Following established protocols (Abud et al., 2017; McQuade et al., 2018), we differentiated one isogenic pair, up to seven control and eight C9orf72 ALS/FTD patient iPSC lines into microglia (Figure 1A, Supplementary Tables S1–3). At a mature stage of 40 days in vitro (DIV), both control and C9orf72 ALS/FTD iPSC-MG display a typical ramified microglia morphology (Figure 1B) and express classic microglial marker proteins including the myeloid transcription factor PU.1, purinergic surface receptor P2RY12, and C-X3-C Motif Chemokine Receptor 1 (CX3CR1), as confirmed via fluorescent immunostaining (Figure 1C, Supplementary Figure S1A). Differentiated microglia uniformly expressed the triggering receptor expressed on myeloid cells 2 (TREM2) and transmembrane protein 119 (TMEM119), further demonstrating a commitment to microglial fate, with no differences detected between C9orf72 and control iPSC-MG (Figure 1C, Supplementary Figures S1A, B). To validate microglial lineage, we performed Illumina paired-end deep RNA sequencing analysis (RNA seq) on four control and seven C9orf72 ALS/FTD lines of iPSC-MG (Figures 1D, E). At a transcriptional level, control and C9orf72 iPSC-MG clustered together and presented with a unique transcriptome compared to iPSCs differentiated into forebrain cortical neurons (iPSC-CN) (Figure 1D). Principal component analysis of this RNA seq dataset revealed a highly similar gene expression profile within iPSC-MG and showed a significant difference from iPSC-CNs confirming that these populations are distinct from each other (PC1, 94.87% variance; PC2, 1.66% variance; Figure 1E). To verify the absence of other cell types in the differentiated iPSC-MG cultures, we assessed the expression of known cell type-specific markers which revealed high expression of microglial-enriched genes and low expression of gene transcripts unique to astrocytes, oligodendrocyte precursor cells (OPC), oligodendrocytes, neurons, endothelial cells, and pericytes (Figure 1F) (Zhang et al., 2014). We observed the expression of endothelial and pericyte marker genes FTL1, GPC3, and FMOD; however, these genes are known to also be expressed in microglia (https://www.proteinatlas.org/). These data confirm that the differentiation of microglia from iPSC results in the production of brain-like microglia that express microglial-enriched genes and proteins, distinct from other cell types, yet no inherent differences were observed between control and human C9orf72 ALS/FTD patient iPSC-MGs.

To test if C9orf72 microglia exhibit intrinsic disease-mediated gene expression alterations, we analyzed our RNA sequencing data to determine differentially expressed genes (DEGs) between mature control and C9orf72 ALS/FTD iPSC-MG. This analysis revealed a few DEGs with significantly altered expression in C9orf72 ALS/FTD iPSC-MG [20 genes; log₂ fold change (FC) ± 1, unadjusted p < 0.005 Figure 2A, Supplementary Figure S2A]. The topmost downregulated genes included HOGA1, SEC14L5, FPR3, TSPAN18, and SEC14L5. Among the topmost upregulate genes are KCNK17, L1TD1, CLEC12A, and VAV3. We next determined if 881 RNA transcripts previously reported to be enriched in cortical microglia were differentially expressed in our dataset but found no significant dysregulation of these select genes (Figure 2B) (Gosselin et al., 2017). Additionally, RNA sequencing data from iPSC-MG cells did not show differences in homeostatic, interferon, activated, or NF-κB response genes (Lall et al., 2021), although varied gene expression was observed among patient lines (Supplementary Figures S3A–D). These data suggest that the C9orf72 HRE in iPSC-MG induces subtle changes in the overall cellular transcriptome including select genes associated with neuroinflammatory pathways. No significant alterations were detected in microglial-enriched genes when C9orf72 iPCS-MG were grown in mono-cultures without surrounding CNS cell types.

To test this hypothesis, we analyzed existing RNA sequencing datasets obtained through the Target ALS consortium and the New York Genome Center. We evaluated gene expression changes of the 881 microglial-enriched RNA transcripts in the frontal cortex and motor cortex brain tissues of C9orf72 ALS/FTD patients (Figures 2C, D, Supplementary Figure S2B, Supplementary Table 4). We also quantified gene expression changes in the occipital cortex as a brain region considered to be less affected by C9orf72 disease pathology (Supplementary Figures S2C, D, Supplementary Table S4). From the 881 microglial-enriched genes, 12 were dysregulated in the frontal cortex: LILRA6, LRRCA, MYO1G, RAB39A, ACSM5, RNF125, and RGS2; 14 microglial-enriched genes were dysregulated in the motor cortex (e.g., CD69, CEACAM21, GPR84, LIPG, and MYO1G) and 21 in the occipital cortex (e.g., CD69, EGR2, GPR183, RGS1, and SPP1; Figures 2C, D, Supplementary Figures S2B–D). The RNA sequencing data from C9orf72 ALS/FTD human tissue, similar to C9orf72 ALS/FTD iPSC-MG, did not show apparent differences in genes related to homeostatic, interferon, activated, or NF-κB pathways (Supplementary Figures S3E–H). The minimal changes observed in microglia-enriched genes from the bulk RNA seq datasets support the need for single-cell or single-nuclei RNA sequencing approaches to provide a more definite answer regarding microglia-specific differentially expressed genes in C9orf72 disease.

