Alzheimer's disease–associated PLCG2 variants alter microglial state and function in human induced pluripotent stem cell–derived microglia‐like cells

Abstract

INTRODUCTION

Variants of phospholipase C gamma 2 (PLCG2), a key microglial immune signaling protein, are genetically linked to Alzheimer's disease (AD) risk. Understanding how PLCG2 variants alter microglial function is critical for identifying mechanisms that drive neurodegeneration or resiliency in AD.

METHODS

Induced pluripotent stem cell (iPSC) –derived microglia carrying the protective PLCG2^P522R or risk‐conferring PLCG2^M28L variants, or loss of PLCG2, were generated to ascertain the impact on microglial transcriptome and function.

RESULTS

Protective PLCG2^P522R microglia showed significant transcriptomic similarity to isogenic controls. In contrast, risk‐conferring PLCG2^M28L microglia shared similarities with PLCG2^KO microglia, with functionally reduced TREM2 expression, blunted inflammatory responses, and increased proliferation and cell death. Uniquely, PLCG2^P522R microglia showed elevated cytokine secretion after lipopolysaccharide (LPS) stimulation and were protected from apoptosis.

DISCUSSION

These findings demonstrate that PLCG2 variants drive distinct microglia transcriptomes that influence microglial functional responses that could contribute to AD risk and protection. Targeting PLCG2‐mediated signaling may represent a powerful therapeutic strategy to modulate neuroinflammation.

Highlights

• The impact of Alzheimer's disease protective‐ and risk‐associated variants of phospholipase C gamma 2 (PLCG2) on the transcriptome and function of induced pluripotent stem cell (iPSC) –derived microglia was investigated.

• PLCG2 risk variant microglia exhibited a basal transcriptional profile similar to PLCG2‐deficient microglia but significantly different from isotype control and the transcriptionally similar PLCG2 protective variant microglia.

• PLCG2 risk variant and PLCG2‐deficient microglia show decreased levels of triggering receptor expressed on myeloid cells 2 (TREM2).

• The differential transcriptional pathways of protective and risk‐associated PLCG2 variant microglia functionally affect proliferation, apoptosis, and immune response.

• Protective PLCG2 microglia show resilience to apoptosis and increased cytokine/chemokine secretion upon exposure to lipopolysaccharide (LPS).

1. INTRODUCTION

Alzheimer's disease (AD) is the most common neurodegenerative disease, clinically characterized by progressive cognitive decline and pathologically defined by the presence of extracellular amyloid plaques and intracellular neurofibrillary tangles composed of aggregated and hyperphosphorylated tau. Microglia, the resident immune cells of the brain, play a crucial role in both neuroprotection and neurodegeneration ¹ . Microglial dysfunction contributes to AD pathology through processes such as impaired debris clearance, chronic inflammation, and abnormal synaptic pruning ² , while homeostatic functions can provide resilience from pathological insult. Genetic studies have identified numerous AD risk variants in immune‐ and microglial‐associated genes, highlighting the importance of microglia function in disease progression. ³ , ⁴ , ⁵ , ⁶ , ⁷

Among these microglia‐specific genes associated with altered AD risk is phospholipase C gamma 2 (PLCG2) ⁸ , ⁹ , ¹⁰ . PLCG2 encodes an enzyme predominately expressed in microglia within the central nervous system. ¹¹ , ¹² PLCG2 is involved in regulating microglial activation and function by mediating receptor signaling. Its role downstream from triggering receptor expressed on myeloid cells 2 (TREM2) is particularly significant, given the well‐established risk variants of TREM2. ¹³ Upon receptor activation and subsequent signaling cascade, PLCG2 associates with the membrane and mediates the hydrolysis of phosphatidylinositol 4, 5‐bisphosphate (PIP2) into two important second messengers: inositol 1, 4, 5‐trisphosphate (IP3) and diacylglycerol (DAG). ¹⁴ , ¹⁵ IP3 triggers calcium release from intracellular stores, while DAG activates protein kinase C (PKC) and other effector proteins, ultimately influencing microglial processes such as migration, phagocytosis, cytokine secretion, and metabolic responses. ¹³ , ¹⁶ , ¹⁷

Recently, rare PLCG2 variants have been linked to the risk of developing late‐onset AD. The mildly hypermorphic P522R variant (PLCG2^P522R) is associated with resilience to AD as well as to other neurodegenerative disorders, including frontotemporal dementia and Lewy body dementia, ⁸ , ⁹ and is overrepresented among centenarians. ¹⁸ Conversely, the partial loss‐of‐function M28L variant (PLCG2^M28L) confers increased risk for AD. ¹⁹ , ²⁰ While the slightly hypermorphic P522R variant is protective in neurodegenerative diseases, mutations in PLCG2 that constitutively activate the enzyme contribute to autoinflammatory disorders and resistance mechanisms in cancer treatment. ²¹ , ²² , ²³ Several studies have examined the role of PLCG2^P522R in myeloid cells and reported that this variant modulates phagocytosis, inflammatory responses, lipid profiles, and mitochondrial function. ¹¹ , ¹⁹ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ PLCG2^M28L, though less studied, has been shown to modulate phagocytic capability and alter inflammatory responses. ¹⁹ , ²⁰ In AD mouse models, we recently reported that PLCG2^P522R and PLCG2^M28L transcriptionally induce differential subpopulations of microglia, highlighting their impact on microglial phenotypes. ¹⁹

Despite the known role of PLCG2 in AD, the functional consequences of PLCG2 variants in human microglia are only starting to be understood. In human microglia, PLCG2^P522R has been shown to increase enzymatic activity ³⁰ and lead to enhanced mitochondrial function, ²⁷ , ³¹ increased lysosomal biogenesis, ²⁷ increased motility, ²⁷ reduced cholesterol ester accumulation, ¹³ increased amyloid beta (Aβ) uptake, ²⁷ , ³⁰ increased dextran ²⁷ uptake, decreased Escherichia coli uptake, ³⁰ decreased zymosan uptake, ²⁷ , ³⁰ decreased synaptosome uptake, ²⁷ and decreased synaptic pruning. ²⁷ Little is known about PLCG2^M28L in human microglia, but it has been reported to decrease Aβ uptake in primary murine microglia. ¹⁹ However, it remains unclear how human microglia expressing these PLCG2 variants affect cellular function in relation to microglia that are deficient in PLCG2 or bear the common PLCG2 allele.

Although PLCG2^P522R and PLCG2^M28L are rare variants, they can provide critical guidance to inform the development of AD therapeutics. By comparing the functional signatures of protective and risk variants to wild‐type (WT) and PLCG2 knockout (KO) microglia, pharmacodynamic benchmarks can be established. These benchmarks can be used to guide assay design strategies in the development of AD therapeutics to ensure that the therapeutics are inducing protective phenotypes as opposed to risk phenotypes. This study aims to address these gaps in knowledge by using iPSC‐derived microglial models to reveal the impact of PLCG2 variants on transcriptional profiles and by using functional assays to explore pathways implicated by transcriptomic analysis.

