Потеря ADAP1/CentA1 защищает от аутоиммунной демиэлизации.

Introduction

Multiple sclerosis (MS) is an autoimmune demyelinating disorder of the central nervous system (CNS) that represents a primary cause of non‐traumatic brain injury among young adults, with women being twice as likely as men to develop MS [ 1 ]. The hallmarks of MS pathology include focal infiltration of peripheral immune cells into the CNS parenchyma, breakdown of myelin sheaths, gliosis, axonal injury, and neuronal loss [ 2 ]. While MS was traditionally considered a T cell‐mediated disorder, current knowledge recognizes the involvement of multiple cytotypes mediating distinct effects on the autoimmune response and lesion formation, either pathogenic or protective [ 3 ]. The etiology of MS is still unknown, but the interplay between specific genetic and environmental determinants is believed to modulate the predisposition to disease. More than 200 MS risk loci enriched in immunological‐relevant genes have been identified in a large genome‐wide association study (GWAS) [ 4 ]. Viral infections, tobacco smoking, low sun exposure, low vitamin D levels, and obesity are the most relevant environmental factors associated with increased MS susceptibility [ 5 ].

From a clinical standpoint, MS typically starts with a relapsing–remitting course dominated by neuroinflammatory processes and later evolves into a progressive neurodegenerative disease characterized by the irreversible accumulation of neurological disability due to neuro‐axonal damage [ 6 ]. A recent GWAS implicates primarily CNS resilience in MS conversion, in contrast to immune‐driven susceptibility [ 7 ]. Several molecular events have been suggested as plausible triggers for the conversion process, including microglia activation, chronic oxidative injury, accumulation of mitochondrial damage, and dysfunctional axonal transport [ 8 ]. Aberrant cytoskeletal dynamics may additionally play a significant role in MS neurodegeneration due to the critical function of the cytoskeleton in maintaining neuronal morphology and proper synaptic activity [ 9 ]. Consistent with this hypothesis, we have recently shown that the microtubule‐associated protein (MAP) tau exerts neuroprotective activity against disease progression in the MS model experimental autoimmune encephalomyelitis (EAE) [ 10 ]. Notably, aberrant microtubule and actin dynamics have been observed in multiple neurological diseases, suggesting that cytoskeletal dysfunctions might be an underlying mechanism for neurodegeneration [ 11 ].

ArfGAP with dual PH domain‐containing protein 1 (ADAP1), also known as Centaurin alpha‐1 (CentA1), is an evolutionarily conserved ADP‐ribosylation factor GTPase‐activating protein that is highly enriched in neurons and, at lower levels, in astrocytes [ 12 ]. From a molecular standpoint, ADAP1/CentA1 binds phosphatidylinositol (3,4,5)‐triphosphate (PIP3) and F‐actin through two pleckstrin homology (PH) domains and regulates the neuronal actin cytoskeleton [ 13 ]. At the subcellular level, ADAP1/CentA1 is mainly localized to axons, dendrites, dendritic spines, and post‐synaptic density (PSD) regions in the synapses. ADAP1/CentA1 is required during neuronal development for correct dendritic branching and spinogenesis [ 14 ]. Furthermore, ADAP1/CentA1 was found to be upregulated in the Alzheimer's disease (AD) brain, and there is evidence that ADAP1/CentA1 accumulates in neuritic plaques [ 15 ]. In this regard, we have reported that Aβ42‐induced ADAP1/CentA1 upregulation induces a Ras‐dependent association of Elk‐1 with mitochondria, leading to mitochondrial and synaptic dysfunctions [ 16 ]. We further demonstrated that ADAP1/CentA1 ablation protects against amyloid‐β induced hippocampal synaptic loss, and CentA1 ‐null mice display higher hippocampal dendritic spine density, spine structural plasticity, and improved memory performance in the object‐in‐place test [ 16 , 17 ].

Altogether, the existing data pinpoint a mechanistic role for ADAP1/CentA1 in synaptic pathology and neurodegeneration. Thus, we were keen to explore whether this cytoskeletal‐associated protein is also implicated in the neurodegenerative processes underlying MS progression. Surprisingly, we found that CentA1 ‐null mice are partially refractory to EAE induction. Subsequent characterization of the autoimmune response in these animals via immunophenotyping experiments and single‐cell RNA‐seq (scRNA‐seq) profiling narrowed down ADAP1/CentA1 activity in the peripheral immune system to specific cytotypes, namely regulatory T cells (Tregs), monocytes, and natural killer (NK) cells. In summary, these results deepen our current understanding of MS pathophysiology and document a novel non‐canonical role for ADAP1/CentA1 outside of the CNS in the context of autoimmune demyelination.

