Dysregulated Expression of Inflammasome and Extracellular Matrix Genes in C9orf72-ALS/FTD Microglia

Abstract

Hexanucleotide repeat expansion (HRE) in the non-coding region of the gene C9orf72 is the most prevalent mutation in amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD). The C9orf72 HRE contributes to neuron degeneration in ALS/FTD through both cell-autonomous mechanisms and non-cell autonomous disease processes involving glial cells such as microglia. The molecular mechanisms underlying the contribution of C9orf72-HRE microglia to neuron death in ALS/FTD remain to be fully elucidated. In this study, we generated microglia from human C9orf72-HRE and isogenic iPSCs using three different microglia derivation methods. RNA sequencing analysis reveals a cell-autonomous dysregulation of extracellular matrix (ECM) genes and genes involved in pathways underlying inflammasome activation in C9orf72-HRE microglia. In agreement with elevated expression of inflammasome components, conditioned media from C9orf72-HRE microglia enhance the death of C9orf72-HRE motor neurons implicating microglia-secreted molecules in non-cell autonomous mechanisms of C9orf72 HRE pathology. These findings suggest that aberrant activation of inflammasome-mediated mechanisms in C9orf72-HRE microglia results in a pro-inflammatory phenotype that contributes to non-cell autonomous mechanisms of motor neuron degeneration in ALS/FTD.

Keywords: Amyotrophic lateral sclerosis; C9orf72; RNA sequencing; extracellular matrix; inflammasome; microglia.

Figure 1.

Generation and characterization of iPSC-derived microglia using three separate methods. (A–C) Schematics of three microglia differentiation protocols, referred to as “protocol P1” adapted from Douvaras et al. (A), “protocol P2” adapted from McQuade et al. (B), and “protocol P3” adapted from Haenseler et al. (C). (D–F) Representative images of microglia preparations differentiated from iPSC lines CS52-C9n6-ISO (CS52) and CS29-C9n1-ISO (CS29) with protocol P1 (D), P2 (E), or P3 (F) and immunostained with anti-IBA1 antibody (red) and counterstained with Hoechst (blue). Scale bar, 50 µm. (G–I) Heatmaps showing known microglia/macrophage marker gene expression in undifferentiated iPSCs (obtained from publicly available datasets) or microglia preparations differentiated from the CS52 and CS29 iPSC lines with protocol P1 (G), P2 (H) or P3 (I). Each column of the heatmap corresponds to one sample (undifferentiated iPSCs (hiPSCs) or isogenic iPSC-derived microglia (derived MG), as indicated at the bottom, and each line corresponds to one gene. Subgroups of genes with correlated expression patterns are portrayed by the dendrogram on the left side of each heatmap. Relative levels of gene expression are indicated by a color scale, with cool colors corresponding to low expression and warm colors corresponding to increased/high expression. Hierarchical clustering of rows was applied to the heatmaps to group genes with correlated expression patterns, resulting in different gene lists, depending on the samples analyzed in the given heatmap. The gene list used for this analysis was as follows: HEXB, P2RY12, CXC3CR1, P2RY13, TREM2, S100A8, TMEM119, S100A9, RNASE4, GPR34, FCRLS, SIGLECH, OLFML3, FOS, SCLO2B1, TGFBR1, SLC2A5, CAMP, ITGB5, CRYBB1, SYNGR1, GPR56, NGP, CMRF35, HPGD, GPR34, MERTK, C1QA, PROS1, GAS6, ITGAM, ITGB2, CSF1R, PTPRC, AIF1, ADORA3, LGMN, GPR84, CCR7, BCL2A1D, TNF, NCF1, GDF15, OSM, LMC25, H2-OA, CD83, CCL3, GNA15, IL1B, PFAU, CCL9, TMEM119, C1QA, LRF8, CXCL16, CH25H, HCK, CCL12, PTAFR, CD300A, LRT5, STPI1, SELPIG, SASH3, BSG, TLR2, P2RY6, CDI4, BCL2A1A, BCL2A1C, RUNX1, and SPI1.

