Muscle Cathepsin B Treatment Improves Behavioral and Neurogenic Deficits in a Mouse Model of Alzheimer's Disease

Abstract

Increasing evidence indicates skeletal muscle function is associated with cognition. Muscle-secreted protease Cathepsin B (Ctsb) is linked to memory in animals and humans, but has an unclear role in neurodegenerative diseases. To address this question, we utilized an AAV-vector-mediated approach to express Ctsb in skeletal muscle of APP/PS1 Alzheimer's disease (AD) model mice. Mice were treated with Ctsb at 4 months of age, followed by behavioral analyses 6 months thereafter. Here we show that muscle-targeted Ctsb treatment results in long-term improvements in motor coordination, memory function, and adult hippocampal neurogenesis, while plaque pathology and neuroinflammation remain unchanged. Additionally, in AD mice, Ctsb treatment normalizes hippocampal, muscle, and plasma proteomic profiles to resemble that of wildtype (WT) controls. In AD mice, Ctsb increases the abundance of hippocampal proteins involved in mRNA metabolism and protein synthesis, including those relevant to adult neurogenesis and memory function. Furthermore, Ctsb treatment enhances plasma metabolic and mitochondrial processes. In muscle, Ctsb treatment elevates protein translation in AD mice, whereas in WT mice mitochondrial proteins decrease. In WT mice, Ctsb treatment causes memory deficits and results in protein profiles across tissues that are comparable to AD control mice. Overall, the biological changes in the treatment groups are consistent with effects on memory function. Thus, skeletal muscle Ctsb application has potential as an AD therapeutic intervention.

Keywords: Alzheimer's disease; cathepsin B; memory; muscle; neurogenesis; proteomics.

FIGURE 1

Experimental design and timeline. (a) Four‐month‐old male mice (either AD or WT mice) were injected into the tail vein with Ctsb (pAAV9‐tMCK‐mCTSB‐IRES‐eGFP) or Control (Con) vector (pAAV9‐tMCK‐IRES‐eGFP): WT‐Con (N = 10), WT‐Ctsb (N = 9), AD‐Con (N = 8), and AD‐Ctsb (N = 14). Mice were subjected to a battery of behavioral tests from 10 months of age. The behavioral assessment consisted of activity box (open field), rotarod, Morris water maze, and fear conditioning paradigm. At 12 months of age, mice were deeply anesthetized and euthanized for tissue and plasma collection. (b) One hemisphere of the brain was sectioned and used for histological evaluation of adult neurogenesis, amyloid beta plaque deposition, and neuroinflammation. Hippocampus from the other brain hemisphere, gastrocnemius muscle, and plasma were used for proteomic assays. Briefly, hippocampal and muscle tissues were lysed in 4% SDS buffer, followed by reduction and alkylation steps, and protein aggregation capture method. Cleaned peptides were injected into LC–MS/MS (Evosep HPLC 44 min coupled to timsTOF Pro 2 mass spectrometer). Plasma proteins were solubilized directly in Tris buffer. Cleaned peptides were injected into LC–MS/MS (Evosep HPLC 21 min coupled to timsTOF Pro 2 mass spectrometer).

FIGURE 2

Ctsb treatment differentially affects locomotion, motor coordination and memory function in WT and AD mice. (a) Ctsb treatment increased distance traveled in WT mice during a 60 min activity box session. (b) AD‐Con mice had a shorter latency to fall than all the other groups on an accelerating rotarod (4–40 rpm). (c) Acquisition curves in the Morris water maze, analyzed over 4‐day bins (during a 20‐day training protocol). WT mice perform better than AD mice throughout the acquisition phase. (d, e) Morris water maze probe trials to test retention of spatial memory. (d) AD‐Con mice and WT‐Ctsb mice had impaired retention of spatial memory performance during the 24 h probe trial. Only WT‐Con and AD‐Ctsb mice preferred the platform quadrant compared to 25% chance. (e) Considering the performance in the 4 h and 24 h probe trials as a whole, WT‐Con (100%), AD‐Ctsb (57%), AD‐Con (25%) and WT‐Ctsb (33%) mice spent more than 25% of time in the target quadrant in both trials. Solid green lines represent 25% chance. (f) In the fear conditioning paradigm WT‐Ctsb mice froze less during the conditioning session and (g) the tone‐cued phase, as compared to the other groups. (h) AD‐Con mice displayed increased CFR to context as compared to the other groups. (N; WT‐Con = 10, WT‐Ctsb = 9, AD‐Con = 8, AD‐Ctsb = 14). ^# p < 0.05 ad compared to WT in the same Treatment. *p < 0.05 Ctsb compared to Con in the same Genotype. ⁺ p < 0.05 of being equal or lesser than 25% chance (one‐tailed Wilcoxon signed rank exact test). ^& p < 0.05 AD compared to WT (Genotype main effect). ^$ p < 0.05 ad‐Ctsb compared to WT‐Ctsb. ^@ p < 0.05 ad‐Con compared to WT‐Con. Data were analyzed by GLM (a, b, d, g) or GLMM (c, f, h) and are presented as Estimated Marginal Means and their respective 95% CI. In (e), solid lines represent the Estimated Marginal Means while dashed lines represent 95% CI.