Materials and Methods

Mouse Strains

Mice constitutively lacking ADAP1/CentA1 expression ( CentA1 ‐null mice) were generated via homologous recombination as previously described [ 17 ]. The line was maintained on a pure C57BL/6 J background to eliminate possible confounding effects of mixed genetic background. Mice were housed in the ECU vivarium in a specific pathogen‐free (SPF) environment. All procedures on animals were performed in compliance with experimental guidelines approved by the East Carolina University Institutional Animal Care and Use Committee (IACUC).

EAE Induction

Active EAE was induced following previously published procedures [ 18 ]. In brief, 8–10‐week‐old CentA1 ‐null and wild‐type mice of both sexes were injected subcutaneously with 100 μg of myelin oligodendrocyte glycoprotein 35–55 (MOG35‐55) peptide (EZBiolab) in complete Freund's adjuvant (CFA) containing 4 mg/mL
Mycobacterium tuberculosis
(DIFCO Laboratories). Mice also received 400 ng of pertussis toxin (LIST Biological Laboratories) intraperitoneally at the time of immunization and 48 h later. All experimental animals were observed daily up to 30 days post‐immunization (dpi), and clinical signs were assessed blindly to the genotypes as follows: 0, no signs; 1, decreased tail tone; 2, mild monoparesis or paraparesis; 3, severe paraparesis; 4, paraplegia; 5, quadriparesis; and 6, moribund or death.

Western Blotting

Mouse brain tissues were homogenized in 10 volumes of cold RIPA buffer (1% Triton X‐100, 50 mM Tris pH 7.5, 75 mM NaCl) supplemented with complete protease inhibitors (Sigma) and phosphatase inhibitors (Roche). Brain lysates were separated by SDS‐PAGE and then blotted onto nitrocellulose membranes (Immobilion) at 100 V for 30 min. Membranes were blocked with a 5% blotting grade blocker (Bio‐Rad) in Tris‐buffered saline with 0.05% Tween‐20 (TBS‐T) for 1 h at room temperature (RT). After blocking, membranes were incubated with primary antibody, either anti‐ADAP1/CentA1 ( EPR14285 , Abcam) or anti‐beta actin (D6A8, Cell Signaling Technology) diluted in blocking solution (1:1000) and left overnight at 4°C. The day after, membranes were washed 3 times with TBS‐T and incubated with horseradish peroxidase (HRP)‐conjugated secondary antibodies (1:10 000) in blocking solution for 1 h at RT. After washing, the chemiluminescent signals were detected with the Supersignal West Dura reagent (Thermo Scientific) using a C300 Imaging System (Azure Biosystems).

Histopathology

Experimental mice were intracardially perfused with 4% paraformaldehyde (PFA), and the dissected spinal cords were post‐fixed for an additional 48 h before processing for paraffin embedding. Transversal sections of 5 μm were then cut from representative anatomical regions of the spinal cord (cervical, thoracic, and lumbar) and stained with hematoxylin and eosin (H&E) for immune cell infiltration and with a myelin basic protein (MBP) antibody for demyelination. Digital images were acquired with a NanoZoomer scanner (Hamamatsu Photonics), and neuroinflammation was assessed as the number of cell foci (groups of > 20 cells) in the spinal cord parenchyma, whereas the extent of myelin loss was quantified by assigning standard demyelination scores to MBP stains. All histopathological analyses were carried out blindly to the genotypes and were outsourced at Hooke Laboratories (Lawrence, MA).

Immunophenotyping

Splenocytes were isolated from CentA1 ‐null and wild‐type mice, either naive or primed with 100 μg MOG35‐55 in CFA for 10 days, and the abundance and activation status of the main immune cell populations were assessed by flow cytometry. Briefly, 3 million splenocytes were blocked with 5 μL True Stain FCX blocker and 5 μL True Stain Monocyte blocker (BioLegend) in PBS with 2% bovine serum albumin (BSA) for 20 min. Then, cells were stained with an 11‐color antibody panel for 9 different immune cell markers and 2 activation markers in Brilliant Stain buffer (BD Biosciences) for 30 min (Table  S1 ). Cells were washed twice with 3 mL of PBS and resuspended in 300 μL PBS for flow cytometry. LIVE/DEAD UV dye (Invitrogen) was added to the resuspended cells to mark dead cells. The resuspended cells were then acquired on an Aurora 5‐laser cytometer equipped with SpectroFlo Software (Cytek Biosciences) at the ECU Flow Cytometry Core. Data analysis was performed using FlowLogic software (Inivai Technologies).