2. METHODS

2.1. Maintenance and expansion of iPSCs

Human iPSC lines utilized in this study included ADRC5 (WT), homozygous PLCG2 knockout (PLCG2^KO), and homozygous PLCG2^P522R lines (all generously provided by Dr. Mathew Blurton‐Jones at the University of California, Irvine), as well as the homozygous PLCG2^M28L line, which was generated in‐house (see below). All iPSC lines were generated from the ADRC5 parental line, providing a well‐characterized isogenic control background, thus eliminating potential differences in individual genomes. Each line underwent karyotyping and routine mycoplasma screening using a commercially available kit (Millipore‐Sigma MP0035‐1KT). iPSCs were cultured on six‐well plates coated with hESC‐qualified Matrigel (Fisher Scientific 8774552) and grown in mTeSR PLUS (STEMCELL Technologies 100‐0276) media. Media were changed every 2 days. Colonies were grown to ∼60–70% confluency before being passaged with ReLeSR (STEMCELL Technologies 100‐0484), an enzyme‐free method for colony dissociation. Colonies were then either replated at a 1:200 dilution for expansion or used for differentiation into human microglia‐like cells (MGLs).

2.2. Generation of PLCG2^M28L cell line

CRISPR/Cas9‐mediated gene editing, following previously described methods, ³² was used to introduce the PLCG2 M28L single‐nucleotide polymorphisms (SNP) into the ADRC5 parental cell line (Figure S1a). In brief, a gRNA sequence targeting the M28L position of PLCG2 was designed and electroporated into iPSCs alongside a homology‐directed repair template containing the M28L SNP and a Cas9–green fluorescent protein (GFP) plasmid using the Neon Transfection System (Life Technologies MPK10096) (Figure S1b). Electroporated iPSCs were then plated onto Matrigel‐coated six‐well plates in the presence of 10 µM Rho‐associated kinase (ROCK) inhibitor Y27632 (Reprocell 04‐0012‐02). At 2 days post‐electroporation, transient GFP‐positive cells carrying the Cas9 plasmid were sorted using fluorescence‐activated cell sorting (FACS). After an additional passage, iPSCs were plated at low density to facilitate single‐cell colony formation. Clonally‐derived colonies were manually selected and transferred to a 96‐well plate coated with Matrigel (Fisher Scientific 8774552). Colonies were cultured in mTeSR PLUS (STEMCELL Technologies 100‐0276) media for further expansion, followed by screening for the expected loss of the Xho1 restriction enzyme site (Figure S1c). Clones were sequenced to identify successfully edited iPSCs (Figure S1d). Additionally, the top five predicted CRISPR off‐target binding sites were sequenced to confirm that no off‐target mutations were present and that each site remained identical to the isogenic control (Figure S1e).

RESEARCH IN CONTEXT

1. Systematic review: The authors performed a comprehensive review of existing literature, using PubMed as the primary search platform. Current research highlights phospholipase C gamma 2 (PLCG2) as a key player in neuroimmune signaling in microglia, with emerging evidence connecting specific PLCG2 variants to Alzheimer's disease (AD) susceptibility. Studies have demonstrated that PLCG2 variants (protective PLCG2^P522R and risk‐conferring PLCG2^M28L) can exert contrasting influences on amyloid deposition and microglial activity in AD. However, no study has directly determined the impact of microglia bearing protective PLCG2^P522R and risk‐associated PLCG2^M28L on cellular function in the context of isogenic controls and PLCG2‐deficient microglia, leaving an important gap in our understanding of the role of PLCG2 within microglial biology.

2. Interpretation: Our findings indicate that AD variants of PLCG2 significantly impact the transcriptomic profile of microglia. This includes a potential role for PLCG2 in regulating microglial expression of triggering receptor expressed on myeloid cells 2 (TREM2). PLCG2 variants differentially regulate microglial phenotypes and functions such as proliferation, cell survival, and immune response. PLCG2^P522R microglia uniquely display protection from apoptosis and increased cytokine secretion upon exposure to lipopolysaccharide (LPS).

3. Future directions: Our findings suggest a nuanced role for PLCG2 in microglial function. Further studies are necessary to pinpoint the mechanism of protection for the protective PLCG2^P522R variant. Additionally, future studies in multicellular and disease‐specific models will enable a more precise understanding of the impact of PLCG2 variants in AD. Finally, activation, not inhibition, of PLCG2 should be explored as a therapeutic mechanism for treating AD.

2.3. Differentiation of iPSC‐derived microglia (MGLs)

Differentiation of iPSCs into MGLs was performed following previously established protocols. ³³ In brief, iPSC colonies were cultured on Matrigel until they reached approximately 75% confluency. Then, enzyme‐free dissociation with ReLeSR reagent was used to detach cells from the plate. The colonies (150–200 microns/colony) were then plated onto Matrigel‐coated six‐well plates at a density of 35–40 colonies per well. Differentiation into hematopoietic progenitor cells (HPCs) was carried out using the STEMdiff Hematopoietic Kit (STEMCELL Technologies 5310). On day 12 of differentiation, HPCs were collected from the media, resuspended in Bambanker Freeze (Fujifilm Irvine Scientific 302‐14686) at a density of 500,000 cells/mL, and cryopreserved in liquid nitrogen to create an HPC stock for subsequent MGL differentiation.

To differentiate HPCs into MGLs, cryopreserved vials of PLCG2^WT, homozygous PLCG2^P522R, homozygous PLCG2^M28L, and homozygous PLCG2^KO HPCs were thawed and transitioned into Microglia Basal Media, which consisted of DMEM/F12 (Life Technologies 11330‐057), 2% Insulin‐Transferrin‐Selenium (Life Technologies 41400045), 2% B27 (Life Technologies 17504001), 1% Antibiotic‐Antimycotic (Life Technologies 15240062), 1% MEM NEAA (Life Technologies 11140050), 1% Glutamax (Life Technologies 35050061), 0.5% N2 (Life Technologies 17502048), 0.2 mg/mL insulin (Millipore‐Sigma I2643), and 400 µM monothioglycerol (Millipore‐Sigma M1753), supplemented with 100 ng/mL interleukin (IL) ‐34 (Peprotech 200‐34‐500 µg), 50 ng/mL transforming growth factor‐beta1 (TGF‐β1; Peprotech 100‐21‐100 µg), and 25 ng/mL macrophage colony stimulating factor (M‐CSF; Peprotech 300‐25‐100 µg). HPCs in Microglia Basal Media were plated in parallel into two wells of a six‐well plate. Media was replenished every 48 h. After 25–28 days of maturation, MGLs were collected for downstream assays. An isogenic set of PLCG2^WT, homozygous PLCG2^P522R, homozygous PLCG2^M28L, and homozygous PLCG2^KO lines was used for subsequent microglia assays. All assays were conducted with at least three independent differentiations.

