Microtubules and Gαo-signaling modulate the preferential secretion of young insulin secretory granules in islet β cells via independent pathways

Abstract

For sustainable function, each pancreatic islet β cell maintains thousands of insulin secretory granules (SGs) at all times. Glucose stimulation induces the secretion of a small portion of these SGs and simultaneously boosts SG biosynthesis to sustain this stock. The failure of these processes, often induced by sustained high-insulin output, results in type 2 diabetes. Intriguingly, young insulin SGs are more likely secreted during glucose-stimulated insulin secretion (GSIS) for unknown reasons, while older SGs tend to lose releasability and be degraded. Here, we examine the roles of microtubule (MT) and Gαo-signaling in regulating the preferential secretion of young versus old SGs. We show that both MT-destabilization and Gαo inactivation results in more SGs localization near plasma membrane (PM) despite higher levels of GSIS and reduced SG biosynthesis. Intriguingly, MT-destabilization or Gαo-inactivation results in higher secretion probabilities of older SGs, while combining both having additive effects on boosting GSIS. Lastly, Gαo inactivation does not detectably destabilize the β-cell MT network. These findings suggest that Gαo and MT can modulate the preferential release of younger insulin SGs via largely parallel pathways.

Fig 1. MTs are dispensable for insulin SGs localization near β-cell periphery.

(A) MT-destabilization by NOC-treatment [A1, 2.8 mM glucose (G2.8) plus DMSO; A2, G2.8 + NOC; A3, (G20 + DMSO); A4, (G20 + NOC)]. Images shown on the top row are maxi-projections of whole mount stained islets, showing co-staining of insulin (red), E-cadherin (E-Cad, blue), and α-tubulin (green). The bottom row includes high-magnification of boxed areas of the top panels, highlighting MT-destruction by NOC. (B) GSIS in islets treated with NOC for 12-hours, scatter plots with mean shown. (C-F) Insulin–subcellular localization (red) in β cells. In (C1-F1), DAPI and Ecadherin signals locate nuclei and cell membrane. A few Ins-negative cells were marked in C, D, and F (white broken circles). (C2, D2) The scheme to assay relative insulin levels in β cells: clusters of β cells showing clear nuclear signals were selected [green circles in C2 (cells 1–4), D2 (cells 5–7)] after excluding the nuclear areas, white circle. (E2, F2) The scheme to examine subcellular insulin localization. Β-cell zones along the PM (±1 micrometer, grey bands) were marked and the portion of insulin signals that localize within these zones were determined. (G, H) Relative insulin levels and the portions of insulin that localized near PM, shown as mean + SEM. In (B), (G), (H), #: p>0.12, **: p<0.01, ***: p<0.001, calculated using multi-comparison ANOVA. Please refer to S1 Table for original numbers used in plots.

Fig 2. TEM assay of insulin SG in β cells in the absence of MTs.

Islets were isolated from wild-type (WT) adult mice and were treated with DMSO or 10 μg/ml NOC for ~12 hour at G2.8 or G20 mM. TEM was used to examine the locations and density of SGs in two batches of islets prepared on different days. (A-C) TEM images and quantification of SGs that localize near PM from DMSO- and NOC-treated islets in the presence of 2.8 mM glucose. (D-F). Images and quantification of SG density in microscopic fields used in (A-C). (G-J) Images and SG quantification as in (A-F), except 20 mM glucose was used. Scales in (D, E, G, H) are the same, labeled in panel (H). In (C), (F), (I), and (J), mean + SEM were presented (*: p<0.05, ***: p<0.001, from two-tailed type II t-test). In all panels, “n” indicates the number of microscopic fields counted (with 3–4 different β cells included in each field). In all images, red dashed lines mark recognizable β cell membrane. Please refer to S1 Table for original numbers used in plots.

Fig 3. Newly synthesized SGs are dispensable for MT-destabilization-enhanced GSIS.

(A—D) Glu-tubulin was stained to verify the effective disruption of MTs by NOC (A, B), with β cells identified by insulin staining (C, D, showing the corresponding fields in A, B, respectively). Arrows in (A, B), primary cilia. Scale bar = 10 μm. (E) Total radioactive amino acid (3H-Leu/Ile) incorporation after 3-hours CHX-treatment. The radioactivity was normalized against DNA levels, compared via real-time PCR. Also see S1 Fig for uncliped western images. (F) GSIS from islets treated with combination of 10 μg/ml NOC and 10 μM CHX. Note that the presence of CHX did not eliminate the NOC-potentiated GSIS (compare columns 2 and 4). (G) The radioactive insulin that were secreted (count/ng insulin) in control and NOC-treated islets, following a 4-hour radiolabeling process. (H) % of total insulin secretion in samples used in panel (G). In (E-H) *: p<0.05; **: p<0.01; ***: p<0.001, values from Holm-Sidak’s multiple comparisons (E, G, H) or Turkey’s multiple comparisons (F). The p-values, not shown between low-glucose samples, are all above 0.3. Please refer to S1 Table for original numbers used in plots.

Fig 4. Gαo and MT regulate SG secretion via parallel pathways.

(A-E) Immunofluorescence and western blot showing Gαo inactivation in Gαo^F/F; Pdx1^Cre islet cells. Note that Cre-mediated Gαo^F deletion will yield a mRNA that translates a short N-terminal Gαo peptide, recognized by the antibody (red) but has no detectable biological effect [18]. Full length funtional Gαo is membrane-bound, while the N-terminal fragment is cytoplasmic, allowing for ready verification of Pdx1^Cre-mediated Gαo inactivation in insulin+ (green) cells. DAPI (blue) stained for nuclei. Scale bar, 20 μm. In (E), note the disapearance of Gαo full length protein and the appearance of a short Gαo fragment in the Gαo^F/F; Pdx1^Cre islet samples. Alpha-tubulin was used as loading controls for the western blot. (F) The levels of ³⁵S-labeling in secreted insulin from control and *: p<0.05; **: p<0.01; ***: p<0.001)β cells, prelabeled for four hours. (G) Insulin secretion from Gαo^F/F; Pdx1^Cre islets, with or without NOC-treatment, induced by basal G2.8, stimulating G20, and KCl-induced depolarization [(A, B) G20 + K25 (25 mM KCl)]. P values, results from multi-comparison ANOVA or one-way ANOVA (#: p>0.3. *: p<0.05; **: p<0.01; ***: p<0.001). Note the red font “#” or “*” in (G) highlight the differences between control and Gαo^F/F; Pdx1^Cre islets. Please refer to S1 Table for original numbers used in plots.

Fig 5. Inactivating Gαo does not alter MT stability and density in β cells.

(A-E) The MT density in control and Gαo^F/F; Pdx1^Cre β cells, stained for tubulin (green), insulin (red), and E-cadherin (blue). Single cells attacehd to coverslips (A, C) or intact islets (B, D) were used. Insets in (A), (C), insulin staining to identify β-cells. Note that in (B, D), a single α-tubulin and a composite channel are presetented. The quantification data in E is MT density in single β cells, which measured the average distances between microtubules at any raodom direction. (F-H) The density of Glu-tubulin in control and Gαo^F/F; Pdx1^Cre β cells. Glu-tubulin (green), E-Cadherin (red), and DAPI (blue) signals via immunofluorescence are shown. Presented data in (E) and (H) are (mean + SEM). #: p>0.2, results from two-tail type II t-test. Scale bars, 5 μm. Please refer to S1 Table for original numbers used in plots.