Glial cells have been shown to exhibit C9orf72-associated phenotypes in postmortem tissue, including HRE-associated RNA foci and DPR proteins, albeit to a lesser extent than neurons (Ash et al., 2013; Lagier-Tourenne et al., 2013; Mackenzie et al., 2013; Mizielinska et al., 2013; Sareen et al., 2013; Zhao et al., 2020). To evaluate C9orf72 iPSC-MG for C9orf72-specific disease mechanisms, we first tested C9orf72 iPSC-MGs for haploinsufficiency by examining the levels of the C9orf72 transcript in C9orf72 ALS/FTD and control iPSC-MG. Many studies have reported C9orf72 haploinsufficiency in patient tissue; however, variable results have been observed in human patient-derived C9orf72 iPSC-differentiated cells, including neurons and astrocytes (Donnelly et al., 2013; Sareen et al., 2013; Zhao et al., 2020). No significant differences in C9orf72 transcript levels were detected between control and C9orf72 groups in our RNA seq dataset (Figure 3A) or by quantitative RT-PCR of C9orf72 (Figure 3B), which is consistent with previously published data (Sareen et al., 2013; Zhao et al., 2020). To measure C9orf72 protein levels, we performed quantitative Western blot analysis from control and C9orf72 ALS/FTD iPSC-MG lysates, which revealed a significant reduction of C9orf72 protein in C9orf72 ALS/FTD iPSC-MGs and was rescued in an isogenic control line (Figures 3C, D).

We then performed a repeat-hybridization chain reaction (R-HCR) in mature C9orf72 iPSC-MG, a more sensitive approach to detect the presence of intranuclear RNA foci (Glineburg et al., 2021) (Figure 3E). A high percentage of mutant C9orf72 iPSC-MG contained repeat associated intranuclear RNA foci (Figure 3F) with an average number of two to five foci per cell for C9orf72 iPSC-MG and an average number of fewer than one foci per cell for control iPSC-MG (Figure 3G). A notable and significant difference was observed for C9-4 iPSC-MG compared to its isogenic control (Figures 3E–G).

Sense and antisense GGGGCC/CCCCGG repeat-associated non-AUG (RAN) translation produces five different DPR proteins that accumulate in cells and are proposed to contribute to neuronal toxicity and cellular dysfunction (Zu et al., 2013; O'Rourke et al., 2015; Jiang and Cleveland, 2016; Zhang et al., 2016b; Freibaum and Taylor, 2017; Gendron et al., 2017). Here, we assessed the presence of poly-(GP) in mature iPSC-MG cultures by measuring poly-(GP) abundance in cell lysates using a customized immunoassay (Andrade et al., 2020). We detected a significant increase in poly-(GP) levels in C9orf72 ALS/FTD iPSC-MGs compared to controls, including the isogenic pair C9–4 (Figure 3H), showing for the first time that the C9orf72 HRE translates into DPR proteins in endogenous C9orf72 HRE expressing microglial mono-cultures.

Cytoplasmic TDP-43 inclusions are one of the hallmark pathologies of C9orf72 ALS and FTD and have been reported in glial cells of C9orf72 ALS/FTD postmortem tissues (Al-Sarraj et al., 2011; Cooper-Knock et al., 2012; Schipper et al., 2016). To evaluate whether iPSC-MG mono-cultures exhibit cytoplasmic TDP-43 inclusions, we performed immunocytochemistry for TDP-43 and measured the nucleocytoplasmic (N/C) ratio of TDP-43 using confocal microscopy. No significant difference in the TDP-43 N/C ratio was observed between control and C9orf72 ALS/FTD microglia at 40 DIV (Supplementary Figures S4A, C). One hypothesis for TDP-43 mislocalization in ALS and FTD is a defect in nucleocytoplasmic trafficking leading to the accumulation of TDP-43 protein in the cytoplasm (Zhang et al., 2015; Chou et al., 2018; Moore et al., 2020). Our laboratory has recently shown that another RNA binding protein, the RNA editing enzyme adenosine deaminase acting on double-stranded RNA 2 (ADAR2), is mislocalized to and accumulates in the cytoplasm of motor neurons in C9orf72 ALS/FTD (Moore et al., 2019). We, therefore, wondered whether C9orf72 microglia would display similar ADAR2 mislocalization. Like TDP-43, immunocytochemistry of C9orf72 ALS/FTD iPSC-MG for ADAR2 revealed no nucleocytoplasmic mislocalization resulting in an unchanged N/C ratio in microglia mono-cultures (Supplementary Figures S4B, D). These data suggest that while C9orf72 ALS/FTD iPSC-MG mono-cultures do exhibit C9orf72 pathobiological phenotypes as shown in C9orf72 ALS/FTD iPSC neurons, there are no microglial cytoplasmic inclusions of TDP-43 or ADAR2.

