G protein-coupled receptor ADGRG1 drives a protective microglial state in Alzheimer’s disease through MYC activation

Summary

Germline genetic architecture of Alzheimer’s disease (AD) indicates microglial mechanisms of disease susceptibility and outcomes. However, the mechanisms enabling protective microglial responses remain elusive. Here, we investigate the role of microglial ADGRG1, an adhesion G protein-coupled receptor (aGPCR) specifically expressed in yolk sac-derived microglia, in AD pathology using the 5xFAD mouse model. Transcriptomic analyses reveal that ADGRG1 activates the transcription factor MYC, leading to upregulation of genes involved in homeostasis, phagocytosis, and lysosomal functions, thereby promoting a protective microglial state. We demonstrate that deletion of Adgrg1 in microglia impairs MYC activation, resulting in increased amyloid-beta deposition, exacerbated neuronal loss, and cognitive deficits. Functional assays in mouse models and human embryonic stem cell-derived microglia confirm that ADGRG1 is required for Aβ phagocytosis. These findings uncover a GPCR-mediated pathway that drives a protective microglial state via MYC activation, suggesting potential therapeutic strategies to alleviate AD progression by enhancing microglial functional competence.

Graphical Abstract

In brief, microglial ADGRG1 drives a protective state in Alzheimer’s disease by activating the transcription factor MYC, which upregulates genes associated with homeostasis, phagocytosis, and lysosomal functions. This leads to increased Aβ clearance, improved neuronal health, and preserved cognitive function.

Introduction

Alzheimer’s disease (AD) is a prevalent form of late-life dementia, characterized by progressive cognitive decline and memory impairment.^1–5 Pathologically, AD is defined by the accumulation of extracellular amyloid beta (Aβ) senile plaques and intracellular neurofibrillary tangles composed of hyperphosphorylated tau.^1–3,6 These proteinaceous aggregates cause extensive neuritic dystrophy, synaptic and neuronal loss, and widespread neuroinflammation, cumulatively leading to neurodegeneration.^1–4,7,8

Microglia, the resident innate immune cells of the central nervous system (CNS), play important roles in maintaining CNS homeostasis and responding to pathological insults.^9–11 In the context of AD, microglia are activated and accumulate around Aβ plaques, where they can exhibit both neuroprotective and neurotoxic activities.¹² The functional state of microglia is closely linked to their gene expression profiles, with recent studies highlighting distinct activation states with specific functions in neurodegenerative diseases.^13–15 For instance, the expression of CLEC7A, a c-type lectin receptor, has been associated with a protective microglial state that enhances myelin debris clearance.^16,17 However, much of our current understanding of microglial state changes in AD is derived from transcriptomic analyses, and the mechanisms driving these state transitions remain incompletely understood.^18–20 Elucidating how microglial states contribute to AD progression is crucial for developing targeted therapeutic strategies.

Microglia are derived from yolk sac macrophage progenitors that infiltrate the developing CNS during early embryogenesis, distinguishing them from other CNS-resident macrophages and peripheral immune cells.^21–24 This unique ontogeny imparts specialized properties to microglia, including their ability to dynamically respond to environmental cues and modulate neuronal function.^20,25–27 Notably, yolk sac-derived microglia expressing P2RY12, a purinergic receptor involved in microglial chemotaxis and surveillance, are particularly responsive to AD pathology and have been implicated in influencing disease progression.^26,28,29

ADGRG1 (also known as GPR56) is an adhesion G protein-coupled receptor (aGPCR) that is selectively expressed in yolk sac-derived microglia, but not in other microglia-like cells or peripheral macrophages.^21,22,30 ADGRG1 plays critical roles in brain development, and loss-of-function mutations in ADGRG1 result in severe cortical malformations, such as bilateral frontoparietal polymicrogyria in humans.^31,32 In mice, deletion of microglial Adgrg1 impairs synaptic pruning and interneuron development in early postnatal stages, highlighting its importance in microglial function and CNS development.^33,34 Given that ADGRG1 defines yolk sac-derived microglia and is crucial for their developmental functions, it serves as an excellent model to study the role of microglia in AD and to understand how microglial state changes impact disease progression. Moreover, recent single-nucleus RNA sequencing (snRNA-seq) studies have identified ADGRG1 as one of the top five upregulated genes in microglia from individuals with mild cognitive impairment (MCI) compared to those from individuals with no AD or severe AD, suggesting a potential role in disease resilience.³⁵

In this study, we investigate the role of microglial ADGRG1 in AD pathology using the 5xFAD mouse model, which recapitulates key features of Aβ deposition and cognitive deficits observed in AD.³⁶ We hypothesize that ADGRG1 mediates a protective microglial state that promotes Aβ clearance and supports neuronal health. By selectively deleting Adgrg1 in microglia, we aim to elucidate the mechanisms by which ADGRG1 influences microglial state transitions and functions in the context of AD. Our findings reveal that microglial Adgrg1 deletion leads to altered microglial states characterized by reduced expression of genes involved in homeostasis, phagocytosis, and lysosomal function. This shift results in impaired microglial phagocytosis of Aβ, increased plaque burden, neuronal loss, and cognitive deficits. Furthermore, we identify the transcription factor MYC as a downstream effector of ADGRG1 signaling that regulates microglial state transitions and Aβ clearance. Collectively, our study underscores the importance of microglial state changes in AD progression and highlights ADGRG1 as a critical mediator of protective microglial functions. Understanding the role of ADGRG1 in microglial biology may offer new therapeutic avenues for modulating microglial responses in AD.

Results

Adgrg1 deletion alters microglial states in 5xFAD mice at the transcriptomic level.

To elucidate the role of microglial ADGRG1 in AD pathology, we employed Cx3cr1^Cre(Jung) transgenic mice,³⁷ based on our past success with this Cre driver.^33,34 Although Cx3cr1^Cre/+ induces recombination in all myeloid cells, Adgrg1 is selectively expressed in microglia, but not in other tissue mononuclear phagocytes.^21,30 We crossed Adgrg1^fl/fl;Cx3cr1^Cre/+ with 5xFAD mice to generate conditional knockout mice 5xFAD;Adgrg1^fl/fl;Cx3cr1^Cre/+ (5xFAD-cKO) and age-matched controls, 5xFAD;Adgrg1^+/+;Cx3cr1^Cre/+ mice (5xFAD-Con) (Figure S1A). We confirmed specific deletion of Adgrg1 in microglia (Figures S1B–E).

To investigate how ADGRG1 regulates microglial responses to amyloid deposition, we conducted snRNA-seq on nuclei isolated from the neocortex of 6-month-old Adgrg1^+/+;Cx3cr1^Cre/+ (Con), Adgrg1^fl/fl;Cx3cr1^Cre/+ (cKO), 5xFAD-Con, and 5xFAD-cKO mice (Figures 1A and S1F–J). Given that microglia are often under-represented in snRNA-seq datasets,³⁸ we enriched for glial cells by staining nuclei with DAPI and NeuN, followed by flow cytometry sorting for DAPI⁺ and DAPI⁺/NeuN⁻ populations (Figure 1A). We sequenced 191,292 single nuclei, with a target of 50,000 reads per cell, resulting in the detection of 1500–2500 genes per cell on average. After quality control, doublet removal, and batch integration, eleven clusters were identified based on their transcriptomic profile, as shown in the uniform manifold approximation projection (UMAP) plot (Figure 1B). The clusters were annotated based on known cell-type marker genes (Figure S1K). Microglial populations were identified by shared enrichment of Csf1r, Spi1, P2ry12, and Cx3cr1 (Figure S1L).

Figure 1. The deletion of Adgrg1 in microglia altered microglial transcriptomic profiles.

(A) Schematic of the snRNA-seq experimental design on 6-month-old mice. Schematic created with BioRender.com.

(B) UMAP plots of all single nuclei and their annotated cell types and proportions.

(C) UMAP plots of microglia subclusters split by genotype and colored according to subclusters.

(D) UMAP plots of mouse microglia snRNA-seq datasets from Zhou et al. (3,712 nuclei)⁴³ re-analyzed and projected onto our microglial subclustering as in (C).

(E) Bar plots showing the percentage of cells in each cluster across genotypes. Statistical comparisons used Propeller unpaired two tailed t-test analysis between 5xFAD; Adgrg1^+/+; Cx3cr1^Cre/+ and 5xFAD; Adgrg1^fl/fl; Cx3cr1^Cre/+ mice.

(F) Heatmap of the top differentially expressed genes defining in each microglia subcluster.

(G) Dot plot showing the expression of lipid processing-related genes in microglia subclusters 3 and 4.

See also Figure S1, S2 and Table S1.

To further dissect microglial heterogeneity, we performed iterative clustering on 11,398 myeloid cells and initially identified eight transcriptomic subclusters. During quality control, one cluster (cluster 7, comprising 1.87% of myeloid cells) was found to exhibit abnormally high RNA content (nFeature_RNA and nCount_RNA), suggesting the presence of potential doublets or hybrid cells (Figure S1M). After excluding this cluster, the remaining cells showed a more coherent nFeature_RNA versus nCount_RNA distribution across clusters (Figure S1N). Therefore, cluster 7 was excluded from all downstream analyses. Following this exclusion, we identified seven transcriptomic subclusters (Figures 1C and S2A). These subclusters were labeled based on their marker gene expression profiles: homeostatic microglia (Tmem119, Cx3cr1, P2ry12; clusters 0, 1), neuronal interacting microglia (Nrg1³⁹, Nrg3⁴⁰; cluster 2), disease-associated microglia (DAM; clusters 3, 4), and two macrophages clusters, including border-associated macrophages (Lyve1, Mrc1, Cd163; cluster 5) and a CCR2⁺ macrophages cluster (cluster 6) with Ccr2 and major histocompatibility complex II (MHC-II) genes (H2-Eb1, H2-Aa, and H2-Ab1) expression (Figures 1C–F, Figures S2D–H, Table S1). Adgrg1 expression was broadly detected across microglial subclusters, albeit lower in DAM clusters (clusters 3, 4), and was not detected in macrophages (clusters 5, 6), consistent with previous studies^41,42 (Figures S2I–K). Notably, DAMs were consistently found in both 5xFAD-Con and 5xFAD-cKO mice (Figure 1C), indicating that Adgrg1 deletion does not impaire microglia progress to the full DAM stage, in contrast to Trem2-dependent DAM1 to DAM2 transition in 5xFAD mice^13,43 (Figure 1D).