T Cell Polarization Analysis

The propensity of T cells to polarize toward pro‐inflammatory (Th1/Th17) and tolerogenic (Treg) lineages was assessed by flow cytometry. Splenocytes isolated from MOG35‐55‐primed mice, either CentA1 ‐null or wild‐type, were cultured in RPMI 1640 media supplemented with 10% fetal bovine serum (FBS) and 50 μM β‐mercaptoethanol for 4 h at 37°C with 5% CO 2 . Cytokine production and intracellular accumulation were induced with Cell Stimulation cocktail (Invitrogen) in combination with GolgiStop protein transport inhibitor (BD Biosciences). After fixation and permeabilization with the Cytofix/Cytoperm kit (BD Biosciences), Th1 and Th17 lineages were assessed by staining CD4+/CD3e+ cells respectively for interferon‐gamma (IFNγ) and interleukin 17 (IL‐17) while Treg polarization was assessed via FOXP3 staining (Table  S1 ). Flow cytometry data acquisition and analysis were performed as for immunophenotyping experiments.

Single‐Cell RNA ‐Seq Analysis

Single‐cell suspensions of splenocytes from CentA1 ‐null or wild‐type mice were subjected to library construction using the Chromium Single Cell 3’ Gene Expression kit (10x Genomics), according to the manufacturer's protocol. The resulting libraries were sequenced on NextSeq 2000 platforms (Illumina) at the Brody Integrative Genomics Core. Raw sequencing reads were aligned to the reference mouse genome (mm10) and then quantified using cellranger (v8.0.0) count. scRNA sequencing captured > 7000 cells per sample (average: 9179) with > 10 000 mean reads per cell (average: 25 913), and > 1500 median genes per cell (average: 2179) (Figure  S1A–C ). Subsequently, the Seurat (v5.1.0) package in RStudio (Build 369) with R (v4.4.1) was used for data analysis. Briefly, single cells were filtered with nFeature_RNA (more than 500 but less than 6000) and percentage of mitochondrial reads (percent. mt, less than 10%). Doublets or multiplets were eliminated using DoubletFinder (v2.0.4), and batch effects were corrected using Harmony (v1.2.0). Gene expression data were then aggregated, normalized, and variance stabilized with sctransform regularized negative binomial regression (SCTransform in Seurat). Dimensionality reduction was performed by principal component analysis (PCA) and uniform manifold approximation and projection (UMAP) embedding (dims = 1:30) with RunPCA and RunUMAP, respectively. Cell clusters were subsequently identified with FindNeighbors (dims = 1:30) and FindClusters (resolution = 0.5) and annotated based on the expression of cell type–specific gene biomarkers. Differentially expressed biomarkers among cell types were assessed with FindAllMarkers or FindMarkers for genes detected in at least 10% of cells using a log(foldchange) threshold at 0.25. Bonferroni‐adjusted p ‐values were used to determine significance at a false discovery rate (FDR) < 0.05. Gene ontology (GO) and transcriptional regulatory relationships unraveled by sentence‐based text mining (TRRUST) enrichment analysis were done using Metascape, which provides comprehensive gene annotation based on GO processes, KEGG pathways, and the Reactome gene set [ 19 ].

Statistical Analyses

The Fisher's exact test was used to test for significance between the frequency of EAE occurrence. The two‐tailed Mann–Whitney U test was used to test for significance between EAE clinical scores and flow cytometry data. Data were expressed as mean ± standard error of the mean (SEM). All statistical tests were conducted in GraphPad Prism 8 software.

Results

ADAP1 / CentA1 Deficiency Protects Against Developing EAE

CentA1 ‐null mice are viable and do not display any overt phenotype [ 17 ]. Hence, we EAE immunized 8‐to‐10‐week CentA1 ‐null and wild‐type mice of both sexes to test whether ADAP1/CentA1 plays a mechanistic role in autoimmune demyelination (Figure  1A ). Significantly fewer CentA1 ‐null mice developed the disease compared to wild‐type littermates (40% vs. 80% incidence, p  = 0.0034) (Figure  1B ). This effect was more pronounced in CentA1 ‐null females ( p  = 0.0113) than in males ( p  = 0.550) (Figure  S2A ). Additionally, the disease course of those CentA1 ‐null mice that became sick was significantly milder around the disease peak (18–20 dpi; p
18dpi  = 0.0184; p
19dpi  = 0.0088; p
20dpi  = 0.0142) while no differences were observed in disease onset between genotypes ( p  = 0.8114) (Figure  1C ). Also, this effect on disease manifestation was driven mainly by female mice ( p
18dpi  = 0.0522; p
19dpi  = 0.0284) (Figure  S2B,C ). Detailed histopathological analysis confirmed the absence of inflammatory foci and demyelinated lesions in the spinal cord parenchyma of the CentA1 ‐null mice refractory to disease (Figure  1D–F ). Conversely, EAE pathology was comparable between genotypes in mice that developed the disease (Inflammatory foci numbers: p  = 0.7529; Demyelination scores: p  = 0.9074) (Figure  1G,H ). Altogether, these data suggest that loss of ADAP1/CentA1 exerts a protective effect on both disease initiation and progression in the context of CNS autoimmunity.

