR activation. Therefore, conditions that increase PP2A activity in ß- cells will decrease Myc levels, and potentially impair adaptive ß-cell proliferation.

Aging decreases not only the basal proliferative response of ß-cells to mitogens but also the adaptive mitogenic response to partial pancreatectomy, streptozotocin injection, GLP-1 administration and HFD feeding (42,61). Here, we observed that Myc protein is upregulated in ß- cells of one-year old mice in which adaptive proliferation does not occur and in which the islet transcriptome signature does not reflect upregulation of cell cycle genes. This suggested impairment of the adaptive action of Myc in ß-cells in one-year-old mice when confronted with overnutrition. Is this the result of aging, or changes induced by the diet, or both together? As observed in the current study, mild Myc overexpression remarkably enhanced proliferation of ß- cells in islet cell cultures from one-year-old mice, complimenting the studies of Puri et al. showing that mild chronic Myc expression (one-year) in transgenic mice results in enhanced ß-cell proliferation (26). These findings indicate that aging alone does not provide the “brakes” on Myc mitogenic action in ß-cells. Since Myc is upregulated in islets from one-year-old mice acutely fed a HFD, but this fails to increase ß-cell proliferation, it appears that the combination of aging together with HFD underlie the failure of adaptive proliferation. In support of this hypothesis, in islets from old mice fed a HFD, the normal DNA hypomethylation of Myc binding regions in cell cycle gene promoters/enhancers appears to be lost. Whereas aging promotes increased DNA methylation in metabolically active tissues including pancreatic islets (62,63), HFD induces global DNA hypomethylation in liver and adipose tissue of young rodents when compared to rodents fed a RD (64,65). Interestingly, 72 h treatment with 5-azadeoxycytidine was able to partially rescue the binding of Myc to promoter regions of cell cycle genes and induce mild ß-cell proliferation in islets from short-term HFD-fed one-year-old mice in a Myc-dependent way. This confirms that DNA methylation of Myc target genes impair Myc action and induction of ß-cell proliferation in HFD-fed one-year-old mice. Collectively, these observations suggest that DNA hypomethylation and enhanced promoter/enhancer availability are a normal feature in tissues of young rodents fed HFD. Conversely, in aged mice, this DNA hypomethylation and “gene access” are further impaired by HFD, events that may be mediated by abnormal levels or function of DNA methylases or demethylases. Further studies to decipher the mechanisms of altered DNA methylation in aging in response to HFD are warranted.

In summary, these studies demonstrate that short-term overnutrition and increased insulin demand upregulates Myc by mechanisms involving the PKCζ-ERK1/2-mTOR-PP2A signaling pathway; that Myc upregulation is required for adaptive ß-cell proliferation in situations of overnutrition and increased insulin demand; and that epigenetic alterations can potentially impair Myc-induced compensatory ß-cell response to overnutrition in older mice. Modification of epigenetic alterations in islets in type 2 diabetes may be of therapeutic value for ß-cell expansion in diabetes.

12 Page 13 of 40 Diabetes

Acknowledgements

The authors thank Martin Walsh, Rupangi C. Vasavada and David Dominguez-Sola (Icahn School of Medicine at Mount Sinai), Benjamin Hubert (New York Genome Center) and Rosalie Sears (Oregon Health & Science University) for helpful comments during the development of these studies. We thank Juan Carlos Alvarez-Perez, Lucy Li and Gabriel Brill (Icahn School of Medicine at Mount Sinai) for technical help. This work was supported in part by grants from the National Institutes of Health-National Institute of Diabetes and Digestive and Kidney Diseases and the Human Islet Research Network (HIRN) (DK-113079, DK-105015, DK-077096, DK-110156, DK- 108905, DK-104211, DK-116873) and the National Cancer Institute (CA-174713), the American Diabetes Association (1-17-IBS-116), the Juvenile Diabetes Research Foundation (1-INO-2016- 212-A-N, and 2-SRA-2015-62), and the Mindich Child Health and Development Institute Pilot and Feasibility Grant. We would also like to thank the New-York Genome Center for RNAseq performance and analysis, the Epigenomics Core Facility of Weill Cornell Medicine for DNA methylation profiling and the Human Islet and Adenovirus Core (HIAC) of the Einstein-Sinai Diabetes Research Center (DK-020541) for generation of adenoviruses.

Author Contributions

C.R., A.K., J.L., P.Z., L.S.K., G.L., L.L., researched data, contributed to discussion, and reviewed and edited the manuscript. A.F.S., E.V.P. contributed to discussion, and reviewed and edited the manuscript. D.K.S. and A.G-O designed the study, contributed to discussion and wrote the manuscript. D.K.S. and A.G.-O. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Data and Resource Availability

The datasets generated during and/or analyzed during the current study are available from the corresponding authors upon reasonable request.

Adenoviruses generated during and/or analyzed during the current study are available from the corresponding authors upon reasonable request. The rest of the resources used during these studies are commercially available and their Research Resource Identifiers (RRIDs) are included in Supplemental Table 1.

13 Diabetes Page 14 of 40

References 1. Aguayo-Mazzucato C, and Bonner-Weir S. Pancreatic ß-cell Regeneration as a Possible Therapy for Diabetes. Cell Metab. 2018;27(1):57-67. 2. Dadon D, Tornovsky-Babaey S, Furth-Lavi J, Ben-Zvi D, Ziv O, Schyr-Ben-Haroush R, Stolovich-Rain M, Hija A, Porat S, Granot Z, Weinberg-Corem N, Dor Y, Glaser B. Glucose metabolism: key endogenous regulator of beta-cell replication and survival. Diabetes, Obesity & Metabolism. 2012;14 Suppl 3(101-8). 3. Dor Y, Brown J, Martinez OI, and Melton DA. Adult pancreatic beta-cells are formed by self-duplication rather than stem-cell differentiation. Nature. 2004;429(6987):41-46. 4. Wang P, Fiaschi-Taesch NM, Vasavada RC, Scott DK, Garcia-Ocana A, and Stewart AF. Diabetes mellitus-advances and challenges in human beta-cell proliferation. Nature Reviews Endocrinology. 2015;11(4):201-212. 5. Bernal-Mizrachi E, Kulkarni RN, Scott DK, Mauvais-Jarvis F, Stewart AF, and Garcia- Ocana A. Human beta-cell proliferation and intracellular signaling part 2: still driving in the dark without a road map. Diabetes. 2014;63(3):819-831. 6. Halban PA, Polonsky KS, Bowden DW, Hawkins MA, Ling C, Mather KJ, Powers AC, Rhodes CJ, Sussel L, and Weir GC. beta-cell failure in type 2 diabetes: postulated mechanisms and prospects for prevention and treatment. The Journal of Clinical Endocrinology and Metabolism. 2014;99(6):1983-92. 7. Talchai C, Xuan S, Lin HV, Sussel L, and Accili D. Pancreatic ß-cell dedifferentiation as a mechanism of diabetic ß-cell failure. Cell. 2012;150(6):1223-1234. 8. Cole MD. The myc oncogene: its role in transformation and differentiation. Annual Review of Genetics. 1986;20:361-384. 9. Dang CV. MYC on the path to cancer. Cell. 2012;149(1):22-35. 10. Hofmann JW, Zhao X, De Cecco M, Peterson AL, Pagliaroli L, Manivannan J, Hubbard GB, Ikeno Y, Zhang Y, Feng B, Li X, Serre T, Qi W, Van Remmen H, Miller RA, Bath KG, de Cabo R, Xu H, Neretti N, Sedivy JM. Reduced expression of MYC increases longevity and enhances healthspan. Cell. 2015;160(3):477-488. 11. Nesbit CE, Tersak JM, and Prochownik EV. MYC oncogenes and human neoplastic disease. Oncogene. 1999;18(19):3004-3016. 12. Evan GI, Wyllie AH, Gilbert CS, Littlewood TD, Land H, Brooks M, Waters CM, Penn LZ, and Hancock DC. Induction of apoptosis in fibroblasts by c-myc protein. Cell. 1992;69(1):119-128. 13. Farrell AS, and Sears RC. MYC degradation. Cold Spring Harbor Perspectives in

