CD2AP deficiency aggravates Alzheimer's disease phenotypes and pathology through p38 MAPK activation

Abstract

Background: Alzheimer's disease (AD) is the most common form of neurodegenerative disorder, which is characterized by a decline in cognitive abilities. Genome-wide association and clinicopathological studies have demonstrated that the CD2-associated protein (CD2AP) gene is one of the most important genetic risk factors for AD. However, the precise mechanisms by which CD2AP is linked to AD pathogenesis remain unclear.

Methods: The spatiotemporal expression pattern of CD2AP was determined. Then, we generated and characterized an APP/PS1 mouse model with neuron-specific Cd2ap deletion, using immunoblotting, immunofluorescence, enzyme-linked immunosorbent assay, electrophysiology and behavioral tests. Additionally, we established a stable CD2AP-knockdown SH-SY5Y cell line to further elucidate the specific molecular mechanisms by which CD2AP contributes to AD pathogenesis. Finally, the APP/PS1 mice with neuron-specific Cd2ap deletion were treated with an inhibitor targeting the pathway identified above to further validate our findings.

Results: CD2AP is widely expressed in various regions of the mouse brain, with predominant expression in neurons and vascular endothelial cells. In APP/PS1 mice, neuronal knockout of Cd2ap significantly aggravated tau pathology, synaptic impairments and cognitive deficits. Mechanistically, the knockout of Cd2ap activated p38 mitogen-activated protein kinase (MAPK) signaling, which contributed to increased tau phosphorylation, synaptic injury, neuronal apoptosis and cognitive impairment. Furthermore, the phenotypes of neuronal Cd2ap knockout were ameliorated by a p38 MAPK inhibitor.

Conclusion: Our study presents the first in vivo evidence that CD2AP deficiency exacerbates the phenotypes and pathology of AD through the p38 MAPK pathway, identifying CD2AP/p38 MAPK as promising therapeutic targets for AD.

Keywords: Alzheimer’s disease; CD2AP; P38 MAPK; Synaptic injury; Tau.

Fig. 1

Expression of CD2AP in the brains of mice. a, b Immunoblots showed protein expression of CD2AP in the different mouse brain regions of 3-month-old mice. OB, Olfactory bulb; CTX, Cerebral cortex; HIP, hippocampus; CB, cerebellum; BS, brainstem; STR, striatum; HY, hypothalamus. c, d Immunoblots showed protein expression of CD2AP in whole-brain lysates from mice at the indicated stages of development. e Representative confocal images for CD2AP, NEUN, CD31, GFAP, IBA1 and DAPI in mouse cortex and hippocampus from 3-month-old mice. f Immunofluorescence image of CD2AP in the 1-month-old KO mouse brain as a negative control. g Quantitative analysis of fluorescence colocalization using Mander’s overlap coefficient methods. h The mean fluorescence intensity of CD2AP within the boundaries of the different cellular markers. All data are presented as mean ± SEM. One-way ANOVA with Turkey’s multiple comparison tests for multiple comparisons (b, d, g, h). *P < 0.05, **P < 0.01

Fig. 2

Neuronal Cd2ap deletion mice were not cognitively impaired. a Scheme of the generation of a conditional Cd2ap knockout mouse line. b RT-PCR analysis of CD2AP mRNA levels in whole-brain lysates from 12-month-old WT and CKO mice. c, d Immunoblots showing decreased protein expression of CD2AP in the brains of 12-month-old CKO mice compared to age-matched WT mice. e, f No significant difference in total distance moved or time spent in the center area was observed in the open field test. n = 18 (12 month-WT, female n = 5, male n = 13), n = 22 (12 month-CKO, female n = 13, male n = 9), n = 7 (15 month-WT, all male), n = 5 (15 month-CKO, all males). g No significant difference in time spent in the open arm was observed in the elevated plus maze. n = 18 (12 month-WT, female n = 5, male n = 13), n = 22 (12 month-CKO, female n = 13, male n = 9), n = 7 (15 month-WT, all males), n = 5 (15 month-CKO, all males). h There was no difference in preference for the novel object between WT and CKO both at 12 months and 15 months. n = 13 (12 month-WT, all male), n = 9 (12 month-CKO, all male), n = 7 (15 month-WT, all male), n = 5 (15 month-CKO, all males). i-k In the Morris Water maze, CKO and WT mice showed no difference in escape latency, the number of crossings of the platform or time spent in the target area. n = 18 (12 month-WT, female n = 5, male n = 13), n = 22 (12 month-CKO, female n = 13, male n = 9), n = 7 (15 month-WT), n = 5 (15 month-CKO). l, m In the fear conditioning test, 12-month-old CKO mice showed a similar freezing time to WT mice. n = 14 (WT, all male), n = 9 (CKO, all male). All data are presented as mean ± SEM. Unpaired t-test with two-tailed analysis was conducted separately for 12- and 15-month-old mice. *P < 0.05, **P < 0.01