However, within the DAM population, microglial Adgrg1 deletion shifted the proportion of microglia from cluster 4 to cluster 3 (Figures 1C, 1E, S2A–C). Both clusters expressed core DAM markers (Myo1e, Igf1, Axl), but cluster 4 cells, which were reduced in 5xFAD-cKO, were uniquely enriched for genes involved in phagocytosis and lysosomal function, including complement (C1qa-c) pathway genes, tetraspanin genes (Cd9 and Cd81), Trem2, the Cathepsin gene family (Ctsb), actin (Actb), and lysosomal genes (Lyz2, Tyrobp, Cst7) (Figures 1F, S2E and S2F). Additionally, homeostatic microglia (P2ry12, Cx3cr1, Tmem119) with enhanced phagocytic function (Cd81, C1qa-c) were less abundant in 5xFAD-cKO mice compared to 5xFAD-Con mice (cluster 0, Figures 1C, 1E, 1F and S2D).

To further dissect the functional divergence between DAM subclusters 3 and 4, we focused on the expression of lipid-processing genes, given increasing evidence that lipid metabolism is a key component of DAM states.^44–46 Cluster 3 microglia were enriched for genes involved in fatty acid activation for intracellular transport and metabolism (Acsl3, Acsl4, Acsl6)⁴⁷, cholesterol efflux (Abca1⁴⁸, Npc1⁴⁹), and lipid-sensing transcriptional regulators (Ppara⁵⁰, Ppard⁵¹, Nr1h2 and Nr1h3⁵²). In contrast, cluster 4 microglia exhibited higher expression of genes associated with lipoprotein processing and uptake (Lpl⁵³, Apoe⁵⁴, Lrp1⁵⁵) and lysosomal glycolipid degradation (Hexa⁵⁶, Hexb⁵⁷) (Figure 1G). These findings suggest that while both DAM subclusters engage lipid metabolism pathways, they are functionally specialized: cluster 3 is enriched for lipid activation, transport and regulatory genes, whereas cluster 4 is more associated with lipid uptake and lysosomal degradation. Taken together, these results highlight that microglial Adgrg1 deletion leads to altered microglial states characterized by reduced expression of genes related to homeostasis, phagocytosis, lysosomal function and lipid clearance, suggesting a loss of protective microglial responses to AD pathology.

Deletion of Adgrg1 diminishes microglial phagocytosis of Aβ in vivo and in human ESC-derived microglia.

To further investigate how ADGRG1 influences microglial response to Aβ pathology, we performed differential gene expression (DEG) analysis. We found a downregulation of lysosomal function-associated genes (Cd68, Grn, Ctsb, Ctsd, Ctss, Ctsl) in 5xFAD-cKO microglia compared to 5xFAD-Con (Figure 2A, Table S2). In contrast, several genes previously implicated in exacerbating neurodegenerative phenotypes were upregulated. Notably, Csf3r, the receptor for granulocyte colony-stimulating factor (G-CSF), has been shown to drive microglia toward a pro-inflammatory state.⁵⁸ Mtor, a central metabolic regulator, is frequently hyperactivated in AD and promotes Aβ accumulation and cognitive impairment^59,60 (Figure 2A). Gene ontology (GO) pathway analysis highlighted significant alterations in lysosome and phagocytosis pathways due to Adgrg1 deficiency (Figure 2B).

Figure 2. Microglial ADGRG1 regulates Aβ phagocytosis in AD mouse models and hESC-derived microglia.

(A) Volcano plot showing significant differentially expressed genes in microglia of 5xFAD; Adgrg1^+/+; Cx3cr1^Cre/+ versus 5xFAD; Adgrg1^fl/fl; Cx3cr1^Cre/+.

(B) Gene ontology analysis showing hallmark pathways in microglia.

(C) Schematic of WT and Adgrg1 KO primary microglia treated with pHrodo-labeled Aβ and recorded using the Incucyte S3 live imaging system.

(D) Representative Incucyte images of WT and KO primary microglia phagocytosing pHrodo-labeled Aβ after 1-hour and 24-hour incubation. Scale bar, 100 μm.

(E) pHrodo-labeled Aβ signal (total integrated intensity) measured hourly over 24 hours.

Quantification of area under curve. n=3 biological replicates per genotype. *** p=0.0008. Unpaired two-tailed t-test.

(F) 3D reconstruction of microglia from 6-month-old 5xFAD; Adgrg1^+/+; Cx3cr1^Cre/+ and 5xFAD; Adgrg1^fl/fl; Cx3cr1^Cre/+ mice, showing engulfed MOAB2⁺ Aβ inside Iba1⁺ microglia. Scale bar, 10 μm.

(G) Percentage of the engulfed Aβ volume to the total Aβ volume. Each circle represents one animal. * p=0.0400. Unpaired two-tailed t-test. Filled circles, male mice; open circles, female mice.

(H) Representative FACS plots of methoxy-X04⁺ plaques and CD11b⁺ microglia in the cortex of 6-month-old 5xFAD; Adgrg1^+/+; Cx3cr1^Cre/+ and 5xFAD; Adgrg1^fl/fl; Cx3cr1^Cre/+ mice. Q2 represents the Aβ plaques that were engulfed by microglia within the past three hours.

(I) Percentage of methoxy-X04⁺ microglia. Each circle represents one animal. **** p<0.0001. Unpaired two-tailed t-test. Filled circles, male mice; open circles, female mice.

(J) Schematic showing the differentiation, maturation, treatment, and Incucyte recording of human WT and ADGRG1 KO hESC-derived microglia.

(K) pHrodo-labeled Aβ signal (total integrated intensity) in WT and ADGRG1 KO hESC-derived microglia measured hourly over 24 hours. Quantification of area under curve. n=8 for WT, n=9 for ADGRG1 KO. n represents the number of independent microglial treatments. **** p<0.0001. Unpaired two-tailed t-test.

Data are presented as mean ± SEM. Schematic created with BioRender.com.

See also Figure S3 and Table S2.

To connect these transcriptomic changes to functional outcomes, we asked whether ADGRG1 deficiency affected microglial phagocytosis. We evaluated the uptake of pHrodo-labeled oligomeric Aβ_1–42 in primary cultured microglia (Figure 2C). We confirmed over 98% purity of our isolated primary microglia (Figure S3A) and the absence of ADGRG1 protein in Adgrg1 KO microglia (Figure S3B). Using the Incucyte live-cell imaging system, we recorded RFP signals hourly over a 24-hour duration. Adgrg1-deficient microglia phagocytosed significantly less pHrodo-Aβ_1–42 than wild-type (WT) microglia (Figures 2D and 2E). We next performed 3D reconstructions of Iba1, CD68, and MOAB2⁶¹, an antibody labels Aβ plaques, on 6-month-old 5xFAD-Con and 5xFAD-cKO brains. Microglia from 5xFAD-cKO mice engulfed significantly less Aβ compared to 5xFAD-Con microglia (Figures 2F and 2G). To quantify Aβ uptake in vivo, we injected 6-month-old mice intraperitoneally (i.p.) with methoxy-X04, a fluorescent dye that specifically binds to Aβ fibrillar β-sheet structures,^62,63 and harvested mouse brain cortices 3 hours later. Microglia were labeled with CD11b antibody for subsequent flow cytometry analysis (Figures S3C and S3D). 5xFAD-cKO mice had significantly fewer methoxy-X04⁺ microglia, suggesting reduced Aβ engulfment due to the deletion of microglial Adgrg1 (Figures 2H and 2I).

To determine whether the phagocytosis impairment observed in Adgrg1-deficient mouse microglia also occurs in human microglia, we generated an ADGRG1 KO human embryonic stem cell (ESC) line using CRISPR-Cas9 (Figure S3E) and differentiated both KO and isogenic control human ESCs into microglia-like cells (iMG) (Figures 2J and S3F). Upon treatment with pHrodo-oligomeric Aβ, ADGRG1 KO iMG showed significantly reduced Aβ engulfment compared to WT iMG, as monitored hourly for a time course of 24 hours using the Incucyte system (Figure 2K). These findings demonstrate that ADGRG1 deficiency impairs microglial phagocytosis of Aβ in both mouse and human microglia, highlighting the importance of ADGRG1 in mediating protective microglial functions in AD.

Transcription factor MYC activation is a downstream effector of microglial ADGRG1 that drives microglial Aβ phagocytosis.

Our studies indicate that deleting Adgrg1 in 5xFAD mice results in a distinct microglial state characterized by the downregulation of genes associated with phagocytosis and lysosomal functions (Figures 1 and 2). The transcriptional state of a cell emerges from an underlying gene regulatory network in which a limited number of transcription factors (TFs) and co-factors regulate each other and their downstream target genes. To uncover potential TFs mediating these changes, we used the ChIP Enrichment Analysis (ChEA) database,⁶⁴ based on DEGs identified in microglia from 5xFAD-Con vs 5xFAD-cKO (Table S2). Among the top 10 identified TF entries, MYC appeared in multiple datasets (Figure 3A, Table S3), consistent with the literature that MYC regulates genes such as Cd68, Cd81, Actb, and members of the cathepsin family^65,66. C-Myc is an oncogene that encodes the protein MYC, a transcription factor involved in cell proliferation, growth, apoptosis and metabolism.^67,68 While transcript levels of c-Myc were not significantly different between 5xFAD-Con and 5xFAD-cKO microglia (Figure S4A), the phosphorylation of MYC at Serine 62 (MYC^pS62) enhances its stability and transcriptional activity.^69–72 Kinases such as ERK⁷², CDK2⁷³, CDK5⁷⁴, and Rho-Kinase⁷⁵ are known to induce MYC phosphorylation at the Serine 62 site. Given that ADGRG1 activates Rho signaling pathways,^76–80 we hypothesized that microglial ADGRG1 promotes MYC phosphorylation at Serine 62 site through activated RhoA, thereby regulating downstream gene transcription (Figure 3B).

Figure 3. Transcription factor MYC activation is a downstream effector of microglial ADGRG1 that regulates microglial Aβ phagocytosis.

(A) Top 10 enriched transcription factors from the ChEA database based on significant downregulated DEGs in microglia from 5xFAD; Adgrg1^fl/fl; Cx3cr1^Cre/+ mice compared to 5xFAD; Adgrg1^+/+; Cx3cr1^Cre/+.

(B) Schematic representation of the hypothesized pathway where microglial ADGRG1 mediates MYC activation through the RhoA signaling pathways.

(C) Representative images of the cortex from 6-month-old 5xFAD; Adgrg1^+/+; Cx3cr1^Cre/+ and 5xFAD; Adgrg1^fl/fl; Cx3cr1^Cre/+ mice stained for Iba1 (red), GTP-RhoA (green) and DAPI. Scale bar, 10 μm. Box indicating the zoom-in view. Scale bar, 5 μm.

(D) Quantification of the percentage of GTP-RhoA⁺ microglia. Each circle represents one animal. * p=0.0161. Unpaired two-tailed t-test.