Microglia can exacerbate or promote neurodegeneration by releasing pro-inflammatory cytokines, including interleukin-1 (IL1), interleukin-6 (IL6), and tumor necrosis factor-α (TNFα) or anti-inflammatory cytokines such as IL4 and IL10 (Smith et al., 2012; Moreno-Martinez et al., 2019; Olesen et al., 2020). As a result, neuroinflammation is considered a major contributor to neuronal dysfunction in neurodegenerative diseases, including ALS/FTD (Lall and Baloh, 2017; Beers and Appel, 2019; McCauley and Baloh, 2019; Olesen et al., 2020). To investigate if C9orf72 ALS/FTD iPSC-MG respond differently to extracellular stimuli, we treated four controls and three C9orf72 ALS/FTD lines of iPSC-MGs with the bacterial toxin, LPS, commonly used to evoke a general non-specific immune response in vitro and in vivo (Beurel and Jope, 2009; Abud et al., 2017; Furube et al., 2018; Hong et al., 2020). Following LPS treatment, we tested for the presence of released chemokines and cytokines in the iPSC-MG cell culture supernatants, as described (Abud et al., 2017). Upon LPS stimulation, both control and C9orf72 ALS/FTD iPSC-MG mono-cultures exhibited increased release of IL1α, IL1β, IL6, and TNFα compared to basal conditions but no differences between the experimental groups were observed (Figures 4A–D; individual cell response shown in Supplementary Figures S5A–D). These results confirm the ability of C9orf72 ALS/FTD patient-derived iPSC-MG mono-cultures to respond to extracellular stimuli and support previous in vitro studies showing no significant changes in inflammatory responses in murine microglia cell lines overexpressing C9orf72 HRE (Rostalski et al., 2020). We then treated five controls and four C9orf72 ALS/FTD lines of iPSC-MG with LPS and analyzed gene expression alterations by RNA sequencing. Both, healthy control and C9orf72 ALS/FTD microglia within their experimental group responded to LPS treatment with significant gene expression changes (Supplementary Figures S6A–B; log₂ fold change (FC) ± 1, p < 0.005). In both experimental groups, differential expression was found in genes involved in inflammatory processes and interferon-gamma signaling pathways, e.g., CXCL8, CXCL10, TNIP3, CCL4L2, CCL20, IFIT2, IFIT3, IL23A, IL12B, and IL6. Among the top LPS-induced significantly dysregulated disease-specific genes is the cytokine IL1α, which is highly expressed and secreted by microglia and is known to be a strong inducer of an A1 reactive astrocyte phenotype (Figure 4E, Supplementary Figure S5E; log₂ fold change (FC) ± 1, p < 0.005) (Zhang et al., 2014; Bennett et al., 2016; Liddelow et al., 2017).

A major function of microglia is to phagocytose unwanted toxic substances and cell debris that can negatively impact brain function. To determine whether control and C9orf72 ALS/FTD iPSC-MG differ in their phagocytic activity and their ability to degrade toxic products, we transiently treated iPSC-MGs with a fluorescently tagged synthetic amyloid beta (Aβ) cargo, known to be phagocytosed by microglia (Paresce et al., 1996; Parnaik et al., 2000; Paolicelli et al., 2017). We treated five healthy control and seven C9orf72 ALS/FTD lines of iPSC-MGs for 5 min with 1 μM Aβ (1-40)-TAMRA or vehicle and then assessed engulfment and clearance at 5 min and after 1 h. To quantify phagocytic activity, iPSC-MG were fixed and immunostained for TREM2, a known immune receptor selectively expressed in microglia (Figures 5A–F). After confocal microscopy and image analyses, we calculated the percentage of iPSC-MG surface area covered by Aβ (1-40)-TAMRA protein in individual cells at both time points. We observed varying levels of retention of phagocytosed Aβ depending on the patient lines (Figures 5H, J). At the 5 min time point, only one C9orf72 line retained significantly more Aβ (1-40)-TAMRA when compared to healthy controls. At the 1 h time point, three out of seven C9orf72 patient lines showed significantly increased area covered by Aβ when compared to healthy control lines. Increased retention of intracellular Aβ levels could suggest potential dysfunction of the endosomal–lysosomal pathways and an inability to adequately process phagocytosed cargo in select patient lines (Figures 5H, J). When grouped, no differences in phagocytosed content between four C9orf72 ALS/FTD and healthy control iPSC-MG were noted (Figures 5I, K). To take into account possible changes in cell size, we measured the total microglia cell surface. No significant differences in the surface area between experimental groups were observed (Figure 5G). Finally, we wondered whether microglia exposed to Aβ would alter their transcriptional profile. Therefore, a separate set of five control and four C9orf72 ALS/FTD lines of iPSC-MGs were treated with Aβ (1-40)-TAMRA and analyzed for gene expression changes via RNA sequencing analyses. Despite lowering the stringency of our DEG analyses to a p < 0.05, neither control nor disease iPSC-MG showed large numbers of dysregulated genes when we compared treated to untreated cells within each experimental group (Figures S6C, D; log₂ fold change (FC) ± 1, p < 0.05). However, when we examined DEG in Aβ-treated C9orf72 ALS/FTD iPSC-MG in reference to Aβ-treated control microglia, we noted several significantly differentially expressed genes [69 genes; log₂ fold change (FC) ± 1, p < 0.005; Figure 5L, Supplementary Figure S6E]. Among the upregulated genes were FCGR3A and DOCK, which are genes known to play a role in phagocytosis (Wu and Horvitz, 1998; Sivagnanam et al., 2010). Additional upregulated genes, such as RAB15, AHNAK, and SDC3, play a role in endocytic trafficking, cytoskeletal rearrangement, or cell migration (Carey, 1996; Zuk and Elferink, 2000; Dumitru et al., 2013). These results suggest that C9orf72 ALS/FTD iPSC-MG responds to Aβ stimuli with an altered transcriptional response of select gene pathways.

