C9orf72-ALS human iPSC microglia are pro-inflammatory and toxic to co-cultured motor neurons via MMP9

Abstract

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease characterized by progressive motor neuron loss, with additional pathophysiological involvement of non-neuronal cells such as microglia. The commonest ALS-associated genetic variant is a hexanucleotide repeat expansion (HRE) mutation in C9orf72. Here, we study its consequences for microglial function using human iPSC-derived microglia. By RNA-sequencing, we identify enrichment of pathways associated with immune cell activation and cyto-/chemokines in C9orf72 HRE mutant microglia versus healthy controls, most prominently after LPS priming. Specifically, LPS-primed C9orf72 HRE mutant microglia show consistently increased expression and release of matrix metalloproteinase-9 (MMP9). LPS-primed C9orf72 HRE mutant microglia are toxic to co-cultured healthy motor neurons, which is ameliorated by concomitant application of an MMP9 inhibitor. Finally, we identify release of dipeptidyl peptidase-4 (DPP4) as a marker for MMP9-dependent microglial dysregulation in co-culture. These results demonstrate cellular dysfunction of C9orf72 HRE mutant microglia, and a non-cell-autonomous role in driving C9orf72-ALS pathophysiology in motor neurons through MMP9 signaling.

Fig. 1. C9orf72-ALS patient-derived iPSC microglia show key pathological features associated with the hexanucleotide repeat expansion in C9orf72.

a Schematic overview showing experimental setup and analyses performed on C9orf72 mutant (C9) iPSC-derived microglia in monoculture (pMGL) in this study, unstimulated or treated with LPS (100 ng/mL, 48 h). b Live cell (left) and immunofluorescent images (right, insets) of pMGL differentiated from healthy control (HC), isogenic control (IC), and C9 iPSC lines showing microglial morphology (images from one representative differentiation). Scale bars: 200 μm (left), 25 μm (right), 15 μm (insets). c Left: Western blot against C9orf72 comparing HC motor neurons (MNs) and pMGL. Right: Quantification normalized to the housekeeping gene GAPDH (n = 3 lines per cell type; n = 1 differentiation for MN, n = 2 differentiations for pMGL, indicated by I/II). d RT-qPCR for C9orf72 in unstimulated and LPS-primed pMGL normalized to the housekeeping gene TBP (n = 3 lines per condition, n = 1 differentiation). e Left: exemplar Western blot against C9orf72 in unstimulated and LPS-primed HC, IC, and C9 pMGL. Right: quantification normalized to the housekeeping gene b-actin. C9orf72 haploinsufficiency is found in C9 pMGL compared with HC (n = 3 lines per condition, n = 3 differentiations each) but not in the IC-C9-2 comparison (n = 1 line per condition, n = 3 differentiations each). f, g Analysis of Poly(GA)/Poly(GP) expression by MSD ELISA in HC, IC, and C9 MNs and pMGL. Both DPRs are detectable in C9 MNs and pMGL but not controls, while pMGL appear to have lower expression (n = 1 differentiation for n = 3 lines for top graphs, n = 1 line for bottom graphs). h Left: detection of nuclear sense (G₄C₂) and anti-sense (G₂C₄) RNA foci by RNAscope in C9 pMGL but not controls (n = 1 differentiation for n = 3 lines for top graph, n = 1 line for bottom graph). Scale bars: 25 μm, 2 μm (insets). Right: quantification showing anti-sense foci are more common than sense foci in C9 pMGL. Single data points and means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Two-tailed unpaired t-test (c). One-way ANOVA with Tukey’s post-hoc test (d, e, h). No statistical tests for f and g due to the small sample size with high variability. The graphics for panel a were created using Biorender.com. Source data are provided as a Source Data file.

Fig. 2. RNA sequencing of C9orf72 mutant iPSC microglia identifies dysregulation of pathways associated with immune cell activation, cyto/-chemokines, lysosomes, and the ER.

a Schematic overview showing the experimental setup for RNA sequencing of C9orf72 mutant (C9) and healthy control (HC) iPSC-derived microglia in monoculture (pMGL), unstimulated or treated with LPS (100 ng/mL, 48 h). b Principal component analysis (PCA) plot based on the top 500 most variable genes with biggest variance in unstimulated and LPS-stimulated (100 ng/mL, 48 h) HC and C9 pMGL showing PC2 and PC6 (n = 3 lines per condition, n = 3 differentiations each), with response to LPS treatment on PC2 and separation of genotypes on PC6. For downstream analyses shown in this figure, the different differentiations for each line were merged. c Venn diagram showing differentially expressed genes (DEGs), defined as |log₂ fc| > 0.5 and adjusted p-value < 0.05, and overlap for unstimulated and LPS-stimulated comparisons. d Bar plots showing top 10 significantly enriched GO biological process terms after over-representation analysis (ORA) using the clusterProfiler package for the DEGs between LPS-primed C9 and HC pMGL. x-axis showing Gene Ratio. The Benjamini-Hochberg corrected p-values for pathway enrichment for each term is indicated by the color of the bars. e Box plots of normalized counts from RNA seq analysis for CXCL1 and CXCL6, two DEGs in LPS-primed C9 pMGL versus HC (n = 3 lines per condition). Box plots show the median (center line), upper and lower quartiles (box limits), 1.5x interquartile range (whiskers), and outliers (points). f Bar plots showing top 10 significantly different GO biological process terms for LPS-stimulated C9 pMGL versus HC based on gene set enrichment analysis (GSEA) using the clusterProfiler package with the whole transcriptome ranked by log₂fc. x-axis showing normalized enrichment score (NES). The Benjamini–Hochberg corrected p-values for pathway enrichment for each term is indicated by the color of the bars. g Summary schematic illustrating the pathways with most prominent enrichment in C9 pMGL in the RNA seq analysis. The graphics for panel a, g were created using Biorender.com.