14 Page 15 of 40 Diabetes

19. Chen G, Liu C, Xue Y, Mao X, Xu K, and Liu C. Molecular mechanism of pancreatic beta- cell adaptive proliferation: studies during pregnancy in rats and in vitro. Endocrine. 2011;39(2):118-27. 20. Demirci C, Ernst S, Alvarez-Perez JC, Rosa T, Valle S, Shridhar V, Casinelli GP, Alonso LC, Vasavada RC, and Garcia-Ocana A. Loss of HGF/c-Met signaling in pancreatic beta- cells leads to incomplete maternal beta-cell adaptation and gestational diabetes mellitus. Diabetes. 2012;61(5):1143-52. 21. Horn S, Kirkegaard JS, Hoelper S, Seymour PA, Rescan C, Nielsen JH, Madsen OD, Jensen JN, Kruger M, Gronborg M, Ahnfelt-Rønne J. Research Resource: A Dual Proteomic Approach Identifies Regulated Islet Proteins During beta-Cell Mass Expansion In Vivo. Mol Endocrinol. 2016;30(1):133-143. 22. Karslioglu E, Kleinberger JW, Salim FG, Cox AE, Takane KK, Scott DK, and Stewart AF. cMyc is a principal upstream driver of beta-cell proliferation in rat insulinoma cell lines and is an effective mediator of human beta-cell replication. Mol Endocrinol. 2011;25(10):1760- 1772. 23. Laybutt DR, Weir GC, Kaneto H, Lebet J, Palmiter RD, Sharma A, and Bonner-Weir S. Overexpression of c-Myc in beta-cells of transgenic mice causes proliferation and apoptosis, downregulation of insulin gene expression, and diabetes. Diabetes. 2002;51(6):1793-1804. 24. Pelengaris S, Khan M, and Evan GI. Suppression of Myc-induced apoptosis in ß-cells exposes multiple oncogenic properties of Myc and triggers carcinogenic progression. Cell. 2002;109(3):321-34. 25. Wang P, Alvarez-Perez JC, Felsenfeld DP, Liu H, Sivendran S, Bender A, Kumar A, Sanchez R, Scott DK, Garcia-Ocana A, Stewart AF. A high-throughput chemical screen reveals that harmine-mediated inhibition of DYRK1A increases human pancreatic ß-cell replication. Nature Medicine. 2015;21(4):383-8. 26. Puri S, Roy N, Russ HA, Leonhardt L, French EK, Roy R, Bengtsson H, Scott DK, Stewart AF, and Hebrok M. Replication confers ß-cell immaturity. Nature Communications. 2018;9(1):485. 27. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, and Gingeras TR. STAR: ultrafast universal RNA-seq aligner. Bioinformatics (Oxford, England). 2013;29(1):15-21. 28. Bray NL, Pimentel H, Melsted P, and Pachter L. Near-optimal probabilistic RNA-seq quantification. Nature Biotechnology. 2016;34(5):525-527. 29. Wang L, Wang S, and Li W. RSeQC: quality control of RNA-seq experiments. Bioinformatics (Oxford, England). 2012;28(16):2184-2185. 30. Love MI, Huber W, and Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology. 2014;15(12):550. 31. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545-15550. 32. Wicksteed B, Brissova M, Yan W, Opland DM, Plank JL, Reinert RB, Dickson LM, Tamarina NA, Philipson LH, Shostak A, Bernal-Mizrachi E, Elghazi L, Roe MW, Labosky PA, Myers MG Jr, Gannon M, Powers AC, Dempsey PJ. Conditional gene targeting in mouse pancreatic ss-Cells: analysis of ectopic Cre transgene expression in the brain. Diabetes. 2010;59(12):3090-3098. 33. de Alboran IM, O'Hagan RC, Gartner F, Malynn B, Davidson L, Rickert R, Rajewsky K, DePinho RA, and Alt FW. Analysis of C-MYC function in normal cells via conditional gene- targeted mutation. Immunity. 2001;14(1):45-55.

15 Diabetes Page 16 of 40

34. Lakshmipathi J, Alvarez-Perez JC, Rosselot C, Casinelli GP, Stamateris RE, Rausell- Palamos F, O'Donnell CP, Vasavada RC, Scott DK, Alonso LC, Garcia-Ocana A. PKC- zeta is essential for pancreatic ß-cell replication during insulin resistance by regulating mTOR and cyclin-D2. Diabetes. 2016; 65(5):1283-1296. 35. Mellado-Gil J, Rosa TC, Demirci C, Gonzalez-Pertusa JA, Velazquez-Garcia S, Ernst S, Valle S, Vasavada RC, Stewart AF, Alonso LC, Garcia-Ocana A. Disruption of hepatocyte growth factor/c-Met signaling enhances pancreatic beta-cell death and accelerates the onset of diabetes. Diabetes. 2011;60(2):525-536. 36. Pedersen KB, Zhang P, Doumen C, Charbonnet M, Lu D, Newgard CB, Haycock JW, Lange AJ, and Scott DK. The promoter for the gene encoding the catalytic subunit of rat glucose-6-phosphatase contains two distinct glucose-responsive regions. American Journal of Physiology Endocrinology and Metabolism. 2007;292(3):E788-801. 37. Stamateris RE, Sharma RB, Hollern DA, and Alonso LC. Adaptive beta-cell proliferation increases early in high-fat feeding in mice, concurrent with metabolic changes, with induction of islet cyclin D2 expression. American Journal of Physiology Endocrinology and Metabolism. 2013;305(1):E149-159. 38. Dang CV. c-Myc target genes involved in cell growth, apoptosis, and metabolism. Mol Cell Biol. 1999;19(1):1-11. 39. Metukuri MR, Zhang P, Basantani MK, Chin C, Stamateris RE, Alonso LC, Takane KK, Gramignoli R, Strom SC, O'Doherty RM, et al. ChREBP mediates glucose-stimulated pancreatic beta-cell proliferation. Diabetes. 2012;61(8):2004-2015. 40. Alonso LC, Yokoe T, Zhang P, Scott DK, Kim SK, O'Donnell CP, and Garcia-Ocana A. Glucose infusion in mice: a new model to induce beta-cell replication. Diabetes. 2007;56(7):1792-1801. 41. Oropeza D, Jouvet N, Budry L, Campbell JE, Bouyakdan K, Lacombe J, Perron G, Bergeron V, Neuman JC, Brar HK, Fenske RJ, Meunier C, Sczelecki S, Kimple ME, Drucker DJ, Screaton RA, Poitout V, Ferron M, Alquier T, Estall JL. Phenotypic Characterization of MIP-CreERT1Lphi Mice With Transgene-Driven Islet Expression of Human Growth Hormone. Diabetes. 2015; 64(11):3798-3807. 42. Velazquez-Garcia S, Valle S, Rosa TC, Takane KK, Demirci C, Alvarez-Perez JC, Mellado-Gil JM, Ernst S, Scott DK, Vasavada RC, Alonso LC, Garcia-Ocaña A. Activation of protein kinase C-zeta in pancreatic beta-cells in vivo improves glucose tolerance and induces beta-cell expansion via mTOR activation. Diabetes. 2011;60(10):2546-2559. 43. Monick MM, Carter AB, Flaherty DM, Peterson MW, and Hunninghake GW. Protein kinase C zeta plays a central role in activation of the p42/44 mitogen-activated protein kinase by endotoxin in alveolar macrophages. Journal of Immunology 2000;165(8):4632-4639. 44. Tschen SI, Dhawan S, Gurlo T, and Bhushan A. Age-dependent decline in beta-cell proliferation restricts the capacity of beta-cell regeneration in mice. Diabetes. 2009;58(6):1312-1320. 45. Hu Y, Shi P, He K, Zhu YQ, Yang F, Yang M, He BS, and Mao XM. Methylation of Tcf712 promoter by high-fat diet impairs beta-cell function in mouse pancreatic islets. Diabetes Metabolism Research and Reviews. 2018;34(4):e2980. 46. Zhang P, Chu T, Dedousis N, Mantell BS, Sipula I, Li L, Bunce KD, Shaw PA, Katz LS, Zhu J, Argmann C, O'Doherty RM, Peters DG, Scott DK. DNA methylation alters transcriptional rates of differentially expressed genes and contributes to pathophysiology in mice fed a high fat diet. Molecular Metabolism. 2017;6(4):327-339. 47. Wang H, Bender A, Wang P, Karakose E, Inabnet WB, Libutti SK, Arnold A, Lambertini L, Stang M, Chen H, Kasai Y, Mahajan M, Kinoshita Y, Fernandez-Ranvier G, Becker TC, Takane KK, Walker LA, Saul S, Chen R, Scott DK, Ferrer J, Antipin Y, Donovan M, Uzilov AV, Reva B, Schadt EE, Losic B, Argmann C, Stewart AF. Insights into ß-cell regeneration

