The association of circulating amylin with β-amyloid in familial Alzheimer's disease

Abstract

Introduction: This study assessed the hypothesis that circulating human amylin (amyloid-forming) cross-seeds with amyloid beta (Aβ) in early Alzheimer's disease (AD).

Methods: Evidence of amylin-AD pathology interaction was tested in brains of 31 familial AD mutation carriers and 20 cognitively unaffected individuals, in cerebrospinal fluid (CSF) (98 diseased and 117 control samples) and in genetic databases. For functional testing, we genetically manipulated amylin secretion in APP/PS1 and non-APP/PS1 rats.

Results: Amylin-Aβ cross-seeding was identified in AD brains. High CSF amylin levels were associated with decreased CSF Aβ₄₂ concentrations. AD risk and amylin gene are not correlated. Suppressed amylin secretion protected APP/PS1 rats against AD-associated effects. In contrast, hypersecretion or intravenous injection of human amylin in APP/PS1 rats exacerbated AD-like pathology through disruption of CSF-brain Aβ exchange and amylin-Aβ cross-seeding.

Discussion: These findings strengthened the hypothesis of circulating amylin-AD interaction and suggest that modulation of blood amylin levels may alter Aβ-related pathology/symptoms.

Keywords: amylin; amyloid; familial Alzheimer's disease; islet amyloid polypeptide; sporadic Alzheimer's disease.

FIGURE 1

Brain amylin–amyloid beta (Aβ) relationship in PSEN1 and APP mutation carriers. A, Schematic amylin–Aβ cross‐seeding hypothesis. B‐E, Amylin and Aβ₄₂ levels and the relationship between amylin and Aβ₄₂ concentrations in temporal cortex specimens from PS1 and APP mutation carriers (n = 11 and n = 7, respectively) and cognitively unaffected individuals (CU; n = 12). F‐M, Representative images of immunohistochemistry analysis using anti‐amylin (brown) and anti‐Aβ (green) antibodies on temporal cortex slices from familial Alzheimer's disease (fAD) brains. Amylin deposition in neurons (F, G; arrowheads) or forming homogenous plaques (F, G; arrows; H, I) and amylin–Aβ plaques (J‐M) are shown. N‐Q, Confocal microscopic analysis of a fAD brain section triple stained with Thioflavin S (ThioS, green; N), anti‐amylin antibody (red; O), and anti‐Aβ antibody (magenta; P). R‐U, representative images of consecutive fAD brain sections stained with Congo red (R), anti‐Aβ (S), anti‐amylin (T), or a combination of anti‐amylin (brown) and anti‐Aβ (green; U) antibodies. V‐Y, Confocal microscopic analysis of a fAD brain section triple stained with anti‐amylin antibody (green; V), anti‐Aβ antibody (red; W) and anti‐α smooth muscle cell (SMC) actin antibody (blue; X). Type of PS1 or APP mutation is labeled on each image; three sections/brain, n = 27 fAD brains and n = 27 CU brains. Scale bars, 50 μm (F‐M, R‐U), 10 μm (N‐Q) and (V‐Y). Data are means ± standard error of the mean; P ≤ 0.01 **, P ≤ 0.0001 ****; one‐way analysis of variance with Dunnett's multiple comparisons test (B, C); correlation analysis (D, E)

FIGURE 2

Cerebrospinal fluid (CSF) amylin–amyloid beta (Aβ)₄₂ relationship in Alzheimer's disease (AD), mild cognitive impairment (MCI), and cognitively unaffected (CU) humans. A, B, Levels of CSF amylin (A) and CSF Aβ₄₂ (B) in patients with sporadic AD (sAD, n = 28); individuals with MCI (n = 70), and in CU individuals (n = 117). C‐E, Correlation analysis of CSF amylin levels versus CSF Aβ₄₂ levels in patients from (A, B). Data are presented as box and whisker plot (A, B) and as individual values (C‐E); P ≤ 0.05 *, P ≤ 0.01 **, P ≤ 0.001 ***, P ≤ 0.0001 ****; two‐tailed, unpaired t test (A, B); correlation analysis (C‐E)

FIGURE 3

Behavior deficits in APP/PS1 and non‐APP/PS1 rats with genetically manipulated amylin secretion. A‐C, Schematic summary of rat models generated for this study and time points of behavior and molecular analyses. Genetic manipulation of circulating amylin was performed in rats with and without Alzheimer's disease (AD)‐like pathology through transgenic expression of human amylin in the pancreas (red color code, A) and by deletion of the amylin gene (light blue color code, B). The time points for behavioral studies were chosen based on previous reports that amylin dyshomeostasis develops at ≈12 months of age in human amylin expressing rats ⁷ , ¹⁵ and that amyloid beta (Aβ) pathology is developed in APP/PS1 rats at about 16 months of age ¹⁶ (C). D‐G, Longitudinal neurological scores of cognitive function (assessed from Novel Object Recognition and water maze performance tests) versus plasma amylin levels (both rat and human amylin; D, E) and versus age (F, G) in APP/PS1/HIP versus APP/PS1 rats, and HIP versus wild‐type (WT) rats (n = 10 rats/group). Data for individual tests are shown in Figure S6 in supporting information. H, Plasma amylin levels in APP/PS1 rats versus APP/PS1 rats with suppressed amylin expression (APP/PS1/AKO rats) at 16 months of age. I, Novel object recognition in APP/PS1 rats versus APP/PS1/AKO rats at 8, 12, and 16 months of age (n = 10 rats/group). J, Longitudinal neurological scores of cognitive function in APP/PS1 versus APP/PS1/AKO rats (n = 10 rats/group). Data are means ± standard error of the mean. P ≤ 0.05 *, P ≤ 0.01 **, *, P ≤ 0.001 ***; correlation analysis (D, E); two‐way analysis of variance with Sidak post hoc (F, G, J); two‐tailed, unpaired t test (H, I)

