Acute anti-Aβ antibody exposure induces microglial changes and significantly alters chemokine signaling

Abstract

Introduction: While the anti-amyloid-lowering immunotherapies provide the first disease-targeting therapies for the treatment of Alzheimer's disease, there remain harmful adverse events causing hesitation among patients, families, physicians, and regulatory bodies. Though these drugs have been repeatedly proven to lower brain amyloid plaque burden, the specific cellular mechanisms and pathways by which the immunotherapy impacts the brain remain unclear.

Methods: This study aimed to transcriptionally profile the brain's immediate immune response to anti-amyloid beta (anti-Aβ) antibodies. To evaluate acute cellular priorities, we intracranially injected anti-Aβ antibody (3D6) or an isotype-matched control immunoglobulin G (IgG) antibody and performed single-cell sequencing analysis after 3 days.

Results: We found reduced numbers of a motile microglia cluster and homeostatic microglia in the 3D6 antibody-injected cortex compared to the IgG-injected cortex. It was also found that chemokine/cytokine signaling was enriched across homeostatic-proinflammatory microglia, interferon-responding microglia, and disease-associated microglia 2 (DAM2) following 3D6 antibody injection. We explored "CCL" signaling, which suggested a change in outgoing signaling coordinated by all microglia types targeting homeostatic microglia, surprisingly not targeting DAM1 or DAM2. We then analyzed enriched signaling pathways clustered by k-means river plots and identified pathways enriched and dampened with acute 3D6 treatment.

Discussion: Together these data supply evidence for significant involvement of microglia in the anti-Aβ response in the brain after just 3 days. Most interestingly, there are changes in cytokine/chemokine signaling across microglia subtypes, specifically with communication in CCL pathways targeting homeostatic microglia and T cells. These acute signaling changes provide novel insights and generate unique hypotheses on the brain's immediate immune reaction to anti-Aβ antibody.

Highlights: Intracranially injected anti-amyloid beta (anti-Aβ) antibody promotes an immediate microglial response.3D6 exposure resulted in fewer homeostatic and motile microglia subtypes.Acute 3D6 enriches for chemokine/cytokine signaling pathways.

Keywords: CCL; anti‐Aβ; anti‐amyloid; chemokine; microglia; single‐cell sequencing.

FIGURE 1

Fewer homeostatic and motile microglia after acute 3D6 exposure. (A) Experimental design showing SAA mice aged to 17 months when surgery for intracranial injections of 3D6 or IgG antibody into the frontal cortex was performed. Single‐cell harvest was performed 3 days post‐surgery for acute timepoint assessment. (B) Demographics of the mice utilized for single‐cell sequencing. (C) UMAP of 26 clusters. (D) Immune cells (microglia, macrophages, and other innate/adaptive immune cells) were annotated by both PanglaoDB (not shown) and gene expression based on previous studies and canonical markers. (E) Cluster 7 upregulated genes GO term enrichment aided in annotation of this cluster as motile/mobile microglia. (F) Bar graph of percentage of cells per cluster captured for our single‐cell analysis. Each cluster has an n = 4 per treatment and t‐test between treatments per cluster were performed. (#: 0.10 > p > 0.05, *: 0.05 > p > 0.01, **: 0.01 > p > 0.001). GO, gene ontology; IgG, immunoglobulin G

FIGURE 2

3D6 results in upregulation of cytokine/chemokine pathways in various microglia subtypes. (A) Number of DEgenes for 3D6 upregulated and downregulated genes compared to IgG across each cluster. (B) KEGG enrichment from upregulated genes in Cluster 2 homeostatic‐proliferative microglia. (C) CNET plot of genes contributing to top KEGG pathways in B. (D) Foldchanges of genes contributing to cytokine/chemokine‐related KEGG pathway terms in C. (E) KEGG enrichment from upregulated genes in Cluster 8 IRM microglia. (F) CNET plot of genes contributing to top KEGG pathways in E. (G) Foldchanges of genes contributing to cytokine/chemokine‐related KEGG pathway terms in F. (H) KEGG enrichment from upregulated genes in Cluster 9 DAM2 microglia. (I) CNET plot of genes contributing to top KEGG pathways in H. (J) Foldchanges of genes contributing to cytokine/chemokine‐related KEGG pathway terms in I. (K) KEGG enrichment from upregulated genes in Cluster 10 DAM2 microglia. (L) CNET plot of genes contributing to top KEGG pathways in K. (M) Foldchanges of genes contributing to cytokine/chemokine‐related KEGG pathway terms in M. IgG, immunoglobulin G

FIGURE 3

Cellular communication is dampened with acute 3D6 treatment. (A) Outgoing versus incoming interaction strength scatter plots for 3D6 (left) and IgG (right). (B) Bar graph of total number of interactions for incoming (left) and outgoing (right) for each cluster. Bars in teal represent 3D6 treated, while bars in purple represent IgG treated. (C) Heatmap of the differential number of interactions (left) and interaction strength (right) when 3D6 and IgG are compared. Teal squares indicate increased communication measures between clusters favoring 3D6, while purple squares indicate increased communication measures favoring IgG. IgG, immunoglobulin G

FIGURE 4

Subtle differences in CCL communication after acute 3D6 exposure. (A) Heatmap showing signaling roles (sender, receiver, mediator, and influencer) of each cluster in the CCL signaling pathway for 3D6 cellular communication. CellChat analysis of sender and receiver refers to the weighted out‐degree or in‐degree of summed communication probabilities. Mediators are calculated with the flow‐betweenness score of a cluster on communication between other clusters. Influencers are calculated by information centrality and refer to global information flow importance. (B) Signaling roles (sender, receiver, mediator, and influencer) of each cluster in the CCL signaling pathway for IgG cellular communication. (C) Chord plot of CCL signaling in 3D6‐treated brains. Outer ring represents each cluster. Inner ring colors nested within the outer ring represent the interactions outgoing from the outer ring to the inner ring, each with a chord across the middle of the circle to denote active communication pathways. Differences can be noted in the absence/presence of chords and complexity changes in inner ring colors for each cluster. (D) Chord plot of CCL signaling in IgG‐treated brains. (E) Hierarchical plots showing CCL signaling in 3D6‐treated brains. Closed circles and labels on outside are source clusters, while open circles are receivers. Left part of panel displays all microglia as receivers (open circles in the middle), while right part of panel displays all other cell types as receivers (open circles in the middle). (F) Hierarchical plots showing CCL signaling in IgG‐treated brains. Closed circles and labels on outside are source clusters while open circles are receivers. Left part of panel displays all microglia while right part of panel displays all other cell types. (G) Inferred relative contributions of CCL signaling ligands and receptors for 3D6‐treated brains. (H) Inferred relative contributions of CCL signaling ligands and receptors for IgG‐treated brains. IgG, immunoglobulin G

FIGURE 5

Changes in global communication patterns after acute 3D6 treatment. (A) River plot flow diagram of 3D6‐treated brain global communication. Left part of panel displays clusters contributing to pattern with outgoing signaling. Right part of panel displays CellChatDB‐signaling pattern enriched and colored by pattern. (B) River plot flow diagram of IgG‐treated brain global communication. Left part of panel displays clusters contributing to pattern with outgoing signaling. Right part of panel displays CellChatDB‐signaling pattern enriched and colored by pattern. (C) Global signaling patterns in CellChatDB that differ between 3D6‐ and IgG‐treated brains. IgG, immunoglobulin G