16 Page 17 of 40 Diabetes

for diabetes via integration of molecular landscapes in human insulinomas. Nature Communications. 2017;8(1):767. 48. Avrahami D, Li C, Zhang J, Schug J, Avrahami R, Rao S, Stadler MB, Burger L, Schubeler D, Glaser B, Kaestner KH. Aging-Dependent Demethylation of Regulatory Elements Correlates with Chromatin State and Improved ß-cell Function. Cell Metabolism. 2015;22(4):619-632. 49. Dhawan S, Tschen SI, Zeng C, Guo T, Hebrok M, Matveyenko A, and Bhushan A. DNA methylation directs functional maturation of pancreatic ß-cells. The Journal of Clinical Investigation. 2015;125(7):2851-2860. 50. Christman JK. 5-Azacytidine and 5-aza-2'-deoxycytidine as inhibitors of DNA methylation: mechanistic studies and their implications for cancer therapy. Oncogene. 2002;21(35):5483-5495. 51. Porat S, Weinberg-Corem N, Tornovsky-Babaey S, Schyr-Ben-Haroush R, Hija A, Stolovich-Rain M, Dadon D, Granot Z, Ben-Hur V, White P, Girard CA, Karni R, Kaestner KH, Ashcroft FM, Magnuson MA, Saada A, Grimsby J, Glaser B, Dor Y. Control of pancreatic ß-cell regeneration by glucose metabolism. Cell Metabolism. 2011;13(4):440- 449. 52. Tschen SI, Georgia S, Dhawan S, and Bhushan A. Skp2 is required for incretin hormone- mediated beta-cell proliferation. Mol Endocrinol. 2011;25(12):2134-2143. 53. El Ouaamari A, Dirice E, Gedeon N, Hu J, Zhou JY, Shirakawa J, Hou L, Goodman J, Karampelias C, Qiang G, Boucher J, Martinez R, Gritsenko MA, De Jesus DF, Kahraman S, Bhatt S, Smith RD, Beer HD, Jungtrakoon P, Gong Y, Goldfine AB, Liew CW, Doria A, Andersson O, Qian WJ, Remold-O'Donnell E, Kulkarni RN. SerpinB1 Promotes Pancreatic ß-cell Proliferation. Cell Metabolism. 2016;23(1):194-205. 54. Georgia S, Hinault C, Kawamori D, Hu J, Meyer J, Kanji M, Bhushan A, and Kulkarni RN. Cyclin D2 is essential for the compensatory beta-cell hyperplastic response to insulin resistance in rodents. Diabetes. 2010;59(4):987-996. 55. Rachdi L, Balcazar N, Osorio-Duque F, Elghazi L, Weiss A, Gould A, Chang-Chen KJ, Gambello MJ, and Bernal-Mizrachi E. Disruption of Tsc2 in pancreatic ß-cells induces ß- cell mass expansion and improved glucose tolerance in a TORC1-dependent manner. Proc Natl Acad Sci U S A. 2008;105(27):9250-9255. 56. Shirakawa J, Fernandez M, Takatani T, El Ouaamari A, Jungtrakoon P, Okawa ER, Zhang W, Yi P, Doria A, and Kulkarni RN. Insulin Signaling Regulates the FoxM1/PLK1/CENP-A Pathway to Promote Adaptive Pancreatic ß-cell Proliferation. Cell Metabolism. 2017;25(4):868-82.e5. 57. Zarrouki B, Benterki I, Fontes G, Peyot ML, Seda O, Prentki M, and Poitout V. Epidermal growth factor receptor signaling promotes pancreatic beta-cell proliferation in response to nutrient excess in rats through mTOR and FOXM1. Diabetes. 2014;63(3):982-993. 58. Cheung L, Zervou S, Mattsson G, Abouna S, Zhou L, Ifandi V, Pelengaris S, and Khan M. c-Myc directly induces both impaired insulin secretion and loss of beta-cell mass, independently of hyperglycemia in vivo. Islets. 2010;2(1):37-45. 59. Kaneto H, Sharma A, Suzuma K, Laybutt DR, Xu G, Bonner-Weir S, and Weir GC. Induction of c-Myc expression suppresses insulin gene transcription by inhibiting NeuroD/BETA2-mediated transcriptional activation. J Biol Chem. 2002;277(15):12998- 13006. 60. Bonal C, Thorel F, Ait-Lounis A, Reith W, Trumpp A, and Herrera PL. Pancreatic inactivation of c-Myc decreases acinar mass and transdifferentiates acinar cells into adipocytes in mice. Gastroenterology. 2009;136(1):309-19.e9.38. 61. Rankin MM, and Kushner JA. Adaptive beta-cell proliferation is severely restricted with advanced age. Diabetes. 2009;58(6):1365-1372.

17 Diabetes Page 18 of 40

62. Davegardh C, Garcia-Calzon S, Bacos K, and Ling C. DNA methylation in the pathogenesis of type 2 diabetes in humans. Molecular Metabolism. 2018. 63. Bacos K, Gillberg L, Volkov P, Olsson AH, Hansen T, Pedersen O, Gjesing AP, Eiberg H, Tuomi T, Almgren P, Groop L, Eliasson L, Vaag A, Dayeh T, Ling C. Blood-based biomarkers of age-associated epigenetic changes in human islets associate with insulin secretion and diabetes. Nature Communications. 2016;7:11089. 64. Cheng J, Song J, He X, Zhang M, Hu S, Zhang S, Yu Q, Yang P, Xiong F, Wang DW, Zhou J, Ning Q, Chen Z, Eizirik DL, Zhou Z, Zhao C, Wang CY. Loss of Mbd2 Protects Mice Against High-Fat Diet-Induced Obesity and Insulin Resistance by Regulating the Homeostasis of Energy Storage and Expenditure. Diabetes. 2016;65(11):3384-3395. 65. Zhang Y, Wang H, Zhou D, Moody L, Lezmi S, Chen H, and Pan YX. High-fat diet caused widespread epigenomic differences on hepatic methylome in rat. Physiological Genomics. 2015;47(10):514-523.

18 Page 19 of 40 Diabetes

Figure Legends

Figure 1. Adaptive proliferation is accompanied by increases in Myc and downstream effectors. Eight-week old male C57Bl/6N mice were fed a regular (RD) or high fat diet (HFD) for 1 week. (A) Pancreata from RD or HFD mice were stained for insulin and Ki67, which was quantified in (B), along with body weight, blood glucose and plasma insulin levels. Data are means ± SEM, n=8 experiments, *, P<0.05. RNAseq analysis of islets isolated from the 2 groups of mice: (C) Volcano plot with genes that are significantly (P<0.05) changed (>1.5- and <0.67-fold) in islets from HFD fed mice vs. RD fed mice marked in red. (D) Gene ontology (GO) analysis shows mainly significant changes in processes involved in cell replication being the two most significant changes in cell cycle and cell division. P values are indicated at the end of the bar. (E) Heat map of top differentially expressed genes (>1.5-fold; P<0.05) between islets from HFD-fed and RD-fed mice of log2 normalized RNAseq count data. (F) Hallmark Myc targets that were significantly enriched in HFD-fed mouse islets identified using gene set enrichment analysis (GSEA). (G) GSEA analysis of differentially expressed genes shows enrichment of genes involved in cell cycle in islets from HFD-fed mice. (H) Quantitative PCR of representative cell cycle genes differently expressed in islets from HFD-fed mice. Data are means ± SEM, n=3 mice, *, P<0.05. (I) Immunoblot analysis of protein extracts prepared from islets from HFD- and RD-fed mice. Blots were probed with antibodies to Myc and GAPDH as the loading control and quantitation is shown as means ± SEM, n=4 mice, *, P<0.05. (J) Immunofluorescent staining of pancreata from RD and HFD mice using antibodies against insulin (red), DAPI (blue) and Myc (green).