Fig. 3

Neuronal Cd2ap deletion aggravated cognitive function and pathological features in APP/PS1 mice. a Scheme of the experimental mouse timeline. Briefly, a series of behavioral tests were performed in 4-month-old mice, and subsequent pathological analyses were conducted in 4.5-month-old mice. b No significant difference in time spent in the center area was observed in the open field test. n = 13 (WT, female n = 5, male n = 8), n = 8 (APP/PS1, female n = 3, male n = 5), n = 10 (CKO, female n = 5, male n = 5), n = 7 (CKO × APP/PS1, female n = 4, male n = 3). c CKO × APP/PS1 mice spent less time in the novel arm in the Y-maze novel arm preference test compared to WT mice. n = 13 (WT, female n = 5, male n = 8), n = 8 (APP/PS1, female n = 3, male n = 5), n = 10 (CKO, female n = 5, male n = 5), n = 7 (CKO × APP/PS1, female n = 4, male n = 3). d, e CKO × APP/PS1 mice showed significantly decreased contextual and cue-related freezing compared to WT mice. n = 13 (WT, female n = 5, male n = 8), n = 8 (APP/PS1, female n = 3, male n = 5), n = 9 (CKO, female n = 5, male n = 4), n = 7 (CKO × APP/PS1, female n = 4, male n = 3). f ELISA analysis of Aβ showed that neuronal Cd2ap deletion had no obvious influence on human Aβ level. n = 3 (APP/PS1, female n = 1, male n = 2), n = 5 (CKO × APP/PS1, female n = 3, male n = 2). g ELISA analysis of Aβ showed that neuronal Cd2ap deletion had no obvious influence on murine Aβ level. n = 6 (WT, female n = 2, male n = 4), n = 6 (APP/PS1, female n = 3, male n = 3), n = 4 (CKO, female n = 3, male n = 1), n = 6 (CKO × APP/PS1, female n = 4, male n = 2). h-j In 4.5-month-old mice, Immunoblots revealed that neuronal Cd2ap deletion had no obvious influence on the full-length APP (APP-FL) and APP-CTF proteins. n = 6 (WT, female n = 3, male n = 3), n = 6 (APP/PS1, female n = 3, male n = 3), n = 6 (CKO, female n = 3, male n = 3), n = 6 (CKO × APP/PS1, female n = 3, male n = 3). k, l In 4.5-month-old mice, immunoblots revealed that neuronal Cd2ap deletion led to significantly increased ptau202/205 (AT8) and p-tau396 level, especially in the CKO × APP/PS1 mice. n = 7 (WT, female n = 3, male n = 4), n = 7 (APP/PS1, female n = 3, male n = 4), n = 7 (CKO, female n = 4, male n = 3), n = 7 (CKO × APP/PS1, female n = 4, male n = 3). All data are presented as mean ± SEM. Unpaired t-test with two-tailed analysis (f), one-way ANOVA with Turkey’s multiple comparison tests for multiple comparisons (b-e, g, i, j, k), Kruskal–Wallis tests with Dunn’s multiple comparison tests (l). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Fig. 4