2.4. Immunocytochemistry

After 25–28 days of differentiation, mature MGLs were replated onto coverslips coated with poly‐D‐lysine and vitronectin and allowed to adhere for 24–48 h. Coverslips were then washed with phosphate buffered saline (PBS) and fixed with 4% paraformaldehyde (PFA) for 20 min. After washing with PBS, coverslips were stored in PBS. For immunostaining, coverslips were again washed with PBS, then permeabilized in 0.1% Triton X‐100 in PBS (PBST) for 10 min. If antigen retrieval was to be performed, coverslips were incubated with 10 mM sodium citrate, pH 6 at 85°C for 10 min, then allowed to cool to room temperature (RT). MGLs were then blocked in 5% bovine serum albumin (BSA) in PBST for 1 h at RT. Following blocking, the coverslips were incubated overnight in 5% BSA in PBST with the following primary antibodies: IBA1 (Fujifilm Irvine Scientific 019‐19741, 1:500), P2RY12 (Sigma Aldrich HPA014518, 1:50), PLCG2 (Cell Signaling Technology E5U4T, 1:250), TREM2 (R&D Systems AF1828, 1:100), Ki‐67 (Abcam ab15580, 1:500). PLCG2 antibody selection was validated by YCharOS. ³⁴

After overnight incubation, coverslips were washed three times in PBST and incubated for 1 h at RT with the appropriate species‐specific secondary antibodies conjugated with AlexaFluor in 5% BSA in PBST (1:1000, Invitrogen). Coverslips were then washed once with DAPI for 5 min, followed by three times in PBST, and mounted onto slides using Prolong Gold Antifade Mountant (Thermo Fisher Scientific, P36930). Images were acquired on a Leica DM6 microscope or a Nikon A1R confocal microscope using similar exposure and gain settings. Three fields of view (FOVs) were taken per coverslip. Two coverslips per differentiation were used for each iPSC line. Images were analyzed using ImageJ software (National Institutes of Health [NIH], version 1.54j). For protein expression analysis (TREM2 and PLCG2), mean fluorescent intensity was measured.

2.5. E. coli uptake assay

Mature MGLs, after 28 days of differentiation, were replated onto vitronectin‐coated 96‐well plates at a density of 1.5 × 10⁴ cells/well in 200 µL complete media (2 wells/genotype). After 24 h, pHrodo™ Red E. coli BioParticles (Thermo Fisher Scientific P35361) were spiked into each well. The plate was imaged in a Tecan Spark Cyto 600 that was preheated to 37°C with 5% CO₂. Images were taken using the same exposure and gain settings. Four FOVs per well were taken at 10x every hour for 6 h.

2.6. RNA isolation for RNA‐sequencing

In order to obtain RNA from both the adherent MGLs and the MGLs suspended in culture, the media from 28‐day, mature MGLs in a six‐well plate were collected, centrifuged for 5 min at 300 x g, washed once with PBS and recentrifuged for 3 min at 500 x g. Adherent MGLs from the same plate were washed once with PBS. Then RNA was isolated directly from the well using Arcturus PicoPure RNA Isolation Kit (Thermo Fisher Scientific 2970541) according to the manufacturer's instructions. The adherent and nonadherent microglia were combined in RNA extraction buffer. One well per line was collected over four batches of differentiation. Initial RNA concentration and purity were measured using a Nanodrop spectrophotometer.

2.7. Bulk RNA‐sequencing

Library preparation, sequencing, and alignment were performed by the Center for Medical Genomics at Indiana University School of Medicine. RNA integrity was assessed using the Agilent TapeStation. RNA samples with a RIN > 8 were used for library construction. RNA libraries were prepared from 100 ng of total RNA using the Illumina Stranded mRNA Prep, Ligation kit (Illumina), following the manufacturer's instructions. Each resulting uniquely dual‐indexed library was quantified and quality accessed by Qubit and Agilent TapeStation. Multiple libraries were pooled in equal molarity. The pooled libraries were sequenced with 2 × 150 bp paired‐end configuration on an Illumina NovaSeq X PLUS sequencer.

2.8. RNA‐sequencing data analysis

Sequencing reads were first quality checked using FastQC (v.0.11.5, Babraham Bioinformatics, Cambridge, UK) for quality control. The sequence data were then mapped to the human reference genome hg38 using the RNA‐seq aligner STAR (v.2.7.10a). Reads mapped to each gene were counted using featureCounts (v. 2.0.3). ³⁵ , ³⁶ , ³⁷ , ³⁸ Differential gene analysis (differentially expressed genes [DEGs]) and principal component analysis (PCA) were conducted using edgeR (3.38.4). ³⁹ DEGs were defined as false discovery rate (FDR) –adjusted p value ≤ 0.05 and an absolute Fold Change ≥ 1.2. Enrichment analysis of the DEGs and Functional Annotation Clustering were conducted using DAVID. ⁴⁰ , ⁴¹ Enrichments were considered to be significant if p value ≤ 0.05. Heatmaps were generated using Morpheus. ⁴² Volcano plots were generated in edgeR using the EnhancedVolcano tool. Bubble plots were generated using SRplot. ⁴³ Raw and processed data have been uploaded to GEO (GSE308813).

2.9. EdU proliferation assay

After 28 days of differentiation, mature MGLs were replated onto vitronectin‐coated 96‐well plates at a density of 1.5 × 10⁴ cells/well in 200 µL complete media (2 wells/genotype). To measure proliferative cells, a Click‐iT Plus EdU Cell Proliferation Kit (Thermo Fisher Scientific C10637) was used following the manufacturer's recommendations. Briefly, after allowing the cells to rest overnight, half of the media was replaced with 2x EdU solution (10 µM final concentration). After 24 h incubation, cells were fixed using 4% PFA for 15 min. Cells were then washed twice with PBS followed by permeabilization with 0.5% Triton X‐100 in PBS for 20 min. The cells were then incubated in 1x Click‐iT Plus reaction cocktail for 30 min. After washing with PBS, nuclei were stained using DAPI. Images were acquired on a Leica DM6 microscope or a Nikon A1R confocal microscope using similar exposure and gain settings. Three FOVs were taken per well. Three independent differentiations were used. The percentage of proliferative cells was determined using ImageJ and calculated as the number of proliferative nuclei in proportion to the total number of nuclei in each image (NIH, version 1.54j).

2.10. Ki‐67 proliferation assay

After 28 days of differentiation, mature MGLs were replated onto vitronectin‐coated 96‐well plates at a density of 1.5 × 10⁴ cells/well in 200 µL complete media (2 wells/genotype). After allowing the cells to rest overnight, half of the media was replaced with fresh media (to mimic the EdU proliferation assay). After 24 h incubation, cells were fixed using 4% PFA for 15 min. Cells were then washed twice with PBS followed by permeabilization with 0.5% Triton X‐100 in PBS for 20 min. The cells were then stained for Ki‐67 and DAPI as described above. Images were acquired on a Leica DM6 microscope or a Nikon A1R confocal microscope using similar exposure and gain settings. Three FOVs were taken per well. Three independent differentiations were used. The percentage of proliferative cells was determined using ImageJ software (NIH, version 1.54j).

2.11. Cell death assay

Mature MGLs were replated into vitronectin‐coated 96‐well plates at a density of 1.5 × 10⁴ cells/well in 200 µL complete media (15 wells/genotype). To measure live and dead cells, the LIVE/DEAD Cell Imaging Kit (488/570) (Thermo Fisher Scientific R37601) was used following the manufacturer's recommendations. Briefly, after letting the cells rest overnight, the media of three wells per genotype were changed to cytokine‐free media (48 h). Subsequent wells were switched to cytokine free media at 0, 6, 12, and 24 h prior to imaging. The 2x working solution (1 mL live stain mixed with 1 µL dead stain) was then added to the media for 15 min. Cells were then imaged immediately to prevent skewed results. Images were acquired on a Leica DM6 microscope or a Nikon A1R confocal microscope using similar exposure and gain settings. Three FOVs were taken per well. Three independent differentiations were used. Percentage of dead cells was determined using ImageJ software (NIH, version 1.54j).