Figure 2. Myc is required to preserve glucose tolerance and adaptive ß-cell proliferation and expansion after a short-term HFD. (A) Scheme showing the experimental design of ß-cell- specific conditional deletion of Myc. (B) PCR showing tamoxifen-specific excision of exons 2 and 3 of the Myc gene. (C) Immunoblot analysis of protein extracts prepared from control and βMycKO mouse islets. Blots were probed with antibodies to Myc and GAPDH as the loading control. (D) Pancreata from βMycKO mice stained with insulin and Myc demonstrating conditional depletion of Myc. (E,F) Ad-lib fed blood glucose and plasma insulin measured after RD or HFD in the presence or absence of Myc. (G, H) Intraperitoneal glucose tolerance test and area under the curve (AUC), respectively. (I) Pancreata from the indicated treatment groups stained for insulin and Ki67. (J) Quantification of ß-cell specific Ki67 staining. (K) Quantification of ß-cell mass. Data are means ± SEM from 6-8 mice/group. *, P<0.05; **, P<0.01

Figure 3. PKC ζ is required for glucose- and HFD-induced upregulation of Myc. (A) Immunoblots of extracts from INS1 832/13 cells transduced with Ad.GFP or Ad.KD-PKCζ following indicated treatments for 24h and quantification (lower panel). (B) Immunoblots from isolated islets of wild type (WT) and KD-PKCζ transgenic mice (TG) on a RD or HFD for one week, and quantification (lower panel). (C) Pancreata from WT or KD-PKCζ TG mice on a RD or HFD for one week and stained for Myc and insulin. (D) Mouse islet cells isolated and dispersed were transduced with the indicated adenovirus and cultured in different glucose concentrations. Twenty-four hours later extracts were processed for immunoblots against Myc and tubulin for loading control and the quantitation of the Myc to tubulin ratios is shown. (E) Isolated mouse islet cells treated as in D were stained with insulin and Ki67, and cell cycle entry was quantified in (F). (G) Immunoblots of extracts from isolated mouse islet cells transduced with Ad.GFP or Ad.CA- PKCζ and treated as indicated using antibodies against Myc and actin. The ratio of Myc to actin

19 Diabetes Page 20 of 40

is quantified in the right panel. (H) Isolated mouse islet cells treated with indicated adenovirus or with 40 μM 10054-F4 were stained with insulin and Ki67 and the percentage of insulin and Ki67 positive cells was determined. The results represent the mean ± SEM of n= 3-6 experiments or mice, *, P<0.05, 2 vs 11 or 20mM glucose or HFD vs. RD. #, P<0.05, 20mM GFP vs 20mM KDPKCζ. &, P<0.05, 11mM LacZ vs 11mM CA-PKCζ. ^, P<0.05, 11mM 10054-F4 or 11mM 10054-F4 +CA-PKCζ vs 11mM LacZ or 11mM CA-PKCζ.

Figure 4. Myc phosphorylation is regulated by PP2A. Immunoblot of extracts from INS1 832/13 cells treated with the indicated adenoviruses and glucose concentrations using antibodies against (A) Myc pSer62 and (B) Myc pT58 and the ratio calculated in (C). (D) Immunoblot of the same extracts using antibodies against pERK1/2, pGSK3β, and GAPDH or tubulin as loading controls. Quantification of the results is shown on the right panels. (E) PP2A activity was determined in INS1 832/13 cells treated for 24h with 2 or 20 mM glucose after transduction with the indicated adenovirus or after treatment with 10nM rapamycin (RAP). (F) Immunoblots of extracts of INS1 832/13 cells after treatment with the indicated adenovirus, glucose concentrations or 5nM endothall (Endo) using antibodies against Myc, Myc pSer62 and GAPDH as a loading control. Quantification is shown on the panels below. (G) Mouse islets were dispersed and treated with 2 mM glucose or 20 mM glucose and Ad.LacZ, Ad.KD-PKCζ, Ad.LacZ and 5nM endothall or Ad.KD-PKCζ. , fixed, stained for insulin, DAPI and Ki67 and ß-cell proliferation was quantitated. Data shown are the mean ± SEM of n=3-6 experiments. *, P<0.05 2 vs 20mM glucose. #, P<0.05 20mM GFP vs 20mM KD-PKCζ or rapamycin. ^, P<0.05 20mM KD-PKCζ vs 20mM KD-PKCζ + endothall.

Figure 5. Upregulation of Myc in ß-cells despite lack of ß-cell proliferation in one year-old HFD-fed mice. (A) One-year-old C57Bl/6N male mice were placed on a regular diet (RD) or a high fat diet (HFD) for one week. Pancreata were collected and proliferation of ß-cells was quantified, as determined by Ki67 and insulin staining. Body weight, blood glucose and plasma insulin levels were also quantified. RNAseq analysis was performed from islets isolated from one- year-old male mice treated for one week with RD or HFD: (B) Volcano plot with genes that are significantly (P<0.05) changed (>1.5- and <0.67-fold) in islets from HFD fed mice vs. RD fed mice marked in red. (C) Gene ontology (GO) analysis shows mainly significant changes in protein processing or negative regulation of biological processes and no cell proliferation processes. P values are indicated at the end of the bars. (D) Heat map of top differentially expressed genes (P<0.05) between HFD-fed and RD-fed islets of log2 normalized RNA-seq count data showing 222 genes increased by 1.5 fold or higher after a HFD (P<0.05). (E) Quantitative PCR of representative cell cycle genes differently expressed in islets from young HFD-fed mice (see Fig. 1). (F) Immunoblot from islets of one-year-old mice after RD or HFD. Quantification shown in the lower panel. (G) Immunofluorescence of insulin (red), DAPI (blue) and Myc (green) in pancreata isolated for one year-old mice fed a RD or HFD for one week. (H) Mouse islet cells isolated and dispersed were transduced with the indicated adenovirus. Twenty-four hours later extracts were processed for immunoblots against Myc and GAPDH for loading control and the quantitation of the Myc to GAPDH ratios is shown. (I) Proliferation of ß-cells as determined by Ki67 and insulin staining of islet cell cultures from one-year-old mice transduced with Ad.GFP or Ad.Myc. Results are the means ± SEM. *, P<0.05, n=3-4 experiments or mice per group.

20 Page 21 of 40 Diabetes

Figure 6. Myc binding to cell cycle gene promoters is impaired in islets from one-year-old mice fed a HFD. (A) ChIP assay in mouse islets from 8-week old mice fed HFD or RD for one week. All the cell cycle genes tested showed significant increase in Myc binding in islets from HFD-fed mice. Binding to the coding regions (Cod) was not detected (ND). (B) ChIP assay in mouse islets from one-year-old mice fed HFD or RD for one week. All the cell cycle genes tested did not show a significant increase in Myc binding in islets from HFD-fed mice and binding in some of them was not detected (ND). (C) Differential methylation across all CpG dinucleotides of the target region spanning a total of about 1 Mbp. DNA methylation for each CpG dinucleotide was averaged in each group and then subtracted as follows: eight-week-old mice fed HFD (Y1HFD)– eight-week-old mice fed RD (Y1RD) (blue); one-year-old mice fed HFD (O1HFD) – one-year-old mice fed RD (O1RD) (yellow). Only those CpG dinucleotides with >2-fold methylation change for each comparison were used for the calculation. (D) Differential methylation in promoters, gene bodies and not otherwise mapped regions. CpG dinucleotides were mapped to promoters by considering, for each isoform of each gene, a region extending from 500 bp upstream the transcription starting site (TSS), to 1,500 bp downstream the TSS. The promoter methylation was then calculated by averaging those CpGs >2-fold methylation change when compared across groups as for panel C. For gene bodies, the whole length of each isoform of each gene was considered irrespective of the intron/exon composition. The gene methylation was then calculated by averaging those CpGs >2-fold methylation change when compared across groups as for panel C. For the not mapped regions we used the same approach as for panel C limited to those CpG that did not map in promoters or gene bodies. (E) Differential methylation in E-Boxes. CpG dinucleotides were mapped to the E-Boxes (±300 bp) in the 1Mbp region. Differential methylation was calculated as for panels C and D. Enhanced hypomethylation was observed with HFD in islets from young but not old mice. (F) Differential methylation in E-Boxes ChIP peaks in the promoters of the genes Ccna2, Ccnb1, Cdc20, Cdca2, Cdk1 studied in the ChIP analysis in A-B. CpG dinucleotides were mapped to the E-Boxes (±250 bp) within promoters. Differential methylation was calculated as for panels C and D. Enhanced hypomethylation was observed with HFD in islets from young but not old mice. Four mice per age and feeding group were used for these studies. Data is mean ± SEM of the five genes. *P<0.05 vs O1RD vs O1HFD. (G) 5- methylcytosine relative to total DNA (5-mC %) from islets obtained from one-week HFD-fed one- year old mice and treated with vehicle or 30 μM 5-azadeoxycytidine for 72h. Each point represents the value in islets from an individual mouse. (H) ChIP assay in mouse islets from one-year-old mice fed HFD or RD for one week and (I) proliferation of ß-cells as determined by Ki67 and insulin staining of islet cell cultures from eight-week-old or one-year-old mice fed RD or HFD for one week. Islets/Islet cells were cultured for 72 h in media containing 11 mM glucose with or without 30 μM 5-azadeoxycytidine with or wihtout 40 μM 10054-F4. Results are the means ± SEM of n=3- 6 experiments or mice per group, *, p<0.05 HFD vs RD.