ADAP1/CentA1 Contributes to autoimmune demyelination. (A) Brain lysates from CentA1 ‐null and wild‐type mice were tested for ADAP1/CentA1 levels by Western blotting. Beta‐Actin levels served as loading controls. (B) EAE was induced using MOG35‐55 peptide in CentA1 ‐null and wild‐type littermates of both sexes. Disease incidence rates were calculated for each genotype from two independent experiments ( N  = 23 WT mice; N  = 25 KO mice). (C) Mean disease scores ±SEM are plotted for mice of each genotype that developed the disease up to 30 dpi. EAE scores were significantly lower in CentA1 ‐null mice during the disease peak. (D–F) Histopathological analysis of representative spinal cords from CentA1 ‐null (sick and EAE resistant) and wild‐type mice using MBP stain for myelin loss and H&E stain for inflammation. Arrowheads mark active lesions. Panels on the right depict inflammatory foci (arrows) with prominent vasogenic edema and demyelinated lesions at high magnification. (G and H) Quantification of inflammation and demyelination in mice of both genotypes that developed EAE. All results are presented as mean ± SEM. Differences in EAE scores and phenotypes were assessed using the Mann–Whitney U test, while differences in disease incidence were assessed using the Fisher's exact test. * p  ≤ 0.05, ** p  ≤ 0.01.

ADAP1 / CentA1 Modulates the Peripheral Immune Response

One possible explanation for the diminished susceptibility of CentA1 ‐null mice to EAE is that ADAP1/CentA1 directly controls the magnitude of the autoimmune response. Multiple mouse and human studies have indeed confirmed that ADAP1/CentA1 is physiologically expressed in the peripheral immune system [ 20 , 21 ].

Hence, we performed a thorough immunophenotyping of the main immune cell populations present in the spleens of naïve and MOG35‐55 immunized mice, namely CD4+ and CD8+ T cells, B cells, monocytes, and NK cells (Figure  S3 ). Under baseline conditions, monocytes were moderately elevated in CentA1 ‐null mice compared to wild‐type controls ( p  = 0.0288) (Figure  2A ). At the same time, they displayed a significant reduction in MOG35‐55‐primed CentA1 ‐null animals ( p  < 0.0001) (Figure  2F ). Also, NK cells from CentA1 ‐null mice underwent a significant expansion in response to MOG35‐55 immunization ( p  = 0.0005) (Figure  2G ). Conversely, no differences were found in the abundance of the other cell populations between genotypes in both conditions (Figure  2B–E,H–K ). Likewise, we did not detect any significant changes in the naïve, central memory, and effector subsets of CD4+ and CD8+ T cells before or after MOG35–55 priming (Figure  S4A–N ).

ADAP1/CentA1 regulates the expansion of specific immune cell subsets. (A–E) Percentages of CD11b+ monocytes, CD49b+/CD355+ NK cells, CD19+ B cells, CD4+ and CD8+ T cells in naïve mice, either CentA1 ‐null or wild‐type. Only the monocyte population is slightly elevated in CentA1 ‐null animals ( N  = 10 mice/genotype). (F–J) Percentages of the same immune cell populations in mice primed with 100 μg MOG35‐55 for 10 days. Monocytes are significantly lower, whereas NK cells are greatly increased in CentA1 ‐null mice compared to wild‐type animals ( N  = 13 WT mice; N  = 11 KO mice). All results are presented as mean ± SEM. Differences were assessed by Mann–Whitney U test. * p  ≤ 0.05, *** p  ≤ 0.005, **** p  ≤ 0.0001.