An important aspect of microglia–neuron communication in neurodegeneration is the role of microglia in the maintenance and refinement of synaptic networks through the selective pruning of synapses. This process occurs predominantly during development (Stevens et al., 2007; Tremblay et al., 2011; Bialas and Stevens, 2013). However, synaptic pruning pathways are known to be re-activated in neurodegeneration leading to synapse loss and contributing to cognitive impairments (Henstridge et al., 2016; Hong et al., 2016; Lui et al., 2016; Colom-Cadena et al., 2020). To determine if C9orf72 ALS/FTD iPSC-MG can phagocytose synapses or exhibit altered phagocytosis due to the presence of C9orf72 HRE, we exposed iPSC-MG mono-cultures to synaptoneurosomes (hSN) derived from healthy control human brain and assessed synaptoneurosomes engulfment via live confocal microscopy. Fresh frozen control human postmortem tissues were used to prepare hSN (Hesse et al., 2019; Tzioras et al., 2019). Synaptic fractions were enriched for pre-synaptic protein synaptophysin and the post-synaptic density 95 (PSD-95) compared to total brain homogenate and nuclear marker histone 3 (Supplementary Figures S7A, B). Control and C9orf72 ALS/FTD iPSC-MGs were labeled with the live cell nuclear marker Hoechst (Figures 6A–I) to identify individual cells followed by treatment with hSN fluorescently tagged with pHrodo succinimidyl ester (hSN-rodo; Figures 6A, J–Q). An increase in pHrodo fluorescence is indicative of the uptake of hSN-rodo into acidic intracellular compartments of iPSC-MGs. After an initial 6-h live cell imaging to determine an optimal time point of internalization of hSN-rodo in control and C9orf72 ALS/FTD iPSC-MGs, we observed that more than 60% of cells exhibit phagocytic activity at 2 h (Figure 6R). IPSC-MG phagocytosis was reduced significantly for both control and C9orf72 ALS/FTD groups in the presence of cytochalasin-D, an actin polymerization inhibitor known to inhibit phagocytosis (Figures 6R, S). We then treated iPSC-MGs (control, n = 7 lines; C9orf72, n = 6 lines; three replicates; six to seven images per replicate) with hSN-rodo and performed fluorescent live cell imaging for 2 h, capturing images every 10 min (Supplementary Figure S7C, Supplementary Video S1). The mean intensity of hSN-rodo was similarly increased in C9orf72 ALS/FTD iPSC-MGs compared to controls (Figures 6T, U). Together, these data suggest that C9orf72 ALS/FTD iPSC-MGs exhibit typical microglia phagocytic activity as shown by the engulfment of healthy human brain synaptoneurosomes. Additional studies are required to determine if C9orf72 ALS/FTD iPSC-MGs respond differently to diseased human brain synaptoneurosomes, which may contain specific signaling molecules necessary for microglia activation and elimination of synapses.

The majority of the iPSC lines used in our studies were purchased from Cedars Sinai-induced pluripotent stem cell core https://www.cedars-sinai.edu/research/areas/biomanufacturing/ipsc.html or obtained from collaborators who have characterized the lines in previous publications. All purchased iPSC lines from Cedars Sinai are characterized for (1) positive staining of pluripotency markers (Oct3/4, NANOG, SOX2, TRA-1-60, TRA-1-81, and SSEA4); (2) karyotyping to reveal no chromosomal abnormalities; (3) determine a pluripotency score via PrimeView global gene expression profile assay (PluriTest); (4) trilineage differentiation potential via Taqman hPSC scorecard assay to confirm appropriate expression of ectodermal, endodermal, and mesodermal factors in the iPSC cells; (5) cell line authentication analysis—STR Analysis—to confirm identity/profile matching score with primary tissue. A Certificate of Analysis (COA) for these lines is available upon request. Please see Supplementary Table S1 for comprehensive information on the lines used in these studies.

IPSCs were differentiated into microglia following an established protocol (McQuade et al., 2018). In brief, iPSCs were maintained in mTeSR Plus Kit (Stemcell Technologies # 05825) in 10 cm dishes. At DIV−1 of HPCs differentiation, 5- to 7-day-old iPSC cultures were then used to generate a cluster of differentiating 43 positive (CD43⁺) hematopoietic progenitor cells (HPCs) following a 12-day commercially available kit (STEMdiff Hematopoietic Kit; Stemcell Technologies # 05310). IPSCs were cleaned and 1/3 dish was gently dissociated with dispase (Stemcell Technologies # 07923) for 12–15 min at 37°C. IPSCs were then collected and spun down at 500 rpm for 1–2 min and resuspended in 2 mL of mTeSR Plus media with 20 μM ROCK inhibitor Y-27632 (Stemcell Technologies # 72304). Matrigel, hESC-Qualified Matrix (Corning # 354277), coated six-well plates were used to start the HPC differentiation. Using a 5-mL serological pipette, one drop or two drops of iPSCs were seeded per well into matrigel. Then, the next day, on DIV 0 of HPCs differentiation, wells with 80 small colonies per well were selected to start HPCs differentiation. IPSCs were fed following the manufacturer's instructions. On DIV 12 of HPCs differentiation, only the non-adherent HPCs were transferred to a new six-well plate to start microglia differentiation (DIV 12/0).

Matrigel, GFR (growth factor reduced) Membrane Matrix (Corning # 356231), coated six-well plates were prepared to start microglia differentiation (DIV12/0). HPCs were differentiated into microglia for 28 days using serum-free media conditions. HPCs in suspension were collected and spun down at 300 xG for 6 min and resuspended into microglia basal media (MBM, 2 mL/well) containing: DMEM/F12 no phenol (Gibco # 11-039-021), 2% Insulin Transferin Selenite (Gibco # 41400045), 2% B27 (Gibco # 17504-044), 0.5% N2 (Gibco # 17502-048), 1% Glutamax (Gibco # 35050-061), 1% NEAA (Gibco # 11140-050), 1% Pen/Strep (Gibco # 15140-122), 400 μM 1-Thioglycerol (Sigma # M1753), 5 μg/mL human insulin (Sigma # I2643), and supplemented with three growth factors (GFs): 100 ng/mL human recombinant interleukin-34 (IL-34; Peprotech # 200-34), 25 ng/mL macrophage colony-stimulating factor (M-CSF; Gibco # PHC9501), and 50 ng/mL transforming growth factor β1 (TGFβ1; Miltenyi Biotec # 130-108-969) (MBM + 3GFs); cytokines and growth factors are known to be essential for the development of microglia (Chen et al., 2002; Abutbul et al., 2012; Greter et al., 2012; Wang et al., 2012; Butovsky et al., 2014; Elmore et al., 2014; Abud et al., 2017; Pandya et al., 2017). Microglia differentiation starts (DIV 12/0) once the HPCs are transferred and plated at a density of 200 000 cells per well of a six-well plate. Cells will predominantly grow in suspension. On DIV 2, 4, 6, 8, and 10 of microglia differentiation, 1 mL of MBM + 3GFs media was added to each well. On DIV 12, a partial media change was done. IPSC-MGs from one six-well plate were spun down at 300 xG for 6 min and resuspended into MBM + 3 GFs and split back into the same six-well plate. On DIV 14, 16, 18, 20, 22, and 24, 1 mL of media was added per well. On DIV 25, MBM + 3GFs was changed to maturation media composed of MBM supplemented with the five growth factors (MBM +5 GFs): 100 ng/mL IL34, 25 ng/mL M-CSF, 50 ng/ml TGFβ1, 100 ng/mL cluster of differentiation 200 (CD200; Novoprotein # C31150UG), and 100 ng/mL fractalkine chemokine C-X3-C motif ligand 1 (CX3CL1; Peprotech # 300-31) (MBM + 5 GFs). The presence of CD200 and CX3CL1 in the culture media, both glial and neuronal molecules, is critical for microglia maturation and maintenance of an in vivo-like microglia resting state phenotype in an in vitro setting (Barclay et al., 2002; Kim et al., 2011; Kierdorf and Prinz, 2013) (Figure 1A). At DIV 28, IPSC-MG reached maturation. Moreover, 1 mL of MBM + 5GFs media was added to the cultures every other day. Mature cells were used for experimentation within 10 days (DIV 28–DIV38).