Figure 7. Schematic representation of Myc function in adaptive ß-cell replication in young and aged mouse ß-cells.

21 Diabetes FigurePage 1 22 of 40

A Insulin-Ki67-DAPI B * * * *

REGULAR HFD DIET C D E

RD HFD 1 2 3 1 2 3

Myc c-MycTargets Targets

H * * * F

Cell Cycle

1.37 <0.000001 G I Islets RD HFD

Myc

GAPDH

1.78 <0.000001 J Myc-Insulin-DAPI *

REGULAR DIET HFD RD HFD Page 23 of 40 Diabetes Figure 2

A B C βMycKO

Myc

GAPDH

Tam : - - + +

Veh : + + - - D E Myc-Insulin-DAPI * F * * Vehicle Blood glucose (mg/dl) Plasma insulin (ng/ml) Tam

RD HFD RD HFD RD HFD RD HFD REGULAR DIET HFD βMycKO βMycKO βMycKO βMycKO Veh Tam Veh Tam

βMycKO-Veh-RD ** G 700 * * βMycKO-Tam-RD H ** * βMycKO-Veh-HFD 600 * * βMycKO-Tam-HFD 500

400 * AUC 300 *

200

100 % initial% of blood glucose RD HFD RD HFD 0 0 15 30 45 60 75 90 105 120 βMycKO βMycKO Veh Tam Time (min) I J K * * * * Beta Cell Mass (mg) Mass Cell Beta %(Ki67+insulin+)/insulin+ RD HFD RD HFD RD HFD RD HFD βMycKO βMycKO βMycKO βMycKO Veh Tam Veh Tam Diabetes FigurePage 243 of 40

2mM 20mM A Glucose Glucose C Myc-Insulin-DAPI B RD HFD KD KD Ad. GFP PKCζ GFP PKCζ WT TG WT TG Myc

Tubulin WTMICE

* *

# TGMICE ζ # PKC KD - Myc/Tubulin Myc/Tubulin REGULAR DIET HFD

GFP KD-PKCζ GFP KD-PKCζ WT TG WT TG * * * 2mM 20mM RD HFD 20mM 20mM 2mM Glucose Glucose Glucose D KD Myc + Ad. GFP GFP PKCζ Myc KD-PKCζ #

Myc Myc/Tubulin Tubulin E DAPI

2mM 20mM 20mM 20mM 20mM Ad.GFP + + - - - 2 mM G 20 mM G 20 mM G 20 mM G 20 mM G + + Insulin/ Ki67/ Ad.Myc - - - LacZ LacZ KD-PKCζ LacZ KD-PKCζ - + + Myc + Myc Ad.KD-PKCζ - - +

6 ζ F # G ζ H *,# *, 2.5 PKC PKC - - &,* GFP CA GFP CA 2 Myc * 4 1.5 Actin 1 ^ * * ^ 2 # 0.5

% Ki67+ Ins+/Ins+ Ins+/Ins+ Ki67+ Cells % 0 % Ki67+ Ins+/Ins+ Ins+/Ins+ Ki67+ Cells % 2mM 11mM Ad.LacZ + + - + - 0 Myc/Tubulin 10058-F4 - - - + + 2mM 20mM Ad.CA-PKCζ - - + - + Ad.LacZ + + - + - Ad.Myc - - - + + Ad.KD-PKCζ - - + - + GFP CA-PKCζ Page 25 of 40 Diabetes 2mM 20mM 2mM 20mM Figure 4 A Glucose Glucose B Glucose Glucose KD KD KD KD Ad. GFP PKCζ GFP PKCζ Ad. GFP PKCζ GFP PKCζ

pS62-Myc pT58-Myc

GAPDH GAPDH * C *

# -Myc/GAPDH

-Myc/GAPDH # pS62 pT58 Myc Ratio - Myc pSer62/T58

GFP KD-PKCζ GFP KD-PKCζ GFP KD-PKCζ GFP KD-PKCζ GFP KD-PKCζ GFP KD-PKCζ 2mM G 20mM G 2mM G 20mM G 2mM G 20mM G 2mM 20mM * D Glucose Glucose * * KD KD Ad. GFP PKCζ GFP PKCζ pERK1/2 #

GAPDH pERK1/2/GAPDH pGSK3ß/Tubulin pGSK3ß

Tubulin GFP KD-PKCζ GFP KD-PKCζ GFP KD-PKCζ GFP KD-PKCζ 2mM G 20mM G 2mM G 20mM G 2.0 # 20mM E 20mM 20mM +KD F 2 mM 20 mM +KD +GFP PKCζ 1.5 +GFP +GFP PKCζ +Endo +Endo #, * * Myc

1.0 GAPDH

0.50 pS62-Myc

0 GAPDH PP2A activity ( - fold of activity 2mMPP2A GFP) GFP KD-PKCζ GFP KD-PKCζ RAP 4.0 2mM 20mM * *,^ # G 3.0 *

2.0 2.5 *,^ *,^ * 1.0 2 Myc/GAPDH 0 GFP GFP KD GFP+ Endo+ 1.5 2mM PKCζ Endo KD # PKCζ 20mM 1 15 * *,^ * 0.5 10 % Ki67+ Ins+/Ins+ Cells Ins+/Ins+ Ki67+ % Myc/GAPDH

- # 0 5 2mM 20mM pS62 LacZ + + - + - Endothall - - - + + 0 GFP GFP KD GFP+ Endo+ KD-PKCζ - - + - + 2mM PKCζ Endo KD PKCζ 20mM Diabetes FigurePage 26 of 405 A D RD HFD 1 2 3 4 1 2 3 4 Selp Zbtb16 * Inhbb * * Adamts4 Gpr6 Sele Rnd1 Hcar2 Lrrtm2 Tnfrsf10b Herpud1 Rnf39 Stc1 Nfil3 C2cd4a Plaur Arrdc3 Rassf10 Mx2 Tubb6 Snai2 Fzd5 Kcnj2 Hspa5 8430408G22Rik Ubc Rsad2 Ppp1r3c Klf6 Msl3l2 Fam46a 1810009J06Rik Btg2 Lrrc32 Krt8 Usp2 Arid5a Phlda1 Adora2b Hbegf Zbtb10 Trib1 Cdkn1a Dusp4 Ticam1 Flrt3 Dnajb4 4930426L09Rik Dusp8 Ppp1r15a Krt18 Chac1 Klf10 Gem Ifrd1 Ifit1 Arl4a Osgin1 Gadd45g Hspa1l B E Tuba1c Nfkbiz - fold) * Myc Mfi2 Hspa1b Alb Ier2 Irs2 Elf3 Maff 1810055G02Rik Hspa1a Arid5b Hic2 Rasd1 Irf2bpl Epha2 mt-Tw Bbc3 Jun Zc3h12c Ripk4 Ddit3 Stbd1 Bag3 Dusp6 Slc2a6 Errfi1 Plekhf2 Tiparp Prrg4 Kcnk5 Dusp5