Figure 2.

Comparison of microglia/macrophage gene expression in iPSC-derived microglia obtained with three different protocols. Heatmap showing expression of known myeloid/macrophage/microglial cell marker genes in microglia preparations differentiated from the CS52-C9n6-ISO (CS52) and CS29-C9n1-ISO (CS29) iPSC lines with protocols P1, P2, or P3. Each column corresponds to one isogenic sample (examined at the indicated differentiation day) and each line corresponds to one gene. Subgroups of genes with correlated expression patterns are portrayed by the dendrogram on the left side of the heatmap. Relative levels of gene expression are indicated by a color scale, with cool colors corresponding to low expression and warm colors corresponding to increased/high expression. The gene list used for this analysis was the same as the list shown in the legend to Figure 1, with addition of the following macrophage marker genes: EMILIN2, SELL, HP.

Figure 3.

Differential gene expression between corrected isogenic and C9orf72-ALS microglia derived using protocol P1. (A–H) Differential gene expression and gene ontology analysis of CS52-C9n6-M versus CS52-C9n6-ISO microglia (CS52 collectively) (A–D) or CS29-C9n1-M versus CS29-C9n1-ISO microglia (CS29 collectively) (E–H). (A and E) Principal component analysis plots showing the relationship between the transcriptomic profiles of CS52 (A) or CS29 (E) iPSC-derived microglia. Each plotted data point represents one biological replicate, and the different symbols refer to the different batches. Blue data points correspond to corrected isogenic samples (Isogenic); red data points correspond to C9orf72-ALS samples (ALS). The percentages of variance explained for dimensions 1 and 2 are indicated along the axes. (B and F) Volcano plots showing the log Fold-Change (X‐axis) and -log10 p value (Y‐axis), highlighting the differentially expressed genes (DEGs) in C9orf72-ALS microglia compared to isogenic microglia differentiated from the CS52 (B) or CS29 (F) iPSC lines. Each plotted data point represents one gene: grey data points correspond to genes that are not DEGs, red data points correspond to up-regulated DEGs, and blue data points correspond to down-regulated DEGs (p‐adjusted <0.05, with no Fold Change cutoff). The numbers of up- and down-regulated DEGs are indicated on top of each volcano plot. (C and G) Heatmaps of the topmost DEGs in C9orf72-ALS microglia compared to isogenic microglia differentiated from the CS52 (C) or CS29 (G) iPSC lines. For each heatmap, each column corresponds to one sample (Isogenic or ALS, as indicated at the bottom) and each line corresponds to one gene. Relative levels of gene expression are indicated by a color scale, with cool colors corresponding to low expression and warm colors corresponding to increased/high expression. (D and H) Gene ontology analysis tables showing KEGG pathways, biological processes, cellular components and molecular functions that are significantly over- or under-represented in C9orf72-ALS compared to isogenic microglia differentiated from either the CS52 (D) or CS29 (H) iPSC line. (I) Table showing the 20 DEGs common to CS52-C9n6-M and CS29-C9n1-M microglia. The p value and log Fold Change are indicated for each gene.

Figure 4.