Finally, to assess the possible regulatory effects of ADAP1/CentA1 on cell activation, we measured by flow cytometry the cell surface expression of the well‐known activation markers CD44 and CD80 in the same cell subsets and experimental conditions of the previous analyses [ 22 , 23 ]. Notably, we found that CD80 levels were significantly decreased in NK and CD4+ T cells from CentA1 ‐null mice upon MOG35‐55 stimulation ( p
NK  = 0.0056; p
CD4  = 0.0257) (Figure  3A,B ). Further analysis of the different CD4+ T cell lineages narrowed down the defective activation phenotype to the memory and effector subsets ( p
Effector  = 0.0059; p
Memory  = 0.0352) (Figure  3C,D ).

ADAP1/CentA1 controls immune cell activation. (A and B) Surface levels of CD44 and CD80 in different immune cell populations (monocytes, NK cells, B cells, CD4+, and CD8+ T cells) before and after priming with 100 μg MOG35‐55 for 10 days. Lower CD80 levels were detected in NK and CD4+ T cells from MOG35‐55 immunized CentA1 ‐null mice. (C and D) CD44 and CD80 levels in different CD4+ and CD8+ T cell subsets (naive, effector, and memory) at baseline and after MOG35‐55 priming. Effector and memory CD4+ T cells from MOG35‐55 immunized CentA1 ‐null mice show lower levels of CD44 and CD80, respectively. In the heatmaps, each column represents a cytotype, and each row the average mean fluorescence intenstiy (MFI) value of a specific marker for all the mice in the group ( N  = 10 mice/genotype for baseline; N  = 13 WT mice and N  = 11 KO mice for MOG35‐55 priming experiments). Differences were assessed by Mann–Whitney U test. * p  ≤ 0.05, ** p  ≤ 0.01.

ADAP1 / CentA1 Controls T Cell Polarization

Because EAE is a T cell‐mediated disease, we next assessed the effect of ADAP1/CentA1 loss on the propensity of CD4+ T cells to polarize toward the pro‐inflammatory Th1/Th17 or the tolerogenic T regulatory (Treg) lineages in response to encephalitogenic challenges. Splenocytes isolated from MOG35‐55 immunized CentA1 ‐null and wild‐type mice were cultured with stimulatory cocktails to induce their activation in vitro, and the size of the different CD4+ T cell lineages was then analyzed by flow cytometry via staining for IFNγ (Th1), IL‐17 (Th17), and FOXP3 (Treg). There was no significant difference in the amount of Th1 and Th17 cells between genotypes ( p
Th1  = 0.3402; p
Th17  = 0.5358) (Figure  4A,B ). On the contrary, we detected a significant increase of the Treg cell lineage in CentA1 ‐null mice compared to wild‐type littermates ( p  = 0.0130) (Figure  4C ). As Treg cells can restrain the immune system from mounting an adequate autoimmune response to start EAE, the expansion of this lineage in CentA1 ‐null mice might explain their lower susceptibility to developing the disease.

ADAP1/CentA1 regulates CD4+ T cell polarization. (A and B) Proportion of Th1 and Th17 CD4+ T cells from CentA1 ‐null and wild‐type mice immunized with 100 μg MOG35‐55 for 10 days. No differences were detected in the amount of both cell lineages between genotypes. (C) In contrast, Treg cells are increased in CentA1 ‐null mice compared to wild‐type animals. All results are presented as mean ± SEM ( N  = 7 WT mice; N  = 12 KO mice). Differences were assessed by Mann–Whitney U test. * p  ≤ 0.051.

ADAP1 / CentA1 Is Differentially Expressed in the Immune System

While our data corroborates the direct activity of ADAP1/CentA1 on the autoimmune response, very little is known about the specific molecular pathways regulated by this protein in the different immune populations in a pro‐inflammatory milieu. To fill this gap in knowledge, we exploited the power of single‐cell RNA‐seq (scRNA‐seq) technology to profile the transcriptomes of splenocytes from MOG35‐55‐primed CentA1 ‐null and wild‐type mice at single‐cell resolution.

Unsupervised clustering found 29 unique cell clusters in both genotypes comprised of 11 B cell clusters, 8 T cell clusters, 5 monocyte/macrophage clusters, 1 granulocyte, 1 endothelial, and 1 NK cell cluster (Figures  5A–D and S7 , Tables  [Link] , [Link] ). Adap1 expression was found highly enriched in the monocyte lineage, clusters 15 and 17, by marker analysis ( p
Cluster 15  = 6.28E‐152, p
Cluster 17  = 5.34E‐293) (Figures  5F and S8 , Table  S4 ). A low level of Adap1 expression was also localized to the follicular B cell subset (cluster 0) but was largely absent from the T cell compartment.

