Human islet amyloid polypeptide-induced β-cell cytotoxicity is linked to formation of α-sheet structure

Abstract

Type 2 diabetes (T2D) results from insulin secretory dysfunction arising in part from the loss of pancreatic islet β-cells. Several factors contribute to β-cell loss, including islet amyloid formation, which is observed in over 90% of individuals with T2D. The amyloid is comprised of human islet amyloid polypeptide (hIAPP). Here we provide evidence that early in aggregation, hIAPP forms toxic oligomers prior to formation of amyloid fibrils. The toxic oligomers contain α-sheet secondary structure, a nonstandard secondary structure associated with toxic oligomers in other amyloid diseases. De novo, synthetic α-sheet compounds designed to be nontoxic and complementary to the α-sheet structure in the toxic oligomers inhibit hIAPP aggregation and neutralize oligomer-mediated cytotoxicity in cell-based assays. In vivo administration of an α-sheet design to mice for 4 weeks revealed no evidence of toxicity nor did it elicit an immune response. Furthermore, the α-sheet designs reduced endogenous islet amyloid formation and mitigation of amyloid-associated β-cell loss in cultured islets isolated from an hIAPP transgenic mouse model of islet amyloidosis. Characterization of the involvement of α-sheet in early aggregation of hIAPP and oligomer toxicity contributes to elucidation of the molecular mechanisms underlying amyloid-associated β-cell loss.

Keywords: islet amyloid; islet amyloid polypeptide; toxic soluble oligomers; type 2 diabetes; α-sheet.

FIGURE 1

Characterization of α‐sheet involvement in aggregation. (a) PBS‐treated amyloidogenic human IAPP (hIAPP) samples were monitored with undisturbed endpoint measurements at two different concentrations (15 and 25 μM) at 25°C. PBS‐treated rodent IAPP (rIAPP; 25 μM) was included as a non‐amyloidogenic control. hIAPP followed the same process of conformational conversion at both concentrations, but the kinetics were faster at the higher concentration. rIAPP did not exhibit conformational conversion. Each time point measurement represents mean and standard deviation. N = 3 per condition. (b) Circular dichroism spectroscopy analysis of PBS‐treated 15 μM hIAPP sampled at different times during aggregation. Spectra for 0 h incubation hIAPP (black) are consistent with a disordered, random coil conformation for the native monomeric state. 2.5 h (red, dotted) and 3.5 h (red, solid) incubation times result in a lift of the curves to essentially featureless “null” spectra. Additional incubation produces a drop near 218 nm consistent with β‐sheet structure. (c) CD spectra of 15 μM rIAPP as a function of time, overlaid spectra with different colors that remain random coil over time. The spectrum for the α‐sheet AP407 model compound (magenta) is provided for comparison along with its NMR structure. (d) Toxicity time‐course of hIAPP. MTT cell viability assay of INS‐1 pancreatic β‐cells exposed to pre‐incubated hIAPP from different incubation duration (15 μM in PBS at 25°C diluted to 5 μM for application to cells). Toxicity is normalized, the maximum toxicity point corresponded to 60% cell viability. N = 2–6 per timepoint. The following abbreviations are used in the figure: rc = random coil; αs = α‐sheet; βs = β‐sheet.