2.12. Caspase 3/7 detection assay

Mature MGLs were replated into vitronectin‐coated 96‐well plates at a density of 1.5 × 10⁴ cells/well in 200 µL complete media (Four wells/genotype). After 24 h, the media of a set of two wells per genotype was replaced with fresh complete media spiked with 1x Caspase Assay Reagent (Thermo Fisher Scientific C10432). The media of a different set of two wells per genotype was replaced with cytokine free media spiked 1x with Caspase Assay Reagent. The plate was then imaged in a Tecan Spark Cyto 600 that was preheated to 37°C with 5% CO₂. Images were taken using the same exposure and gain settings. Four FOVs per well were taken at 10x every 6 h for 48 h. Three independent differentiations were used. Spark Cyto software “ImageAnalyzer” (version 1.2) was used to determine the percentage of activated caspase 3/7‐positive cells (apoptotic cells).

2.13. Lipopolysaccharide induction / cytokine secretion assay

Mature MGLs were replated into vitronectin‐coated 96‐well plates at a density of 1.5 × 10⁴ cells/well in 200 µL complete media (four wells/genotype). After 24 h, the media of two wells per genotype were replaced with fresh culture media without TGF‐β (to allow for cellular activation) spiked with ultrapure lipopolysaccharide (LPS) derived from E. coli serotype O55:K59(B5)H‐ (InvivoGen tlrl‐pb5lps) (150 ng/mL final concentration). The media of the remaining two wells per genotype were replaced with fresh culture media without TGF‐β spiked with vehicle. The cells were then incubated for 24 h, after which the conditioned media was collected and frozen at −80°C. To assess cytokine concentrations, samples were sent to Eve Technologies Corporation for multiplex analysis (Human Cytokine/Chemokine Panel A 48‐Plex Discovery Assay Array (HD48A), Eve Technologies) using the Luminex 200 system with Bio‐Plex Manager software (Bio‐Rad Laboratories Inc.) as per the manufacturer's instructions (MILLIPLEX Human Cytokine/Chemokine/Growth Factor Panel A Magnetic Bead Panel Cat. # HCYTA‐60K). Three independent differentiations were used.

2.14. Statistical analysis

Data are expressed as the mean ± SEM. Statistical analyses were performed using GraphPad Prism (GraphPad Software, version 10.2.2). Analyses were performed using a non‐parametric one‐way or two‐way analysis of variance (ANOVA) with Tukey's or Sidak's post hoc test. p < 0.05 was considered significant.

3. RESULTS

3.1. iPSC‐derived human MGLs expressing PLCG2 variants display hallmark characteristics of microglia

To investigate the effects of AD‐associated variants of PLCG2 in microglia, we analyzed human MGLs lines carrying the P522R and M28L PLCG2 variants (Figure S1), as well as PLCG2‐deficient knockouts (KO) and isogenic controls (WT). We first generated MGLs from induced pluripotent stem cells (iPSCs) following a directed differentiation protocol (Figure 1A). iPSCs were differentiated into hematopoietic progenitor cells (HPCs), followed by exposure to IL‐34, TGF‐β1, and M‐CSF to drive microglial differentiation and maturation. Brightfield imaging of HPCs at day 12 of hematopoiesis revealed a dense, uniform population of HPCs across all genotypes (Figure 1B). After 28 days of microglial differentiation and maturation, MGLs displayed characteristic ramified morphology of mature microglia across all genotypes (Figure 1C). To confirm the identity of differentiated MGLs, we performed immunofluorescence staining for Iba1, a microglia marker. All genotypes showed robust Iba1 staining, indicating successful differentiation (Figure 1D, left). Similarly, staining for the homeostatic microglial marker P2RY12 showed strong fluorescence across all genotypes (Figure 1D, right). Next, to test functional responses, we assessed the ability of the MGLs to uptake debris by exposing them to pHrodo‐labeled E. coli particles for 6 h (Figure 1E). All genotypes demonstrated strong uptake of bacterial particles, as indicated by widespread red fluorescence within the cells. Overall, these findings suggest that the PLCG2 variants did not affect the ability of iPSCs to become and behave as MGLs, as all genotypes were successfully differentiated into functionally‐phagocytic MGLs.

FIGURE 1.

iPSC‐derived human MGLs expressing PLCG2 variants display hallmark characteristics of microglia. (A) Schematic depicting the iPSC differentiation protocol to generate MGLs. (B) Representative brightfield images of HPCs at day 12 of hematopoiesis (20×). (C) Representative brightfield images of MGLs at day 28 of differentiation (20×). (D) Representative immunofluorescent images of MGLs stained for microglial markers IBA1 (green, left) and P2RY12 (green, right) at day 28 of differentiation (40×). (E) Representative images of PLCG2 variant MGLs (brightfield) competently internalizing pHrodo‐labeled E. coli (red) after 6 h (20×). HPCs, hematopoietic progenitor cells; iPSC, induced pluripotent stem cell; MGLs, microglia‐like cells.

3.2. PLCG2 variant MGLs are transcriptionally distinct

To investigate the impact of PLCG2 and its AD‐associated variants on microglial transcriptional states, we performed RNA sequencing on PLCG2^WT, PLCG2^P522R, PLCG2^M28L, and PLCG2^KO MGLs. PCA revealed distinct clustering of samples correlated to genotype (Figure 2A), indicating that the PLCG2^P522R and PLCG2^M28L mutations lead to significantly divergent transcriptomes in MGLs. The first principal component (PC1) accounts for 27% of the variance, clearly separating WT from the variant groups. The P522R variant, while forming a distinct cluster, grouped near the PLCG2^WT group. PC2 represents an additional 21% of the variance, with the M28L and KO groups clustering separately, suggesting differential gene expression profiles for the partial loss‐of‐function variant and total PLCG2 deficiency.

FIGURE 2.

PLCG2 variant MGLs are transcriptionally distinct. (A) PCA showing variance in gene expression (normalized counts per million) of indicated MGL genotypes from all expressed genes obtained by RNAseq (n = 4 independent differentiations, which are each represented by one dot). Dot color represents the genotype of the cell. (B) Euler diagrams showing the number of shared and unshared DEGs (FDR < 0.05, Fold Change > 1.2) between WT versus P522R, WT versus KO, and WT versus M28L MGLs. (C) Overview heatmap showing mRNA expression of the top 500 genes. Hierarchical clustering (one minus Pearson correlation) shows grouping of genotypes. Heatmap colors are mapped by average expression values for each gene. (D) Volcano plots showing significant DEGs (FDR < 0.05, fold change > 1.2) between WT versus P522R, WT versus KO, and WT versus M28L MGLs. (E) Bubble plots showing functional annotation clustering of WT versus P522R, WT versus KO, and WT versus M28L MGLs. Significant pathways from R, KW, KEGG, and WP databases were clustered by similar gene profiles. Bubble size represents gene count, and color denotes statistical significance (−log₁₀ adjusted p value). DEGs, differentially expressed genes; FDR, false discovery rate; GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; KO, knockout; KW, Uniprot Keywords; PCA, principal component analysis; PLCG2, phospholipase C gamma 2; R, Reactome; RNAseq, RNA sequencing; WP, Wikipathways; WT, wild‐type.