Fig. 3. C9orf72 mutant iPSC microglia are pro-inflammatory with consistently upregulated expression and release of MMP9.

a ELISA quantification of MMP9 in supernatants from unstimulated and LPS-stimulated healthy control (HC), isogenic control (IC), and C9orf72 mutant (C9) microglia in monoculture (pMGL) shows significantly increased MMP9 release in the C9-HC (n = 3 lines per condition, n = 5 differentiations each) and IC-C9-2 pMGL comparisons (n = 1 line per condition, n = 6 differentiations each). b Fluorometic activity assay for active MMP9 in supernatants from unstimulated and LPS-stimulated HC, IC, and C9 pMGL shows significantly increased MMP9 activity in the C9-HC (n = 3 lines per condition, n = 5 differentiations each) and IC-C9-2 pMGL comparison (n = 1 line per condition, n = 6 differentiations each). c RT-qPCR for MMP9 expression in unstimulated and LPS-stimulated HC, IC, and C9 pMGL normalized to the housekeeping gene TBP showing significantly increased MMP9 expression in the C9-HC (n = 3 lines per condition, n = 3 differentiations each) and numeric increase in IC-C9-2 pMGL comparison (n = 1 line per condition, n = 3 differentiations each). d Top: exemplar Western blot against the pro-form and active form of MMP9 in unstimulated and LPS-stimulated HC, IC, and C9 pMGL. Bottom: quantification shows significantly increased expression of pro-MMP9 in the C9-HC pMGL comparison and a numeric increase in the IC-C9-2 pMGL comparison. Active MMP9 is significantly increased in both comparisons. Both normalized to the housekeeping gene b-actin (n = 3 lines per condition for C9-HC comparison, n = 1 line for the IC-C9-2 comparison, n = 4 differentiations each). Single data points and means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. One-way ANOVA with Tukey’s post-hoc test (a–d). Source data are provided as a Source Data file.

Fig. 4. Unstimulated C9orf72 mutant iPSC microglia do not cause overt toxicity to co-cultured healthy motor neurons but show an altered cytokine profile and upregulate DPP4 release.

a Experimental setup for motor neuron (MN)-microglia (MG) co-culture. b Percentage of MNs/MG in co-cultures of healthy control (HC) MNs and HC, isogenic control (IC), and C9orf72 mutant (C9) MG showing a ~ 1:1 ratio (n = 3 microglial lines (C9-HC comparison), n = 1 microglial line (IC-C9-2 comparison), n = 3 differentiations). c Left: Exemplar images showing the apoptosis marker cleaved caspase 3 (CC3) in HC MNs in co-culture with MG. Scale bars: 15 μm. Right: Quantification of relative fold change of CC3 expression (n = 3 microglia lines (C9-HC comparison), n = 1 microglial line (IC-C9-2 comparison), n = 3 differentiations). d, e Patch clamping analysis of resting membrane potential and capacitance of HC MNs in co-culture with MG (datapoints are single MNs from co-cultures with n = 3 microglial lines (d, +HC: n = 18, +C9: n = 17; e, both n = 17), n = 1 differentiation). f Multi-electrode array analysis (MEA) of mean firing rate of HC MNs in co-culture with MG (n = 3 microglial lines (C9-HC comparison), n = 1 microglial line (IC-C9-2 comparison), n = 2 wells per n = 3 differentiations). g Left: exemplar images of neurites of HC MNs in co-culture with MG. Right: quantification of neurite outgrowth (n = 3 microglial lines (C9-HC comparison), n = 1 microglial line (IC-C9-2 comparison), n = 3 differentiations). Scale bars: 50 μm. h Left: exemplar images of synaptophysin (sy-physin) staining in co-culture of HC MNs with MG. Right: quantification of the number of synaptophysin particles (n = 3 microglial lines (C9-HC comparison), n = 1 microglial line (IC-C9-2 comparison), n = 3 differentiations). Scale bars: 50 μm. i Relative fold change in co-culture supernatant for selected targets using a cytokine array (pooled from co-cultures of HC MNs with n = 3 microglial lines (C9-HC comparison), n = 1 microglial line (IC-C9-2 comparison), n = 3 differentiations). j ELISA quantification of DPP4 in co-culture supernatants from HC MNs and MG (n = 3 microglial lines for C9-HC comparison, n = 1 microglial line for IC-C9-2 comparison, n = 3 differentiations). Single data points and means ± SEM. **P < 0.01; ***P < 0.001; ns not significant. Two-tailed unpaired t-test (c–e, g, h, j), two-way ANOVA with Tukey’s post-hoc test (f). The graphics for panel a were created using Biorender.com. Source data are provided as a Source Data file.

