Predisposition to Proinsulin Misfolding as a Genetic Risk to Diet-Induced Diabetes

Abstract

Throughout evolution, proinsulin has exhibited significant sequence variation in both C-peptide and insulin moieties. As the proinsulin coding sequence evolves, the gene product continues to be under selection pressure both for ultimate insulin bioactivity and for the ability of proinsulin to be folded for export through the secretory pathway of pancreatic β-cells. The substitution proinsulin-R(B22)E is known to yield a bioactive insulin, although R(B22)Q has been reported as a mutation that falls within the spectrum of mutant INS-gene-induced diabetes of youth. Here, we have studied mice expressing heterozygous (or homozygous) proinsulin-R(B22)E knocked into the Ins2 locus. Neither females nor males bearing the heterozygous mutation developed diabetes at any age examined, but subtle evidence of increased proinsulin misfolding in the endoplasmic reticulum is demonstrable in isolated islets from the heterozygotes. Moreover, males have indications of glucose intolerance, and within a few weeks of exposure to a high-fat diet, they developed frank diabetes. Diabetes was more severe in homozygotes, and the development of disease paralleled a progressive heterogeneity of β-cells with increasing fractions of proinsulin-rich/insulin-poor cells as well as glucagon-positive cells. Evidently, subthreshold predisposition to proinsulin misfolding can go undetected but provides genetic susceptibility to diet-induced β-cell failure.

Figure 1

Insulin B-chain substitution R(B22)E. A: Schematic representation of insulin chains showing the respective disulfide bonds and the R(B22)E substitution. B: Transfection of Min6 cells with empty vector (EV), untagged WT human proinsulin, hPro-R(B22)E-CpepMyc, or hPro-C(A7)Y-CpepMyc (Akita-Myc). The media-bathing transfected cells were collected overnight, and both cell lysates and media were resolved by SDS-PAGE under reducing conditions and immunoblotting with anti-human-proinsulin or cyclophilin B (CypB) (loading control) (bottom panels). Cell lysates were similarly analyzed under nonreducing conditions to reveal disulfide-linked proinsulin complexes (top panel). C: Transfection of Min6 cells as in B (first four lanes) or cotransfection with both untagged WT human proinsulin and hPro-R(B22)E-CpepMyc (last lane). Proinsulin-R(B22)E is secreted from Min6 cells, and under these conditions, proinsulin-R(B22)E does not block the secretion of coexpressed untagged WT proinsulin.

Figure 2

Heterozygous (Het) Ins2-proinsulin-R(B22)E mice fed normal chow (NC) or HFD. A: Weekly body weight measurements in male mice (n = 4–8 per group). B: Intraperitoneal glucose tolerance test in 5.5-week-old males on NC (n = 4–5 per group). C: Area under the curve (AUC) from the data in B. D: Weekly random blood glucose in males on NC or HFD (n = 4–8 per group), with Ins2-proinsulin-R(B22)E-Het on NC shown in blue and on HFD in red. E: Intraperitoneal glucose tolerance test in 11-week-old males on NC or HFD for 6 weeks (n = 3–5 per group), with Ins2-proinsulin-R(B22)E-Het mice on NC in blue and on HFD in red. F: AUC from the data in E. G: Random serum insulin level in 11-week-old WT (purple) or Ins2-proinsulin-R(B22)E-Het (red) males on HFD for 6 weeks (n = 5–7 per group). H: The data from G used to calculate serum insulin-to-glucose ratio. I: Insulin secretion from isolated islets of WT (purple) or Ins2-proinsulin-R(B22)E-Het (red) males on HFD at 2.8 mmol/L (unstimulated) and 16.7 mmol/L (stimulated) glucose (n = 3–4 per group). J: Isolated islets from male WT and Ins2-proinsulin-R(B22)E-Het mice fed NC or HFD were lysed and analyzed by reducing SDS-PAGE and immunoblotting with monoclonal antibody anti-proinsulin (top panel), guinea pig anti-insulin (middle panel), and anti-cyclophilin B (CypB) (loading control) (bottom panel). K: Weekly body weight measurements in female Ins2-proinsulin-R(B22)E-Het mice fed NC (blue) or HFD (green) (n = 4 per group). L: Weekly random blood glucose measurements in female Ins2-proinsulin-R(B22)E-Het mice on NC (blue) or on HFD (green) (n = 4 per group). Data are mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001 by one-way ANOVA.

Figure 3

Homozygous (Hom) Ins2-proinsulin-R(B22)E mice develop spontaneous diabetes. A: Random blood glucose at 5 weeks of age in WT (black), heterozygous (Het) (blue), and Hom (red) Ins2-proinsulin-R(B22)E mice fed normal chow (n = 3–7 per group). Males and females shown separately as well as combined. B: The same mice from A remeasured 8 weeks postpartum. C: Serum insulin level measured at 5.5 weeks of age in WT (black), Het (blue), or Hom (red) males fed normal chow (n = 5 per group). Males and females shown separately as well as combined. D: Insulin-to-glucose ratio calculated from the samples in B and C, respectively. E: GSIS in vivo at 0 and 15 min poststimulation from WT (black), Het (blue), or Hom (red) Ins2-proinsulin-R(B22)E mice fed normal chow (6-week-old males, n = 5–6 per group). F: GSIS from islets isolated from WT (black), Het (blue), or Hom (red) Ins2-proinsulin-R(B22)E mice fed normal chow at 2.8 mmol/L (unstimulated) and 16.7 mmol/L (stimulated) glucose concentration (6-week-old males, n = 5–7 per group). G: Serum proinsulin level from the same animals measured in C (males and females combined). H: Proinsulin-to-insulin ratio from the animals in G (males and females combined). Data are mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001 by one-way ANOVA.