Given the separation between PLCG2 variant MGLs, differential expression and functional enrichment analyses were conducted to characterize the transcriptional landscape of each genotype. DEGs were visualized on an Euler plot to show how many genes are uniquely regulated by each PLCG2 variant (Figure 2B). Upregulated and downregulated genes were analyzed separately to account for oppositely regulated DEGs between genotypes. When compared to WT MGLs, PLCG2^M28L MGLs exhibited the highest number of unique DEGs (628 upregulated, 596 downregulated), whereas PLCG2^KO MGLs had fewer unique DEGs (293 upregulated, 315 downregulated), and PLCG2^P522R had the least unique DEGs (14 upregulated, 8 downregulated). Additionally, relative to the WT MGLs, there was some overlap between the KO and M28L groups (256 upregulated, 385 downregulated DEGs), indicating shared gene expression that was likely due to partial loss‐of‐function in PLCG2^M28L MGLs. Unsupervised hierarchical clustering of the top 500 most variable DEGs showed distinct transcriptional signatures for each genotype (Figure 2C). WT samples form a separate cluster distinct from the PLCG2‐deficient and variant MGLs. KO and M28L samples displayed more pronounced shifts in gene expression relative to WT samples, while P522R MGLs were more closely related. Volcano plots display the magnitude (log₂FC) and significance (−log₁₀FDR) of gene expression changes for each genotype comparison (Figure 2D). Notably, genes associated with microglial reactive states (TREM2, APOE, GPNMB, CGAS) are differentially regulated. Finally, to understand how the distinct transcriptomic profiles of each genotype may lead to functional changes, functional annotation clustering was performed to discover enriched gene groups in PLCG2^P522R, PLCG2^KO, and PLCG2^M28L MGL relative to WT MGL (Figure 2E). Examples of gene pathways that are significantly enriched among the PLCG2 variants are related to cellular proliferation, cell death/apoptosis, and immune response. These results demonstrate that these PLCG2 variants result in differential microglial gene expression signatures in the absence of added stimuli.

3.3. PLCG2 expression correlates with TREM2 expression in MGLs

To further understand how PLCG2 variants and deficiency affect the expression profile of MGLs, a heatmap with the top 50 genes correlating with PLCG2 expression was generated (Figure 3A). TREM2 is among these highly correlated genes, indicating a potential co‐regulation or a shared pathway in microglia. We then compared the normalized counts per million (CPM) of TREM2 and PLCG2 between the genotypes (Figure 3B). As expected, PLCG2^KO MGLs exhibit a significant reduction in PLCG2 relative to WT cells. Notably, TREM2 expression is also significantly reduced in PLCG2^KO MGLs, demonstrating that PLCG2 deficiency negatively impacts TREM2 expression. Interestingly, the P522R and M28L PLCG2 variants do not significantly alter PLCG2 or TREM2 expression levels. To confirm these findings at the protein level, we used immunocytochemistry to visualize PLCG2 and TREM2. Representative images indicate that PLCG2 expression (green) is abundant in PLCG2^WT and PLCG2^P522R MGLs but significantly reduced in PLCG2^KO and PLCG2^M28L cells (Figure 3C). Mean fluorescent intensity (MFI) quantification further supports this, showing a statistically significant decrease in MFI for PLCG2^KO and PLCG2^M28L MGLs compared to WT cells (Figure 3D). Following a similar pattern, representative TREM2 stains show PLCG2^WT and PLCG2^P522R MGLs exhibit strong fluorescent staining for TREM2 (red), whereas PLCG2^KO and PLCG2^M28L cells display reduced staining (Figure 3E). MFI quantification confirms significant reductions in TREM2 in PLCG2^KO and PLCG2^M28L MGLs (Figure 3F). Overall, these results suggest a strong correlation between PLCG2 and TREM2 expression, with PLCG2 likely playing a role in regulating TREM2 levels.

FIGURE 3.

PLCG2 expression correlates with TREM2 expression in MGLs. (A) Heatmap of top 75 genes with expression profiles that correlates strongest with the expression profile of PLCG2 (Pearson correlation) in bulk RNAseq analysis of PLCG2 variant MGLs. (B) Bar graph showing PLCG2 and TREM2 mRNA expression as measured by bulk RNAseq of MGLs. Expression is represented as fold change of CPM relative to WT cells (n = 4 for each genotype). (C) Representative images (20×) and (D) MFI quantification of PLCG2 (green) in MGLs. MGL nuclei were stained with DAPI (blue) (n = 4 independent differentiations for each genotype, data point represents the mean of three technical replicates). (E) Representative images (20x) and (F) MFI quantification of TREM2 (red) in MGLs. MGL nuclei were stained with DAPI (blue) (n = 3 or 4 independent differentiations for each genotype, data point represents the mean of three technical replicates). Data are presented as mean ± SEM analyzed by ANOVA followed by Tukey's multiple comparisons test. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001; ns. ANOVA, analysis of variance; CPM, counts per million; DAPI, 4′,6‐diamidino‐2‐phenylindole; MFI, mean fluorescence intensity; MGL, microglia‐like cell; ns, not significant; PLCG2, phospholipase C gamma 2; TREM2, triggering receptor expressed on myeloid cells 2

3.4. PLCG2 variants affect MGL proliferation

The transcriptomic differences between PLCG2 variant MGLs indicated that microglial functions such as proliferation, apoptosis, and inflammatory response might be affected. Thus, we sought to determine whether these responses are altered by PLCG2 KO and variants. First, we examined effects on MGL proliferation. Hierarchical clustering of proliferation‐related genes (gene ontology [GO]: 0042129) revealed significant transcriptional changes in genes across the different PLCG2 genotypes, suggesting that PLCG2 influences pathways regulating cell proliferation (Figure 4A). To functionally test MGL proliferation, we performed Click‐iT EdU incorporation assays, which measure DNA synthesis over 24 h. Quantification of the percentage of fluorescently‐labeled, proliferative cells, as seen in the representative images (Figure 4B), showed that PLCG2^KO and PLCG2^M28L MGLs exhibited increased proliferation rates, with a significantly greater percentage of EdU⁺ cells (green) compared to PLCG2^WT and PLCG2^P522R MGLs (Figure 4C). To further validate these findings, we assessed Ki‐67 expression (red), a marker of actively cycling cells, via immunostaining (Figure 4D) (see Figure S2 for higher magnification and single channel images of Figures 4B and 4D). Quantification of Ki‐67⁺ cells also showed low levels of proliferation in PLCG2^WT and PLCG2^P522R MGLs but a significant increase in the percentage of proliferating cells in PLCG2^KO MGLs compared to WT cells. Taken together, these results indicate that PLCG2^KO and PLCG2^M28L MGLs show increased proliferation compared to PLCG2^WT MGLs, while proliferation in PLCG2^P522R MGLs is similar to PLCG2^WT cells.