FIGURE 4

Amylin–amyloid beta (Aβ) cross‐seeding and impaired cerebrospinal fluid (CSF)–brain Aβ exchange by high blood amylin levels (both rat and human amylin) in APP/PS1 rats. A‐E, Representative immunohistological images of co‐staining with anti‐amylin (brown) and anti‐Aβ (green) antibodies in brain sections from 16‐month‐old APP/PS1/HIP rats (A, C, E) and APP/PS1 rats (B, D; n = 5 rats/group). Amylin–Aβ plaques (A‐D) and amylin depositions within neurons (E) are shown. F, Analysis of total plaques including Aβ plaque, amylin plaque, and amylin‐core Aβ plaque depositions in APP/PS1 and APP/PS1/HIP rat brains, assessed from immunohistological images (n = 5 rats/group; data were normalized to total imaging area). G, H, Relative distribution of amylin, Aβ, and amylin–Aβ in plaques in the gray matter of APP/PS1/HIP rat brains (G; n = 5) and APP/PS1 rat brains (H; n = 5). I‐K, CSF total Aβ (I; n = 6 rats/group), brain amylin (Triton‐soluble fractions of tissue homogenate; n = 5 rats/group; J), and CSF amylin (K; n = 6 rats/group) levels in APP/PS1/HIP versus APP/PS1 rats at 16 months of age. Scale bars, 200 μm (A, B), 50 μm (C‐E). Data are means ± standard error of the mean; P ≤ 0.05 *, P ≤ 0.01 **, P ≤ 0.001 ***; two‐tailed, unpaired t test

FIGURE 5

Decreased brain amylin amyloid‐core plaques in APP/PS1 rats with genetically suppressed amylin secretion. A‐F, Representative images of immunostaining for amylin (brown) and amyloid beta (Aβ) (green) in brain tissues from 16‐month‐old APP/PS1/AKO rats (A‐C) and APP/PS1 littermates (D‐F). Arrows indicate amylin‐core Aβ plaques (CP). G, Counts of amylin‐core Aβ plaques in APP/PS1 and APP/PS1/AKO rat brains assessed from immunohistological images (n = 5 rats/group; data were normalized to total imaging area). H, I, Amylin and Aβ levels in brain homogenates from 16‐month‐old APP/PS1/AKO rats and APP/PS1 littermates. J, Counts of Aβ plaques in the same APP/PS1 and APP/PS1/AKO rat brains as in (G; n = 5 rats/group). Scale bars, 50 μm (A, B, D, E), 200 μm (C, F). Data are means ± standard error of the mean. P ≤ 0.05 *; P ≤ 0.01 **; two‐tailed, unpaired t test

FIGURE 6

Circulating aggregated amylin cross‐seeds with brain amyloid beta (Aβ) in young APP/PS1 rats. A, Experimental protocol for intravenous amylin infusion in rats. B, Plasma amylin levels in APP/PS1 rats intravenously infused with human amylin and age matched APP/PS1 control rats at baseline and after injections (n = 5 rats/group). C, D, Immunohistological analysis with anti‐amylin (brown) and anti‐Aβ (green) antibodies showing amylin‐Aβ plaques (C; arrow) and amylin accumulation in neurons (D; circles) of APP/PS1 rats intravenously infused with human amylin and age‐matched APP/PS1 control rats (n = 3 rats/group). E, Western blot analysis of amylin and Aβ levels in the brain homogenate extracted using phosphate‐buffered saline or sodium dodecyl sulfate from APP/PS1 rats intravenously infused with human amylin and age‐matched APP/PS1 control rats (n = 3‐4 rats/group). F, Schematic summary of the results. Circulating aggregated amylin affects brain and cerebrospinal fluid (CSF) Aβ levels via brain amylin accumulation and amylin‐Aβ cross‐seeding. Amylin gene deletion in APP/PS1 rats blocks amylin accumulation in the brain and amylin–Aβ cross‐seeding, providing protection against AD‐associated effects. Scale bars, 50 μm (C, D). Data are means ± standard error of the mean; two‐tailed, paired t test (B), P ≤ 0.05 *; two‐tailed, unpaired t test (E)