PDIA6 regulates insulin secretion by selectively inhibiting the RIDD activity of IRE1

Abstract

Protein disulfide isomerase A6 (PDIA6) interacts with protein kinase RNA-like endoplasmic reticulum kinase (PERK) and inositol requiring enzyme (IRE)-1 and inhibits their unfolded protein response signaling. In this study, shRNA silencing of PDIA6 expression in insulin-producing mouse cells reduced insulin production (5-fold) and, consequently, glucose-stimulated insulin secretion (3-4-fold). This inhibition of insulin release was independent of the PDIA6-PERK interaction or PERK activity. Acute inhibition of PERK did not change the short-term response of β cells to glucose. Rather, PDIA6 affected insulin secretion by modulating one of the activities of IRE1. At 11 mM glucose and lower, the regulated IRE1-dependent decay (RIDD) of the mRNA activity of IRE1 was activated, but not its X-box binding protein (XBP)-1 splicing activity. In the absence of PDIA6, RIDD activity toward insulin transcripts was enhanced up to 4-fold, as shown by molecular assays in cultured cells and the use of a fluorescent reporter in intact islets. Such physiologic activation of IRE1 by glucose contrasted with IRE1 activation by chemical stress, when both IRE1 activities were induced. Thus, whereas the stimulus determines the quality of IRE1 signaling, PDIA6 attenuates multiple enzymatic activities of IRE1, maintaining its signaling within a physiologically tolerable range.

Keywords: GSIS; unfolded protein response; β-cell metabolism.

Figure 1.

PDIA6 depletion affects significantly glucose-stimulated insulin secretion. A) INS1 832/13 cells expressing a nontargeting (shCtrl, short hairpin control) or a PDIA6-targeting shRNA were obtained from 2 different infection events. Lysates were immunoblotted to detect the expression level of PDIA6. ERp72 served as a specificity control and 14.3.3 as a loading control. There was progressive recovery of PDIA6 expression over time. B) INS1 cells from (A) were exposed to the indicated glucose concentrations for 3 h. Supernatants were analyzed by RIA to detect human insulin. Each measurement was normalized to the protein content. Data are means ± SD of 3 independent experiments. *P ≤ 0.05; **P ≤ 0.01; Student’s t test; shPDIA6 vs. corresponding shCtrl condition.

Figure 2.

Loss of PDIA6 augments PERK signaling. A) Short hairpin control (shCtrl) and shPDIA6 INS1 832/13 cells were exposed to 500 nM TG for the indicated times. Cell lysates were then immunoblotted to detect PERK, KDEL, total or Ser⁵¹-phosphorylated eIF2α, Phospho-signal/total eIF2α ratios (normalized to eIF2α) are showed in the gel (n = 2). B) shCtrl or shPDIA6 MIN6 cells were exposed to 1 μM TG for the indicated times. Cell lysates were then immunoblotted to detect total or Ser⁵¹-phosphorylated eIF2α. In both experiments, untreated shCtrl samples served as internal reference. Means ± sd of the phosphosignal/total eIF2α ratios are plotted (n = 3). *P ≤ 0.05. C) Immunoblotting of cells from (B) to show the level of PDIA6 depletion. D) INS1 832/13 cells were exposed to 500 nM TG for the indicated times and in parallel, either pretreated or treated, with the GSK2606414 at the indicated concentrations. PERK, total, or Ser-51-phosphorylated eIF2α levels were assessed by immunoblotting.

Figure 3.

Glucose-stimulated insulin secretion is not mediated by PERK. A) Short hairpin control (shCtrl) or shPDIA6 INS1 cells were exposed to the indicated glucose levels and the human insulin measured in the supernatants after 3 h of exposure by RIA. Each measurement was normalized to the protein content. B and C) Samples in (A) were subjected to immunoblot analysis, to detect PERK, PDIA6, total, or Ser⁵¹-phosphorylated eIF2α levels. The second blot in (C) shows cells treated with the GSK 2606414. Data are means ± sd of 3 independent experiments. D) INS1 832/13 cells were either pretreated or treated with the GSK inhibitor during the glucose stimulation. Human insulin was measured as in (A). E) Cell lysates from (D) were analyzed by immunoblot, as in (B–C).