(E) Quantification of the percentage of GTP-RhoA⁺ microglia and non-microglial cells. Each circle represents one animal. ** p=0.0083 for Iba1⁺ cells; ** p=0.0068 for Iba1⁻ cells. Two-way ANOVA with Bonferroni’s multiple comparisons test.

(F) Representative images of cortices from 6-month-old 5xFAD; Adgrg1^+/+; Cx3cr1^Cre/+ and 5xFAD; Adgrg1^fl/fl; Cx3cr1^Cre/+ mice stained for MYC^pS62 (red), Iba1 (green), MOAB2 (gray) and DAPI. Scale bar, 10 μm.

(G) Quantification of MYC^pS62 positive plaque-associated and non-plaque-associated microglia in the cortices of 6-month-old mice. Each circle represents one animal. **** p<0.001. Two-way ANOVA with Bonferroni’s multiple comparisons test.

(H) Total integrated intensity of pHrodo-labeled Aβ signal measured every 3 hours over 24 hours in primary WT and Adgrg1 KO microglia, with or without 20 μM MYC inhibitor pre-treatment. Quantification of area under curve. n=4 biological replicates per genotype. *** p=0.0005, **** p<0.0001, One-way ANOVA with Bonferroni’s multiple comparisons test.

(I) Schematic of intracerebroventricular (ICV) injection of vehicle or MYC inhibitor (0.05 mg/kg per hemisphere) into 8–9 month old 5xFAD mice. Mice were euthanized 48 hours later for in vivo Aβ engulfment assay and plaque burden analysis.

(J) Representative flow cytometry plots showing methoxy-X04⁺ plaques and CD11b⁺ microglia from 5xFAD mice treated with a vehicle or the MYC inhibitor.

(K) Quantification of methoxy-X04⁺ microglia. Each circle represents one animal. ** p=0.0047. Unpaired two-tailed t-test.

(L) Representative tile scan images of Aβ staining (MOAB2) in 5xFAD mice following ICV administration of vehicle or MYC inhibitor. Scale bar, 1 mm.

(M) Quantification of the MOAB2⁺ area in upper and deeper cortical layers. Each circle represents one animal. *** p=0.0003. Two-way ANOVA with Bonferroni’s multiple comparisons test.

Filled circles, male mice; open circles, female mice. Data are presented as mean ± SEM.

See also Figure S4 and Table S3.

We first examined the activation of RhoA in microglia by co-staining for Iba1, GTP-RhoA (the active form of RhoA), and MOAB2 in 6-month-old mice. A significantly higher level of GTP-RhoA signal was observed in microglia of 5xFAD-Con mice compared to 5xFAD-cKO mice (Figures 3C and 3D). Interestingly, we saw significantly higher percentage of GTP-RhoA positive non-microglial cells in 5xFAD-Con brains as well (Figures 3C and 3E). Triple immunostaining of Iba1, GTP-RhoA, and GFAP showed that many of those Iba1 negative GTP-RhoA positive cells were astrocytes (Figure S4B). Next, we assessed MYC activation by immunostaining for MYC^pS62, Iba1 and MOAB2, and found a significantly higher percentage of MYC^pS62 positive microglia in 5xFAD-Con mice compared to 5xFAD-cKO mice (Figures 3F and 3G). Importantly, MYC^pS62 positive microglia were predominantly localized to plaque-associated regions, where microglia express CLEC7A, a marker of DAM,^13,16 suggesting a role for MYC in regulating Aβ phagocytosis (Figures S4C–E).

To further investigate whether Aβ induces MYC activation in a manner dependent on ADGRG1, we treated primary WT and Adgrg1 KO microglia with 1uM of fibrillar Aβ for 24 hours. Immunofluorescence and Western blot analyses revealed that Aβ treatment resulted in a significantly higher level of MYC^pS62 in WT microglia, but not in Adgrg1 KO microglia (Figures S4F–I), suggesting that ADGRG1 is required for Aβ-induced MYC^pS62 production in microglia.

To determine whether MYC contributes to Aβ phagocytosis, we performed an in vitro microglia engulfment assay using pHrodo-labeled oligomeric Aβ, with or without pre-treatment of a MYC inhibitor (10058-F4)⁸¹ (Figure S4J). In WT microglia, MYC inhibition significantly reduced Aβ uptake (Figure 3H and S4K). Consistent with the result shown in Figure 2E, we observed significantly impaired Aβ phagocytosis in Adgrg1 KO microglia. To extend these findings in vivo, we administered MYC inhibitor 10058-F4 and its carrier solution as a control intracerebroventricularly (ICV) into 8- to 9-month-old 5xFAD mice bilaterally. At 45 hours after ICV injection, which was three hours prior to brain harvest, mice were injected with methoxy-X04 to label Aβ plaques (Figure 3I). We used one hemisphere of each of the cortices for the in vivo Aβ engulfment assay, and the other hemispheres for Aβ burden evaluation by immunostaining. Flow cytometry analysis revealed a significant reduction in the proportion of Aβ-phagocytosing CD11b-positive microglia following MYC inhibition (Figures 3J, 3K and S4L). Moreover, MYC inhibitor-treated mice displayed a significant increase in cortical Aβ burden compared to vehicle controls (Figures 3L, 3M and S4M). Taken together, these data suggest that microglial ADGRG1 regulates Aβ phagocytosis through a MYC-dependent mechanism, indicating a novel pathway by which microglial state changes influence AD pathology.

Microglial Adgrg1 deletion exacerbates Aβ plaque load and impairs microglia-plaque association.

Given the impaired phagocytosis function of microglia lacking ADGRG1 in vitro, we hypothesized that deleting microglial Adgrg1 will lead to failure of plaque clearance. To test this hypothesis, we examined Aβ plaque accumulation in vivo. We observed a significant increase in Aβ deposition in the cortex and hippocampal CA1 region of 6-month-old 5xFAD-cKO mice compared to 5xFAD-Con mice (Figures S5A–E). Further analysis using Thioflavin S (ThioS)^82,83 and MOAB2 double staining revealed that plaques in 5xFAD-cKO mice were more diffuse compared to those in 5xFAD-Con mice (Figures S5F and S5G), suggesting altered plaque morphology in the absence of microglial Adgrg1.

To address potential developmental effects of microglial Adgrg1 deletion^33,34, we generated an inducible microglial-specific knockout line by employing P2ry12^CreER/+;Ai14 mice⁸⁴ to generate 5xFAD;Adgrg1^fl/fl; P2ry12^CreER/+;Ai14 (AD-icKO) and the 5xFAD;Adgrg1^+/+;P2ry12^CreER/+;Ai14 mice (AD-iCon) mice. Tamoxifen was administered at P31-P35 to induce Adgrg1 deletion prior to amyloid plaque formation.³⁶ Five consecutive intraperitoneal tamoxifen injections (100 mg/kg) achieved approximately 80% recombination efficiency (Figures S5H–M). Our recombination rate exceeded what was observed in published reports, likely due to a favorable inter-loxP distance.^85,86 Importantly, we observed increased Aβ load in 5xFAD-icKO mice compared to age-matched 5xFAD-iCon mice, recapitulating the phenotype observed in constitutive microglial Adgrg1 deletion (Figures S5N–P). To examine how microglial Adgrg1 expression changes over the course of disease progression, we reanalyzed a published scRNA-seq dataset from 5xFAD mice.¹³ Adgrg1 mRNA levels showed a modest but non-significant negative correlation with age. In contrast, P2ry12, a canonical homeostatic microglial marker, exhibited a more pronounced age-dependent decline (Figure S5Q). To study whether the impact of microglial Adgrg1 deletion on Aβ pathology is influenced by the timing of gene ablation, we administered tamoxifen at 1, 3, or 5 months of age, followed by tissue harvest at 6 months (Figure 4A). Across all induction timepoints, 5xFAD-icKO mice exhibited significantly increased Aβ plaque burden compared to 5xFAD-iCon mice (Figures 4B–E), suggesting that microglial ADGRG1 remains critical for limiting plaque accumulation even after disease onset.

Figure 4. Conditional deletion of Adgrg1 in microglia exacerbated amyloid deposition, diminished microglial capacity to respond to Aβ plaques and increased neuronal loss.

(A) Schematic for tamoxifen-inducible microglial Adgrg1 deletion.

(B) Representative images of cortices from 6-month-old 5xFAD; Adgrg1^+/+; P2ry12^CreER/+ and 5xFAD; Adgrg1^fl/fl; P2ry12^CreER/+ mice following tamoxifen induction at 1, 3, and 5 months, stained for MOAB2 (red) and DAPI. Scale bar, 50 μm.

(C-E) Quantification of MOAB2⁺ plaque area in upper and deeper cortical layers in mice with tamoxifen induction at 1 month (C), 3 months (D), or 5 months (E). Each circle represents one animal. ** p=0.0090 (C), ** p=0.0071 (D), ** p=0.0072 (E). Two-way ANOVA with Bonferroni’s multiple comparisons test.

(F-G) Representative images of cortices from 6-month-old 5xFAD; Adgrg1^+/+; P2ry12^CreER/+ and 5xFAD; Adgrg1^fl/fl; P2ry12^CreER/+ mice with tamoxifen induction at 1 month and stained for MOAB2 (red), Iba1 (green), and DAPI. Zoom-in images indicating the upper (F’ and G’) and deeper (F” and G”) cortical layers. Scale bar, 50 μm.

(H) Quantification of microglia density within a 30-μm radius from the plaque cores and its relationship to the plaque size. Each circle represents a single plaque analysed. 379 plaques for 5xFAD; Adgrg1^+/+; P2ry12^CreER/+, 430 plaques for 5xFAD; Adgrg1^fl/fl; P2ry12^CreER/+, from n=3 animals per genotype. ** p=0.0012, simple linear regression of slopes.

(I) Representative images of NeuN staining in the cortices of 6-month-old 5xFAD; Adgrg1^+/+; P2ry12^CreER/+ and 5xFAD; Adgrg1^fl/fl; P2ry12^CreER/+mice. Scale bar, 50 μm.

(J) Quantification of NeuN-positive neurons. Each circle represents one animal. * p=0.0384. Two-way ANOVA with Bonferroni’s multiple comparisons test.

(K) Representative images of the cortex with TBR1 (red), CTIP2 (green) and DAPI staining. Scale bar, 50 μm.

(L) Quantification of the CTIP2-positive neurons in cortical layer V and layer VI. Each circle represents one animal. * p=0.0479. Two-way ANOVA with Bonferroni’s multiple comparisons test.

Data are presented as mean ± SEM.

See also Figures S5.