Proteomic analyses in Cd2ap knockout mice and neuronal Cd2ap deletion led to synaptic loss. a GO annotation analysis based on differentially expressed proteins was presented as the top six in cellular component (CC), and the top ten in biological process (BP) and molecular function (MF). b The top ten enriched pathways from KEGG pathway analysis of differentially expressed proteins. c Gene set enrichment analysis (GSEA) revealed that the reactive oxygen species pathway was upregulated in the Cd2ap knockout group. d, e Immunoblots showed a decreased synaptophysin level in the 4.5-month-old CKO and CKO × APP/PS1 mice. n = 6 (WT, female n = 3, male n = 3), n = 6 (APP/PS1, female n = 3, male n = 3), n = 5 (CKO, female n = 3, male n = 2), n = 7 (CKO × APP/PS1, female n = 4, male n = 3). f Immunoblots showed a decreased PSD95 level in the 4.5-month-old CKO × APP/PS1 mice. n = 6 (WT, female n = 3, male n = 3), n = 6 (APP/PS1, female n = 3, male n = 3), n = 6 (CKO, female n = 3, male n = 3), n = 5 (CKO × APP/PS1, female n = 2, male n = 3). g, h Immunofluorescence analysis showed significantly decreased synaptophysin in the hippocampus in 4.5-month-old CKO and CKO × APP/PS1 mice. n = 4 (WT, all male), n = 3 (APP/PS1, all male), n = 3 (CKO, all male), n = 4 (CKO × APP/PS1, all male). All data are presented as mean ± SEM. One-way ANOVA with Turkey’s multiple comparison tests for multiple comparisons (e, h), Kruskal–Wallis tests with Dunn’s multiple comparison tests (f). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Fig. 5

Neuronal Cd2ap deletion led to synaptic deficits in 4.5-month-old APP/PS1 mice. a Representative traces of fEPSP during LTP recording. b Time-course of fEPSP slopes during LTP recording. c The normalized average fEPSP slopes during the last 10 min. n = 10 from 3 mice in WT group, n = 9 from 3 mice in APP/PS1 group, n = 7 from 3 mice in CKO group, n = 8 from 3 mice in CKO × APP/PS1 group, one-way ANOVA with Turkey’s multiple comparison tests for multiple comparisons. d Schematic diagram showing the localization of stimulating and recording electrodes in the hippocampal slice. e Representative I/O curves for fEPSPs recorded in hippocampal CA1 subregion from 4 groups. f Quantitative analyses of I/O curves. n = 10 from 3 male mice in WT group, n = 9 from 3 male mice in APP/PS1 group, n = 9 from 3 male mice in CKO group, n = 9 from 3 male mice in CKO × APP/PS1 group, two-way ANOVA. HFS: High-frequency stimulation. All data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Fig. 6

Neuronal Cd2ap knockout activated p38 MAPK signaling and apoptosis. a–c Immunoblots showed significantly increased phosphorylated p38 MAPK in 4.5-month-old CKO and CKO × APP/PS1 mice and significantly increased CC3 level in 4.5-month-old CKO × APP/PS1 mice. n = 6 (WT, female n = 3, male n = 3), n = 6 (APP/PS1, female n = 3, male n = 3), n = 6 (CKO, female n = 3, male n = 3), n = 6 (CKO × APP/PS1, female n = 3, male n = 3). d–f Immunoblots showed significantly increased p-p38 MAPK and CC3 levels in CD2AP stable knockdown cells and especially in KD cells with APPSwe co-expression. g Knockdown of CD2AP with APPSwe expression significantly increased caspase-3 activity in SH-SY5Y cells. h, i Analysis of flow cytometry showed that the number of apoptotic cells was significantly increased in CD2AP knockdown cells and especially in CD2AP knockdown cells with APPSwe co-expression. The vertical axis represented the cells labeled with 7-AAD and the horizontal axis indicated those stained with Annexin V (eFluor™ 450). The percentage of cells labeled with Annexin V but without 7-AAD was calculated and compared. j, k The apoptotic cells were detected with the TUNEL assay with nuclei stained with DAPI. The percentage of cells was quantified in different groups. All data are presented as mean ± SEM, n = 3 per group (d–k). One-way ANOVA with Turkey’s multiple comparison test for multiple comparisons. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Fig. 7