FIGURE 3

Ctsb treatment improves adult neurogenesis in AD mice. (a–d) Representative photomicrographs of the dentate gyrus derived from (a) WT‐Con; (b) WT‐Ctsb; (c) AD‐Con; (d) AD‐Ctsb brain tissue sections subjected to DCX staining. (e) Treatment with Ctsb in AD mice restored DCX cell numbers to WT‐Con levels. (N; WT‐Con = 10, WT‐Ctsb = 8, AD‐Con = 8, AD‐Ctsb = 13). Scale bars: 50 μm overview; 25 μm inset. ^# p < 0.05 AD compared to WT in the same Treatment. *p < 0.05 Ctsb compared to Con in the same Genotype. Data in (e) was analyzed by GLM and are presented as Estimated Marginal Means and their respective 95% CI.

FIGURE 4

Ctsb treatment does not modify neuroinflammation in AD mice. Representative photomicrographs of dorsal hippocampal sections subjected to immunofluorescent double‐labeling for GFAP (green) and Iba1 (red), derived from (a) WT‐Con, (b) WT‐Ctsb, (c) AD‐Con and (d) AD‐Ctsb mice. Nuclei were stained with 4′,6‐diamidino‐2‐phenylindole (DAPI) blue. (e, f) Astrocyte and microglia area density. Increased hippocampal density of (e) Iba1⁺ microglia but not of (f) GFAP⁺ astrocytes was observed in the AD groups. (N; WT‐Con = 9, WT‐Ctsb = 9, AD‐Con = 8, AD‐Ctsb = 12). Scale bars: 100 μm. *p < 0.05 for Genotype effect (WT vs. AD). Data were analyzed by GLM and are presented as Estimated Marginal Means and their respective 95% CI.

FIGURE 5

Ctsb treatment induces cytoskeletal reorganization in WT mice, whereas transcription and RNA processing in AD mice. (a) Volcano plot displaying average effect of AD. Gene set enrichment analysis (GSEA) based on (b) biological processes (BP) and (c) cellular components (CC). (d) Abundance of proteins associated with the lysosome ontology that have log fold changes > 0.50. (e) Volcano plots showing significant proteins, if present, in each pairwise group comparison. (f) GSEA of biological processes summarizing changes observed in all groups. (g) Mean abundance of proteins associated with mRNA metabolic process ontology. (h) Volcano plot of interaction between AD and Ctsb treatment. (i) Mean abundance of proteins associated with myofibril and cytosolic ribosome ontologies, which were found to be enriched by GSEA of interaction data using cellular components. (j) Abundance of selected proteins involved in neurogenesis and neural plasticity. Multiple hypothesis testing was performed by the Benjamini‐Hochberg method for each differential expression analysis, and proteins with FDR < 0.10 were considered statistically significant. For GSEAs, a cut‐off of FDR < 0.05 was used. Data for the pairwise comparisons are annotated as follows: AD‐WT(Ctsb), AD‐Ctsb vs WT‐Ctsb; AD‐WT(Con), AD‐Con vs WT‐Con; Ctsb‐Con(WT), WT‐Ctsb vs WT‐Con; Ctsb‐Con(AD), AD‐Ctsb vs AD‐Con.

FIGURE 6

Effects of Ctsb treatment on muscle and plasma proteome. Muscle proteome: (a) GSEA plot displaying Wnt signaling pathway as one of the most affected gene sets in AD. GSEA of average effect of Ctsb treatment based on (b) biological processes (BP). (c) GSEA using biological processes summarizing changes observed in each group. Plasma proteome: GSEA of (d) biological processes summarizing changes observed in the average effect of AD. (e) GSEA of biological processes summarizing changes observed in all groups. (f) Mean abundance of proteins involved in metabolic processes and cytoskeletal organization, selected from GSEA of interaction data. (c, e, f) Data for the pairwise comparisons are annotated as follows: AD‐WT(Ctsb), AD‐Ctsb vs WT‐Ctsb; AD‐WT(Con), AD‐Con vs WT‐Con; Ctsb‐Con(WT), WT‐Ctsb vs WT‐Con; Ctsb‐Con(AD), AD‐Ctsb vs AD‐Con.