CFTR Influences Beta Cell Function and Insulin Secretion Through Non-Cell Autonomous Exocrine-Derived Factors

Abstract

Although β-cell dysfunction in cystic fibrosis (CF) leads to diabetes, the mechanism by which the cystic fibrosis transmembrane conductance regulator (CFTR) channel influences islet insulin secretion remains debated. We investigated the CFTR-dependent islet-autonomous mechanisms affecting insulin secretion by using islets isolated from CFTR knockout ferrets. Total insulin content was lower in CF as compared with wild-type (WT) islets. Furthermore, glucose-stimulated insulin secretion (GSIS) was impaired in perifused neonatal CF islets, with reduced first, second, and amplifying phase secretion. Interestingly, CF islets compensated for reduced insulin content under static low-glucose conditions by secreting a larger fraction of islet insulin than WT islets, probably because of elevated SLC2A1 transcripts, increased basal inhibition of adenosine triphosphate-sensitive potassium channels (K-ATP), and elevated basal intracellular Ca2+. Interleukin (IL)-6 secretion by CF islets was higher relative to WT, and IL-6 treatment of WT ferret islets produced a CF-like phenotype with reduced islet insulin content and elevated percentage insulin secretion in low glucose. CF islets exhibited altered expression of INS, CELA3B, and several β-cell maturation and proliferation genes. Pharmacologic inhibition of CFTR reduced GSIS by WT ferret and human islets but similarly reduced insulin secretion and intracellular Ca2+ in CFTR knockout ferret islets, indicating that the mechanism of action is not through CFTR. Single-molecule fluorescent in situ hybridization, on isolated ferret and human islets and ferret pancreas, demonstrated that CFTR RNA colocalized within KRT7+ ductal cells but not endocrine cells. These results suggest that CFTR affects β-cell function via a paracrine mechanism involving proinflammatory factors secreted from islet-associated exocrine-derived cell types.

Copyright © 2017 Endocrine Society.

Figures

Figure 1.

GSIS from CF and WT ferret islets. (A) Islet perifusion assay of insulin secretion from CF and WT cultured neonatal ferret islets in response to a glucose shift from 1.67 mM (G1.67) to 16.7 mM (G16.7). (B, C) Insulin area under the curve (AUC) from the data shown in (A) depicts the (B) first and (C) second phases of insulin secretion. (D) Islet perifusion assay of glucose and cAMP amplified insulin secretion triggered by tolbutamide (100 µM) from CF and WT cultured neonatal ferret islets in the presence of DAZ (100 µM) and forskolin (1 µM, Frsk). (E) Insulin AUC from the data shown in (D) depicts the amplifying phase of insulin secretion. (F–H) Effect of 100 µM DAZ on insulin secretion from cultured CF and WT neonatal ferret islets at low glucose (1.67 mM) under 1-hour static conditions. (F) Percentage of insulin secretion (total secreted insulin/total secreted insulin + islet insulin content), (G) total secreted insulin, and (H) total islet insulin. Data for all graphs show the mean ± standard error of the mean for the number (N) of experiments in each panel. Each experimental set of islets was derived from three or four independent animals. Statistical analysis was performed by (B, C, and E) two-tailed Student t test and (F–H) one-way analysis of variance and Bonferroni post hoc test. Asterisks mark genotypic comparisons, with significance of *P = 0.0273, **P = 0.0072, and ****P = 8 × 10−5. All other P values are given in the panels.

Figure 2.

Fura-2-acetoxymethyl ester–Ca2+ imaging and hormone secretion from ferret and human islets in response to CFTR-KO or CFTRinh-172. (A–C) Neonatal CF and WT ferret islets were loaded with Fura-2 and then equilibrated in 1.67 mM glucose (G1.67) followed by a sequential challenge with 16.7 mM glucose (G16.7) and 40 mM KCl. (A) Absolute 340/380-nm Fura-2 ratio from WT and CF ferret neonatal islets. (B) Absolute 340/380-nm Fura-2 ratios at baseline (Base), peak, and 5 minutes after sequential glucose and high K+ stimulation. (C) Data from (A) were used to calculate the percentage increase over baseline for the 340/380-nm ratios at peak and 5 minutes after glucose addition. (D) CFTRinh-172 (20 µM) reduces insulin secretion from CF and WT neonatal ferret islets in the presence of 16.7 mM glucose. Panels show from left to right, percentage insulin secretion, total secreted insulin, and total islet insulin. (E) CFTRinh-172 reduces peak and 5-minute intracellular Ca2+ after glucose challenge of CF neonatal ferret islets. Data from (B) were normalized to baseline (Rmin = 0) and high K+ (Rmax = 1) 340/380-nm ratios for each islet’s response to glucose for WT and CF islets in the absence of CFTRinh-172. Similar studies were performed with CF islets treated with 20 µM CFTRinh-172 and the high K+ normalized 340/380-nm ratios calculated. Shown are the changes in the high K+ normalized 340/380-nm Fura-2 ratios at peak and 5 minutes after glucose challenge for each group of islets. (F, G) CFTRinh-172 (20 µM) reduces insulin secretion by (F) adult ferret and (G) human islets in the presence of 16.7 mM glucose but did not alter glucagon or PP secretion. (B, C, and E) Data for all graphs show the mean ± standard error of the mean for the indicated number (N) of islets with the given number of independent islet preparations or (D, F, G) N of independent experiments. Each neonatal islet experiment generated islets from three or four newborn ferrets of each genotype, whereas each adult islet experiment used one donor. Statistical analysis was performed by two-tailed Student t test, with significance of *P < 0.05, **P < 0.01, and ***P < 0.001. Comparison in (E) is between CF and CF + CFTRinh-172. DMSO, dimethyl sulfoxide.

