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

ALS features progressive degeneration of motor neurons and contributions from non-neuronal cells, notably microglia. The most common genetic cause is a hexanucleotide repeat expansion (HRE) in C9orf72, which in neurons produces haploinsufficiency, RNA foci, dipeptide repeats (DPRs), and TDP-43 mislocalization. Microglia express high levels of C9orf72 and can adopt pro-inflammatory states in response to stimuli like LPS. Prior mouse studies of C9orf72 knockout showed pro-inflammatory microglia and neuronal toxicity, but whether human iPSC-derived microglia carrying the pathogenic HRE show disease-associated phenotypes and exert non-cell-autonomous effects on motor neurons remained unclear. This study aimed to characterize human C9orf72-HRE iPSC-derived microglia, define transcriptomic and secretory changes (especially under LPS priming), test their effects on co-cultured healthy iPSC-derived motor neurons, and assess whether MMP9 mediates any observed neurotoxicity. The study also assessed DPP4 as a potential secreted marker of microglial dysfunction.

The paper situates ALS as a disorder with cell-autonomous neuronal degeneration and non-cell-autonomous contributions from glia. Microglia activation correlates with ALS progression and is especially evident in C9orf72-HRE patients. Mechanisms in neurons include C9orf72 haploinsufficiency, RNA foci, DPRs, and TDP-43 pathology. In myeloid cells and microglia, C9orf72 loss-of-function in mice induces pro-inflammatory states and toxicity to neurons, implicating endolysosomal and autophagy pathway disturbances. However, knockout models do not capture gain-of-function HRE effects. Recent iPSC-microglia studies reported pathological C9orf72 features (RNA foci, DPRs) with moderate intrinsic phenotypes or autophagy-related dysfunction, but comprehensive analysis of human C9orf72-HRE microglia, their secretome, and non-cell-autonomous impact on motor neurons, particularly under pro-inflammatory priming, required investigation.

• Cell models: Human iPSC lines from three C9orf72-ALS patients (HRE carriers), three sex- and age-matched healthy controls, and one isogenic control line (CRISPR HDR expansion-corrected) were used. iPSCs were cultured in mTeSR1 and differentiated into microglia precursors using an embryoid body protocol with BMP4, VEGF, SCF, then matured into iPSC-derived microglia (pMGL) with IL-34, BDNF, GDNF, and Laminin for ~14 days. iPSCs were also differentiated into spinal motor neurons (MNs) via patterning with retinoic acid and smoothened agonist.
• Microglia characterization: Morphology and marker expression (IBA1, TMEM119, P2RY12) assessed by immunofluorescence, Western blot, and RT-qPCR. Phagocytosis measured using pHrodo zymosan particles with inhibitor controls (Cytochalasin D, Bafilomycin A1). Viability with MTS assay. TDP-43 localization quantified by imaging.
• HRE pathology in microglia: C9orf72 expression measured by RT-qPCR/Western; DPRs Poly(GA) and Poly(GP) quantified by MSD ELISA; RNA foci (sense and antisense) detected by RNAscope.
• LPS and cytokine priming: Microglia monocultures treated with LPS (100 ng/mL, 48 h; or repeated on days 0, 2, 4). Alternative priming with TNF (50 ng/mL) + IL1β (20 ng/mL) for 48 h.
• RNA-seq: Unstimulated and LPS-primed microglia from three C9orf72 and three control lines, three independent differentiations per line. PolyA-selected, strand-specific libraries (NEBNext Ultra II), paired-end 150 bp on NovaSeq6000 (≥30M reads/sample). Alignment with HISAT2 to GRCh38.p13; counts with featureCounts; DE analysis with DESeq2 including sex, with |log2fc|>0.5 and FDR<0.05 as DEG threshold; ORA and GSEA with clusterProfiler.
• Co-culture: Healthy iPSC-MNs co-cultured with microglia at 1:1 ratio for ≥14 days. LPS co-culture paradigms: (i) 3 doses over 10 days (DIV25–DIV35); (ii) prolonged 10 doses over 20 days without BDNF/GDNF. In some experiments, microglia cultured in inserts to prevent direct contact.
• MMP9 modulation: Recombinant human active MMP9 (30 ng/mL) applied to MNs ± MMP9 inhibitor I (3 μM). During LPS paradigms, MMP9 inhibitor I (3 μM) or DMSO vehicle was co-administered.
• Readouts: Western blot for C9orf72, MMP9 (pro and active), IBA1, ChAT; ELISAs for MMP9, DPP4, CXCL10; active MMP9 fluorometric activity assay; cytokine/protease arrays of supernatants; MN viability (MTS); apoptosis marker cleaved caspase-3 (CC3) in MNs by ICC and image quantification; neuron counts and microglia:neuron ratios; neurite outgrowth (TUJ1 area), synaptophysin puncta; MN electrophysiology (whole-cell patch clamp of resting membrane potential, capacitance, currents, AP properties) and MEA mean firing rate; NfL release by MSD.
• Statistics: t-tests for two-group, ANOVA with post hoc tests for multiple groups; significance at P<0.05. Biological replicates as specified per figure; data shown as mean ± SEM.