Figure 4.

Insulin transcripts decrease in the absence of PDIA6. A, B) INS1 832/13 cells were exposed to 500 nM TG for 8 h in the presence or absence of 30 μM 4μ8C. Ins1 and XBP1 splicing mRNAs were measured by qPCR, with actin as the housekeeping gene. C, D) Ins1 and Ins2 transcripts were measured as above in PDIA6-deficient (shPDIA6) and -sufficient (shCtrl, short hairpin control) INS1 cells. Data are means ± sd of 3 independent experiments. E) Samples in (C, D) were immunoblotted to estimate PDIA6 depletion. F, G) Cells in (C, D) were treated with 500 nM TG for the indicated times, and Ins1 and Ins2 transcripts measured by qPCR. Data are means ± sd of 3 independent experiments. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; Student’s t test; shPDIA6 vs. the corresponding shCtrl condition. H) INS1 832/13 cells were pretreated with 1 μg/ml actinomycin D (Act D) for 30 min and then stressed with 1 μM TG for 5 h, in the continued presence of actinomycin D. Ins2 mRNA was measured by qPCR, with actin as the housekeeping gene. I) Bloc1S1 and Pmp22 transcripts were measured as above in PDIA6-deficient and -sufficient INS1 cells, before and after TG treatment.

Figure 5.

Degradation of insulin transcripts in the absence of PDIA6 is mediated by IRE1. A) PDIA6-depleted INS1 832/13 cells were treated with different concentrations (10, 30, and 90 μM) of 4μ8C for 8 h. Untreated shCtrl samples served as an internal reference. B) PDIA6-depleted INS1 cells were transiently transfected with empty vector (EV) or wild-type (WT) or cysteine-free V5-tagged PDIA6. Ins2 transcript was measured by qPCR at 48 h after transfection. C) Samples in (B) were immunoblotted to detect the level of exogenous PDIA6 over the endogenous. D) PDIA6-sufficient and -depleted INS1 cells were treated with 500 nM TG for 6 h. In addition to TG, one sample was pretreated with 4μ8C and another with the GSK inhibitor as the negative control. Unspliced (u) and spliced (s) XBP1 mRNAs were amplified by RT-PCR. β-Actin served as control for RNA recovery. The percentage of spliced XBP1 out of the total is reported under each lane. E) PDIA6-sufficient and -depleted INS1 cells were treated with 500 nM TG for the indicated times, and XBP1 splicing was analyzed by qPCR.

Figure 6.

PDIA6 depletion exacerbates chronic glucose exposure–induced RIDD activity. A, B) PDIA6-sufficient and -depleted INS1 832/13 cells were exposed to the indicated concentrations of glucose for 2 days. Ins2 (A) and Ins1 (B) transcripts were measured by qPCR. Each value was normalized to shCtrl cells in 11 mM glucose. Data are means ± sd of 3 independent experiments. *P ≤ 0.05; **P ≤ 0.01; Student’s t test; shPDIA6 vs. the corresponding shCtrl condition. C) XBP1 splicing was measured from samples as in (A) by qPCR. Cells were also treated with 500 nM TG for 6 h as a positive control for induction of XBP1 splicing. D) INS1 832/13 cells were exposed to the indicated concentrations of glucose for 3 days in 2 independent experiments. Unspliced (u) and spliced (s) XBP1 mRNAs were amplified by RT-PCR. β-Actin served as the control for RNA recovery. E) Mouse islets were transfected with the tdTomato XBP1 splicing reporter and then exposed to different glucose concentrations. TG treatment served as the positive control to identify transfected cells, given that only a portion of islet cells are responsive to ER stress.

Figure 7.

Intracellular insulin content is compromised in the absence of PDIA6. A) PDIA6-sufficient and -deficient INS1 cells were lysed and analyzed by RIA for the human intracellular insulin. Each measurement was normalized to the protein content. Data are means ± sd of 3 independent experiments. *P ≤ 0.05; **P ≤ 0.01; Student’s t test; shPDIA6 vs. the corresponding shCtrl condition. B) Immunostaining of total insulin in either PDIA6-sufficient or -deficient INS1 cells. C) Quantitation of total fluorescence intensity from cells in (B).