An siRNA screen in pancreatic beta cells reveals a role for Gpr27 in insulin production

Abstract

The prevalence of type 2 diabetes in the United States is projected to double or triple by 2050. We reasoned that the genes that modulate insulin production might be new targets for diabetes therapeutics. Therefore, we developed an siRNA screening system to identify genes important for the activity of the insulin promoter in beta cells. We created a subclone of the MIN6 mouse pancreatic beta cell line that expresses destabilized GFP under the control of a 362 base pair fragment of the human insulin promoter and the mCherry red fluorescent protein under the control of the constitutively active rous sarcoma virus promoter. The ratio of the GFP to mCherry fluorescence of a cell indicates its insulin promoter activity. As G protein coupled receptors (GPCRs) have emerged as novel targets for diabetes therapies, we used this cell line to screen an siRNA library targeting all known mouse GPCRs. We identified several known GPCR regulators of insulin secretion as regulators of the insulin promoter. One of the top positive regulators was Gpr27, an orphan GPCR with no known role in beta cell function. We show that knockdown of Gpr27 reduces endogenous mouse insulin promoter activity and glucose stimulated insulin secretion. Furthermore, we show that Pdx1 is important for Gpr27's effect on the insulin promoter and insulin secretion. Finally, the over-expression of Gpr27 in 293T cells increases inositol phosphate levels, while knockdown of Gpr27 in MIN6 cells reduces inositol phosphate levels, suggesting this orphan GPCR might couple to Gq/11. In summary, we demonstrate a MIN6-based siRNA screening system that allows rapid identification of novel positive and negative regulators of the insulin promoter. Using this system, we identify Gpr27 as a positive regulator of insulin production.

Figure 1. siRNA screening system to identify regulators of insulin promoter activity.

A. Schema of reporter constructs integrated into the screening MIN6 cell line. These are pSico lentiviruses containing the proximal 362 bases of the human insulin promoter driving destabilized GFP and the rous sarcoma virus (RSV) promoter driving mCherry. B. After transfection with siRNAs, GFP and mCherry mean fluorescence intensity (MFI) are measured by flow cytometry. If the siRNA targets a positive regulator, GFP/mCherry MFI decreases. If the siRNA targets a negative regulator, the GFP/mCherry MFI increases. C. Control (green), Pdx1(red), or GFP(blue) siRNAs were transfected into the MIN6 reporter cell line. After 5 days, GFP (left panel) and mCherry (right panel) fluorescence were measured by flow cytometry. D. As in C. The ratio of GFP geometric mean fluorescence intensity (MFI) to mCherry MFI was calculated for each sample. The ratio was normalized to that of control siRNA transfected cells. Data shown are averages and standard error (n = 3). ** p<0.01 versus control siRNA.

Figure 2. siRNA screen hit selection and initial confirmation.

A. Heat map showing the top six putative positive regulators of the insulin promoter clustered based on RSA determined p value, and the base 10 log of the fragments per kilobase of exon model per million mapped reads (FPKM) value from three independent mRNA-seq data sets. B. As in B, but analyzed for negative regulators of insulin promoter activity. Islet-Merck refers to SRA008619 submitted by Merck, Islet-non preg refers to islets from non-pregnant mice, and Islet-preg refers to islets from pregnant mice . C. Confirmation of hits from A and B. The indicated siRNAs were transfected into the screening cell line. GFP and mCherry was measured by flow cytometry. Error bars show standard error of three biological replicates performed (n = 3). * indicates p<0.05 versus >6 of the negative control siRNAs.

Figure 3. Gpr27 is required for mouse insulin promoter activity and glucose stimulated insulin secretion.

A. MIN6 cells were infected with either Ad-control or Ad-shGpr27. Three days after infection, RT-qPCR was performed for the indicated genes. The data are plotted as % expression compared to control adenovirus with standard error (n = 3). B. As in A, but dispersed primary mouse islets were infected with Ad-control or Ad-shGpr27. After 3 days, infected cells were sorted by flow cytometry of GFP positive cells and RT-qPCR was performed for the indicated genes (n = 3). C. MIN6 cells were infected with either Ad-control or Ad-shGpr27. Three days after infection, glucose stimulated insulin secretion was measured by ELISA after 1 hour of static incubation at either 2 mM or 20 mM glucose. Data are represented as the average of fractional insulin secretion with standard error (n = 9). *p<0.05, **p<0.005 versus Ad-control.

Figure 4. Gpr27 knockdown affects multiple transcription factors and requires Pdx1 for its effect on the insulin promoter.

A. MIN6 cells were infected with Ad-control or Ad-shGpr27. Three days after infection RT-qPCR was performed for the indicated genes. Expression level normalized to that of Ad-control are plotted with standard error (n = 6). B. Schematic of the human insulin promoter with selected regulatory sequences and transcription factors that bind to these elements. C. MIN6 cells were cotransfected with the indicated siRNA, insulin promoter firely luciferase construct and thymidine kinase renilla luciferase construct. Two days after transfection, firefly and renilla luciferase activity were measured. The ratio of firefly to renilla luciferase was normalized to the control siRNA. Average and standard error are plotted (n = 6–9) D. As in C using human insulin promoter −362 firefly luciferase (n = 6–9). E. MIN6 cells were transfected with the indicated siRNAs and glucose stimulated insulin secretion was measured after 5 days. Fractional insulin secretion is shown (n = 12). * p<0.05 ** p<0.005 *** p<0.0005 versus control siRNA or control adenovirus (at high glucose part E).

Figure 5. Gpr27 positively regulates inositol phosphate levels.

A. Schema of canonical GPCR signaling pathways and resulting expected changes in second messengers cAMP and IP3. B. HEK293T cells were transiently transfected with either GFP plasmid (control) or FLAG-Gpr27 plasmid. 24 hours after transfection cells were analyzed by flow cytometry for extracellular FLAG. C. As in B, but cells were lysed and assayed for IP1. D. As in C, but lysates were analyzed for cAMP. E. MIN6 cells were infected with control or Gpr27 knockdown adenovirus and 3 days later, cells were lysed and IP1 was measured. F. As in E, but cAMP was measured. For C–F, average and standard error is plotted (n = 9); * p<0.005 versus control.