Human amygdala involvement in Alzheimer's disease revealed by stereological and dia-PASEF analysis

Abstract

Alzheimer's disease (AD) is characterized by the accumulation of pathological amyloid-β (Aβ) and Tau proteins. According to the prion-like hypothesis, both proteins can seed and disseminate through brain regions through neural connections and glial cells. The amygdaloid complex (AC) is involved early in the disease, and its widespread connections with other brain regions indicate that it is a hub for propagating pathology. To characterize changes in the AC as well as the involvement of neuronal and glial cells in AD, a combined stereological and proteomic analysis was performed in non-Alzheimer's disease and AD human samples. The synaptic alterations identified by proteomic data analysis could be related to the volume reduction observed in AD by the Cavalieri probe without neuronal loss. The pathological markers appeared in a gradient pattern with the medial region (cortical nucleus, Co) being more affected than lateral regions, suggesting the relevance of connections in the distribution of the pathology among different brain regions. Generalized astrogliosis was observed in every AC nucleus, likely related to deposits of pathological proteins. Astrocytes might mediate phagocytic microglial activation, whereas microglia might play a dual role since protective and toxic phenotypes have been described. These results highlight the potential participation of the amygdala in the disease spreading from/to olfactory areas, the temporal lobe and beyond. Proteomic data are available via ProteomeXchange with identifier PXD038322.

Keywords: BM88 antigen (BM88); antioxidant protein 2 (AOP2); calpactin II; calpactin-1 heavy chain (CAL1H); centaurin-alpha-1 (CENTA1); endonexin II (ENX2); nuclear chloride ion channel 27 (NCC27).

FIGURE 1

Amygdaloid volume reduction is specific to the Co and BLA, in particular the La. Nissl staining of the non‐AD (A) and AD (B) in the AC with delimitation of the amygdaloid nuclei studied. The global AC volume (C) and volume of the Co and BLA were significantly reduced in AD. In the BLA, volume was reduced specifically in the La (the graphs show the volume mean ± SEM, **p value <0.01, ***p value <0.001). AC, amygdaloid complex (Co, BLA); Co, cortical nucleus; BLA, basolateral complex (BM, BL, La); BM, basomedial nucleus; BL, basolateral nucleus; La, lateral nucleus. Scale bar = 1000 μm.

FIGURE 2

Generalized astrogliosis in the amygdaloid nuclei in AD. Immunohistochemical staining for MAP2 (A,B), Iba‐1 (D,E), and GFAP (G,H) in the BL in non‐AD and AD samples represents neurons, microglia, and astrocytes, respectively. The number of MAP2‐positive cells (C), Iba‐1‐positive cells (F), and GFAP‐positive cells (I) in the global AC and in the different nuclei are shown (the graphs show the mean ± SEM, *p value <0.05). Note that neither the number of neurons nor microglia was altered, and the number of astrocytes was increased in the whole AC. AC, amygdaloid complex (Co, BLA); Co, cortical nucleus; BLA, basolateral complex (BM, BL, La); BM, basomedial nucleus; BL, basolateral nucleus; La, lateral nucleus. Scale bar = 50 μm.

FIGURE 3

The cortical region is the most affected by pathology in AD. Aβ (A) and Tau (B) immunohistochemical staining of AD samples. Detail of Aβ (A', A'') and Tau (B', B'') staining pattern observed in Co and La, respectively. The area fractions of Aβ (C) and Tau (D) in the global AC and the different nuclei are shown (the graphs show the mean ± SEM, *p value <0.05, **p value <0.01, ***p value <0.001). Note that both Aβ and Tau appeared as a gradient with higher levels in medial (Co) regions than in lateral regions. Co: Cortical nucleus, BM: Basomedial nucleus, BL: Basolateral nucleus, La: Lateral nucleus. Scale bar = 1000 μm in (A,B); and 100 μm in (A',A''; B',B'').