We next evaluated the density of microglia surrounding amyloid plaques. The number of microglia within a 30-μm radius from plaques was significantly reduced in both 5xFAD-icKO and 5xFAD-cKO mice at 6 months, compared to 5xFAD-iCon (Figures 4F–H) and 5xFAD-Con mice (Figures S6A–C). Sholl analysis of microglial morphology revealed no significant differences between 5xFAD-Con and 5xFAD-cKO mice (Figures S6D and S6E). Taken together, these findings suggest that microglial ADGRG1 is necessary for effective microglial recruitment of amyloid plaques and for limiting amyloid deposition, highlighting the critical role of ADGRG1 in modulating microglial responses to Aβ pathology.

Deleting microglial Adgrg1 results in increased neuronal and synaptic loss and impaired neuritic integrity in 5xFAD mice.

It has been reported that fibrillar Aβ_1–42 contributes to neurotoxicity.⁸⁷ To assess neuronal density across cortical layers, we performed NeuN⁸⁸ immunostaining and observed a significant reduction of layer V neurons in 5xFAD-cKO mice compared to 5xFAD-Con (Figures S6F and S6G). To further confirm layer V neurons were mostly affected, we conducted double-labeling of CTIP2 for layer V⁸⁹ and TBR1 for layer VI⁹⁰, and found a significant reduction in CTIP2 density in layer V in the absence of microglial Adgrg1 (Figures S6H and S6I). This phenotype was consistent between constitutive (Figures S6F–I) and inducible (Figures 4I–L) microglial Adgrg1 deletion mouse models.

Diffuse Aβ plaques are associated with increased neuritic dystrophy and axonal spheroids.^91,92 To investigate the impact of microglial Adgrg1 deletion on neuritic integrity, we performed MAP2^93,94 staining in 6-month-old mouse brains and observed significantly shorter dendrites when Adgrg1 was deleted (Figures 5A and 5B, S6J and S6K). LAMP1 staining, which marks amyloid plaque-associated axonal spheroids,^92,95,96 showed increased immunoreactivity in both 5xFAD-icKO and 5xFAD-cKO mice compared to their respective controls (Figures 5C and 5D, S6L and S6M), indicating worsened neuritic dystrophy due to microglial Adgrg1 deletion. At the synaptic level, co-labeling with the pre-synaptic marker vGlut2 and the post-synaptic marker PSD95 revealed a significant reduction in synaptic density in the cortical peri-plaque regions of 5xFAD-icKO mice and 5xFAD-cKO mice compared to their respective controls (Figures 5E and 5F, S6N and S6O). Taken together, these findings suggest that microglial ADGRG1 is crucial for maintaining neuronal health and synaptic integrity in the context of AD pathology.

Figure 5. Loss of ADGRG1 in microglia impaired neuritic integrity, synaptic density, and spatial learning and memory.

(A) Representative images of MAP2 staining in the cortices of 6-month-old 5xFAD; Adgrg1^+/+; P2ry12^CreER/+ and 5xFAD; Adgrg1^fl/fl; P2ry12^CreER/+ mice. Scale bar, 50 μm.

(B) Quantification of MAP2⁺ dendritic length in the cortices of 6-month-old mice. * p=0.0438. Unpaired two-tailed t-test.

(C) Representative images of MOAB2 and LAMP1 staining in 6-month-old 5xFAD; Adgrg1^+/+; P2ry12^CreER/+ and 5xFAD; Adgrg1^fl/fl; P2ry12^CreER/+ mice. Scale bar, 50 μm.

(D) Percentage of cortical LAMP1⁺ coverage area. * p=0.0450. Unpaired two-tailed t-test.

(E) Representative images of ThioS (blue), vGlut2 (green) and PSD95 (red) in the cortices of 6-month-old 5xFAD; Adgrg1^+/+; P2ry12^CreER/+ and 5xFAD; Adgrg1^fl/fl; P2ry12^CreER/+ mice. Box indicating the zoomed-in view of overlapped vGlut2 and PSD95 as synapses. Scale bar, 10 μm.

(F) Synapse density quantification. * p=0.0178. Unpaired two-tailed t-test.

(G) Schematic of behavioral experimental design on 4-month-old mice. Schematic created with BioRender.com.

(H) Rotarod test on 4-month-old mice for each genotype, shown as the maximum time that mice remained on the accelerating rotating rod. * p=0.0499, ** p=0.0050 (Adgrg1^fl/fl; Cx3cr1^Cre/+ vs. 5xFAD; Adgrg1^+/+; Cx3cr1^Cre/+), ** p=0.0061 (Adgrg1^fl/fl; Cx3cr1^Cre/+ vs. 5xFAD; Adgrg1^fl/fl; Cx3cr1^Cre/+). One-way ANOVA with Bonferroni’s multiple comparisons test.

(I) 24-hour probe of Morris water maze test on 4-month-old mice for each genotype. Heat map images represent weighted occupancy across the entire 60-second trial. Hot colors indicate longer dwell times. The platform area is marked with a square. The star represents the drop/pick up point.

(J) Percentage of time in target quadrant without submerged platform or in other quadrants. *** p=0.0001 (Adgrg1^+/+; Cx3cr1^Cre/+), *** p=0.0003 (Adgrg1^fl/fl; Cx3cr1^Cre/+), *** p=0.0009 (5xFAD; Adgrg1^+/+; Cx3cr1^Cre/+). Two-way ANOVA with Bonferroni’s multiple comparisons test.

Sample size of behavioral tests: 4-month-old Adgrg1^+/+; Cx3cr1^Cre/+ (n=29), Adgrg1^fl/fl; Cx3cr1^Cre/+ (n=34), 5xFAD; Adgrg1^+/+; Cx3cr1^Cre/+ (n=33), and 5xFAD; Adgrg1^fl/fl; Cx3cr1^Cre/+(n=31).

Each circle represents one animal. Filled circles, male mice; open circles, female mice. Data are presented as mean ± SEM.

See also Figure S6.

Deletion of Adgrg1 in microglia accelerates memory deficits in 5xFAD mice.

5xFAD mice manifest cognitive deficits at 6 months of age.^36,97,98 Based on the worsened pathology associated with microglial Adgrg1 deletion and the similar phenotypes between the constitutive and tamoxifen-inducible knockout lines, we performed a series of behavioral tests in 4-month-old Con, cKO, 5xFAD-Con, and 5xFAD-cKO mice (Figure 5G). We observed reduced body weight in male 5xFAD-cKO mice compared to Con mice (Figures S6P and S6Q). Motor learning and function, as assessed by the rotarod test, were unaffected by Adgrg1 deletion (Figure 5H). Consistent with the literature,⁹⁹ 5xFAD mice exhibited longer latencies to fall from the rotarod than WT mice (Figure 5H).

Spatial learning and memory were evaluated using the Morris water maze. During the memory recall session without a submerged platform, 4-month-old 5xFAD-cKO mice spent approximately equal amounts of time in the target and non-target quadrants, indicating impaired memory retention. In contrast, mice from other genotypes spent significantly more time in the target quadrant, demonstrating intact spatial memory (Figures 5I and 5J). Consistent with this result, plotting the absolute time spent in the target quadrant showed a similar trend of reduction in 5xFAD-cKO mice compared to other genotypes (Figure S6R). Furthermore, 5xFAD-cKO mice exhibited a notably higher swimming speed compared to cKO mice, suggesting motor hyperactivity (Figure S6S). However, the average distance from the platform remained consistent across genotypes (Figure S6T). Taken together, these behavioral tests support that, while sparing body weight and motor skills, microglial Adgrg1 deletion accelerates the onset and significantly compromises spatial learning and memory capabilities in 5xFAD mice.

ADGRG1 is highly expressed in individuals with mild cognitive impairment and positively correlates with phagocytosis-related genes in human microglia.

Our findings thus far demonstrate that microglial ADGRG1 is critical for limiting Aβ plaque burden and maintaining cognitive function in AD models. Previous snRNA-seq analyses have identified ADGRG1 as one of the top five upregulated genes in microglia from individuals with early-stage AD compared to those with late-stage AD and non-pathology controls,³⁵ raising the possibility that individuals who upregulate microglial ADGRG1 may exhibit resilience and survive to advanced age with mild AD pathology. To correlate our mouse model findings with human AD, we re-analyzed this dataset and confirmed that microglial ADGRG1 expression is higher in early-stage AD individuals compared to those without AD or late-stage AD (Figure S7A). We further evaluated microglial ADGRG1 protein levels in human brain tissues (Table S4). We performed immunostaining for Iba1, CG4 (a monoclonal antibody against ADGRG1),¹⁰⁰ and the Aβ antibody H31L21¹⁰¹ on human middle temporal gyrus tissue sections, a brain region vulnerable to AD pathological changes.^102–106 Microglia from individuals with mild cognitive impairment due to AD (MCI, AD Thal phase ≤4, Braak stage ≤4, AD neurologic change low to intermediate) exhibited prominent CG4-positive signals compared to those from AD patients, suggesting a potential resilient function of microglial ADGRG1 in individuals with MCI (Figures 6A and 6B).

Figure 6. ADGRG1 expression and correlation with phagocytosis-related genes in human microglia.

(A) Representative images of human temporal gyrus sections stained for Iba1 (red), CG4 (ADGRG1, green), H31L21 (Aβ, gray), and DAPI (blue) in aged-matched no pathology, mild-cognitive impairment (MCI) and AD subjects. Scale bar, 10 μm.

(B) Quantification of percentage of Iba1⁺ CG4⁺ microglia. Each circle represents an individual subject. * p=0.0159. One-way ANOVA with Bonferroni’s multiple comparison. Filled circles, male; open circles, female.

(C) Histogram of Pearson correlation coefficient demonstrating the linear relationship between microglia ADGRG1 expression and other microglial genes.

Data are presented as mean ± SEM.

See also Figure S7 and Table S4.

Given the lack of natural loss-of-function mutations in ADGRG1 in existing human AD samples, we conducted a transcriptomic meta-analysis across 590 human brain tissues using published snRNA-seq, scRNA-seq^{35,43,107,108} and the Religious Orders Study and Rush Memory and Aging Project (ROSMAP) studies.¹⁰⁹ We found that microglial ADGRG1 expression positively correlated with the expression of phagocytosis-related genes, such as CD68, CTSD, and CTSB, and negatively correlated with PPARD and ITSN1, aligning with findings from our mouse models (Figure 6C). Additionally, analysis of GWAS data on AD risk genes (GWASdb) revealed that ADGRG1 is among the top 300 significantly altered genes in AD.¹¹⁰ Comparing microglial DEGs from our snRNA-seq data with known human AD risk genes, we found that 24 AD risk genes overlapped with DEGs upregulated in 5xFAD-Con microglia, whereas 97 risk genes overlapped with those upregulated in 5xFAD-cKO, suggesting a substantial increase in the overlap of DEGs with known AD risk genes due to ADGRG1 deficiency (Figure S7B). Taken together, these data demonstrate a linkage between microglial ADGRG1 and phagocytosis related genes and AD risk genes in humans, indicating ADGRG1’s potential role in modulating disease resilience via microglial functional properties.