Fig. 5. LPS-stimulated C9orf72 mutant iPSC microglia increase apoptotic marker expression in co-cultured healthy iPSC motor neurons and upregulate DPP4 release via an MMP9-dependent mechanism.

a Experimental setup for LPS-stimulated motor neuron (MN)-microglia (MG) co-cultures. b Percentage of MNs/MG in LPS-stimulated co-cultures of healthy control (HC) MNs and HC, isogenic control (IC), and C9orf72 mutant (C9) MG showing a ~ 30% microglia content (n = 3 microglial lines, n = 4 differentiations (C9-HC comparison), n = 1 microglial line, n = 5 differentiations (IC-C9-2 comparison)). c Exemplar Western blot against IBA1, ChAT, and MMP9 in LPS-stimulated co-cultures of HC MNs with MG. d Quantification of blot shown in c demonstrating pro-MMP9 and active MMP9 normalized to the housekeeping gene b-actin are significantly increased in LPS-stimulated co-cultures of HC MNs with C9 MG compared with controls (n = 3 microglial lines (C9-HC comparison), n = 1 microglial line (C9-2-IC comparison), n = 3 differentiations). e Exemplar images of apoptosis marker cleaved caspase 3 (CC3) in HC MNs in LPS-primed co-cultures with MG plus MMP9 inhibitor 1 (3 μM, MMP9i) or DMSO as vehicle control. Scale bars: 25 μm. f Quantification showing significantly increased relative fold change in CC3 expression in LPS-primed co-cultures with C9 MG compared with HC MG. Administration of MMP9i ameliorates microglial neurotoxicity. The C9-2-IC comparison shows the same pattern but does not reach statistical significance (n = 3 microglial lines, n = 4 differentiations (C9-HC comparison, n = 1 microglial line, n = 4 differentiations (IC-C9-2 comparison)). g Relative fold change in LPS-stimulated co-culture supernatant for selected targets using a cytokine array. Supernatant samples from LPS-primed co-cultures of HC MNs with C9 MG showing a dysregulated cytokine profile compared with HC MG, which is ameliorated by MMP9i treatment (pooled samples from n = 3 microglial lines with n = 3 differentiations each). h ELISA quantification of DPP4 in co-culture supernatants from LPS-stimulated co-cultures of HC MNs and MG showing significantly increased DPP4 levels, which is significantly reduced after MMP9i treatment (n = 3 microglial lines (C9-HC comparison), n = 1 microglial line (IC-C9-2 comparison), n = 3 differentiations). Single data points and means ± SEM. *P < 0.05; **P < 0.01; ****P < 0.0001; ns: not significant. Two-tailed unpaired t-test (d) and one-way ANOVA with Tukey’s post-hoc test (f, h). The graphics for panel a were created using Biorender.com. Source data are provided as a Source Data file.

Fig. 6. C9orf72 mutant iPSC microglia cause non-cell-autonomous toxicity to co-cultured healthy iPSC motor neurons after prolonged LPS treatment via an MMP9-dependent mechanism.

a Experimental setup for experimental setup for prolonged treatment of motor neuron (MN)-microglia (MG) co-cultures with LPS. b Quantification showing significantly increased relative fold change in cleaved caspase 3 (CC3) expression in co-cultures of healthy control (HC) MNs with C9orf72 mutant (C9) MG compared with HC MG after prolonged LPS treatment. Treatment with MMP9 inhibitor 1 (3 μM, MMP9i) ameliorates microglial neurotoxicity (n = 3 microglial lines per genotype, n = 3 differentiations). c Top: exemplar images showing co-cultures of HC MNs with HC and C9 MG after pro-longed LPS treatment. Bottom: quantification showing significantly reduced relative MN survival in co-cultures of HC MNs with C9 MG compared with HC MG after prolonged LPS treatment. Treatment with MMP9i ameliorates microglial neurotoxicity (n = 3 microglial lines per genotype, n = 4 differentiations). Scale bars: 100 μm. Single data points and means ± SEM. *P < 0.05; ns: not significant. One-way ANOVA with Tukey’s post-hoc test (b, c). The graphics for panel a were created using Biorender.com. Source data are provided as a Source Data file.