Differential gene expression between corrected isogenic and C9orf72-ALS microglia derived using protocols P1 and P2. (A–H) Differential gene expression and gene ontology analysis of CS52-C9n6-M versus CS52-C9n6-ISO microglia (CS52 collectively) generated using protocol P2 (A–D) or CS29-C9n1-M versus CS29-C9n1-ISO microglia (CS29 collectively) generated using protocol P3 (E–H). (A and E) Principal component analysis plots showing the relationship between the transcriptomic profiles of CS52 microglia generated with protocol P2 (A) or CS29 microglia generated with protocol P3 (E). Each plotted data point represents one biological replicate, and the different symbols refer to the different batches. Blue data points correspond to isogenic samples (Isogenic); red data points correspond to C9orf72-ALS samples (ALS). The percentages of variance explained for dimensions 1 and 2 are indicated along the axes. (B and F) Volcano plots showing the log Fold-Change (X‐axis) and -log10 p value (Y‐axis), highlighting the differentially expressed genes (DEGs) in C9orf72-ALS microglia compared to isogenic microglia differentiated from the CS52 (B) or CS29 (F) iPSC lines using protocols P2 or P3, respectively. Each plotted data point represents one gene: grey data points correspond to genes that are not DEG, red data points correspond to up-regulated DEGs, and blue data points correspond to down-regulated DEGs (p‐adjusted <0.05, logFC >0.26). The numbers of up- and down-regulated DEGs are indicated on top of each volcano plot. (C and G) Heatmaps of the topmost DEGs in C9orf72-ALS microglia compared to isogenic microglia differentiated from either the CS52 iPSC line with protocol P2 (C) or the CS29 iPSC line with protocol P3 (G). For each heatmap, each column corresponds to one sample (Isogenic or ALS, as indicated at the bottom) and each line corresponds to one gene. Relative levels of gene expression are indicated by a color scale, with cool colors corresponding to low expression and warm colors corresponding to increased/high expression. (D and H) Gene ontology analysis tables showing KEGG pathways, biological processes, cellular components and molecular functions that are significantly over- or under-represented in C9orf72-ALS compared to isogenic microglia differentiated from either the CS52 iPSC line with protocol P2 (D) or CS29 iPSC line with protocol P3 (H).

Figure 5.

Differentially secreted proteins in C9orf72-ALS microglia identified by multiplex immunoassay. (A-C) Heatmaps showing the top 50 most differentially secreted proteins in CS29 (A) or CS52 (B) iPSC-derived microglia cultures differentiated with protocol P1, and CS52 iPSC-derived microglia cultures derived with protocol P2 (C). Levels of 180 secreted proteins in cell culture supernatants were measured using the nELISA bead-based multiplex immunoassay platform. For each heatmap, columns correspond to isogenic or C9orf72-ALS microglia samples and each line corresponds to one protein. Subgroups of proteins with correlated level patterns are portrayed by the dendrogram on the left side of each heatmap. Relative levels of proteins are indicated by a color scale, with cool colors corresponding to low secreted levels and warm colors corresponding to high secreted levels.

Figure 6.

Effect of C9orf72-ALS microglia conditioned media on C9orf72-ALS motor neurons. (A) Representative immunocytochemistry images of motor neurons (Motoneurons) derived from isogenic (left column) or C9orf72-ALS (CS29-ALS) (right column) iPSCs and maintained in culture for 3 weeks in the absence of microglia conditioned media (MCM) (top images: No MCM) or in the presence of conditioned media from isogenic microglia (middle images: isogenic MCM) or C9orf72-ALS (CS29-ALS) MCM (bottom images). Cultures were immunostained with anti-CC3 antibody (magenta), anti-CHAT antibody (cyan), and counterstained with Hoechst (blue). Scale bars, 50 µm. (B) Plot showing the quantification of the percentage of apoptotic CC3-positive motor neurons in isogenic or CS29 motor neuron preparations after 3 weeks of culture in each of the following three conditions: absence of MCM (No MCM); presence of isogenic microglia conditioned media (ISO MCM); or presence of CS29 MCM (C9-ALS MCM). Statistics were conducted at the level of n = 3 biological replicates per group and per condition (black squares for Isogenic MNs; red dots for ALS MNs). For each biological replicate, the percentages were computed on three images of >300 cells in three random fields (grey squares for Isogenic MNs; pink dots for ALS MNs). Two-way ANOVA and Tukey’s multiple comparisons post-test; n.s. = non-significant; * = p value < 0.05; **p value < 0.005; ***p value < 0.0005.