Gene/Actin Gene Exp. ( Exp. Gene Gene/Actin Txnip Manf Sox7 RD HFD RD HFD RD HFD RD HFD C2cd4b Trib3 1810011O10Rik Rasgef1b Egr1 Trim16 Myc Ccna2 Cdk1 Ccnb2 Amigo3 Jund Ptp4a1 4930539E08Rik Map3k6 Hspb1 Ppp1r15b Igfbp3 Nrarp Fabp4 Egr4 C F 52-week-old Sgk1 Thbs1 Nr4a2 Hmgcs2 Sik1 RD HFD Fos Hba-a1 Ankrd34a Pbk Rhob Id1 Myc Zc3h12a Gprc5a Mxd1 Nr1d1 mt-Tl1 GAPDH Tgif1 Prdm1 Actb Cldn23 Rasd2 Ttc30b Mettl21b Mafk Tnfrsf23 * Rem2 Gm11769 Mum1l1 E2f7 Cxcl10 Adamts9 Slc26a2 Ednrb Trp53inp1 Fosb Actg1 Dusp1 Ptger4 Ifi44 Sfn P2ry1

Myc/GAPDH Ankrd34b Siah2 Cd14 Skil Amotl2 Prc1 Krtap17-1 Zfp941 Chd7 RD HFD Hdc Piga G Gm12346 Rassf1 H I Cbx4 Myc Krt23 2.0 Dyrk3 Gm15459 Mmp12 * Srf GAPDH Fam84a Rgs16 * Tob1 Gm15747 Adv: Aldh1b1 LacZ Myc Hbb-bs 1.5 Zfp36

REGULAR DIETREGULAR Iffo2 Dffb 4.0 Clca3a1 Fosl2 Gmppb * * Rgs4 Slc25a33 DAPI

- Prickle1 fold) Trpc4 - 3.0 1.0 Junb Derl3 Nr0b2 Hsp90aa1 Hbb-bt Knstrn 2.0 Fam26e Insulin Rnf138rt1

- Rell1 Socs3

/GAPDH ( /GAPDH 0.5 # Irf2bp2 Ki67+ Ins+ Cells (%) Klhl38 Dnaja4 Nedd9

Myc 1.0 Hspd1 Myc Atf3 Zfp622 c- Zfand2a Rtp4

HFD 0 Egr2 0 Foxj1 Dnajb1 Hsp90ab1 Hspa8 Akap12 Sertad3 Cldn3 Isg15 Arc Adv: LacZ Myc A830010M20Rik Adv: LacZ Myc Tnfrsf12a Per1 Dusp10 Page 27 of 40 Diabetes Figure 6 A Eight-week-old B One-year-old Myc binding Myc binding 0.10 0.10 RD RD HFD 0.08 * HFD 0.08 IgG IgG 0.06 0.06 *

- Input % 0.04 - Input % 0.04 * * 0.02 * 0.02

0 0

RD vs HFD RD vs HFD RD vs HFD 8 wks old Mice 8 wks old Mice E 8 wks old Mice C 1 yr old Mice D 1 yr old Mice 1 yr old Mice

1 Mbp DNA Region Promoters Genes Not mapped E-Boxes

F P<0.05 N.S. G P<0.05 mC (%) 5-

HFD HFD + 5-AZA RD HFD RD HFD I Eight-week-old One-year-old P<0.02 Ccna2, Cdc20, Cdk1, Cdca2, Ccnb1 E-Boxes P<0.01 N.S. P<0.01 H 0.08 P<0.05

0.06 N.S. IgG P<0.01 P<0.01 0.04 P<0.05 N.S.

- Input % N.S. 0.02 Ki67+ Insulin+/Insulin + cells (%) RD HFD RD HFD HFD HFD HFD 0 + + + RD HFD RD HFD RD HFD RD HFD RD HFD RD HFD RD HFD RD HFD 8-week-old Mice 5-AZA 10058-F4 5-AZA +5-AZA +5-AZA +5-AZA +5-AZA + Cod CCNA2 CCNA2 Cod CDCA2 CDCA2 10058-F4 One-year-old Mice Diabetes Page 28 of 40 Figure 7

EIGTH-WEEK-OLD ONE-YEAR-OLD Short-term HFD Short-term HFD ß-Cell High Glucose High Glucose Ki67+ß-Cell α-Cell

HFD High Glucose

PI3K/PKC ζ/ERK/mTOR PI3K/PKC ζ/ERK/mTOR

PP2A PP2A

Stability MYC Degradation Stability MYC Degradation

MYC 3 3 3 CH CH - - CH Cell Cycle Genes -

Cell Cycle Genes

Proliferation Proliferation ß-Cell ß-Cell Suppl.Page 29 of 40Figure 1 Diabetes

A B RD1 RD2 RD3 HFD1 HFD2 HFD3

* *

C

* *

* *

D Diabetes Page 30 of 40 Suppl. Figure 2

A lox/lox B undeleted Myc allele Myclox/lox islets

Exon 1 LoxP Exon 2 Exon 3 LoxP Flox Myc

500bp ∆Myc P3 P1 P2 gapdh Cre expression

recombined Myc allele (ΔMyc)

Exon 1 LoxP

Ki67-insulin 700bp P1 P3 D C 823/13 Myclox/lox 5mM cells islets +Ad.LacZ Myc

β-actin

20mM E +Ad.LacZ *

# 20mM + Ad.Cre % Ki67+ Ins+/Ins+ Cells Ins+/Ins+ Ki67+ %

Glucose 5mM 20mM F 100 Ad.LacZ + + - * Ad.Cre - - + Myclox/lox Islet Cells 80

60

40

20 % % Myc+ Ins+/Ins+ Cells

0 RD HFD RD HFD βMycKO βMycKO Veh Tam Page 31 of 40 Diabetes Suppl. Figure 3 A B * * * *

C * * D

E F * *

* * Diabetes Suppl. FigurePage 324 of 40

2mM 20mM KD KD GFP PKCζ GFP PKCζ

PIN1

GAPDH

B56α

GAPDH B56α/GAPDH PIN1/GAPDH

GFP KD-PKCζ GFP KD-PKCζ GFP KD-PKCζ GFP KD-PKCζ 2mM G 20mM G 2mM G 20mM G Page 33 of 40 Diabetes Suppl. Figure 5

A 2mM 20mM * Veh PD98059 Veh PD98059 *

pERK1/2 pS62-Myc # #

# - Myc/GAPDH # GAPDH pERK1/2/GAPDH Myc pSer62 Veh PD Veh PD Veh PD Veh PD GAPDH 2mM G 20mM G 2mM G 20mM G *

# # /GAPDH Myc

Veh PD Veh PD 2mM G 20mM G

B 2mM 20mM * Veh Rapamycin Veh Rapamycin * p-p70S6K

GAPDH # /GAPDH

Myc Myc

p70S6K/GAPDHp- # # pS62-Myc Veh Rapa Veh Rapa Veh Rapa Veh Rapa GAPDH 2mM G 20mM G 2mM G 20mM G

*

- Myc/GAPDH # pS62

Veh Rapa Veh Rapa 2mM G 20mM G Diabetes Page 34 of 40 Suppl. Figure 6 Page 35 of 40 Diabetes

Supplemental Figure Legends

Supplementary Figure 1. Downregulated genes in islets after 1 week of a high fat diet in mice and verification of Myc binding to target genes in beta cells. Eight-week-old C57Bl/6N male mice were placed on a regular diet (RD) or a high fat diet (HFD) for one week. (A) Islets were isolated and RNA seq analysis was performed. Shown is a heatmap of genes significantly (padj < 0.05) downregulated more than 40%. n=3 mice per group. (B) INS1 832/13 cells were treated for 18 h with either 2 or 20 mM glucose, cells were fixed and extracts were subjected to chromatin immunoprecipitation using antibodies against Myc or a control IgG and specific primers (P) for E-boxes or the coding (cod) region. The data are presented as the percent input after subtraction of the IgG signal. Data are means ± SEM of n=6 experiments. No binding in ccna2 was detected with P3. (C) One-year-old C57Bl/6N male mice were placed on a regular diet (RD) or a high fat diet (HFD) for one week. Islets were isolated and RNA seq analysis was performed. Shown is a heatmap of genes significantly (padj < 0.05) downregulated more than 40%. n=4 mice per group. (D) Gene ontology (GO) analysis of biological processes significantly downregulated in islets from one-year-old HFD-fed mice.