FIGURE 4.

PLCG2 variants affect MGL proliferation. (A) Heatmap visualizing level of MGL gene expression of top 50 genes from GO:0042129 (Regulation of T Cell Proliferation). The heatmap is arranged by gene and genotype by hierarchical clustering (one minus Pearson correlation) of bulk RNAseq data. (B) Representative images (10 ×) and (C) quantification of an EdU proliferation assay with PLCG2 variant MGLs. The percentage of proliferative cells was quantified by dividing the number of proliferating cells (EdU, green) by the total number of cells (DAPI, blue) (n = 3 independent differentiations for each genotype, data point represents the mean of three technical replicates). (D) Representative images (20×) and (E) quantification of Ki‐67⁺ cells among PLCG2 variant MGLs. The percentage of proliferative cells was quantified by dividing the number of proliferating cells (Ki‐67, red) by the total number of cells (DAPI, blue) (n = 3 independent differentiations for each genotype, data point represents the mean of three technical replicates). Data are presented as mean ± SEM analyzed by ANOVA followed by Tukey's multiple comparisons test. * p < 0.05, ** p < 0.01, *** p < 0.001; ns: not significant. ANOVA, analysis of variance; DAPI, 4′,6‐diamidino‐2‐phenylindole; MGL, microglia‐like cell; PLCG2, phospholipase C gamma 2

3.5. PLCG2 variants alter MGL susceptibility to cell death

Our enrichment analysis implicated apoptosis pathways as observed in the hierarchical clustering of apoptosis‐related genes (GO: 0043066), which revealed distinct gene expression patterns across PLCG2^WT, PLCG2^P522R, PLCG2^KO, and PLCG2^M28L MGLs (Figure 5A). Several pro‐apoptotic genes, such as FAS and MMP9, were upregulated in PLCG2^M28L MGLs, whereas PLCG2^WT cells exhibited lower expression of these genes. In contrast, anti‐apoptotic genes such as BCL2A1 were more highly expressed in PLCG2^WTand PLCG2^P522R cells compared to PLCG2^KO and PLCG2^M28L MGLs. These results suggest that MGLs carrying PLCG2^KO and PLCG2^M28L may have an increased predisposition to apoptosis. Therefore, we performed functional cell death assays under stress‐induced conditions. Cell death was induced by cytokine starvation (absence of M‐CSF, TGF‐β1, and IL‐34) followed by a live/dead assay (Figure 5B). At baseline (0 h), all genotypes displayed minimal cell death, as indicated by the predominance of live (green) cells. However, after 48 h of cytokine deprivation, PLCG2^KO and PLCG2^M28L MGLs exhibited a striking increase in dead (red) cells (∼80 and 87%, respectively) compared to PLCG2^WTand PLCG2^P522R cells (∼56 and 47%, respectively) (Figure 5B‐C). To establish that the observed cell death was due to apoptosis, we conducted a live‐cell caspase 3/7 activation assay (Figure 5D). Over 48 h of cytokine starvation, PLCG2^KO MGLs displayed a substantial increase in caspase activation (red signal) (26%), whereas PLCG2^P522R cells showed decreased apoptosis relative to PLCG2^WT MGLs (6%) (Figure 5D‐E). Interestingly, PLCG2^WT and PLCG2^M28L MGLs showed similar levels of apoptosis (20% and 16%, respectively). Importantly, MGLs maintained in complete media (with cytokines) exhibited minimal apoptosis across all genotypes, confirming that cytokine deprivation was the primary driver of apoptosis. Together, these results indicate that absence of PLCG2 promotes increased sensitivity to apoptosis under cytokine starvation, likely due to transcriptional upregulation of pro‐apoptotic genes. In contrast, PLCG2^P522R cells exhibit greater resistance to apoptosis.

FIGURE 5.

PLCG2 variants alter MGL susceptibility to cell death. (A) The heatmap depicting level of MGL gene expression of top 60 genes from GO:0043066 (Negative Regulation of Apoptotic Process). Heatmap is arranged by gene and genotype by hierarchical clustering (one minus Pearson correlation) of RNAseq data. (B) Representative images (10 ×) and (C) quantification of live/dead assay. Cells were switched into cytokine free media and starved for 0, 6, 12, 24, or 48 h. The percentage of dead cells was quantified by dividing the number of dead cells (compromised membrane, red) by the total number of cells (sum of the dead cells and the live cells (intact membrane, green)) (n = 3 independent differentiations for each genotype, data point represents the mean of two technical replicates). Data are presented as mean ± SEM analyzed by ANOVA and Tukey's multiple comparisons test. ** p < 0.01, *** p < 0.001, **** p < 0.0001; ns. (D) Representative images (20×) and (E) quantification of caspase−3/7 activation (apoptosis) in PLCG2 variant MGL. The percentage of cells undergoing apoptosis was quantified over 48 h of either cytokine starvation or no starvation by dividing the number of cells undergoing apoptosis (activated caspase 3/7, red) by the total number of cells (determined by imager software analysis, bright field) (n = 3 independent differentiations for each genotype, data point represents mean of two technical replicates). Data are presented as mean ± SEM analyzed by two‐way ANOVA followed by Sidak's multiple comparisons test. Statistics only shown for comparisons between cytokine starved samples. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001; ns: not significant. ns, not significant. ANOVA, analysis of variance; GO, gene ontology; MGL, microglia‐like cell; PLCG2, phospholipase C gamma 2; SEM, standard error of the mean

3.6. PLCG2 variants affect MGL immune response

Transcriptomics analysis also revealed distinct gene expression patterns of innate immune response related genes (GO: 0045087) across PLCG2^WT, PLCG2^P522R, PLCG2^KO, and PLCG2^M28L MGLs (Figure 6A). Interestingly, respective to the other MGL lines, the immune‐related gene signature of PLCG2^P522R MGLs included the unique increase in expression of DAPK1 and the unique decrease in expression of CGAS. Notably, DAPK1 was decreased in PLCG2^M28L MGLs. PLCG2^KO and PLCG2^M28L MGLs showed an increase in immune response genes TRIM22, CX3CR1, NLRP1, and LAG3 relative to PLCG2^WT and PLCG2^P522R MGLs. Conversely, PLCG2^KO and/or PLCG2^M28L showed a decrease in TLR signaling‐related genes SARM1, MR1, TRAF3, LGALS3, and CD14 relative to PLCG2^WT and PLCG2^P522R MGLs. These results suggest that MGLs carrying PLCG2^KO and PLCG2^M28L may prompt a unique immune response, potentially involving TLR signaling. To understand how basal level‐expression of cytokine genes was affected by PLCG2 genotypes, we compared the normalized logCPM from the bulk RNAseq results (Figure 6B). We found that PLCG2^KO and/or PLCG2^M28L MGLs showed modified expression in many cytokines (CXCL8 [IL‐8], IL10, IL18, CCL2, CCL22 [MDC], TGFA) compared to PLCG2^WT and PLCG2^P522R MGLs. To assess the functional consequences of these transcriptional changes, and given that PLCG2 has been shown to mediate LPS‐induced TLR signaling ¹³ , we performed a LPS stimulation assay (Figure 6C). After 24 h of LPS stimulation, the culture media was collected from variant MGLs, and secreted protein levels of cytokines were detected by enzyme‐linked immunosorbent assay (ELISA) (Figure S3). Compared to PLCG2^WT MGLs, PLCG2^P522R MGLs elicited significant increases in secretion of chemokines (CCL2, CCL4, CXCL1, and CXCL9) and pro‐inflammatory cytokines (IL‐6, IL‐8, IL‐12p40, TNFα); however, secretion of the regulatory cytokine IL‐1RA is also increased. The response of PLCG2^KO MGLs to LPS stimulation was more similar to PLCG2^WT MGLs, with the exception of increased CCL2 secretion. PLCG2^M28L MGLs showed a decrease in IL‐3, IL‐9, IL‐17A, IL‐27, and TGF‐α relative to PLCG2^WT MGLs. Taken together, these results show that the protective P522R variant sensitizes MGLs to LPS stimulation, resulting in an enhanced inflammatory response, while partial loss of function (PLCG2^M28L) leads to a generally blunted immune response.