CentA1/Adap1 is highly expressed in monocytes and alters monocyte, B cell, and T cell transcriptomes. (A) Uniform manifold approximation and projection (UMAP) graph showing 29 cell clusters observed in splenocytes upon MOG35‐55 stimulation for 10 days. (B) Bar chart indicating the percentage of cells in each cluster observed in the analysis. (C) Expression UMAP for the T cell subset markers Cd8b1 and Cd4 , NK and cytotoxic T cell marker Nkg7 , and the pan‐B cell marker Ms4a1 . (D) Further expression of CD4+ Th cell effector markers Tbx21 and Il4 and Treg marker Foxp3 . (E) Bar graph showing the distribution of DEGs across all observed cell clusters. (F) Expression UMAP for Adap1 shows restriction primarily to monocytes and B cell compartments.

Next, we analyzed the effects of ADAP1/CentA1 loss on gene expression. Notably, 1111 differentially expressed genes (DEGs) were found across 24 cell clusters ( p
adj < 0.05), with most of the DEGs coming from clusters 0 (follicular B cell 1), cluster 15 ( Ly6c‐ high monocytes), cluster 10 (granulocytes), cluster 5 (mitochondrial‐high B cell 1), and cluster 2 (naïve CD4+ T cells). Only 1 DEG was found in effector CD8+ T cells ( Cd8a+ , Nkg7 +; cluster 11) and no DEGs were found in Treg cells ( Foxp3+ ; cluster 9) or Th cells ( Cd3e +, Il4 +, Tbx21 +, Rora +; cluster 23) suggesting that ADAP1/CentA1 is not directly involved in CD8+ or CD4+ T cell processes (Figure  5E , Table  S5 ).

ADAP1 / CentA1 Controls Distinct Pathways in Different Immune Cell Populations

To elucidate the cellular processes and molecular functions regulated by ADAP1/CentA1 in the different clusters identified by scRNA‐seq profiling, we performed GO enrichment analysis on the DEGs between wild‐type and knockout cells found within each cluster (Table  S6 ). TRRUST analysis was also conducted to identify the transcription factors mediating the different transcriptional changes (Table  S7 ).

In the B cell compartment (C0 and C5), upregulated genes in follicular B cells showed “B cell activation” as the top term ( p  = 6.89E‐12), while “Peptidyl‐serine phosphorylation” was enriched in both clusters ( p  = 3.79E‐7) (Figure  6A,C ). Among downregulated DEGs, “metal sequestration by antimicrobial peptides” was a top enriched ontology in both B cell clusters ( p  = 2.94E‐11). “Cytoplasmic ribosomal protein” processes were enriched in downregulated DEGs for C0 follicular B cells only ( p  = 2.44E‐73) (Figure  6A,C ). TRRUST analysis highlighted top enrichment for transcription factor 3 (TCF3) regulation in follicular B cells ( p  = 0.000316), but nuclear factor kappa B subunit 1 (NFKB1) and signal transducer and activator of transcription 3 (STAT3) were the only significant transcription factors in the mitochondrial high population ( p
NFKB1  = 0.003162; p
STAT3  = 0.00631) (Figure  6B,D ). In the naive CD4+ and naive CD8+ T cell clusters (C2 and C3), GO analysis of upregulated genes revealed an enrichment in the ontologies “T cell differentiation”: ( p  = 3.76E‐11) and “TCR signaling” ( p  = 2.75E‐6) (Figure  6E,G ). Conversely, “Metal sequestration by antimicrobial proteins” was the top enriched term for the downregulated genes in both clusters of T cells (Figure  6E,G ). TRRUST analysis found enrichment of TCF3 in both clusters ( p
CD4  = 0.001585; p
CD8  = 0.000316), as well as ETS proto‐oncogene 1 (ETS1) and runt‐related transcription factor 2 (RUNX2) in CD4+ T cells ( p
ETS1  = 0.000158; p
RUNX2  = 0.003162), and specificity protein 1 (SP1), STAT3, and NFKB1 in CD8+ T cells ( p
SP1  = 0.000794; p
STAT3  = 0.001259; p
NFKB1  = 0.002512) (Figure  6F,H ).

Loss of ADAP1 alters the expression of genes involved in leukocyte activation in B cells, T cells, granulocytes, and monocytes. (A–D) Gene ontology and TRRUST analysis of B cell clusters with highest DEGs, C0 Follicular B cells and C5 Mitochondrial high B cells. (E–H) Gene ontology and TRRUST analysis of T cell clusters with highest DEGs, Naïve CD4, and Naïve CD8 T cells. (I and J). GO and TRRUST analysis of granulocyte cell cluster. (K and L) GO and TRRUST analysis of Ly6c high monocyte cluster. The top 5 most significantly enriched ontologies are shown in each graph.

