BET Proteins Are Required for Transcriptional Activation of the Senescent Islet Cell Secretome in Type 1 Diabetes

Abstract

Type 1 diabetes (T1D) results from the progressive loss of pancreatic beta cells as a result of autoimmune destruction. We recently reported that during the natural history of T1D in humans and the female nonobese diabetic (NOD) mouse model, beta cells acquire a senescence-associated secretory phenotype (SASP) that is a major driver of disease onset and progression, but the mechanisms that activate SASP in beta cells were not explored. Here, we show that the SASP in islet cells is transcriptionally controlled by Bromodomain ExtraTerminal (BET) proteins, including Bromodomain containing protein 4 (BRD4). A chromatin analysis of key beta cell SASP genes in NOD islets revealed binding of BRD4 at active regulatory regions. BET protein inhibition in NOD islets diminished not only the transcriptional activation and secretion of SASP factors, but also the non-cell autonomous activity. BET protein inhibition also decreased the extent of SASP induction in human islets exposed to DNA damage. The BET protein inhibitor iBET-762 prevented diabetes in NOD mice and also attenuated SASP in islet cells in vivo. Taken together, our findings support a crucial role for BET proteins in the activation of the SASP transcriptional program in islet cells. These studies suggest avenues for preventing T1D by transcriptional inhibition of SASP.

Keywords: BET proteins; beta cells; senescence and SASP; type 1 diabetes.

Figure 1

BET protein BRD4 is required for transcriptional activation of senescence-associated secretory phenotype (SASP) genes in nonobese diabetic (NOD) islets. (a) Chromatin immunoprecipitation and quantitative polymerase chain reaction (ChIP and qPCR) for H3K27ac at the enhancer of Mmp2 and promoter of Il6, along with distal negative regions, on islets isolated from two different euglycemic 14-week-old NOD mice (NOD413, NOD414), n = 2 biological replicates per mouse. Female mice are used throughout our study. The schematic shows the positions of the regions in the ChIP. Immunoglobulin G (IgG) ChIP was a negative control. Error bars are SD. (b) Representative IHC for BRD4 and insulin with DAPI on pancreas section from euglycemic NOD mice at eight weeks of age. Scale bar = 20 μm. (c) ChIP and qPCR for BRD4 at the same regulatory regions of Mmp2 and Il6 shown in (a) on pooled islets from euglycemic 14-week-old NOD mice cultured for 18 h with vehicle or iBET-762 (n = 2 biological replicates per group), error bars are SD. IgG ChIP was a negative control. (d) Islets isolated from 14-week-old NOD mice were depleted of CD45⁺ immune cells to enrich for endocrine islet cells and then cultured for 24 h with vehicle (Dimethyl Sulfoxide, DMSO) or iBET-762 followed by quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis. qRT-PCR of SASP genes Mmp2, Il6, Flnb, and Igfbp4 show the average relative expression for vehicle (n = 5 biological replicates) or iBET-762 treated (n = 4 biological replicates) islets cells. Error bars are SD. (e) qRT-PCR for Brd4 and Il6 on 14-week-old NOD islets transfected with non-targeting control or Brd4 siRNAs. Data show average expression levels in siBrd4 relative to siCtrl from n = 3 independent experiments. Error bars are SD. For all panels, * p < 0.05, ** p < 0.005, *** p < 0.0005, two-tailed t-tests.

Figure 2

iBET-762 attenuates SASP secretion and paracrine activities from NOD islets. (a) Islets isolated from 14-week-old NOD mice were cultured in vehicle or iBET-762 for 24h and the conditioned media (CM) was collected for luminex assays. Relative quantification of average secretion of SASP factors IL-6, Igfbp3, Mmp3, and Mmp12 from iBET-762-treated islets versus vehicle-control islets, n = 3 biological replicates per group. Error bars are SD. * p < 0.05, ** p < 0.005, *** p < 0.0005, two-tailed t-tests. (b) Islets isolated from 14-week-old NOD mice were depleted of CD45⁺ cells and cultured in vehicle or iBET-762 for 24 h and the resulting CM was collected used in paracrine senescence induction assays on islets from three-week-old old NOD mice, where qRT-PCR for Cdkn1a was a readout of paracrine senescence induction. Regular islet media (not conditioned by islets) was used as a negative control to set the baseline. Data are average relative expression levels from n = 5 or 6 biological replicates per group, error bars are SD. ** p < 0.005, one-way analysis of variance (ANOVA). (c) CM was collected from islets of 14-week-old NOD mice cultured with vehicle or iBET-762 and used in chemotaxis assays on THP-1 cells. Regular islet media (not conditioned by islets) was used as a negative control to measure the baseline. Data are chemotaxis fold-change, relative to control, from n = 3 biological replicates per group. Error bars are SD. ** p < 0.005, one-way ANOVA.

Figure 3

BET proteins are required for activation of SASP in human islets exposed to DNA damage. (a) Representative IHC for BRD4 and Insulin with DAPI in pancreas sections from nondiabetic (ND), autoantibody-positive (AA+), and Type 1 Diabetes (T1D) donors. Scale bars = 20 μm. (b) Islets from nondiabetic human donors were cultured with bleomycin (bleo) to induce DNA damage, or vehicle (0.1% DMSO) as a control, followed by replacing the media with media containing iBET-762 or vehicle and culture for an additional five days. Then, islets were harvested for qRT-PCR and conditioned media (CM) saved for luminex assays. (c) qRT-PCR analysis of CDKN1A and CDKN2A (INK4A/P16) on vehicle control, Bleo + vehicle, or Bleo + iBET-762 islets isolated from two different donors (n = 3 biological replicates per group for each donor). Error bars are SD. ** p < 0.005, one-way ANOVA. (d) Luminex assay of average protein secretion of SASP factors CXCL1, IL-8, and IGFPB4 from the same islets as in (c). Each dot represents a biological replicate of islets (n = 6 per group, from the two different donors shown in (c). Error bars are SD. * p < 0.05, ** p < 0.005, *** p < 0.0005, one-way ANOVA.

Figure 4

iBET-762 prevents diabetes and attenuates SASP in vivo. (a) Twelve-week-old NOD mice (n = 10 per group) were administered vehicle or iBET-762 intraperitoneally (i.p.) for two weeks and glycemia was monitored up to 30 weeks of age. Diabetes incidence curves show the frequency of hyperglycemic mice at each time point. Gray shaded area shows treatment window (12–14 weeks of age). p = 0.0115, log-rank test. (b) Insulits grading in n = 30 islets taken from two euglycemic vehicle control and two euglycemic iBET-762 mice after the prevention study in (a) at 30 weeks. Error bars are SD. (c) Recent onset hyperglycemic NOD mice were administered vehicle or iBET-762 i.p. for one week and then IHC was performed for SASP markers. Representative IHC of SASP factors IL-6 or Flnb with insulin in pancreas sections from the indicated diabetic mice after treatment. Scale bars = 20 μm. (d) Quantification of IL-6⁺/Ins⁺ cells per islet in the hyperglycemic NOD mice treated as in (a). Each dot represents an islet, where n = 20 islets for a total of 735 beta cells were scored from three vehicle-treated mice, and n = 16 islets for a total of 891 beta cells were scored from three iBET-762-treated mice. * p < 0.05, two-tailed t-test.