Figure 3.

The proinflammatory state of CF ferret neonatal islets affects insulin secretion through IL-6. (A–D) Whole mount α-SMA and desmin immunostaining of (A, B) WT and (C, D) CF newborn ferret pancreata (see Supplemental Materials for staining protocol). Panels (A) and (C) are confocal cross-sections, and panels (B) and (D) are three-dimensional reconstructions. In WT pancreas α-SMA (green) is expressed in vessels between acini. Around acini, there are cells that express desmin (red), which are quiescent PaSCs. By contrast, in CFTR-KO pancreas there are cells that coexpress α-SMA and desmin (yellow), which is a phenotype of activated PaSCs. (E, F) Cultured neonatal islets from (E) WT and (F) CF ferrets have an organized exocrine cellular component. Bright field panels show groups of islets, and the bottom panels show cross-sections of islets immunostained for CK7 (green) with and without the DAPI channel (blue). Micron bars are equal to 20 μm. (G) Enzyme-linked immunosorbent assay analysis of supernatants harvested from cultures of neonatal islets for the indicated cytokines demonstrated a significant elevation of IL-6 in CF islets. (H) Eight-day WT neonatal ferret islets were divided into two equal groups, treated with 1 ng/mL IL-6 or vehicle for 24 hours, and then used for insulin secretion assays in low glucose (1.67 mM) under 1 hour static conditions in the continued presence of 1 ng/mL IL-6 or vehicle. Panels show from left to right, percentage insulin secretion, total secreted insulin, and total islet insulin. Data for all graphs show the mean ± standard error of the mean for the number (N) of experiments in each panel. Each experimental set of islets was derived from three or four independent animals. Statistical analysis was performed by paired two-tailed Student t test, with significance of **P < 0.01.

Figure 4.

Exocrine and endocrine gene expression is altered in cultured neonatal islets from CF ferrets. Cultured CF and non-CF neonatal islets were evaluated for the indicated RNAs and the housekeeping RNA PPIB via multiplex assays that simultaneously assess multiple targets in each sample. The ratio of each target to PPIB RNA was used to calculate the relative fold change in expression between CF and non-CF while normalizing the average of the non-CF sample to 1. (A) Analysis of endocrine marker gene demonstrated that INS and GCG were significantly reduced in CF islets. (B) Analysis of genes that influence β-cell function demonstrated that SLC2A1 (Glut1) was significantly upregulated in CF islets. (C) Analysis of genes involved in fibrosis and remodeling demonstrated that ACTA1 and TGFB1 were significantly upregulated in CF islets, whereas ACTA2 was nonsignificantly increased. (D) Analysis of exocrine cell marker genes demonstrated upregulation of KRT7 (CK7) but lower expression the acinar cell marker CELA3b (fecal elastase) in CF islets. (E) Analysis of genes involved in inflammation demonstrated no significant changes, but CXCL10 and IL6 trended higher in CF. (F) Analysis of genes involved in β-cell proliferation and regeneration demonstrated significant upregulation of SERPINB1, PAX4, NEUROD1, and REG3a in CF islets. (G) Analysis of genes involved in the transcriptional regulation of β-cell maturation demonstrated significant upregulation of NEUROD1 and NEUROG3 in CF islets. Data for all graphs show the mean ± standard error of the mean for the number (N) of islet preparations evaluated, as shown in each panel. Each preparation of islets (N) was derived from three or four independent animals. Statistical analysis was performed by two-tailed Student t test, with significance of *P < 0.01, **P < 0.01.

Figure 5.

CFTR RNA localizes to KRT7+ exocrine cells, but not endocrine cell types, of the newborn ferret pancreas. (A–D) Colocalization of CFTR and INS transcripts using smFISH in (A, C) WT and (B, D) CF newborn ferret pancreas. (A, B) Fluorescent images demonstrating INS (red) and CFTR (green) localization with DAPI as a nuclear stain. (C, D) Bright field images demonstrating INS (blue) and CFTR (red) localization with no counterstain. CF animal lack detectable CFTR transcripts. (E–I) High power fluorescent images of CFTR colocalization with (E) INS, (F) GCG, (G) PPY, (H) SST, and (I) KRT7. Single channels for CFTR and each cellular marker are shown to the right of each merged image. Micron bars: (A, B) 150 μm; (C, D) 60 μm; (E–I) 30 μm. Results are representative of four experiments on independent animals for each genotype.

Figure 6.

CFTR RNA localizes to KRT7+ exocrine cells, but not endocrine cell types, of isolated ferret and human islets. Isolated adult WT ferret and human islets were dissociated with trypsin and used for smFISH localizing CFTR RNA with other islet cell marker RNAs. (A–E) CFTR RNA (green) is colocalization with (A) INS, (B) GCG, (C) PPY, (D) SST, and (E) KRT7 RNA (red) in cells from ferret islets. (F–J) CFTR colocalization with (F) INS, (G) GCG, (H) PPY, (I) SST, and (J) KRT7 in cells from human islets. The boxed regions (i and ii) are shown below each main panel (A–D) and (F–I) as single-channel images for the CFTR. To better demonstrate colocalization of CFTR and KRT7, merged and single-channel images are shown for each transcript in (E) and (J). Micron bars are 10 μm. Results are representative of N > 4 independent islet donors for each species.