Supplementary Figure 2. Acute Myc deletion blocks glucose-induced beta cell proliferation. (A) Diagram shows the localization of LoxP sequences flanking Exons 2 and 3 of the Myc gene, Cre-mediated deletion of these floxed exons and the corresponding structures of floxed Myc and ∆Myc allelic variants of the Myc gene. Location and size of genotyping PCR products are indicated. Arrows indicate PCR primers used for genotyping. (B) PCR analysis of islet DNA obtained from Myclox/lox mouse islets transduced with adenovirus expressing Cre (Ad.Cre) or Ad.LacZ as control showing floxed and deleted allele of Myc gene. (C) Western blot analysis of protein extracts prepared from INS1 832/13 cells and Myclox/lox mouse islets treated with Ad.Cre or Ad.LacZ as control. Blots were probed with antibodies to Myc and β-actin, as the loading control. (D) Dispersed islet cells from Myclox/lox mice were transduced with Ad.LacZ or Ad.Cre to delete the Myc gene and treated with 5 mM or 20 mM glucose for 24 h. Proliferating beta cells were detected by immunostaining with antibodies to insulin and Ki67 and quantified in (E). Data are means ± SEM of n = 5 experiments, *, P<0.05, **, P<0.01. (F) Deletion of Myc protein expression in beta cells of βMycKO mice quantified in islets stained for insulin and Myc as in Fig. 2D in pancreas sections from βMycKO mice treated with Tam or vehicle and fed regular diet or HFD. Data are means ± SEM of n = 5 mice per group, *, P<0.05.

Supplementary Figure 3. Effect of tamoxifen administration on glucose homeostasis and beta cell proliferation in 12 week-old male MIP-creERTAM mice fed short-term HFD or RD. (A) Body weight, (B) non-fasting blood glucose and (C) plasma insulin measured after RD or HFD in mice after tamoxifen injection following the same protocol as depicted in Fig. 2A. (D, E) Intraperitoneal glucose tolerance test and area under the curve (AUC), respectively. (F) Pancreata from the indicated treatment groups were stained for insulin and Ki67 and quantified. Data are means ± SEM from 3 mice/group. *, P<0.05;

Supplementary Figure 4. Glucose and PKCζ do not alter PIN1 and B56α expression in INS1 832/13 cells. Immunoblots of extracts from INS1 832/13 cells transduced with Ad.GFP or Ad.KD- PKCζ with indicated treatments for 24h and quantification (lower panel). Blots were probed with

1

Diabetes Page 36 of 40

antibodies to PIN1, B56α and GAPDH, as the loading control. Data are means ± SEM of n = 3 experiments.

Supplementary Figure 5. Regulation of Myc expression and phosphorylation in Ser62 by ERK1/2, mTOR and PP2A. (A) Immunoblots of extracts from INS1 832/13 cells treated with 10µM PD 098059, inhibitor of ERK1/2 activation, and 2 or 20mM glucose for 24h and quantification (right panels). Blots were probed with antibodies to phERK1/2, phSer62-Myc and Myc and GAPDH, as the loading control. (B) Immunoblots of extracts from INS1 832/13 cells treated with 10nM rapamycin, mTOR inhibitor, and 2 or 20mM glucose for 24h and quantification (right panels). Blots were probed with antibodies to ph-p70S6K, phSer62-Myc, Myc and GAPDH, as the loading control. Data are means ± SEM of n = 3 experiments, *, P<0.05 vs 2mM glucose, #, P<0.05 vs. vehicle and 20mM glucose treated.

Supplementary Figure 6. Schematic representation of the different promoter areas of the cell cycle genes tested in the ChIP assays in Fig. 6. Blue boxes indicate exons. Red boxes indicate E-boxes where Myc binds.

2

Page 37 of 40 Diabetes

Supplemental Table 1. Resources used in this study with their corresponding companies, catalogue number and Research Resource Identifiers (RRID) from https://scicrunch.org/resources (when available).

Product RRID TruSeq® Stranded mRNA Library Prep (48 Samples) (Illumina, cat#20020595, RRID:SCR_014983) RNA isolation kit (Qiagen, cat#74106, RRID:SCR_008539) Quant-iT™ PicoGreen™ dsDNA Assay Kit (Life Technologies, cat#P7598, RRID:SCR_008817)

Qubit™ 4 Fluorometer (Thermo Fisher Scientific, Cat#Q33226, RRID:SCR_008452) 3500 Genetic Analyzer for Resequencing & Fragment (Thermo Fisher Scientific, Cat#44440462, Analysis RRID:SCR_008452)

The NGS Fragment Analyzer Kit (Agilent technologies, Cat#DNF473, RRID:SCR_013575) 2100 Bioanalyzer Instrument (Agilent technologies, Cat#G2939BA, RRID:SCR_013575) HiSeq2500 Sequencing System (Illumina, cat#SY-401-2501, RRID:SCR_014983) Tamoxifen (Sigma-Aldrich, cat# T5648, RRID:SCR_008988) DAPI (Thermo Fisher Scientific, Cat# 50-247-04, Fluorogel-Electro Miscroscope Science RRID:SCR_008452) anti-insulin antibody ELISA kit (Mercodia Cat# 10-1201-01, RRID:AB_2636872) Polyclonal Guinea Pig Anti- Insulin antibody (Agilent Technologies Cat# A056401-2, RRID:AB_2617169), dil 1/1000 (MYC) Monoclonal, Unconjugated, Clone Y69 antibody (LifeSpan Cat# LS-C49460-100, RRID:AB_1191623), dil 1/500 Ki67 antibody (Thermo Fisher Scientific Cat# MA1-80756, RRID:AB_2142242), dil 1/200 Ph-Ser62-Myc antibody (Abcam Cat# ab51156, RRID:AB_869189), dil 1/500

h-Thr58-Myc antibody (Abcam Cat# ab28842, RRID:AB_731667), dil 1/500

PIN1 antibody (GenWay Biotech Inc. Cat# 20-002-35034-0.1 ml, RRID:AB_1022769), dil 1/500 Beta Actin antibody (Sigma-Aldrich Cat# A2066, RRID:AB_476693), dil 1/1000 GAPDH antibody (Sigma-Aldrich Cat# G9545, RRID:AB_796208), dil 1/10000 Alpha tubulin antibody (Millipore Cat# MABT205, RRID:AB_11204167), dil 1/1000 Anti-PP2A-B’ (B56) antibody (Millipore Cat# 07-1221, RRID:AB_11211106), dil 1/500 Myc antibody (Cell Signaling Technology Cat# 9402, RRID:AB_2151827), dil 1/500 Phospho-p70 S6 Kinase (Thr389) antibody (Cell Signaling Technology Cat# 9205, RRID:AB_330944), dil 1/500 Phospho-GSK-3 (Ser9) antibody (Cell Signaling Technology Cat# 9336, RRID:AB_331405), dil 1/500 Amersham ECL Prime Western Blotting Detection (GE Healthcare, Cat#RPN2232, RRID:SCR_000004) Reagent Diabetes Page 38 of 40

Nonfat-Dried Milk (Sigma-Aldrich, Cat#M7409, RRID:SCR_008988) Immobilon-P PVDF membrane (EMD Millipore, Cat#IPVH00010, RRID:SCR008983) Collagen P (Roche, Cat#11213857001, RRID:SCR001326) Cover Glasses Round 12 mm (Thermo Fisher Scientific, Cat#12-545-80, RRID:SCR_008452) Gibco™ Trypsin-EDTA (0.05%), Phenol red (Thermo Fisher Scientific, Cat#25300054, RRID:SCR_008452) Corning™ RPMI 1640 Medium (Mod.) 1X with L- (Thermo Fisher Scientific, Cat#10040CV, Glutamine RRID:SCR_008452) Gibco™ Fetal Bovine Serum, certified, heat inactivated (Thermo Fisher Scientific, Cat#10082147, RRID:SCR_008452) Gibco™ Thermo Fisher Scientific, Cat#15140122, Penicillin-Streptomycin (10,000 U/mL) RRID:SCR_008452)