Discussion

In this study, we uncover a critical role for the adhesion G protein-coupled receptor ADGRG1 in mediating protective microglial state changes in AD pathology. Using the 5xFAD mouse model, we demonstrate that microglial ADGRG1 is essential for effective microglial responses to Aβ deposition. Specifically, ADGRG1 deficiency compromises the ability of microglia to associate with and phagocytose Aβ, leading to increased amyloid plaque burden in the cortex and hippocampus, exacerbated neuronal and synaptic loss, and accelerated cognitive decline.

We selected the 5xFAD model for our investigations due to its aggressive phenotype, which allows for rapid assessment of microglial functions in AD.³⁶ Moreover, the abundance of publicly available transcriptomic data from this model, including studies on Trem2 deletion by Dr. Colonna and colleagues⁴³, provides a valuable framework for comparative analyses. This context enabled us to delineate ADGRG1-specific functions in mediating microglial responses to Aβ and to compare these with the roles of other key microglial receptors.

Microglia are highly heterogeneous immune cells in the brain and display remarkable plasticity in response to environmental cues during development, disease, and aging. They adapt their functions through distinct phenotypic states driven by unique and sometimes overlapping transcriptomic profiles.^25,111 During neurodegenerative disease conditions such as AD, microglia transition into DAM states by first downregulating homeostatic genes like Cx3cr1, P2ry12, and Tmem119 in stage 1 DAM, and subsequently upregulating genes such as Tyrobp, Apoe, and Trem2 in stage 2 DAM to enhance their phagocytic capacity under TREM2-SYK regulation.^13,112–114 Additionally, axon tract microglia (ATM)¹¹⁵ and proliferative region-associated microglia (PAM)¹⁷ are observed during development, playing roles in axonal maintenance, and supporting neurogenesis and oligodendrocyte differentiation, respectively. Our findings reveal that microglial Adgrg1 deletion in 5xFAD mice leads to a unique transcriptomic profile characterized by a downregulation of both homeostatic and phagocytic genes, indicating a dysfunctional microglial state that exacerbates Aβ deposition and neurodegeneration. Deletion of microglial Adgrg1 leads to a significant reduction in the microglial subcluster cluster 0, marked by homeostatic microglia with enhanced phagocytic function. Unlike Trem2 deletions, where microglia fail to fully transition from homeostatic to DAM states,¹³ Adgrg1 deletion shifts microglial function without affecting DAM progression, but alters the transcriptomic profile within this state. Importantly, both DAM clusters, cluster 3 and 4, participate in lipid metabolism but are functionally distinct. Cluster 4, which is reduced in 5xFAD-cKO mice, is enriched for genes involved in lipid uptake and lysosomal lipid degradation (Lpl, Apoe, Lrp1, Hexa, Hexb), whereas cluster 3, expanded upon Adgrg1 deletion, express genes associated with fatty acid activation, cholesterol efflux, and lipid-sensing transcriptional regulation (Acsl3, Acsl4, Acsl6, Abca1, Npc1, Ppard, Nr1h2, Nr1h3). Notably, the human lipid-processing microglia state (MG4) identified by Sun et al. showed the most significantly positive correlation with tangles, amyloid, Braak score, and cognitive decline in humans²⁶. The loss of cluster 4 in our dataset provides a mechanistic link to our observed increase in amyloid deposition and cognitive deficit in 5xFAD-cKO mice. Taken together, our study results advance our understanding of how microglial states are modulated through a GPCR.

Our study also sheds light on the dual functionality of microglial ADGRG1 across different stages of life. During development, inactive synapses externalize phosphatidylserine (PS) on the presynaptic terminal, facilitating microglial ADGRG1-mediated synaptic pruning. Deleting microglial Adgrg1 during this stage results in an accumulation of excess synapses.³⁴ Conversely, in 5xFAD mice, specifically around the peri-plaque regions, we observed a reduction in synaptic density following microglial Adgrg1 knockout, likely attributable to the neurotoxic effects of Aβ plaques in 5xFAD mice. In both 5xFAD-cKO and 5xFAD-icKO brains, a higher amyloid burden is observed due to impaired microglial function, which results in worsened neuritic dystrophy. It is reported that Aβ-related hyperactive synapses externalize PS, resulting in synaptic loss in AD mouse models.^116–118

A key finding of our study is the identification of the transcription factor MYC as a downstream effector of ADGRG1 signaling in microglia. Our results support the idea that the activation of the transcription factor MYC is crucial for microglial ADGRG1-mediated phagocytosis of Aβ, expanding our understanding of MYC beyond its well-established roles in cell proliferation and cancer biology. The phosphorylation of MYC at Serine 62 is a key regulatory step that promotes its nuclear translocation and transcriptional activity.^70–72 Despite extensive research on MYC in cancer cells, there are few studies exploring its role in brain immune cells. Recent work has shown that MYC mediates early-phase microglia proliferation following nerve injury,¹¹⁹ while in macrophages, MYC modulates polarization and metabolic pathways, maintaining the balance between pro- and anti-inflammatory states for efficient phagocytosis.¹²⁰ However, its specific function in microglial biology, particularly in the context of neurodegenerative diseases, has remained largely unexplored. Our results provide the first in vitro and in vivo evidence that MYC activation regulates microglial Aβ phagocytosis through an ADGRG1-dependent mechanism. This discovery opens new avenues for exploring the broader implications of MYC in microglial function and ADGRG1-MYC signaling pathways to modulate microglial states therapeutically.

In summary, our results reveal that microglial ADGRG1 is a critical mediator of protective microglial state changes in influencing microglial responses in AD, providing new insights into how a specific GPCR regulates transcription factors, ultimately shaping microglial behavior and contributing to the protective response in AD pathology. Future research needs to explore whether enhancing ADGRG1 expression or activity in microglia can serve as a therapeutic strategy to promote protective microglial states in AD. By manipulating ADGRG1 expression or its downstream signaling pathways, one can potentially guide microglia towards a protective state in AD.

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Xianhua Piao (Xianhua.piao@ucsf.edu).

Material availability

This study did not generate new unique reagents.

Data and code availability

Original single-nuclei RNA-seq data is deposited at GEO: GSE273690. No original or custom code was developed for this study. All analyses were performed using publicly available software packages, as detailed in the Methods section. Additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

STAR METHODS

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Mice

All mice were handled according to the guideline of the Institutional Animal Care and Use Committee at the University of California, San Francisco (IAUCU# AN194756–01F). The Adgrg1^fl/fl mice were generated in-house and have been reported previously.^34,78 To confirm the presence of loxP sites, the following primers were used: (1) 5’-GGT GAC TTT GGT GTT CTG CAC GAC-3’, (2) 5’-TGG TAG CTA ACC TAC TCC AGG AGC-3’ and (3) 5’-CAC GAG ACT AGT GAG ACG TGC TAC-3’. The Cx3cr1^Cre/+ mice were acquired from the Jackson Laboratory (strain number: 025524). For detecting Cre expression, these primers were employed: (1) 5’-GCA GGG AAA TCT GAT GA AG-3’; (2) 5’-GAC ATT TGC CTT GCT GGA C-3’ and (3) 5’-CCT CAG TGT GAC GGA GAC AG-3’. 5xFAD mice were obtained from the Jackson Laboratory (strain number: 034848), and their genotype was determined using these primers: (1) 5’-CGG GCC TCT TCG CTA TTA C-3’; (2) 5’-ACC CCC ATG TCA GAG TTC CT-3’ and (3) 5’-TAT ACA ACC TTG GGG GAT GG-3’. The P2ry12^CreER/+; Ai14 mice were obtained from Dr. Thomas Arnold at UCSF.⁸⁴ To detect P2ry12^CreER expression, these primers were employed: (1) 5’-AAG AAG GTG GCG AAC CAA G-3’; (2) 5’-CAC CCTG CAG ACT AAG ATT TTT CC-3’ and (3) 5’-GTT CAG CAG GGA ACC ATT TC-3’. To induce CreER activity, one-month-old animals (postnatal day 31, P31) were intraperitoneal injected once daily with 100 mg/kg tamoxifen dissolved in corn oil at a concentration of 20 mg/mL for five consecutive days. All P2ry12^CreER/+; Ai14 mice carry the Ai14 reporter allele, which enables tdTomato (RFP) expression upon Cre-mediated recombination.

For our experimental design, Adgrg1^+/+; Cx3cr1^Cre/+ mice were designated as Con group, Adgrg1^fl/fl; Cx3cr1^Cre/+ mice as cKO mice; 5xFAD; Adgrg1^+/+; Cx3cr1^Cre/+ mice as 5xFAD-Con and 5xFAD; Adgrg1^fl/fl; Cx3cr1^Cre/+ mice as 5xFAD-cKO. Tamoxifen-induced Adgrg1^+/+; P2ry12^CreER/+; Ai14 mice as iCon; 5xFAD; Adgrg1^+/+; P2ry12^CreER/+; Ai14 mice as 5xFAD-iCon; 5xFAD; Adgrg1^fl/fl; P2ry12^CreER/+; Ai14 mice as 5xFAD-icKO mice.

Mice were analyzed at 4 months of age for behavioral testing, 6 months of age for histological and snRNA-seq experiments, and 8–9 month of age for intracerebroventricular injection. Male and female mice were used in approximately equal numbers for all experiments unless otherwise noted. No sex-dependent differences were observed, as indicated in the figures by filled circles (males) and open circles (females).

