Molecular interaction between type 2 diabetes and Alzheimer's disease through cross-seeding of protein misfolding

Abstract

Numerous epidemiological studies have shown a significantly higher risk for development of Alzheimer's disease (AD) in patients affected by type 2 diabetes (T2D), but the molecular mechanism responsible for this association is presently unknown. Both diseases are considered protein misfolding disorders associated with the accumulation of protein aggregates; amyloid-beta (Aβ) and tau in the brain during AD, and islet amyloid polypeptide (IAPP) in pancreatic islets in T2D. Formation and accumulation of these proteins follows a seeding-nucleation model, where a misfolded aggregate or 'seed' promotes the rapid misfolding and aggregation of the native protein. Our underlying hypothesis is that misfolded IAPP produced in T2D potentiates AD pathology by cross-seeding Aβ, providing a molecular explanation for the link between these diseases. Here, we examined how misfolded IAPP affects Aβ aggregation and AD pathology in vitro and in vivo. We observed that addition of IAPP seeds accelerates Aβ aggregation in vitro in a seeding-like manner and the resulting fibrils are composed of both peptides. Transgenic animals expressing both human proteins exhibited exacerbated AD-like pathology compared with AD transgenic mice or AD transgenic animals with type 1 diabetes (T1D). Remarkably, IAPP colocalized with amyloid plaques in brain parenchymal deposits, suggesting that these peptides may directly interact and aggravate the disease. Furthermore, inoculation of pancreatic IAPP aggregates into the brains of AD transgenic mice resulted in more severe AD pathology and significantly greater memory impairments than untreated animals. These data provide a proof-of-concept for a new disease mechanism involving the interaction of misfolded proteins through cross-seeding events which may contribute to accelerate or exacerbate disease pathogenesis. Our findings could shed light on understanding the linkage between T2D and AD, two of the most prevalent protein misfolding disorders.

Figure 1. In vitro Aβ aggregation is enhanced by IAPP heterologous seeding

A: Soluble seed-free Aβ40 (2000 nM) in 100 mM Tris-HCl pH 7.4 (200 μl) containing 5 μM thioflavin T (ThT) was allowed to aggregate at 22°C with intermittent shaking at 500 rpm in the absence or in the presence of either 20 nM Aβ40 seeds (homologous seeding) or 20 nM IAPP seeds (heterologous seeding). The extent of aggregation was periodically monitored by measuring the ThT fluorescence at excitation of 435 nm and emission of 485 nm. Error bars indicate standard deviation (S.D.) B: Transmission electron microscopy microphotography of double immuno-gold labeling of fibrils formed by the addition of IAPP seeds (10% with respect to Aβ) into seed-free Aβ1–42 (37 μM) aggregation solution. After in vitro cross-seeding, fibrils were stained using antibodies coupled to gold nanoparticles. IAPP was detected with an antibody coupled to 12 nm gold particles while Aβ was labeled with 6 nm gold particles. Scale bar: 100 nm. Red arrows point to aggregates showing staining for both peptides.

Figure 2. IAPP^+/+ × APP mice display increased Aβ-immunopositive deposition in the brain

Aβ deposition was analyzed in the brain of transgenic animals over-expressing human IAPP and APP and control animals by immunohistological staining using 4G8 antibody. n=5–10 animals/group; 5 sections/animal. A–B: Brain Aβ burden was quantified as the immune-reactive area per total area analyzed in IAPP^+/+ × APP, IAPP^+/− × APP, APP animals injected with STZ, and untreated APP mice in hippocampal and cortical areas. C–D: Amyloid plaque density was measured as the number of plaques per mm² in the hippocampus and cortices. E: Plaques bigger than 500 μm² were quantified in the brain of experimental and control groups. Data in panel A–E was analyzed by one-way ANOVA, followed by the Tukey’s multiple comparison post-hoc test. *p<0.05; **p<0.01; ***p<0.001. F: Representative pictures of amyloid plaques in the cortical area reactive to human anti-Aβ antibody 82E1 in analyzed groups. G: Double immune-staining of IAPP (in red) and Aβ (in green), and co-localization (yellow) in the cortical area of IAPP^+/+ × APP transgenic mice. Scale bar: 50 μm.

Figure 3. Aβ pathology is aggravated by exogenous addition of pancreas homogenates containing IAPP aggregates

A: Representative microphotography of fluorescent immuno-labeled Aβ in the brain of APP animals intracerebrally inoculated with IAPP-PH, WT STZ-injected PH and untreated APP animals. Scale bar: 25 μm. B–C: Immunohistochemical quantification of Aβ burden (immune-reactive area per total area analyzed) in cortical and hippocampal brain areas in experimental and control groups. D–E: The number of plaques per mm² was quantified to analyze differences in the amyloid plaque density in the brain of APP mice inoculated with IAPP-PH, WT STZ-injected PH and untreated APP animals. F: The density of plaques bigger than 500 μm² per mm² was quantified in the cortical area of experimental and control groups. G-H: Whole brain homogenate of APP mice intracerebrally inoculated with IAPP-PH, WT STZ-injected PH, and untreated APP animals was fractionated by ultracentrifugation. PBS and formic acid soluble fractions containing either oligomeric soluble (panel G) or fibrillar insoluble Aβ (panel H) were extracted and quantified using an ELISA kit to measure Aβ42 protein concentration per gram of tissue. Data in panels B–H was statistically analyzed by One-way ANOVA, followed by the Tukey’s multiple comparison test, *p<0.05; **p<0.01; ***p<0.001. n=5–10 animals/group; 5 sections/animal.

Figure 4. APP animals inoculated with pancreas homogenate containing aggregated IAPP display learning and memory deficits

Learning and memory was measured by Barnes maze, a spatial working memory, hippocampal dependent task that measures the spatial navigation and memory of an animal. A: Learning was measured in experimental and control groups at 8 months old in 5–10 animals for five consecutive days as the primary latency in seconds. Two-way ANOVA, Tukey’s: *p<0.05, **p<0.01 compared to untreated APP; ^#p<0.05, ^##p<0.01 compared to inoculated WT. B: Short term memory was measured as the time in seconds the animals take to reach the escaping box (latency) in the day 5 after the learning period. One-way ANOVA, Tukey’s: **p<0.01.