Inhibition of p38 MAPK rescued learning deficit, synaptic loss, and apoptosis in CKO × APP/PS1 mice. a Scheme of drug treatment and test schedules. b No significant difference in time spent in the center area was observed among 4-month-old mice in the open field test. n = 14 (WT, female n = 8, male n = 6), n = 13 (CKO + Vehicle, female n = 6, male n = 7), n = 11 (CKO + SB, female n = 7, male n = 4), n = 7 (CKO × APP/PS1 + Vehicle, female n = 5, male n = 2), n = 8 (CKO × APP/PS1 + SB, female n = 5, male n = 3). c, d In the Y-maze novel arm preference test, 4-month-old CKO × APP/PS1 mice given SB203580 treatment spent more time in the novel arm than those given the vehicle. n = 11 (WT, female n = 7, male n = 4), n = 13 (CKO + Vehicle, female n = 6, male n = 7), n = 9 (CKO + SB, female n = 6, male n = 3), n = 6 (CKO × APP/PS1 + Vehicle, female n = 4, male n = 2), n = 8 (CKO × APP/PS1 + SB, female n = 6, male n = 2). e, f Four-month-old CKO × APP/PS1 mice given SB203580 treatment showed significantly increased contextual and cue-related freezing time than those given the vehicle. n = 14 (WT, female n = 8, male n = 6), n = 13 (CKO + Vehicle, female n = 6, male n = 7), n = 11 (CKO + SB, female n = 7, male n = 4), n = 7 (CKO × APP/PS1 + Vehicle, female n = 5, male n = 2), n = 9 (CKO × APP/PS1 + SB, female n = 6, male n = 3). g SB203580 treatment did not obviously suppress the phosphorylated p38 MAPK in 4.5-month-old CKO and CKO × APP/PS1 mice. n = 4 (WT, female n = 2, male n = 2), n = 4 (CKO + Vehicle, female n = 2, male n = 2), n = 4 (CKO + SB, female n = 2, male n = 2), n = 6 (CKO × APP/PS1 + Vehicle, female n = 4, male n = 2), n = 6 (CKO × APP/PS1 + SB, female n = 3, male n = 3) h–j SB203580 treatment significantly decreased the hyperphosphorylation of tau in 4.5-month-old CKO and CKO × APP/PS1 mice, n = 4 (WT, female n = 2, male n = 2), n = 4 (CKO + Vehicle, female n = 2, male n = 2), n = 5 (CKO + SB, female n = 2, male n = 3), n = 5 (CKO × APP/PS1 + Vehicle, female n = 3, male n = 2), n = 6 (CKO × APP/PS1 + SB, female n = 3, male n = 3). k, l SB203580 treatment increased synaptophysin protein level in 4.5-month-old CKO and CKO × APP/PS1 mice. n = 4 (WT, female n = 2, male n = 2), n = 4 (CKO + Vehicle, female n = 2, male n = 2), n = 5 (CKO + SB, female n = 2, male n = 3), n = 5 (CKO × APP/PS1 + Vehicle, female n = 3, male n = 2), n = 6 (CKO × APP/PS1 + SB, female n = 3, male n = 3). m, n SB203580 treatment significantly decreased CC3 level in 4.5-month-old CKO × APP/PS1 mice. n = 5 (WT, female n = 3, male n = 2), n = 5 (CKO + Vehicle, female n = 3, male n = 2), n = 5 (CKO + SB, female n = 2, male n = 3), n = 5 (CKO × APP/PS1 + Vehicle, female n = 3, male n = 2), n = 4 (CKO × APP/PS1 + SB, female n = 2, male n = 2). Data are presented as mean ± SEM (b, c, e f, i, l) and box plots with median (g, j, n). One-way ANOVA with Turkey’s multiple comparison tests for multiple comparisons (b-f), One-way ANOVA or Kruskal–Wallis tests for detecting overall differences among multiple groups (g-n), followed by specific comparisons, including CKO + Vehicle vs. CKO + SB, and CKO × APP/PS1 + Vehicle vs. CKO × APP/PS1 + SB, using unpaired t-test with two-tailed analysis (i, l) and Mann–Whitney test (g, j, n), with Bonferroni correction applied. *P < 0.05, **P < 0.01, ***P < 0.001