FIGURE 2

Binding of α‐sheet peptides to hIAPP. (a) Inhibition of hIAPP aggregation by AP5 and AP90 and lack of inhibition with the P1 peptide. Designed peptides were co‐incubated with hIAPP at 1:4 molar ratio (hIAPP: AP peptide). Data bars represent standard deviation. N = 8 for hIAPP, AP5, and AP90 conditions. N = 3 for P1. For all panels, *p < 0.05, **p > 0.01, ***p < 0.001, ns ≥ 0.05, and error bars represent standard deviations. (b) Insulinoma cell viability when incubated with hIAPP in the absence of AP5 (gray) and in the presence of AP5 (red). AP5 significantly neutralized the toxicity of oligomeric hIAPP (2.5:1 molar ratio, 37.5 μM AP5 and 15 μM IAPP) preincubated for 24 (p = 0.0006) and 48 (p = 0.004) hours (p‐values relative to untreated hIAPP). Vertical bars represent standard deviation. N = 3–6 per condition. (c) SH‐SY5Y neuroblastoma cell viability when incubated with hIAPP in the absence of AP5 (black) and the presence of AP5 (red). Oligomeric IAPP (50 μM, 1.5 h preincubation, corresponding to formation of α‐sheet oligomers at this high concentration) reduced cell viability (p = 0.0002 relative to medium with vehicle), and the addition of AP5 (1:1 molar ratio) resulted in significant recovery of cell viability (p = 0.0004, relative to untreated hIAPP). Vertical bars represent standard deviation. N = 3 per condition. (d) The soluble oligomer binding assay of hIAPP sampled at different times during aggregation. The hIAPP samples (15 μM) were incubated and then diluted to 5 μM immediately before the assay. Higher absorbance values denote higher amounts of hIAPP α‐sheet oligomers. hIAPP at 20 h (p = 0.0002) and 40 h (p = 0.0007) showed significant binding compared to 0 h. Vertical bars represent standard deviation. N = 4 per condition. ***p < 0.001 vs. 0‐h sample. (e) Binding of hIAPP α‐sheet oligomers in the SOBA assay is correlated with toxicity (given as the inverse of the cell viability).

FIGURE 3

Histological examination of wild‐type mouse islets co‐incubated without or with α‐sheet peptides. Representative images of wild type mouse islets treated with (a) vehicle, (b) 100 μM AP5 or (c) 100 μM AP90. Blue staining represents cell nuclei and red staining represents insulin. (d) β‐cell area (insulin area/islet area, %) and (e) islet area (mm²) were quantified from wild‐type mouse islets treated without or with α‐sheet peptides. Vertical bars represent standard error of the mean. N = 3 per condition. ns p > 0.05 vs. untreated islets.

FIGURE 4

Immune response to AP90 after in vivo peptide administration. Following in vivo administration of saline (white bars), 100 μM AP90 (light blue bars), or 1 mM AP90 (dark blue bars) to mice for 4 weeks, peripheral blood mononuclear cells were isolated from spleens and incubated alone (unstimulated), with AP90 (stimulated, 6.25, 12.5, 25, or 50 μg/mL), or with concanavalin A (Con A, 10 μg/mL). Unstimulated and stimulated 18 h‐³H‐thymidine incorporation was determined by scintillation counting. Stimulation index (SI) was calculated as mean counts per minute (CPM) stimulated/unstimulated. Vertical bars represent standard error of the mean. **p < 0.01 vs. all other groups.

FIGURE 5

Access of amyloid‐inhibiting peptides to the intra‐islet space. Amyloid‐prone hIAPP transgenic islets were cultured in the absence (left) or presence of 100 μM fluorescein labeled AP5 (center) or AP90 (right) peptide. Following 48 h of culture under these conditions, islets were transferred to media lacking peptide and subjected to brightfield (top) and confocal (bottom) microscopy (Leica SP5, 10× magnification). Fluorescein‐labeled peptides were imaged by excitation at 488 nm.

FIGURE 6

Histological examination of hIAPP transgenic mouse islets expressing hIAPP co‐incubated with or without α‐sheet peptides. Amyloid‐prone hIAPP transgenic mouse islets were cultured in 16.7 mM glucose (an amyloidogenic condition) without (hIAPP) or with 100 μM AP90 (+AP90) for 144 h. (a) Representative images of hIAPP transgenic mouse islets with and without AP90 treatment. Blue staining represents cell nuclei, green amyloid deposits, and red insulin. Islets were subjected to quantitative microscopy for determination of (b) amyloid prevalence (% amyloid positive islets), (c) amyloid severity (amyloid area/islet area, %), (d) β‐cell area (insulin area/islet area, %), and (e) islet area (mm²). Vertical bars represent standard error of the mean. N = 5 per condition. *p < 0.05 or **p < 0.01 or ns p > 0.05 vs. untreated islets.