For cortical neuron differentiation, 70% of confluent iPS cells maintained in mTeSR Plus media on 10 cm dishes were used for embryoid bodies (EBs) formation. The cells were cultured in low attachment six-well plates (Greiner bio-one # 657970) using WiCell Medium containing: DMEM/F12 (Gibco #11330057), 25% knock-out serum replacement (Gibco # 10828-028), 1.3% L-glutamine (Gibco # 35050-061), 1.3% NEAA (Gibco # 11140-050), and 0.1 mM 2-Mercaptoethanol (Sigma# M3148), and placed on a shaker in the incubator for 8 days to allow EBs formation. EBs were then resuspended in Forebrain Neural Induction Media (FB-NIM) containing: DMEM/F12, 1% N2 supplement (Gibco # 17502-048), 1% NEAA, 2 ug/mL heparin (Sigma # H3149), and 10 μg/mL bFGF (Stemcell Technologies # 78003), and plated on to T25 flasks coated with basement membrane matrigel (Corning # 356234) to allow formation of neuronal rosettes. Neuronal rosettes were maintained in FB-NIM for the next 10 days and then, collected and maintained in suspension on a shaker with half FB-NIM media changes every other day to allow for neurosphere growth. Neurospheres were maintained for 20 days in FB-NIM and then resuspended using forebrain neuronal differentiation media (FB-DM) containing: Neurobasal Medium (Corning # 21103-049), 2% B27 (Gibco # 17504-044), 10 ug/mL BDNF (Stem cell technologies # 78005), 10 ug/mL GDNF (Stem cell technologies # 78058), 1 μg/mL laminin (Life technologies # 23017-015), 3.3 μg/mL cAMP (Stem cell technologies # 73884), 3.52 μg/mL Ascorbic acid (Stem cell technologies # 72132), 0.5 mM L-glutamine, and 1% NEAA, and then plated on to T25 flasks. iPSC cortical neurons were harvested on DIV 65-72 for RNA sequencing analysis.

For iPSC-MG DIV 28–33 of microglia differentiation, non-adherent iPSC-MGs were collected and plated onto a 4- or 8-well fibronectin (sigma # F0895; 1:40) coated glass bottom chamber slides (Ibidi # 80427 or #80827) at a seeding density of 250 000 or 125 000 cells per well, respectively. One hour after plating, cells were fixed with 4% paraformaldehyde (PFA; Electron Microscopy Sciences # 15714-S) for 20 min, washed three times in PBS for 5 min, and then blocked with 0.2% Triton X-100 and 5% Normal Goat Serum (Vector # S1000) for 1 h at room temperature. Primary antibodies were prepared in a blocking solution and applied overnight at 4°C. The following primary antibodies were used during our studies: anti-PU.1 (Cell Signaling Technology # 2266S) 1:500; anti-P2RY12 (Sigma # HPA014518) 1:500; anti-CX3CR1 (Biorad/AbD Serotec # AHP1589) 1:500; anti-TREM2 (Abcam # AB209814) 1:500; anti-TMEM119 (Abcam # ab185333) 1:100; anti-LAMP1 (Developmental Hybridoma Bank # H4A3-s) 1:100; anti-EAA1 (BD Biosciences # 610457) 1:700; anti-C9orf72 (Sigma # HPA023873) 1:100; anti-TDP-43 (Cell signaling # 89789, TDP-43 D9R3L) 1:500; and anti-ADAR-2 (Sigma # HPA018277) 1:500. Next, cells were washed in PBS three times for 7 min and then incubated consecutively with respective fluorophores secondary antibodies. Alexa Fluorophores (Invitrogen) were used at a 1:750 and prepared in blocking solution without triton and incubated for 45 min at room temperature. Cells were then washed with PBS three times for 7 min each and DAPI was applied. For nuclear markers TDP-43 and ADAR, wheat germ agglutinin 680 (Invitrogen # W32465) was used to label the iPSC-MG cell surface. Mounting media ibidi (ibidi # 50001) was used on chamber slides.