The top GO term for upregulated genes in the myeloid granulocyte cluster (C10) included “Neutrophil degranulation” ( p  = 6.78E‐12) and “positive regulation of cell migration” ( p  = 7.77E‐11), while downregulated genes were primarily involved in the GO function “cytoplasmic ribosomal processing” ( p  = 1E‐100) (Figure  6I ). Finally, out of the 194 DEGs found in the Ly6c‐ high monocyte cluster (C15), the upregulated genes were enriched in “positive regulation of cell migration” ( p  = 2.39E‐11) and “adaptive immune system” ( p  = 4.98E‐11) while downregulated genes were enriched in “cytoplasmic ribosomal proteins” ( p  = 1.0E‐100) and “spindle elongation” ( p  = 7.89E‐14) (Figure  6K ). TRRUST highlighted significant enrichment in regulatory factor X 5 (RFX5), regulatory factor X associated ankyrin‐containing protein (RFXANK), and regulatory factor X associated protein (RFXAP) transcriptional regulation ( p  = 3.98E‐5) (Figure  6L ).

Discussion

In the present study, we report that ADAP1/CentA1 is mechanistically involved in autoimmune demyelination. Mice lacking ADAP/Cent1 display partial resistance toward EAE, a disease model recapitulating several features of MS pathology at both cellular and molecular levels [ 24 ]. Most previous studies on ADAP/CentA1 functions focused on its structural role in controlling dendritic spine density in neurons via cytoskeletal remodeling. Thus, we initially intended to explore its putative role in modulating neuronal resilience against neuroinflammatory stress. Unexpectedly, our data pinpointed instead a novel function for ADAP1/CentA1 in regulating encephalitogenic immune responses in the periphery.

Only a few studies had previously investigated ADAP/CentA1 pathophysiology outside the CNS. ADAP1/CentA1 was shown to limit cardiomyocyte hypertrophy by reducing integrin cell surface expression [ 25 ]. The protein was also demonstrated to promote carcinoma progression by activating both cell migration and invasion [ 26 ]. Furthermore, ADAP1/CentA1 can promote latent human immunodeficiency virus 1 (HIV‐1) reactivation by amplifying the extracellular signal‐regulated kinase (ERK) pathway in T cells [ 27 ]. Here, we expand this scenario by providing experimental evidence that ADAP1/CentA1 controls the function of multiple immune cell populations that are critical for initiating and progressing the autoimmune response against myelin components.

The Treg lineage is one of the main targets of ADAP1/CentA1 regulation in the context of CNS autoimmunity. Tregs are lymphocytes that play a critical role in maintaining peripheral immune tolerance. They can control the autoimmune response by directly inhibiting the activity of effector T cells or by suppressing antigen‐presenting cell (APC) function [ 28 ]. Likewise, Tregs from MS patients are reported to have lower suppressive capabilities [ 29 ]. In the EAE paradigm, the presence of Tregs in the CNS mediates recovery from the acute phase [ 30 ]. Conversely, Treg ablation via anti‐CD25 antibodies exacerbates the EAE course [ 31 ]. Furthermore, Tregs were shown to prevent disease development or delay onset in spontaneous EAE models [ 32 , 33 ]. Hence, the increase in Tregs and the concomitant decrease in effector T cell activation observed in CentA1 ‐null mice in response to MOG35‐55 stimulation may explain their reduced EAE incidence. Notably, no DEGs were detected in the Treg cluster by scRNA‐seq, suggesting that ADAP1/CentA1 may act on this population in a non‐cell autonomous fashion.