Generation and maintenance of WT and ADGRG1 KO human embryonic stem cells

Experiments using hESC were conducted in accordance with UCSF guidelines and regulations. The hESC line WA09/H9 was obtained from Dr. Arnold Kriegstein’s (UCSF). All hESC were expanded on growth factor-reduced Matrigel-coated plates. hESC were grown in StemFlex Pro Media supplemented with 10 μM ROCK inhibitor Y-27632 for the first day, which was removed the following day if there were at least eight cells per colony for most colonies. Once this criterion was reached, Rock inhibitor was removed. Media were changed every other day, and cell lines were passaged when colonies reached 80% confluency. Stem cells were passaged using ACCUTASE^™ solution (Millipore, SCR005). All lines used for this study were below passage 30.

hESC were edited as previously described with modifications^123,124. In brief, ADGRG1 KO was generated using CRISPR/Cas9-based non-homology end joining, largely following the protocol of Alt-RTM CRISPR-Cas9 System from Integrated DNA Technologies (IDT). The guide RNA (gRNA) sequence (5’-ACACTCTTCCAGAGGACGAA-3’) was selected from Predesigned Alt-R CRISPR-Cas9 gRNA (IDT), targeting exon 7 of the ADGRG1 gene locus. Equal amount of crRNA and ATTO^™550- labeled tracrRNA (IDT, 1075927) were mixed to a final concentration of 100 μM, heated to 95 °C for 5 minutes, and then cooled to room temperature to anneal, followed by forming the RNP complex with Alt-R^™ S.p. HiFi Cas9 Nuclease V3 (IDT,1081061) at room temperature for 20 minutes. The RNP complex was delivered to single stem cell suspension using the Neon^™ electroporation system (1500V, 20ms, 1 pulse) according to the manufacturer’s instructions. After electroporation, ATTO^™550+ cells were selected by FACS after three days of culture and sparsely seeded to form single-cell colonies. A loss-of-function mutation cell line 4D3 was selected by Sanger sequencing, followed by exclusion of any mutations at the top 5 potential off-target sites. Further Sanger sequencing and immunostaining were applied to confirm ADGRG1 KO.

Generation of hESC-derived microglia

Cells were differentiated into iMG following the protocols of the STEMdiff^™ Hematopoietic Kit (05310, STEMCELL), STEMdiff^™ Microglia Differentiation Kit (100–0019, STEMCELL), and STEMdiff^™ Microglia Maturation Kit (100–0020, STEMCELL). On day 28 of maturation, cells were collected and seeded at a density of 20,000 cells per well in PLL-coated 96-well plates and 80,000 cells on PLL-coated coverslips in 24-well plates. Cells were incubated overnight in microglia maturation medium and used for experiments the following day (day 29 of maturation).

Purification of primary microglia

Primary microglia were isolated from mixed glial cultures derived from the cortices of postnatal day 3 to 5 (P3-P5) 5xFAD; Adgrg1^fl/fl; Cx3cr1-Cre^+/− and 5xFAD; Adgrg1^+/+; Cx3cr1-Cre^+/− mice. The cortical tissues were dissected in Hanks’ Buffer Saline Solution (HBSS) and the meninges were carefully removed. Tissues were finely minced on ice using a blade and then transferred into 3 mL of HBSS. After allowing the tissue fragments to settle at the bottom of the tube, the supernatant was gently removed. The resulting cell pellets were transferred in 0.01% poly-L-Lysine (PLL)-coated T75 flasks containing 10 mL of mixed glial culture medium composed of high-glucose DMEM (Cytiva, SH30243.01), supplemented with 20% heat-inactivated fetal bovine serum (FBS; Thermo Fisher, A31604–02), 1% Pen/strep (Thermo Fisher, 15070063), and 20 μg/mL GM-CSF (PeproTech, 315–03). On the third day in vitro (DIV3), the culture medium was completely replaced to remove non-adherent cells and debris. On DIV10, microglia were flushed off the astrocyte layer by shaking at 200 rpm for 2 hours within a standard tissue culture incubator. Subsequently, the detached microglia were collected and centrifuged at 800 rpm for 5 minutes at 4°C. The purified microglia were then seeded at a density of 20,000 cells per well in PLL-coated 96-well plates and 80,000 cells on PLL-coated coverslips in 24-well plates. Cells were incubated overnight in microglia culture medium and used for experiments the following day (DIV 11).

The microglia were cultured in microglia culture medium consisting of DMEM/F12 (Millipore Sigma, D8437), supplemented with 1% Pen/strep, 2mM L-glutamine (Thermo Fisher, 25030081), and 5 μg/mL N-Acetyl Cysteine (NAC; Millipore Sigma, A7250), 5 μg/mL Insulin (Millipore Sigma, I0516), 100 μg/mL Apo-transferrin (Millipore Sigma, T2252), 100 ng/mL Sodium Selenite (Millipore Sigma, S5261), 2 ng/mL TGF-beta (PeproTech, 100–35B), 100 ng/mL IL-34 (R&D Systems, 5195) and 1.5 μg/mL Cholesterol (Millipore sigma, C3045). Microglia were incubated overnight in this medium before being subjected to subsequent treatments.

Human brain tissue samples

Postmortem human brain samples were obtained from the Neurodegenerative Disease Brain Bank at the University of California, San Francisco, under approved ethical guidelines. Formalin-fixed paraffin-embedded sections from middle temporal gyrus from seven individuals diagnosed with AD, six individuals with MCI, and six neurologically healthy control subjects were analyzed by immunostaining. Detailed information regarding the age, sex, postmortem interval (PMI), and other relevant clinical data of the tissue donors are documented in Table S4. Both male and female brain samples were included across diagnostic groups. No sex-associated differences were observed in the quantified outcome, as also indicated in the figures using filled circles for males and open circles for females.

METHOD DETAILS

Intracerebroventricular injection (ICV)

Mice were anesthetized with isoflurane delivered via a vaporizer. Following a midline scalp incision to expose the skull, the lateral ventricle was targeted bilaterally using the following stereotaxic coordinates: anteroposterior (AP) −0.4mm from bregma, mediolateral (ML) ±1.0 mm from the midline, and dorsoventral (DV) −2.2mm from the dura. Two injection sites were prepared, one above each lateral ventricle. A total dose of 0.1mg/kg of 10058-F4 (1mg/ml) was divided equally between hemispheres (0.05 mg/kg per side) and infused using a Hamilton syringe at a rate of 1 μl/minute. The 10058-F4 solution was prepared according to the manufacture’s protocol: 10% DMSO, 40% PEG300, 5% Tween 80, 45% saline (v/v), added sequentially. The solution was sonicated for 5 minutes before injection to ensure complete dissolution.

Mouse brain sample preparation

The brain samples were collected from mice following a standardized and ethically approved procedure. The mice were anesthetized using an isoflurane chamber and subsequently underwent transcardial perfusion with 20 mL of ice-cold 1x PBS. The right hemisphere of each brain was dissected into cortex, hippocampus, mid-brain and cerebellum. These tissues were immediately snap-frozen in dry ice and then stored at −80°C. The left hemisphere was fixed in 4% paraformaldehyde (PFA; Thermo Fisher, J19943-K2) for 48 hours at 4°C. Following fixation, the tissue was cryoprotected by immersion in 30% sucrose in PBS until it sank. The cryoprotected tissue was then embedded in O.C.T. (SCIgen, #4586). The embedded brain tissue was sectioned at thicknesses of either 14 μm or 40 μm using a cryostat (Leica) for the subsequent immunohistochemistry.

Immunofluorescence staining

For the 14-μm brain sections, antigen retrieval was performed for 5 minutes at 95°C using an antigen retrieval buffer (BD Pharmingen, 550524). Subsequently, sections were blocked in a blocking buffer containing 5% goat serum and 1% bovine serum albumin (BSA), and 0.03% Triton X-100 in PBS at room temperature for 1 hour. Slices were incubated with primary antibodies overnight at 4°C. The antibodies used were as follows: For amyloid plaque labeling, anti-Aβ (MOAB2, 1:1000; or H31L21, 1:1000). To quantify microglial cells, anti-Iba1 (1:500). For neuronal health assessment, anti-NeuN (1:500), anti-CTIP2 (1:1000), anti-TBR1 (1:1000), anti-MAP2 (1:2000), anti-LAMP1 (1:200), anti-PSD95 (1:500), anti-vGlut2 (1:500). For enhanced permeability in anti-ADGRG1 (CG4, 1:200) staining, sections underwent a 15-minute treatment with Protease III (ACD) at 40°C before blocking. The following secondary antibodies were applied at a 1:500 dilution at room temperature for 2 hours: goat anti-chicken IgY Alexa Flour 555, goat anti-rat IgG Alexa Flour 555, goat anti-mouse IgG Alexa Flour 488, goat anti-mouse IgG Alexa Flour 647, goat anti-rabbit IgG Alexa Flour 488, goat anti-rabbit IgG Alexa Flour 647, goat anti-guinea pig IgG Alexa Flour 488, goat anti-guinea pig IgG Alexa Flour 555. DAPI (1 μg/mL) was used for nuclear counterstaining, followed by mounting with Fluoromount-G mounting medium. For human brain sections, to reduce autofluorescence, slides were treated with TrueBlack Lipofuscin Autofluorescence Quencher at a 1:20 dilution in 70% ethanol for 30 seconds prior to mounting. Additionally, Thioflavin S staining was performed post-antibody labeling. Briefly, slides were stained with 1% Thioflavin S for 10 minutes, followed by washing in 70% ethanol for 5 minutes once and twice in water for 5 minutes. For analyzing of microglial engulfment of Aβ, 40μm free-floating sections were used. Post-incubation in 10% Triton X-100 in PBS for 30 minutes and subsequent blocking, sections were incubated with primary antibodies (anti-Aβ, anti-Iba1, anti-CD68) for 40–48 hours at 4°C and secondary antibodies were incubated for 2 hours at room temperature. Quantitative analysis of Aβ engulfment by microglia or CD68 co-localization was conducted using 3D rendering of confocal images in Imaris 9.8.0 software.

For immunocytochemistry, primary microglia seeded on coverslips were treated with Protease III at 40°C for 15 minutes for CG4 staining. Post-blocking, cells were incubated with anti-CD11b (1:200) and anti-CG4 (1:200) overnight at 4°C. Secondary antibodies were applied at a 1:500 ratio and incubated at room temperature for 1 hour. Coverslips were mounted using DAPI-Fluoromount-G.

Both mouse and human sections were imaged using a Keyence BZ-X800 microscope with a 20x objective for whole section scans and Leica SP5 confocal microscope with 20x or 63x objectives. The imaging resolution was set to 2048 × 2048 pixels, resulting in a pixel size of 550 μm × 550 μm for 20x and 174.60 μm × 174.60 μm for 63x. Z-stacks of 10 μm were acquired with a 0.5 μm z-step size, using sequential scans with 3x averaging at 488, 555, and 647 nm wavelengths and 1x averaging at 405 nm wavelength. Experiments were blinded to the genotype during image acquisition and processing.

RNAscope in situ hybridization

The RNAscope in situ hybridization was employed on fixed, frozen mouse brain tissue samples using the Multiplex Fluorescence v2 kit (Advanced Cell Diagnostics) according to the manufacturer’s protocol. Probe for mouse-Gpr56 were commercially available from the manufacturer. TrueBlack Lipofuscin Autofluorescence Quencher was applied to the sections prior to mounting.