At DIV 28–30, iPSC-MG from eight C9orf72 ALS/FTD patient lines and four control lines were pelleted and lysed using QIAshredder (QIAGEN-79654) and RNA was isolated with RNeasy Mini Kit (QIAGEN-74104) following the manufacturer's instructions. RNA samples were measured for quantity with Quant-iT Ribogreen RNA Assay (Thermo Fisher, Cat. No. R11490) and quality with Agilent High Sensitivity RNA ScreenTape and buffer (Agilent, Cat. No. 5067-5579 & 5067-5580). For each RNA sample, an indexed, Illumina-compatible, double-stranded cDNA whole transcriptome library was synthesized from 1 μg of total RNA with Takara Bio's SMARTer Stranded Total RNA Sample Prep Kit – HI Mammalian (Takara Bio, Cat. No. 634876) and SMARTer RNA Unique Dual Index Kit (Takara Bio, Cat. No. 634418). Library preparation included ribosomal RNA depletion, RNA fragmentation (94°C for 3 min), cDNA synthesis, and a 12-cycle unique dual indexing enrichment PCR. Each library was measured for size with Agilent's High Sensitivity D1000 ScreenTape and reagents (Agilent, Cat. No. 5067–5584 and 5067–5603) and concentration with KAPA SYBR FAST Universal qPCR Kit (Kapa Biosystems, Cat. No. KK4824). Libraries were then combined into an equimolar pool which was also measured for size and concentration. The pool was clustered onto a paired-end flowcell (Illumina, Cat. No. 20012861) with a 20% v/v PhiX Control v3 spike-in (Illumina, Cat. No. FC-110-3001) and sequenced on Illumina's NovaSeq 6000. The first and second reads were each 100 bases. All Aβ (1–40) and LPS-stimulated cells at DIV 28-30 iPSC-MG from four C9orf72 ALS/FTD patient lines and five control lines were processed as previously described.

We accessed human brain tissue RNA sequencing performed by Target ALS and the New York Genome Center (http://www.targetals.org/research/resources-for-scientists/resource-genomic-data-sets/). Sixteen cases of control frontal cortex, eight C9orf72 ALS/FTD frontal cortex, 15 control motor cortex, 12 C9orf72 ALS/FTD frontal cortex, four control occipital cortex, and five C9orf72 ALS/FTD frontal cortex were evaluated for differential expression of microglia specific genes.

Fastq files were quality- and adapter-trimmed using cutadapt (version 1.14). Adapter-trimmed fastq files were then aligned to the human genome (hg38, gencode v29) using STAR (version 2.6.1d) with default options. RNA count matrices were pulled from aligned BAM files using featureCounts (version 1.6.4). All downstream statistical analysis was done in R (version 3.6.2) using raw count matrices from featureCounts as input. Low-expression genes were filtered such that genes with mean read counts <10 were removed from the analysis. Differential expression analysis was done using DESeq2 (version 1.26.0) using disease status or treatment regime as the model. Volcano plots were generated from DESeq2 output using EnhancedVolcano. Heatmaps were generated using heatmap from z-scores calculated from DESeq2 normalized gene counts. Tissue data from Target ALS was downloaded from the New York Genome Center as raw fastq files and pushed through an identical analysis pipeline as data generated in our lab.

We followed previously established protocols (Renton et al., 2011a) to determine the presence of the hexanucleotide repeat expansion (G₄C₂: >30) in C9orf72 of iPSC and iPSC-MG (Supplementary Table S2).

RNA was isolated using Qiagen RNeasy Micro Kit (Cat #74004) according to the manufacturer's instructions. RNA was reverse transcribed to cDNA with oligo(dT) with the Promega Reverse Transcriptase System (Cat # A3500) and analyzed using SYBR Green Master Mix (Applied Biosystems). C9orf72 (Forward- 5′- CAGTGATGTCGACTCTTTG−3′ and Reverse- 5′ AGTAGCTGCTAATAAAGGTGATTTG−3′) expression was normalized to RPL13A (Forward- 5′ CCTGGAGGAGAAGAGGAAAGAGA-3′ and Reverse- 5′ TTGAGGACCTCTGTGTATTTGTCAA-3′) or B2M (Forward- 5′ TGCTGTCTCCATGTTTGATGTATCT-3′ and Reverse- 5′ TCTCTGCTCCCCACCTCTAAGT-3′).

Microglia cell pellets were homogenized in RIPA lysis and extraction buffer (Thermo Scientific, Cat #89900), supplemented with protease inhibitor cocktail (complete, Roche) and phosphatase inhibitor cocktail (PhosSTOP, Roche). Protein concentration was determined by a BCA assay kit (Thermo Fisher, Cat# 23225). Cell lysates were separated on 4–20% protean TGX precast gels (Biorad, Cat # #4561096) and blotted onto nitrocellulose membranes (Biorad, Cat # 1704159). Membranes were blocked for 60 min with Odyssey blocking buffer (PBS, Li-Cor, Cat #927-40000) and incubated overnight at 4°C with anti-C9orf72 (GeneTex Cat #. GTX634482, 1:1000) and anti-actin (Sigma-Aldrich A5441, 1:5000) antibodies. We used bone marrow-derived macrophages (BMDM) from a C9orf72 knockout mouse (O'Rourke et al., 2016) to validate the antibody used for the Western blot analysis. After washing, membranes were incubated for 60 min with IRDye fluorescent secondary antibodies (Li-Cor). After washing, blots were subsequently analyzed with a Li-COR imaging system (Odyssey CLx). The Western blots were imaged and analyzed using the LICOR Odyssey imaging system which allows visualizing both faint and strong signals in a single image using increased sensitivity and no image saturation. Overexposure of high abundance proteins such as actin can mask the sample-to-sample variation of low-abundant proteins such as C9orf72, which can be overcome by using digital systems like the Licor Odyssey imaging platform.