Figure 4

Proinsulin and insulin double immunofluorescence showing random blood glucose (BG) at the time of euthanasia. A: WT control. B: Ins2-proinsulin-R(B22)E heterozygote (Het) (6-week-old female). C: Normoglycemic 4-week-old female Ins2-proinsulin-R(B22)E homozygote (Hom). D and E: Ins2-proinsulin-R(B22)E-Hom 7-week-old female. F: Quantitation of insulin-positive cytoplasmic area as a percentage of total β-cell cytoplasmic area in independent islets from the genotypes indicated (mean ± SD; P = 0.0002). ***P < 0.001.

Figure 5

Transmission electron microscopy of islet cells from a 4-week-old male Ins2-proinsulin-R(B22)E homozygote (Hom), highlighting heterogeneity in the abundance of insulin secretory granules. A: Lower power view, with β-cells identified. The cell at the bottom and the two cells in the upper left corner exhibit glucagon secretory granules; random blood glucose (BG) was taken at the time of euthanasia. B: The yellow boxed area from A shown at higher power. Vesiculotubular clusters (VTCs) are a pre-Golgi compartment. A more detailed electron microscopy survey is presented in Supplementary Fig. 3. C: For each indicated genotype and diet, total β-cell granule numbers were counted per unit area (25-µm² boxed areas). Those with a classic electron-dense granule core were considered to be mature (green), and those lacking the classic electron-dense core were considered to be underfilled (red). Quantification compared HFD-fed (H) WT and heterozygous (Het) conditions and normal chow (N) diet–fed Het and Hom conditions. Data are mean ± SD. **P < 0.01, ****P < 0.0001. Mito, mitochondrion.

Figure 6

Growing subpopulations of proinsulin-enriched cells (red) and glucagon-enriched cells (blue) and a decline of insulin-enriched cells (green) in Ins2-proinsulin-R(B22)E heterozygotes (Het) and homozygotes (Hom), as identified by triple immunofluorescence. Random blood glucose levels are indicated. A: WT 6-week-old female. B and C: Ins2-proinsulin-R(B22)E-Het females (age 6.5 weeks). D and E: Ins2-proinsulin-R(B22)E-Hom females (ages 4 and 6 weeks, respectively). Proins, proinsulin.

Figure 7

Progressive loss of insulin, accompanied by proinsulin misfolding, in Ins2-proinsulin-R(B22)E homozygotes (Hom). A: Reducing SDS-PAGE and immunoblotting with anti-proinsulin (top panel), anti-insulin (middle panel), and anti-cyclophilin B (CypB) (loading control, bottom panel). Islets were isolated from males (ages 5–7 weeks) with the genotypes shown above; random blood glucose level at the time of euthanasia is also indicated. B: Nonreducing SDS-PAGE and immunoblotting of the identical samples as that shown in A, highlighting aberrant proinsulin disulfide-linked complexes. C: Immunoblotting of p58ipk and BiP from male WT and Ins2-proinsulin-R(B22)E heterozygotes (Het) and Hom (age 5–6 weeks). D: Quantitation of immunoblotting like that shown in C (n = 6 per group). Statistical analysis (mean ± SD) shows that changes in normalized BiP protein levels were nonsignificant. **P < 0.01. Proins, proinsulin.

Figure 8

Biosynthesis of proinsulin and insulin in WT control and Ins2-proinsulin-R(B22)E heterozygous (Het) and homozygous (Hom) males with progression of diabetes (age 4–8 weeks). The random blood glucose of each animal at the time of euthanasia is indicated. Isolated islets were pulse labeled with ³⁵S-amino acids for 30 min and then chased for 2 h, as indicated. Cells (C) and chase media (M) were either combined (so that no protein was lost) or analyzed separately. Samples (normalized to trichloroacetic acid–precipitable counts in the cell lysates) were immunoprecipitated with anti-insulin followed by Tris-tricine-urea-SDS-PAGE under nonreducing or reducing conditions followed by fluorography. A line is drawn separating the nonreduced and reduced samples, but these images show the complete gels, and no lanes have been excised. A–C: Three independent experiments with the genotypes shown (ages 4, 5, and 8 weeks, respectively). D: Quantitative recovery of newly synthesized insulin derived from pulse-labeled proinsulin, as derived from the preceding phosphorimages; genotype and random blood glucose are indicated. Proinsulin bands (at chase time 0) were quantitated from reducing gels; newly synthesized insulin derived from the pulse-labeled samples were quantified from nonreducing gels. (Insulin is a two-chain protein that “falls apart” under reducing conditions; thus, nonreducing gels are preferable for this analysis.) ox, oxidized; R, reduced.