In vivo Aβ phagocytosis assay

In vivo Aβ phagocytosis assay was adapted from a previous study with modification.¹²⁵ 6-month-old 5xFAD; Adgrg1^fl/fl; Cx3cr1-Cre^+/− and 5xFAD; Adgrg1^+/+; Cx3cr1-Cre^+/− mice were intraperitoneally injected with 10 mg/kg of methoxy-X04 in a solution of 10% DMSO, 45% propylene glycol, and 45% PBS. Three hours later, mice were anesthetized using an isoflurane chamber and transcardially perfused with 20 mL of ice-cold 1x PBS. The cortices, dissected on ice, were homogenized in 1 mL of 1xPBS, then centrifuged at 800g for 5 minutes at 4°C. The resulting cell pellet was resuspended in 10 mL of 30% Percoll and overlaid with 2 mL of 1x PBS. This was centrifuged at 800g for 15 minutes at 4°C to separate myelin. After removing myelin, cells were washed in 1 mL of FACS buffer (1% BSA in PBS) and pelleted at 300g for 5 minutes at 4°C. The cells were blocked with FACS buffer containing CD16/CD32 (1:100) for 20 minutes on ice, then incubated with DRAQ7 (1:200), and CD11b-PE (1:50) in FACS buffer for 20 minutes on ice. After a final centrifugation, cells were resuspended in FACS buffer and analyzed by flow cytometry to gate CD11b⁺ microglia and assess methoxy-X04 signal within the gated microglia population.

pHrodo-oligomeric Aβ preparation

Oligomeric Aβ was prepared following a previously published protocol.¹²⁶ Briefly, HFIP pre-treated Aβ was dissolved in DMSO and further diluted in phenol-red free F-12 medium to obtain a final concentration of 100 μM. The samples were sonicated for 10 minutes, followed by a 24-hour incubation at 4°C. The samples were centrifuged at 14,000g for 10 minutes. The resultant supernatant was aliquoted and stored at −80°C. Oligomeric Aβ was then incubated with pHrodo Red- Succinimidyl Ester diluted in 0.1M sodium bicarbonate for 30 minutes covered at room temperature. Following incubation, HBSS was used to wash the samples and remove excess dye. pHrodo Red-labelled Aβ was used immediately in subsequent experiments.

In vitro phagocytosis assay

Microglial cells, both mouse primary microglia (DIV10) and hESC-derived microglia (day 28 of maturation), were seeded in clear bottom black 96-well plates (Corning, 07–200-565). 24 hours later after seeding, the cells were treated with 1 μM of pHrodo Red-labeled oligomeric Aβ. Following the treatment, microglia were immediately subjected to live-cell imaging using the Incucyte S3 Live-Cell Analysis System (Sartorius). Phase contrast and RFP images were captured for each well at hourly intervals over a 24-hour period. For each time point, the integrated intensity of red objects was quantified. This quantification was performed using Red Calibrated Unit (RCU) measurements, multiplied by the area (μm²) of recorded image.

Western blot

Primary microglia culture medium was aspirated and cells were washed once in ice-cold 1xPBS and lysed in radioimmunoprecipitation assay buffer (RIPA buffer, Merck Millipore) supplemented with EDTA-free protease inhibitor and phosphor-stop cocktail (Roche, 1 tablet per 10 mL RIPA extraction buffer). Lysates (25 μg protein) were subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The gel-separated proteins were then transferred to PVDF membrane (Bio-Rad, Hercules, CA, USA). The membranes were blocked in TBS containing 0.1% Tween-20 and 5% skimmed milk or BSA. The blots were rinsed, primary antibodies applied and incubated overnight in room temperature. After rinsing, horseradish peroxidase-conjugated secondary antibodies were added for 1 hour in room temperature followed by a final rinse. Antibody-stained bands were visualized with luminol-based enhanced chemiluminescence HRP substrate (Bio-RAD). The intensity of bands was quantified utilizing ImageJ.

General design of behavioral tests

Behavioral tests were conducted on 4-month-old male and female mice, housed in groups of two to five mice per cage. All procedures were carried out at the UCSF Gladstone Behavioral Core by professional technicians. The genotypes of all mice were blinded to both experiment conductors and data analyzers.

Rotarod

The rotarod test was performed on a rotarod apparatus (Med Associates Inc., Vermont, USA) under normal lighting conditions. The rotarod apparatus was cleaned with Vimoba before and after each use, and sanitized with 70% alcohol between individual trials. During training sessions, groups of up to five mice, segregated by sex, were simultaneously placed on the rotarod. The rotarod was set to rotate at a constant speed of 16 rotations per minute (rpm). A trial was deemed complete when a mouse fell off the rod or after a duration of 5 minutes had elapsed. Each mouse underwent 3 individual trials with an inter-trial interval of 15–20 minutes. For the actual test, the rotarod speed was progressively increased from 4 rpm to 40 rpm, with an increment of 4 rpm every 30 seconds. The mice were subjected to two sessions of three trials each, conducted in the morning and afternoon over two consecutive days.

Morris water maze

The Morris water maze consisted of a circular pool (122 cm diameter, 50 cm height) filled with water maintained at 22± 1°C. The water was made opaque using non-toxic white tempera liquid paint to obscure the submerged platform. Mice underwent hidden platform training over 5 days, with 2 sessions comprising 2 trials each day. The submerged platform (1.5 cm × 1.5 cm) was kept in a consistent location throughout the training period, located 1.5 cm below the water surface and without visible cues. Mice were given a maximum of 60 seconds to locate the platform in each trial and were required to stay on the platform for at least 10–15 seconds before being removed. The starting positions for each trial were randomized, including both “closer” and “further” drop locations relative to the platform. 24 hours following the final day of training, probe trials were conducted with the platform removed. The mice’s swim patterns were recorded for 60 seconds using the Etho VisionXT video tracking system (Noldus, Netherlands). For each trial, mice were introduced into the pool at a position 180° opposite the former platform location. After the 60-second trial, the experimenter places his/her hand where the hidden platform used to be and lets the mouse swim to his/her hand.

Isolation of nuclei from fresh frozen mouse tissues

Flash-frozen mouse neocortex brain tissues were minced on dry ice and transferred to a Dounce homogenizer. The tissues were homogenized in 1.5 mL of Nuclei-Pure Lysis Buffer (Sigma) supplemented with 0.2 U/μL RNase inhibitor (Takara) on ice. The homogenate was filtered through a 40 μm cell strainer to remove large debris and centrifuged at 600g for 5 minutes at 4°C to pellet the nuclei. The pelleted nuclei were washed thrice with 1 mL of nuclei wash buffer, composed of 1% BSA in PBS, 20 mM DTT, and RNase inhibitor, with centrifugation between each wash. The washed nuclei were resuspended in 0.04% BSA in PBS containing RNase inhibitor and incubated with anti-NeuN-Alexa 488 conjugated antibody (1:500) for 40 minutes on ice. Subsequently, the nuclei were stained with DAPI (1:100) for 5 minutes on ice. Nuclei were washed and passed through a 35 μm cell strainer cap of Falcon round-bottom polystyrene test tubes. Using flow cytometry, DAPI⁺ and DAPI⁺ NeuN⁻ nuclei were gated and collected from each animal. The nuclei were then pelleted again at 600g for 5 minutes at 4°C and resuspended in 0.04% BSA in PBS. The nuclei were counted for downstream library preparation procedures.

Single-nucleus RNA sequencing

Following nuclei isolation, the snRNA-seq library was prepared using Chromium Next GEM Single Cell 3′ Reagent Kits v3.1 according to the manufacturer’s protocol (10x Genomics). The generated snRNA-seq libraries were sequenced using NovaSeq 6000 aiming for 50K read pairs per nuclei. Sequencing results were aligned to the mm10 mouse genome (GENCODE vM23/Ensembl 98) using CellRanger v6.1.2 (10x Genomics). Include-introns was used to include pre-mature mRNA in the nucleus. The count matrix then underwent doublet removal step using DoubltFinder v2.0.3 (https://www.cell.com/cell-systems/fulltext/S2405-4712(19)30073-0) accompanied with Seurat v4.1.1 (https://www.sciencedirect.com/science/article/pii/S0092867421005833?via%3Dihub). Nuclei with 500–6000 genes detected, 1000–40,000 unique molecular identifiers (UMIs), and less than 2% reads mapped to the mitochondrial genes were kept for further analysis. Clustering and cell-types markers were identified using Seurat v4.1.1.

Clustering and finding markers

Fast mutual nearest neighbour (MNN) correction was applied for batch effect removal across samples.¹²⁷ The MNN-corrected dimensional reduction output was directly used for clustering and visualization. Specifically, the top 30 MNN components were used with FindNeighbors ( ), FindClusters ( ) with a resolution of 0.1, and RunUMAP ( ) functions in Seurat. Differential gene expression was performed for each cluster against all others to identify both negative and positive markers. For subclustering analysis, nuclei identified as microglia were extracted and re-clustered using the top 7 principle components to construct a shared nearest neighbor graphs. Clustering was then performed at a resolution of 1.

Analysis of gene differential expression

Differential gene expression analysis was performed using the FindMarkers ( ) function in the Seurat package. Gene expression between conditions was compared using a built-in implementation of the Student’s t-test, and p-values were adjusted for multiple testing using false discovery rate (FDR) correction. Genes with a log₂ fold change greater than 0.25 and p-value < 0.05 were used for downstream gene ontology (GO) and transcription factor (TF) enrichment analysis via Enrichr (https://maayanlab.cloud/Enrichr/).

Projecting Trem2 dataset to microglia subcluster UMAP

To compare microglia subclustering in the Trem2 dataset, we projected all microglia (3713 cells) from the Trem2 dataset onto our microglial subclustering using ProjectUMAP ( ) function in Seurat.

Human ADGRG1 expression correlation analysis

To perform transcriptomic meta-analysis, original count matrix of 590 human brain samples were collected from published dataset.^{35,43,107–109} These microglial transcriptomes are pooled by sample, and Pearson’s correlation coefficient is calculated between the expression level of all genes and ADGRG1. The correlation coefficient is then plotted as a histogram to show the distribution. Important genes in the proposed pathway are labeled along with the number of genes in that bin on the histogram.

QUANTIFICATION AND STATISTICAL ANALYSIS

Mean values, SEM values, Student’s t test, one-way ANOVA, and two-way ANOVA were calculated using Prism 10.0 software (GraphPad). p value less than 0.05 were considered significant. ns=not significant, *p < 0.05, ** p< 0.01, *** p<0.001, ****p< 0.0001. Exact n value and sample sizes are reported in the figure legends. Data points representing individual animals are shown in plots, with filled circles indicating males and open circles indicating females. For in vitro assays, n denotes the number of biologically independent replicates.

Supplementary Material

2

Table S1. Full list of marker genes in each brain myeloid cell clusters, related to Figure 1.