We followed previously published methods (Glineburg et al. 2021). All main reagents and probes were purchased from Molecular Instrument, Inc. The negative control probe lacked any binding region with the tail ends of the initiator probe that binds to the hairpins. In brief, on DIV 30 of microglia differentiation, iPSC-MGs were plated onto eight chambers glass bottom slides (Ibidi #80827) coated with human fibronectin (sigma # F0895; 1:40) at a cell density of 125 000–150 000 cells per chamber. Fixation step: One hour after plating, iPSC-MG were washed once with 1x PBS and then fixed with 4% PFA for 10 min at room temperature. We further fix cells in 70% cold ethanol overnight at 4°C. Initiation/Hybridization step: Next day, 70% cold ethanol was removed and 250 μl per well of preheated hybridization buffer was added to cells and incubated at 45°C for 30 min. The hybridization buffer was then replaced with 125 μl of pre-warmed initiation probe and incubated at 45°C overnight (12–16 h). Approximately 4 nM C9 (CCCCGG)₆ and negative scramble initiator probes solutions were prepared in a 45°C pre-warmed hybridization buffer. Amplification step: Six hours before starting the amplification stage, the R-HCR probe washing buffer was thawed and warmed up at 45°C. Cells were then washed 4x with pre-warmed (45°C) HCR probe wash buffer for 5 min at 45°C. Then washed 2x for 5 min with 5x SSC-T (5 X SSC, 0.1 % Tween20) at room temperature. Cells were then left in 5x SSC-T in a humid chamber at room temperature until the amplification stage was started. Approximately 1 h before the start of the amplification stage, we thawed the amplification buffer and hairpin probes B1H1 and B1H2 Alexa Fluor-546 to room temperature. Separately, hairpin B1H1 and hairpin B1H2 were prepared by snap cooling to the appropriate volume of the 3 μM stock. Approximately 15 nM of each hairpin probe was used. Each hairpin was heated up at 95°C for 90 s, placed on ice, and then left in a dark drawer at room temperature (no ice) for 30 min. We combined the snap-cooled B1H1 hairpins and snap-cooled B1H2 hairpins in the appropriate volume of amplification buffer at room temperature. The pre-amplification solution was removed and 125 μl per well of the hairpin solution was added to each chamber. iPSC-MGs were incubated cells with hairpin probes in the dark at room temperature overnight (12–16 h). The next day, cells were washed x5 for 5 min each in 5x SSC-T at room temperature. Then, incubated with DAPI (made in 5x SSC-T) for 10 min at room temperature. Wash cells 2x for 7 min each in 5x SSC-T at room temperature. Mount with Ibidi mounting medium (Ibidi #50001). IPSC-MGs were then visualized and imaged using a Zeiss LSM800 laser scanning confocal microscope. Using Imaris Software from Bitplane, we quantified the percentage of iPSC-MG presenting HRE-associated nuclear foci and the total number of foci per cell.

Levels of poly-(GP) in cell lysates were measured in a blinded fashion using a Meso Scale Discovery (MSD) immunoassay and an MSD QUICKPLEX SQ120 instrument. A purified mouse monoclonal poly-(GP) antibody was used as both the capture and detection antibody (TALS 828.179, Target ALS Foundation). The capture antibody was biotinylated and used to coat 96-well MSD Small Spot Streptavidin plates, whereas the detection antibody was tagged with an electrochemiluminescent label (MSD GOLD SULFO-TAG). Lysates were diluted to the same protein concentration, and each sample was tested in duplicate. For each well, the intensity of emitted light, which is reflective of poly-(GP) levels and presented as arbitrary units, was acquired upon electrochemical stimulation of the plates.

On DIV 29, iPSC-MGs were plated onto four chamber glass slides coated with human fibronectin (sigma # F0895; 1:40) at a cell density of 250 000 cells per chamber. On DIV 30, control and disease iPSC-MGs were treated with LPS (100 ng/mL) for 6 h, based on previous studies (Balez et al., 2016; Muffat et al., 2016). The conditioned media was collected and analyzed for selective cytokine/chemokine profile using the U-PLEX Biomarker Group 1 (Human; Mesoscale #K15067L) Multiplex Assay per manufacturer's protocol.

On DIV 28–30, iPSC-MG from four C9orf72 ALS/FTD patient lines and five control lines were plated onto four chamber glass slides as described above. For LPS stimulation: One hour after plating, iPSC-MGs were treated for 6 h with LPS (100 ng/mL). All iPSC-MG were then collected and sent for RNA sequencing analysis. For Aβ (1-40) stimulation, iPSC-MGs were treated for 2 h with vehicle (DMSO) or 1 μM Aβ (1–40) TAMRA (human Aβ, AnaSpec # AS-60488; Paolicelli et al., 2017) prepared in microglia basal media with growth factors. All iPSC-MGs were collected at 2 h and sent for RNA sequencing analysis.

On DIV 30 of microglia differentiation, iPSC-MGs were plated onto four chamber glass slides coated with human fibronectin (sigma # F0895; 1:40) at a cell density of 250 000 cells per chamber. One hour after plating, iPSC-MGs were treated for 5 min with vehicle (DMSO) or 1 μM of fluorescently labeled Aβ (1–40) TAMRA (human Aβ, AnaSpec # AS-60488; Paolicelli et al., 2017) prepared in microglia basal media with growth factors. All iPSC-MGs were washed at 5 min with microglia basal media supplemented with growth factors. IPSC-MGs were then fixed after 5 min, 30 min, and 1 h and immunostained for TREM2 as described above. Only 5 min and 1 h time points were quantified. Cells were imaged using a Zeiss LSM800 confocal microscope. Using Imaris Software from Bitplane, we determine the cell volume and percentage of the microglia surface area covered by Aβ (1-40) TAMRA at different time points.