FIGURE 6.

PLCG2 variants affect MGL immune response. (A) The heatmap showing level of MGL gene expression of top 75 genes from GO:0045087 (Innate Immune Response). The heatmap is arranged by gene and genotype by hierarchical clustering (one minus Pearson correlation) of RNAseq data. (B) Violin plots showing cytokine mRNA expression as measured by bulk RNAseq of MGLs. Expression is represented as normalized counts per million (logCPM) (n = 4 independent differentiations for each genotype). (C) Schematic of LPS stimulation paradigm. 28‐day MGLs were starved of TGF‐β for 24 h and then stimulated with LPS for 24 h. Culture media was then harvested, and multiplex ELISA performed. (D) Quantification of cytokines secreted into media by PLCG2 variant MGLs over 24 h of LPS stimulation (n = 3 independent differentiations for each genotype, data point represents the mean of two technical replicates tested in duplicate). Data are presented as mean ± SEM analyzed by ANOVA followed by Tukey's multiple comparisons test. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant. ANOVA, analysis of variance; ELISA, enzyme‐linked immunosorbent assay; LPS, lipopolysaccharide; MGL, microglia‐like cell; PLCG2, phospholipase C gamma 2; SEM, standard error of the mean; TGF‐β, transforming growth factor‐beta

4. DISCUSSION

This study investigated the cell intrinsic attributes of microglia bearing rare variants of PLCG2, the P522R protective variant and the M28L risk variant, along with PLCG2‐deficient microglia, on microglial transcriptional landscapes, proliferation, survival, and immune response in human MGLs. Not only have we established a valuable experimental platform for future therapeutic research, but by studying these microglia, we can better understand the mechanisms and functions associated with mitigating AD pathology.

PLCG2 plays a crucial role in signal transduction downstream from cell surface receptors that are important for regulation of microglial transcriptional signatures and functions. Our previous work showed that mice carrying either PLCG2^P522R or PLCG2^M28L induced distinct microglial subpopulations in response to amyloid pathology in vivo ¹⁹ , highlighting the importance of PLCG2 in mediating transitions in microglial phenotype in the context of AD. This study supports these findings in a translationally relevant human cellular context by showing significant basal transcriptional impact on biological processes in PLCG2^KO and the partial loss‐of function‐variant PLCG2^M28L MGLs relative to PLCG2^WT MGLs. Interestingly, we did not observe significant transcriptional effects in MGLs bearing PLCG2^P522R. Since we previously showed that murine microglia carrying PLCG2^P522R exhibited numerous effects on microglial gene expression in the context of amyloid pathology, the mild effect of PLCG2^P522R on basal human MGL gene expression suggests additional stimuli or more complex experimental systems may be required to fully understand its transcriptional impact. This aligns with prior studies showing mild effects of the PLCG2^P522R variant on the microglial transcriptome in vitro or in healthy mice, ²⁷ , ²⁹ but a stronger effect in the presence of pathology. ¹⁹ , ²⁶ , ³¹

Though the basal effect of the PLCG2^P522R variant on microglial gene expression was less marked compared to the risk variant, we identified genes that were uniquely differentially expressed. Of these genes, THEM6 is of interest. THEM6 encodes thioesterase superfamily member six, which is involved in lipid metabolism and homeostasis. ⁴⁴ THEM6 has been shown to be a microglial Aβ response protein, with increased abundance in the presence of amyloid pathology. ⁴⁵ Additionally, APP/PS1 mice showed decreased expression of Them6 after aducanumab anti‐amyloid therapy. ⁴⁶ Given that THEM6 expression was also decreased in PLCG2^P522R microglia, THEM6 and the effect of PLCG2^P522R on microglial lipid homeostasis and stress responses may warrant further investigation.

An interesting observation in this study was the correlation between PLCG2 and TREM2 expression, highlighting a potential co‐regulatory relationship. Transcriptomic and protein analyses showed that PLCG2^KO and PLCG2^M28L MGLs exhibited significant reductions in both PLCG2 and TREM2 levels, whereas PLCG2^WT and PLCG2^P522R MGLs maintained robust levels. Notably, reduction of PLCG2 protein levels in PLCG2^M28L MGLs, despite unchanged mRNA expression, aligns with previous studies in AD mouse models. ¹⁹ This suggests that PLCG2^M28L is post‐translationally degraded or that the mRNA is not translated effectively. We also recently observed that TREM2 was significantly decreased in 5xFAD Plcg2 knockout mice. ⁴⁷ A potential mechanistic link for this observation could be that PLCG2 loss‐of‐function leads to increased TREM2 internalization/degradation. Obst et al. found that PLCG2 knockout iPSC‐derived human macrophages exhibit decreased TREM2 and sTREM2, indicating that loss‐of‐function likely does not lead to increased shedding of TREM2. ²⁸ Loss of TREM2 has been shown to induce age‐related myelin dysregulation, ⁴⁸ thus the effect of PLCG2 on myelin homeostasis may be an important area for further investigation. Given that both PLCG2 and TREM2 are implicated in AD pathology, understanding their interplay is critical for informing the development of therapies targeting these proteins.

Our observations indicate that PLCG2 is also important for microglial functions by balancing proliferation and apoptosis. Prior studies have shown contradictory findings with respect to the role of PLCG2 in proliferation and apoptosis. Studies of various tumor cells, SH‐SY5Y cells, and iPS‐derived macrophages have shown increased ⁴⁹ or decreased ⁵⁰ , ⁵¹ proliferation as well as increased cell death ²⁸ , ⁵⁰ , ⁵¹ , ⁵² due to decreased PLCG2 expression. Overexpression studies in tumor cells have shown increased ⁵⁰ or decreased ⁴⁹ , ⁵³ , ⁵⁴ proliferation as well as increased ⁵³ , ⁵⁵ or decreased ⁵⁰ cell death. Here, EdU and Ki‐67 assays demonstrated that PLCG2^KO and PLCG2^M28L MGLs exhibit increased proliferation compared to PLCG2^WT and PLCG2^P522R MGLs. This hyperproliferative state could reflect a compensatory response to cellular stress or dysfunction. Cytokine starvation revealed that PLCG2^KO and PLCG2^M28L MGLs are more susceptible to cell death overall, while caspase‐based assays showed that PLCG2^KO but not PLCG2^M28L MGLs were susceptible to apoptosis. These results indicate that the increased cell death observed in PLCG2^M28L MGLs stems from a different cell death mechanism. Together, these results suggest that loss of PLCG2 disrupts the equilibrium between pro‐survival and pro‐apoptotic pathways. Conversely, PLCG2^P522R MGLs showed resilience to apoptosis compared to PLCG2^WT MGLs, with no difference in proliferation. This finding supports previous studies showing that the P522R variant enhances cellular survival ²⁹ .