3

Table S2. Full list of differential expressed genes in microglia between 5xFAD-Con and 5xFAD-cKO, related to Figure 2.

4

Table S3. Full list of significant transcription factor identified from the ChEA database, related to Figure 3.

5

Table S4. Clinical and Pathological Data on the Human Individuals Profiled with Immunostaining in the Study, related to Figure 6.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-human GPR56 (mouse monoclonal, clone CG4)	BioLegend	Cat #358202 RRID: AB_2562082
Anti-Iba1 (guinea pig monoclonal)	Synaptic Systems	Cat #234308 RRID: AB_2924932
Anti-β Amyloid (rabbit monoclonal, clone H31L21)	Thermo Scientific	Cat #700254 RRID: AB_2532306
Anti-β Amyloid (mouse monoclonal, clone MOAB2)	NOVUS Biologicals	Cat #NBP2-13075 RRID: N/A
Anti-MAP2 (chicken polyclonal)	Abcam	Cat #ab5392 RRID: AB_2138153
Anti-LAMP1 (rabbit monoclonal)	Abcam	Cat #ab208943 RRID: AB_2923327
Anti-NeuN (guinea pig polyclonal)	Millipore Sigma	Cat # ABN90 RRID: AB_11205592
Anti-NeuN, Alexa Flour 488 (rabbit monoclonal)	Abcam	Cat #ab190195 RRID: AB_2716282
Anti-CD68 antibody (mouse monoclonal, clone KP1)	Abcam	Cat #ab955 RRID: AB_307338
Anti-CD11b, PE (rat monoclonal, clone M1/70)	BD Biosciences	Cat #553311 RRID: AB_394775
DRAQ7	Abcam	Cat #ab109202 RRID: N/A
Anti-CTIP2 (rat monoclonal, clone 25B6)	Abcam	Cat #ab18465 RRID: AB_2064130
Anti-TBR1 (rabbit polyclonal)	Abcam	Cat #ab31940 RRID: AB_2200219
Anti-PSD95 (rabbit polyclonal)	Abcam	Cat # ab18258 RRID: AB_444362
Anti-vGlut2 (guinea pig polyclonal)	Millipore Sigma	Cat # AB2251 RRID: AB_1587626
Anti-RFP (rabbit polyclonal)	Rockland	Cat #600-401-379 RRID: AB_2209751
Anti-CD11b antibody (rat monoclonal, clone M1/70)	Abcam	Cat #ab8878 RRID: AB_306831
Anti-Olig2 (mouse monoclonal, clone 211F1.1)	Millipore Sigma	Cat #MABN50 RRID: AB_10807410
Anti-ALDH1L1 (rabbit polyclonal)	Abcam	Cat #ab87117 RRID: AB_10712968
Anti-MYCpS62 (rabbit monoclonal)	Abcam	Cat # ab185656 RRID: AB_2935659
Anti-GAPDH (mouse monoclonal, clone 6C5)	Abcam	Cat # ab8245 RRID: AB_2107448
Anti-CLEC7A (rat monoclonal, clone R1-8g7)	InvivoGen	Cat #mabg-mdect-2 RRID: AB_2753143
Anti-CD206 (rat monoclonal, clone MR5D3)	BioRad	Cat #MCA2235 RRID: AB_324622
Goat anti-mouse IgG (H+L) Antibody, Alexa Flour 488 Conjugated	Thermo Fisher	Cat #A11029 RRID: AB_2534088
Goat anti-mouse IgG (H+L) Antibody, Alexa Flour 647 Conjugated	Thermo Fisher	Cat #A32728 RRID: AB_2633277
Goat anti-guinea pig IgG (H+L) Antibody, Alexa Flour 488 Conjugated	Thermo Fisher	Cat #A11073 RRID: AB_2534117
Goat anti-guinea pig IgG (H+L) Antibody, Alexa Flour 555 Conjugated	Abcam	Cat #ab150186 RRID: N/A
Goat anti-rabbit IgG (H+L) Antibody, Alexa Flour 488 Conjugated	Thermo Fisher	Cat #A32731 RRID: AB_2633280
Goat anti-rabbit IgG (H+L) Antibody, Alexa Flour 647 Conjugated	Thermo Fisher	Cat #A32733 RRID: AB_2633282
Goat anti-Chicken IgY (H+L) Antibody, Alexa Fluor 555 Conjugated	Thermo Fisher	Cat #A32932 RRID: AB_2762844
Goat anti-rat IgG (H+L) Antibody, Alexa Flour 555 Conjugated	Thermo Fisher	Cat #A21434 RRID: AB_141733
Chemicals, peptides, and recombinant proteins
Thioflavin S	Millipore Sigma	Cat #T1892
pHrodo Red, succinimidyl ester	Thermo Fisher	Cat #P36600
Methoxy-X04	R&D Systems	Cat #492010
Beta-Amyloid (1–42), HFIP	rPeptide	Cat #A-1163-1
Nuclei PURE Lysis Buffer	Millipore Sigma	Cat #L9286
Scigen Tissue-Plus O.C.T Compound	Fisher Scientific	Cat #23-730-571
Sucrose	Millipore Sigma	Cat #S7903
Goat serum	Fisher Scientific	Cat #16-210-064
Triton X-100	Millipore Sigma	Cat #X100
Tween 20	Research Products International	Cat #P20370-0.5
Percoll	Millipore Sigma	Cat #P4937
TrueBlack Lipofuscin Autofluorescence Quencher	Biotium	Cat #23007
DAPI	Thermo Fisher	Cat #D1306
DAPI-Fluoromount-G	SouthernBiotech	Cat #0100-20
Fluoromount-G	SouthernBiotech	Cat #0100-01
Hanks’ Buffer Saline Solution (HBSS)	Thermo Fisher	Cat #14025092
PBS, pH 7.4	Thermo Fisher	Cat #10010049
HyClone DMEM	Cytiva	Cat #SH30243.01
DMEM/F-12, HEPES	Millipore Sigma	Cat #D8437
Neurobasal Medium	Thermo Fisher	Cat #21103049
Fetal Bovine Serum (FBS)	Thermo Fisher	Cat #A31604-02
GlutaMAX Supplement	Thermo Fisher	Cat #35050061
B-27 Plus Supplement	Thermo Fisher	Cat #A3582801
Paraformaldehyde solution, 4% in PBS	Thermo Fisher	Cat #J19943.K2
BD Retrievagen Antigen Retrieval A	Fisher Scientific	Cat #BDB550524
Penicillin-Streptomycin	Thermo Fisher	Cat #15070063
Poly-L-Lysine	Millipore Sigma	Cat #P4832
GM-CSF	PeproTech	Cat #315-03
L-Glutamine	Thermo Fisher	Cat #25030081
N-Acetyl-L-cysteine (NAC)	Millipore Sigma	Cat #A7250
Insulin	Millipore Sigma	Cat #I0516
apo-Transferrin human	Millipore Sigma	Cat #T2252
Sodium selenite	Millipore Sigma	Cat #S5261
Human TGF-beta 2 Recombinant Protein	PeproTech	Cat #100-35B
Recombinant Mouse IL-34 Protein	R&D systems	Cat #5195
Cholesterol	Millipore Sigma	Cat #C3045
Recombinant RNase Inhibitor	Takara	Cat #2313B
10058-F4	MCE	Cat #HY-12702
PEG300	MCE	Cat #HY-Y0873
Tween 80	MCE	Cat #HY-Y1891
Critical commercial assays
RNAscope Multiplex Fluorescent detection Kit v2	Advanced Cell Diagnostics	Cat #323110
RNAscope Pretreatment reagents for sample permealization using Hydrogen peroxide, Protease Plus, Protease III, and Protease IV	Advanced Cell Diagnostics	Cat #322381
RNAscope Pretreatment reagents for sample permealization using Target retrieval	Advanced Cell Diagnostics	Cat #322000
RNAscope Wash buffer	Advanced Cell Diagnostics	Cat #310091
RNAscope Probe-Mm-Gpr56	Advanced Cell Diagnostics	Cat #318241
TSA Plus Cyanine 5 System	Akoya Biosciences	Cat #NEL745001KT
Chromium Next GEM Single Cell 3’ Kit v3.1	10x Genomics	Cat #1000268
STEMdiff™ Hematopoietic Kit	STEMCELL	Cat #05310
STEMdiff™ Microglia Differentiation Kit	STEMCELL	Cat #100-0019
STEMdiff™ Microglia Maturation Kit	STEMCELL	Cat #100-0020
Deposited data
Single-nucleus RNA sequencing of neocortex from 6-month-old Adgrg1+/+; Cx3cr1Cre/+, Adgrg1fl/fl; Cx3cr1Cre/+, 5xFAD; Adgrg1+/+; Cx3cr1Cre/+, and 5xFAD;Adgrg1fl/fl; Cx3cr1Cre/+ mice	This paper	GeneExpression Omnibus (GEO): GSE273690
Experimental models: Organisms/strains
5xFAD mice: B6.Cg-Tg(APPSwFlLon,PSEN1*M146L*L286V)6799Vas/M mjax	The Jackson Laboratory	Cat #034848-JAX; RRID: MMRRC_034848-JAX
Adgrg1fl/fl	Giera et al., 201578	N/A
Cx3Cr1-cre (B6J.B6N(Cg)-Cx3cr1tm1.1(cre)Jung/J	The Jackson Laboratory	Cat #025524; RRID: IMSR_JAX:025524
B6(129S6)-P2ry12em1(icre/ERT2)Tda/J	McKinsey et al., 202084	N/A
Software and algorithms
GraphPad Prism 10	GraphPad	RRID: SCR_002798
Imaris 9.8.0	Imaris	RRID: SCR_007370
Fiji/ImageJ software	Fiji	RRID: SCR_002285
BioRender	BioRender	RRID: SCR_018361
FlowJo	BD Biosciences	https://www.flowjo.com
R Studio	Posit Software	https://posit.co/download/rstudio-desktop/
R 4.3.2	R project	https://www.r-project.org
Cellranger (6.1.2)	10x Genomics	N/A
DoubltFinder (2.0.3)	McGinnis et al., 2019121	RRID: SCR_018771
Seurat (4.1.1)	Seurat v4122	RRID: SCR_016341
Other
Leica SP5 Confocal Microscope	Leica Microsystems	N/A
Keyence BZ-X800 All-in-One Fluorescence Microscope	KEYENCE	N/A

Highlights:

• ADGRG1 activates MYC to induce a protective microglial state in Alzheimer’s disease.

• MYC activation upregulates genes involved in phagocytosis and lysosomal activity.

• Loss of microglial ADGRG1 impairs Aβ clearance and exacerbates AD pathology.