Fluorescently labeled pHrodo (a pH-sensitive dye that fluoresces only in acidic compartments) human brain synaptoneurosomes (hSN-rodo) were generated as described previously (Hesse et al., 2019; Tzioras et al., 2019). On DIV 30 of the microglia differentiation, iPSC-MGs were plated onto eight chamber glass slides that were coated with human fibronectin (sigma # F0895; 1:40) at a cell density of 100 000 cells per chamber. One hour after plating, we labeled live microglia with the nuclear marker Hoechst 33342 (Thermo Fisher # H3570). Then, control and C9orf72 iPSC-MGs were treated with 1:100 dilution of 4 mg/mL hSN-rodo in the presence or absence of 10 μM cytochalasin-D (inhibitor of actin polymerization). Confocal live cell imaging of iPSC-MGs was done using a 20X objective of a Zeiss 800 confocal microscope. IPSC-MGs were imaged every 10 min for up to 6 h for initial phagocytosis studies to determine the percentage of cells that exhibit phagocytic activity. For all other sets of experiments, iPSC-MGs were imaged every 10 min for up to 2 h. All lines had three technical replicates for the hSN-rodo treatment. Six images were taken per well per line for a total of 18 images every 10 min. All images were analyzed using the Imaris Software from Bitplane. To determine the percentage of iPSC-MGs engulfing hSN-rodo, the spots module was used to count the total number of iPSC-MGs based on the Hoechst marker and the total number of phagocytic iPSC-MGs per image. Additionally, we quantified the fluorescence mean intensity of the cargo hSN-rodo per phagocytic cell at the 2 h time point, where more than 60% of the iPSC-MGs were engulfing hSN-rodo. Cytochalasin-D was used as a negative control.

All immunostained iPSC-MGs were visualized and imaged using a Zeiss LSM800 laser scanning confocal microscope. Per staining, all images were taken with the same settings for parallel cultures. For all iPSC-MG immunostainings and Aβ (1-40) TAMRA phagocytic activity assay, a plan Apochromat 63x oil immersion objective was used; Z-stacks were generated with 1024 x 1024 image size, 0.5x XY scan zoom, and 1 μm scaling. For some immunostainings, differential interference contrast (DIC) was used to highlight the iPSC-MG surface area. For live cell imaging of iPSC-MG engulfing synaptoneurosomes, confocal microscopy with differential interference contrast was used. Tiled images were captured using a 20X objective with a 1.0x XY scan zoom and 0.624 μm x 0.624μm scaling. Bright-field images of the iPSC-MG cultures were taken using a Zeiss AxioVert.A1 microscope and a resolve HD Ludesco camera.

All images were processed and analyzed using the Imaris Software 9.5.1 and 9.6 from Bitplane. To obtain volume, area, mean intensity, and sum intensity per cell for a large number of samples, we assigned a randomized color identification per cell followed by the use of the ImarisVantage module to extract multiple numerical values from the created three-dimensional structures. For iPSC-MG marker characterization, the spots module was used to count the total number of cells positive for a specific microglia marker per image while using the DAPI channel as a reference. To calculate the TDP-43 and ADAR-2 nucleocytoplasmic ratio (N/C ratio), we used the surface module to generate iPSC-MG microglia 3-dimensional cellular (based on membrane staining with wheat germ agglutinin) and nuclear surfaces (based on DAPI). The sum intensity of TDP-43 or ADAR-2 was acquired for the nucleus and cytoplasm as well as the volume for each cellular compartment. The following formula was used: (Cell Sum Intensity – Nucleus Sum Intensity) = Cytoplasm Sum Intensity; (Cell Volume – Nucleus Volume) = Cytoplasm Volume. N/C ratio = (Nucleus Sum Intensity/Nucleus volume)/(Cytoplasm Sum Intensity/Cytoplasm volume). Data from each cell were acquired by assigning randomized color identification followed by the use of ImarisVantage. For the Aβ (1-40) TAMRA phagocytic activity, we used Aβ (1-40) TAMRA fluorescence signal and chose an ideal threshold to create the Aβ surfaces inside the IPSC-MGs. TREM2 staining was used to generate the cell surface structures. An algorithm was also generated and used in all iPSC-MG parallel cultures. Randomized color identification and ImarisVantage was also used to extract all data like previous datasets. For hSN-rodo live cell imaging over time, the spots module was used to count the total number of cells over time. Phagocytic cells were manually identified based on the hSN-rodo signal inside iPSC-MG. We set a threshold between 10,000 and 12,000 units above the grayscale as the positive signal indicative of engulfment and lysosomal internalization. At the 2 h time point, the cell surface area of phagocytic cells was manually outlined using the Imaris Software manual creation tool and the hSN-rodo mean intensity value was obtained per cell. For iPSC-MGs EEA1 and Lamp1 analysis, based on the staining pattern, we manually choose an optimal threshold for each protein marker to generate the 3D surfaces. We then stored the surface algorithms and used them in all iPSC-MG parallel cultures. Randomized color identification and ImarisVantage were then used to extract the cell volume and mean intensity of both markers per cell.

Statistical analysis for RNA sequencing was done in R (version 3.6.2) and detailed above. All other statistical analyses were performed using Graphpad Prism 7 and 8. For the comparison of the two groups, we used a two-tailed Student's t-test. Two-way ANOVA was used to analyze the difference between the means of more than two groups. One-way ANOVA with a post-hoc Dunnett's correction was performed to compare every cell line mean to the average of the control group. Two-tailed Mann–Whitney test was performed for the poly-(GP) ELISA assay. All statistical significance was ranked as the following: ^*p ≤ 0.05; ^**p ≤ 0.01; ^***p ≤ 0.001; ^****p ≤ 0.0001; and p > 0.05 not significant. All other statistical details and exact p-values are reported in each Figure legend.