FIGURE 4

Procedure for proteomic data analysis and criteria for protein selection validation. In a first step, dia‐PASEF analysis of human AC samples revealed 2153 proteins. After applying restricted condition of FC > 1.5 and p value <0.05, 178 proteins were identified as DEPs and cell type expression, SynGo and Metascape analyses were performed (data shown in Tables 3, 4, 5). Then, literature review of DEPs was carried out in order to select proteins for validation. Proteins were chosen based on three main criteria: previous evidence linking protein and AD must be reported; proteins widely described in the disease were excluded; and potential relation or expression in the studied cell types (neurons, microglia, and astrocytes) was also considered.

FIGURE 5

Neuronal involvement in the amygdaloid complex nuclei in AD: ADAP1, CEND1, and ANXA2. Triple immunofluorescence against ADAP1 (A–C), CEND1 (D,E), ANXA2 (G–I), and pathological markers. In non‐AD, ADAP1 (A, green) was mainly associated with vesicles in axons and dendrites, although it was also observed in soma. CEND1 (D, green) revealed neuronal expression in non‐AD samples. ADAP1 expression was drastically reduced in AD (B,C), with spatial coexpression with Tau (red) and MAP2 (purple) in the soma (B). Neurons close to Aβ (C, dashed line) presented a reduced number of ADAP1 vesicles in the soma and axon. A reduced number of CEND1‐stained neurons was observed in AD (E,F). CEND1 staining was remarkably associated with Tau deposits (E) compared with neurons near Aβ plaques (F, dashed line). ANXA2 (G, green) expression in neurons was identified in non‐AD samples. In AD, ANXA2 expression was increased close to Aβ (red) deposits (H,I). ANXA2 staining was higher on the outside of the plaques (H) than on the inside (I). Scale bar = 10 μm.

FIGURE 6

Microglial involvement in amygdaloid pathology in AD. Immunofluorescences against CLIC1 (A–C) and ANXA5 (D–F) and pathological markers are shown. In non‐AD samples, CLIC1 (A, green) labeling suggested possible expression in neurons (dashed line). In AD, CLIC1 colocalized intimately with Tau pathology (B, red) and microglia (purple, dashed line). Additionally, CLIC1 expression was observed in the microglial cells nearest to Aβ plaques (red) (C, arrow). ANXA5 (D, green) was related to microglia (purple, arrow) in non‐AD tissue. Microglial ANXA5 expression was increased in AD (E,F, arrow) with a closed spatial expression with Aβ plaques (E, reed) and Tau deposits (F, arrowhead). Scale bar = 10 μm.

FIGURE 7

Astroglial participation in AD. Immunofluorescences against ANXA1 and PRDX6 are shown in non‐AD (A and D, respectively) and AD (B,C and E,F, respectively) samples. In non‐AD samples, ANXA1 (A, green) was expressed in neurons (dashed line) and in astrocytes to a lesser extent (purple, arrow). In AD, ANXA1 expression in astrocytes was increased (B, arrow), and ANXA1 was coexpressed with Tau deposits (B, arrowhead). Frequently, neurons with slight Tau staining expressed increased levels of ANXA1 (C, dashed line). PRDX6 (green) expression by astrocytes (purple) was observed in non‐AD (D) and AD (E,F) samples. PRDX6 was related to Tau (red) (E) and Aβ (red) (F) pathology. Scale bar = 10 μm.

FIGURE 8

Amygdaloid complex as a “switch” in AD. Scheme of the amygdaloid complex (AC) and its main connections with olfactory areas, the hippocampus, and the entorhinal cortex (EC). Different amygdaloid nuclei are represented in grayscale from more (darker) to less (weaker) affected by pathology. Efferences and afferences regarding olfactory areas, CA1 and the EC might act as vehicles for pathology from and to the AC. AG, ambiens gyrus; BL, basolateral nucleus; BM, basomedial nucleus; Co, cortical nucleus; La, lateral nucleus.

FIGURE 9

Synaptic and glial responses against injury. Representative scheme of neuronal and glial responses against pathology in the amygdala according to proteomic data analysis and the literature. Reductions in ADAP1 and CEND1 suggest synaptic dysfunction. To control the disease, ANXA2 might mediate autophagosome‐lysosome fusion. Astrocytes might promote the activation of phagocytic microglia (PRDX6) and mark neurons for their clearance by microglia (ANXA1). Microglia might have a dual role since protective (ANXA5) and neurotoxic (CLIC1) roles have been linked. Created with BioRender.com