In addition to Tregs, the monocyte compartment constitutes another principal target of ADAP1/CentA1 activity in the autoimmune response. Monocytes contribute to MS pathogenesis by acting as professional APCs and releasing pro‐inflammatory cytokines [ 34 ]. They also represent the predominant myeloid cell type promoting inflammation during EAE onset [ 35 ]. Monocyte infiltration into the CNS strongly correlates with progression to the paralytic stage of EAE [ 36 ]. In this light, the decrease in the monocyte pool we detected in MOG peptide‐immunized CentA1 ‐null mice could also contribute to their low EAE incidence by providing a sub‐optimal source of cytokines and chemokines for initiating the disease. At the molecular level, ADAP1/CentA1 is highly expressed in monocytes, and its ablation leads to the upregulation of genes controlling leukocyte migration, possibly representing a homeostatic response to cope with the defective proliferation of this cell population upon MOG35‐55 stimulation. Intriguingly, only the Ly6c ‐high monocyte cluster had a significant number of DEGs, pinpointing possible functional heterogeneity of ADAP1/CentA1 within this cell population. The Ly6c ‐high monocytes represent the “classical” pro‐inflammatory monocyte fraction involved in tissue extravasation and inflammation. They are critical for developing EAE, whereas Ly6c ‐low monocytes primarily perform scavenger functions in the vasculature [ 37 , 38 ]. Loss of ADAP1/CentA1 might also interfere with the monocyte‐mediated presentation of autoantigens to T cells as a perturbation in ARF1/ARF6 activity impairs phagocytosis in the monocyte/macrophage lineage [ 39 , 40 ]. Consistently, Ly6c ‐high monocytes from CentA1 ‐null mice exhibit downregulation in the major histocompatibility complex II (MHCII) genes H2‐Aa , H2‐Ab1 , and H2‐Eb1 .

Lastly, the NK cell population is another immune cell subset that appears to be affected by the lack of ADAP1/CentA1. NK cells are effector lymphocytes that play a regulatory role in EAE by exerting a direct cytotoxic activity on encephalitogenic T cells [ 41 ]. Likewise, NK cell depletion by anti‐NK1.1 monoclonal antibody treatment exacerbates EAE relapses and severity [ 42 ]. Consistent with this scenario, we detected an increased number of NK cells in CentA1 ‐null mice upon MOG35‐55 immunization, and this could contribute to their disease refractory phenotype by more efficient removal of autoreactive T cells in the periphery, thus preventing their infiltration into the CNS parenchyma. Intriguingly, lower NK cell counts have been measured in MS patients compared to healthy individuals, suggesting that their dysfunction is also critical in the etiology of the human disease [ 43 ].

Conclusions

In summary, we propose a working model in which ADAP1/CentA1 is necessary to orchestrate the aberrant immune response against the CNS via balancing pro‐inflammatory and tolerogenic signals. The deficiency of this protein in the immune system results in defective activation and proliferation of key cell subsets, eventually leading to a decrease in the incidence of EAE. Our findings align with previous observations demonstrating limited EAE susceptibility in mice lacking key immune genes [ 44 , 45 , 46 ]. We further highlighted specific genetic networks under the ADAP1/CentA1 control, which might represent novel druggable targets to tackle autoimmune demyelination. However, several important questions remain unanswered in our study. First, the exact contribution of ADAP1/CentA1 to MS pathogenesis is still unclear. The ADAP1 locus was not found to be associated with MS risk in the most recent GWAS [ 4 ]. However, it might be possible that its effect size was below the detection limit of the study or that long‐range regulatory events are involved via chromatin remodeling and other epigenetic mechanisms. Interestingly, ADAP1 has been linked to Crohn's disease susceptibility, possibly underlying a broad involvement of ADAP1/CentA1 in autoimmunity [ 47 ]. Second, the signaling pathways mediating the ADAP1/CentA1‐dependent transcriptional changes in the different immune cell populations are still elusive. The mitogen‐activated protein kinase (MAPK) cascade has been recognized as one of the main mediators of ADAP1/CentA1 function [ 16 , 27 ]. It will be important to explore whether the same pathway is involved in the immune system upon encephalitogenic stress. Previous studies showed that T cell‐specific loss of ARF1/ARF6 protects against developing EAE [ 48 ]. Thus, it may be possible that ADAP1/CentA1's role in CNS autoimmunity is independent from ARF1/ARF6 modulation. Also, since ADAP1/CentA1 was shown to bind nuclear proteins in vitro and in vivo, a direct modulatory activity on gene expression cannot be formally ruled out [ 49 ]. Finally, the heterogeneity of MS clinical manifestations cannot be recapitulated by a single EAE model. In the future, it will be relevant to extend our characterization of ADAP1/CentA1's role in CNS autoimmunity to alternative disease paradigms, including relapsing–remitting and spontaneous EAE models. It will be also interesting to assess whether the sex effect we observed on disease progression is replicated in other EAE models.

Author Contributions

A.D. conceived the study. A.D. and E.M.S. jointly supervised the research. J.J.C., W.P.B., R.R.D., W.H., E.M.S., and A.D. carried out all the experiments and analyzed the data. J.J.C., E.M.S., and A.D. wrote the paper. All authors read and approved the final manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figures S1–S8.

Table S1.

Table S2.

Table S3.

Table S4.

Table S5.

Table S6.

Table S7.