Endothall (Sigma-Aldrich, Cat#E7649, RRID:SCR_008988) Cmyc Inhibitor II CAS413611-93-5 (EMD Millipore, Cat#475956, RRID:SCR008983) 5-Aza-2′-deoxycytidine (Sigma-Aldrich, Cat#A3656, RRID:SCR_008988) (Emsdiasum, Cat#15710, RRID:SCR_008497) Formaldehyde Aqueous Solution, 16%) Rapamycin (Sigma-Aldrich, Cat#R0395, RRID:SCR_008988) PD 98059 (Sigma-Aldrich, Cat#P215) (EMD Millipore, Cat#17-313, RRID:SCR008983) PP2A Immunoprecipitation Phosphatase Assay Kit

Pierce™ BCA Protein Assay Kit Thermo Fisher Scientific, Cat#23227, RRID:SCR_008452) DNeasy Blood & Tissue Kit (Qiagen, cat#69504, RRID:SCR_008539)

Kapa Hyper Prep kit, PCR-free version (Roche (Roche, Cat#KK8505, RRID:SCR001326) Sequencing & Life Sciences). SeqCap Epi Enrichment System (Roche, SEQ#100146, RRID:SCR001326)

(Beckman Coulter, A63881, RRID:SCR_008940) Agencourt AMPure XP

EZ DNA Methylation-Lightning Kit Cite this (Zymo Research, Cat#D5030T, RRID:SCR_008968) (Roche, Cat#07959052001, RRID:SCR001326) KAPA HiFi HotStart Uracil+ DNA Polymerase

MethylFlash Methylated DNA 5-mC Quantification Kit (Epigentek, Cat# P-1034-48)

Page 39 of 40 Diabetes

Supplemental Table 2. Sequences chosen for deep methylome sequencing. Highlighted genes are direct Myc targets upregulated in young mice after a HFD. Other genes act as controls.

Gene Location chromosome 5' 3' total bp Bag3 chr7:135,662,637-135,677,238 chr7: 128,305,752 128,329,116 23,364 Ccna2 chr3:36,452,073-36,482,631 chr3: 36,744,037 36,775,865 31,828 Ccnb1 chr13:101,823,744- chr13: 101,823,744 101,893,475 69,731 Ccnb2 chr9:70,204,248-70,231,977 chr9: 70,204,248 70,231,977 27,729 Ccnd1 chr7:144,738,041-144,754,159 chr7: 144,738,041 144,754,159 16,118 Ccnd2 chr6:127,099,577-127,137,586 chr6: 127,099,577 127,137,586 38,009 cdc20 chr4:117,918,961-117,962,440 chr4: 117,918,961 117,962,440 43,479 Cdca2 chr14:66,626,107-66,672,314 chr14: 66,626,107 66,672,314 46,207 Cdca3 chr6:124,779,663-124,810,838 chr6: 124,779,663 124,810,838 31,175 Cdk1 chr10:68,715,714-68,764,547 chr10: 68,715,714 68,764,547 48,833 Cdkn1a chr17:28,805,683-28,829,132 chr17: 28,805,683 28,829,132 23,449 Cdkn2C chr4:109,165,039-109,178,526 chr4: 109,165,039 109,178,526 13,487 Ect2 chr3:27,329,745-27,363,282 chr3: 27,329,745 27,363,282 33,537 H2afx chr9:44,084,167-44,096,379 chr9: 44,084,167 44,096,379 12,212 Hmmr chr11:40,562,260-40,594,276 chr11: 40,562,260 40,594,276 32,016 Il1b chr2:129,052,454-129,072,058 chr2: 129,052,454 129,072,058 19,604 Ins2-Igf2- chr7:142,344,970-142,502,169 chr7: 142,344,970 142,502,169 157,199 Kif20aH19 chr18:34,741,513-34,766,361 chr18: 34,741,513 34,766,361 24,848 Mki67 chr7:135,525,250-135,560,705 chr7: 135,525,250 135,560,705 35,455 Mlxipl chr5:135,370,143-135,401,612 chr5: 135,370,143 135,401,612 31,469 Nusap1 chr2:119,295,846-119,327,324 chr2: 119,295,846 119,327,324 31,478 Odc1 chr12:17,558,983-17,577,306 chr12: 17,558,983 17,577,306 18,323 Plk1 chr7:121,936,932-121,968,041 chr7: 121,936,932 121,968,041 31,109 Rgs2 chr1:145,756,592-145,777,356 chr1: 145,756,592 145,777,356 20,764 Stmn1 chr4:133,730,851-133,747,134 chr4: 133,730,851 133,747,134 16,283 Top2a chr11:98,831,095-98,861,446 chr11: 98,831,095 98,861,446 30,351 Ttk chr9:83,599,160-83,636,002 chr9: 83,599,160 83,636,002 36,842 Txnip chr3:96,634,081-96,651,635 chr3: 96,634,081 96,651,635 17,554 Ube2c chr2:164,449,664-164,476,429 chr2: 164,449,664 164,476,429 26,765 Diabetes Page 40 of 40

Online Supplemental Data mRNA Library Preparation and Sequencing RNA libraries were prepared using the TruSeqStranded mRNA Library Preparation Kit (Illumina). Purified mRNA from 100ng of total islet RNA underwent first and second strand cDNA synthesis. cDNA was then adenylated, ligated to Illumina sequencing adapters, and amplified by PCR. Final libraries were evaluated using fluorescent-based assays including PicoGreen (Life Technologies), Qubit (Invitrogen), Fragment Analyzer (Advanced Analytics) or BioAnalyzer (Agilent 2100), and were sequenced on an Illumina HiSeq2500 sequencer (v4 chemistry, v2 chemistry for Rapid Run) using 2 x 50bp cycles, aiming for 40 M reads per sample.

Expression Analysis. Reads were aligned to the mm10 mouse reference using STAR aligner (v2.4.2a) (27). Quantification of genes annotated in Gencode vM5 was performed using featureCounts (v1.4.3) and quantification of transcripts using Kalisto (28). QC were collected with Picard (v1.83) and RSeQC (29) (http://broadinstitute.github.io/picard/). Normalization of feature counts and differential expression analysis was done using the DESeq2 package (30). Gene set enrichment analysis was performed from the Broad Institute web site (http://software.broadinstitute.org/gsea/index.jsp) (31).

DNA Methylation Analysis. Probes were designed to capture different regions of the mouse genome (Suppl. Table 2) spanning a total of 1Mbp according to the SeqCap Epi Enrichment System (Roche NimbleGen, Roche Sequencing & Life Sciences) for hybridization-based targeted enrichment of bisulfite-treated DNA. Pre-capture libraries were prepared using the Kapa Hyper Prep kit, PCR-free version (Roche Sequencing & Life Sciences). Samples were barcoded and multiplex-sequenced in a single run and the reads run through a customary DNA methylation pipeline for generating methylation calls at every CpG dinucleotide. Sequencing was performed at the Epigenomics Core Facility of Weill Cornell Medicine. Islet genomic DNA (200 ng/sample) from mice administered RD or HFD was sonicated using a Covaris S220 (Covaris) to approximately 180–220bp fragments. End-repair and A-tailing was performed in a single reaction, followed by ligation of methylated indexed adaptors provided in the Roche SeqCap Epi Enrichment kit. Products were cleaned using Agencourt AMPure XP beads (Beckman Coulter). Bisulfite conversion was carried out at 54°C for 1h using the Zymo EZ DNA Lightning kit (Zymo Research), followed by 12-cycles of ligation-mediated PCR amplification performed with HiFi HotSart Uracil + polymerase (Roche Sequencing & Life Sciences). Multiplex hybridization was performed by using 1μg of bisulfite converted libraries and hybridizing to the custom SeqCap Epi Choice probe pool at 42°C for 72h. Hybridized products were purified with Capture Beads and PCR amplified for 15 cycles to create the final libraries for sequencing. Final yields were quantified in a Qubit and quality of the library was assessed on a DNA1000 Bioanalyzer chip (Agilent Technologies). The percent methylation data across the CpG dinucleotides assessed by the targeted DNA methylation platform were averaged within each group and compared with each other.