• Pathological C9orf72 features in microglia: Compared with healthy microglia, C9orf72-HRE microglia exhibited C9orf72 haploinsufficiency (by Western vs healthy controls), presence of DPRs Poly(GA)/(GP) (lower than MNs from same lines), and sense/antisense RNA foci; no TDP-43 mislocalization observed. C9orf72 expression increased upon LPS stimulation in microglia.
• Transcriptomics: RNA-seq revealed genotype separation and dysregulation in C9orf72-HRE microglia. DEGs vs healthy controls: 34 genes (unstimulated) and 41 genes (LPS-primed), with 9 overlapping. LPS-primed C9orf72 microglia showed enrichment of immune activation and cytokine/chemokine pathways, including upregulation of CXCL1 and CXCL6, and negative enrichment for ER- and lysosome-associated terms.
• Pro-inflammatory secretome with MMP9 upregulation: In LPS-primed microglial monocultures, MMP9 mRNA and protein (pro and active forms) and secreted levels/activity were consistently increased in C9orf72 microglia vs healthy and isogenic controls (ELISA, Western, activity assay). MMP9 was not detectable in MN monoculture supernatants, indicating microglial origin. Alternative priming with TNF+IL1β also increased MMP9 to a lesser extent. Some other proteases (e.g., cathepsin L, MMP-2, TIMP-2) were elevated in supernatants but not at transcript level.
• Unstimulated co-cultures: With healthy MNs, unstimulated C9orf72 microglia did not alter MN survival or function over ~2 weeks: similar CC3 levels, neuron counts, electrophysiological properties (resting membrane potential, capacitance, ionic currents, AP properties), mean firing rate (MEA), neurite outgrowth, synaptophysin puncta, and NfL release. However, supernatant profiling showed dysregulation with markedly increased DPP4 and elevated CXCL1 among others; ELISA confirmed DPP4 upregulation.
• LPS-primed co-cultures (10-day paradigm): Co-cultures with C9orf72 microglia showed significantly increased neuronal CC3 (apoptotic marker) compared with controls; MN viability reduced in insert-based setups, indicating soluble toxicity. Pro- and active MMP9 were significantly increased in C9orf72 co-cultures; MN monocultures lacked MMP9. MMP9 inhibitor treatment reduced CC3 in MNs and normalized cytokine profile, specifically lowering DPP4 and Endoglin in C9orf72 co-cultures but not controls. ELISA confirmed DPP4 reduction with MMP9 inhibition.
• Prolonged LPS exposure (20-day paradigm): C9orf72 microglia induced sustained apoptotic signaling and significant MN loss; co-treatment with MMP9 inhibitor ameliorated neurodegeneration and abolished differences vs healthy microglia.
• Mechanistic support: Exogenous recombinant MMP9 decreased MN viability, which was rescued by MMP9 inhibitor, supporting a causal role for MMP9. DPP4 release tracked with MMP9 activity, serving as an MMP9-dependent marker of microglial dysfunction in co-culture.

The study addresses whether human C9orf72-HRE microglia exhibit disease-associated phenotypes and cause non-cell-autonomous motor neuron injury. C9orf72-HRE iPSC microglia displayed both loss-of-function (haploinsufficiency vs healthy controls) and gain-of-function features (DPRs, RNA foci), implicating intrinsic pathogenic mechanisms in microglia. Transcriptomic analyses, especially under LPS priming, revealed robust enrichment of immune activation and cytokine/chemokine pathways and downregulation of ER/lysosomal terms, consistent with known roles of C9orf72 in endolysosomal/autophagic processes. Functionally, the central finding is that LPS-primed C9orf72 microglia become neurotoxic to healthy motor neurons via MMP9, as MMP9 is consistently upregulated and its inhibition rescues apoptotic signaling and neuron survival. Unstimulated co-cultures did not show overt neurotoxicity, suggesting that C9orf72 microglia possess latent neurotoxic potential requiring pro-inflammatory activation. DPP4 emerged as a sensitive readout of microglial dysregulation that is MMP9-dependent, indicating potential utility as a biomarker. These results refine the understanding of microglial contributions to C9orf72-ALS by pinpointing MMP9 as a key effector and suggesting therapeutic avenues focused on microglial MMP9 inhibition. The findings complement prior knockout-based models by highlighting gain-of-function HRE impacts and defining conditions under which microglial toxicity manifests.

Human iPSC-derived microglia from C9orf72-ALS patients exhibit intrinsic molecular pathology (haploinsufficiency vs healthy, DPRs, RNA foci) and, under pro-inflammatory priming, a pro-inflammatory transcriptome and secretome with consistent MMP9 upregulation. In co-culture with healthy iPSC motor neurons, LPS-primed C9orf72 microglia induce apoptotic signaling and, with prolonged exposure, neuron loss, both ameliorated by MMP9 inhibition, establishing MMP9 as a mediator and potential therapeutic target. DPP4 release serves as an MMP9-dependent marker of microglial dysfunction and may have biomarker potential. Future work should assess MMP9 and DPP4 in biofluids of C9orf72-ALS patients, evaluate microglial responses to C9orf72-mutant neurons and proteinopathies (TDP-43, DPRs), and test microglia-targeted MMP9 inhibition in vivo and in longer-term, more complex human model systems.

• In vitro iPSC-based models with a limited number of patient and control lines (n=3 per group) may not capture full patient heterogeneity.
• Neurotoxicity required exogenous pro-inflammatory stimulation (LPS); relevance to in vivo disease triggers and timing needs validation.
• Co-culture duration was relatively short; longer-term interactions might reveal additional phenotypes (e.g., synaptic changes) not detected here.
• Isogenic control comparisons did not show haploinsufficiency, indicating potential epigenetic or line-specific differences and limiting attribution of some effects purely to loss-of-function.
• The study focused on healthy MNs to isolate microglial effects; interactions with C9orf72-mutant MNs or other CNS cell types were not examined.
• Findings were not validated in patient biofluids or in vivo, so generalizability to clinical ALS remains to be established.