Another key role of microglia is activation in response to pathogen‐associated molecular patterns (PAMPs) and damage‐associated molecular patterns (DAMPs) in AD. We observed that the response to PAMPs like LPS is another function of microglia that is influenced by PLCG2 signal transduction. Transcriptomic and functional assays revealed that PLCG2^KO and PLCG2^M28L MGLs exhibit altered responsiveness to LPS. Importantly, the cytokine secretome of PLCG2^M28L MGLs shared similarities to PLCG2^WT MGLs but showed a decrease in the secretion of some cytokines (IL‐3, IL‐9, IL‐17A, IL‐27, and TGF‐α). The dysregulation observed in PLCG2‐deficient and loss‐of‐function MGLs has profound implications for AD, as ineffective microglial responses to PAMPs and DAMPs may contribute to amyloid plaque accumulation and subsequent neurodegeneration.

Conversely, PLCG2^P522R MGLs showed a more robust pro‐inflammatory cytokine response to LPS stimulation compared to PLCG2^WT MGLs. The increased levels of CXCL1, IL‐6, and TNFα align with prior investigations, ¹⁹ , ²⁴ , ²⁹ although a recent report observed decreased TNFα secretion by monocyte‐derived macrophages derived from PLCG2^P522R carriers. ³¹ Interestingly, PLCG2^P522R MGLs also showed an increase in IL‐1RA, indicating that these MGLs might be simultaneously promoting specific inflammatory responses while also shutting down aberrant signaling. Chemokine secretion was also increased in PLCG2^P522R MGLs, potentially explaining increased clustering of P522R variant microglia around plaques in vivo. ¹⁹ Future work should address whether the protective mechanism imparted by microglia bearing PLCG2^P522R is due to increased responsiveness to insult combined with a built‐in regulatory mechanism via IL‐1RA. Additionally, PLCG2^P522R microglia may be poised to mount a unique immune response characterized by the unique upregulation of DAPK1 and unique downregulation of CGAS as seen in the transcriptomic data. Overall, these results suggest that PLCG2 is essential for proper microglial response to PAMPs, and its loss leads to an attenuated inflammatory response.

While this study does not directly identify which signaling pathways contribute to the functional changes associated with PLCG2 variants, the transcriptomic data suggest that, because there are only relatively small changes in TREM2 signaling‐related genes, PLCG2^P522R may mediate its effects post‐transcriptionally via protein processing, translocation, or posttranslational modifications. The PLCG2^P522R variant may even modulate interactions with other proteins after receptor activation or may subtly alter association with the interleaflet of the plasma membrane.

Taken together, these findings establish PLCG2 as a key regulator of microglial function, with broad implications for AD and neurodegeneration. Loss‐of‐function genotypes, PLCG2^KO and PLCG2^M28L result in profound microglial dysfunction, affecting proliferation, apoptosis, and immune response—all of which can contribute to AD pathogenesis. This study raises critical questions regarding potential therapeutic strategies including: Could modulating microglial proliferation, apoptosis, or other responses through PLCG2 or related pathways mitigate AD progression? Why does a PLCG2 variant associated with protection from AD lead to increases in cytokine levels that are typically thought to be effectors of AD pathology? These insights emphasize that inactivation of PLCG2 is unlikely to be a viable therapeutic strategy for AD. Instead, maintaining PLCG2 activity may hold potential as a neuroprotective strategy, with the PLCG2^P522R variant serving as a valuable benchmark for identifying therapeutic compounds capable of promoting PLCG2 activity. This study also identifies resistance to apoptosis as an altered function unique to the PLCG2^P522R MGLs. Future studies are warranted to identify if PLCG2^P522R confers an anti‐apoptotic phenotype in response to other DAMPs and PAMPs, or if resilience is limited to particular cellular contexts. Finally, investigations of other mechanisms of cell death may reveal other preferential death pathways related to PLCG2 loss of function.

Despite the insights gained, several limitations must be acknowledged. Due to the low throughput and personnel requirements of managing four parallel MGL lines, this study relied on a single isogenic iPSC‐derived line, which may not fully capture phenotypes dependent upon individual genetic variability. Additionally, in vitro models have inherent limitations in recapitulating the in vivo microenvironment. Further studies in animal models will help contextualize MGL phenotypes in vivo. The lack of priming stimuli conditions for most of the experiments in this study may have masked functional differences, particularly for microglia bearing PLCG2^P522R, which might require stimuli to reveal its full effects. Follow‐on functional experiments incorporating varied microglial stimulations will be an important area for future investigation. In these future studies, phagocytic capacity for pathogenic particles should be analyzed and compared. These future avenues of research should also include more complex model systems to better understand the dynamic role of PLCG2 in microglia.

In conclusion, this study provides critical insights into how PLCG2 variants shape microglial biology, influencing their transcriptional profiles, proliferation, survival, and immune competence. The findings underscore the importance of PLCG2‐TREM2 interactions in regulating microglial function and highlight the potential of PLCG2‐targeted therapies in AD. Furthermore, by using human MGLs, we have established a valuable experimental platform for future therapeutic research. Future studies are needed to translate these discoveries into clinical interventions that address the underlying microglial dysfunction driving AD pathology.

AUTHOR CONTRIBUTIONS

Logan M. Bedford, Kaylee D. Tutrow, Timothy I. Richardson, and Stephanie J. Bissel conceived and designed the study. Logan M. Bedford and Kaylee D. Tutrow conducted the experiments and analyzed data. Kaylee D. Tutrow and Melody Hernandez generated the PLCG2^M28L iPSCs. Kaylee D. Tutrow and Karly Hooper performed cell culture and assisted with immunofluorescence experiments. Karly Hooper assisted with apoptosis and LPS experiments. Evan J. Messenger assisted with RNA sequencing data analysis. Logan M. Bedford, Kaylee D. Tutrow, Bruce T. Lamb, Timothy I. Richardson, Jason S. Meyer, and Stephanie J. Bissel interpreted and discussed the results. Logan M. Bedford and Stephanie J. Bissel wrote the manuscript and assembled figures with revisions from Kaylee D. Tutrow, Timothy I. Richardson, Jason S. Meyer, and Bruce T. Lamb.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest. Author disclosures are available in the supporting information.

CONSENT STATEMENT

No human subject materials were used in this study.

Supporting information

Supporting Information

Supporting Information

Supporting Information

Supporting Information

Bedford LM, Tutrow KD, Hooper K, et al. Alzheimer's disease–associated PLCG2 variants alter microglial state and function in human induced pluripotent stem cell–derived microglia‐like cells. Alzheimer's Dement. 2025;21:e70772. 10.1002/alz.70772