Temporal biphasic regulation of photoreceptor degeneration by microglial TREM2: A metabolic-immune nexus in retinitis pigmentosa

Rong Li; Jing Zhang; Jiang-Mei Wu; Jun-Qi Fan; Bin Lin

doi:10.1126/sciadv.adw9299

. 2025 Sep 12;11(37):eadw9299. doi: 10.1126/sciadv.adw9299

Temporal biphasic regulation of photoreceptor degeneration by microglial TREM2: A metabolic-immune nexus in retinitis pigmentosa

Rong Li ^1,^*, Jing Zhang ², Jiang-Mei Wu ², Jun-Qi Fan ², Bin Lin ^1,^2,^3,^*

PMCID: PMC13141908 PMID: 40938987

Abstract

Retinitis pigmentosa (RP), the leading cause of inherited blindness, lacks therapies because of undefined photoreceptor degeneration mechanisms. While microglia/myeloid cells drive RP progression, their phenotype-regulating determinants remain unclear. Using rd10 mice, we reveal TREM2 as a biphasic RP regulator via STAT2–mediated microglial reprogramming. Early TREM2 loss amplifies neuroinflammation through STAT2 hyperactivation, while late deficiency triggers NF-κB/STAT2–driven microglial apoptosis, impairing phagocytosis yet preserving photoreceptors. We uncover a photoreceptor-microglia metabolic axis where apoptotic photoreceptors release arachidonic acid, salicylic acid, and creatinine to induce STAT2-dependent apoptosis in TREM2-deficient cells. Crucially, we identify intermicroglia PSAP/GPR37 signaling as a self-propagating apoptotic mechanism—the first evidence of apoptotic transmission in retinal degeneration. This study establishes three advances: (i) TREM2 exhibits stage-dependent neuroprotective/neurotoxic roles, (ii) photoreceptor metabolites dictate microglia/myeloid cell fate via STAT2, and (iii) microglial apoptosis spreads through membrane signaling complexes. Our findings redefine neuroimmune dynamics in retinal degeneration, propose chronotherapeutic TREM2 targeting, and extend to Alzheimer’s and other microglia-associated CNS disorders via metabolic-immune interplay.

TREM2 deficiency elicits early-stage pathogenesis and late-stage protection by NF-κB/STAT2–driven microglial apoptosis in RP.

INTRODUCTION

Retinitis pigmentosa (RP), a genetically heterogeneous group of neurodegenerative disorders affecting 1:4000 individuals worldwide, represents the most common inherited cause of irreversible blindness (1). While more than 80 causative genes have been identified (2, 3), the unified pathophysiological mechanisms driving photoreceptor demise—initially rods then cones—remain elusive, leaving patients without effective disease-modifying therapies. This knowledge gap persists despite decades of research, underscoring the urgent need to identify conserved regulatory nodes controlling degeneration progression.

Emerging evidence implicates neuroimmune dysregulation as a cardinal feature of RP pathogenesis. Microglia, the retina’s resident macrophages, undergo pathological activation early in disease progression, releasing cytotoxic cytokines that accelerate photoreceptor apoptosis (4–8). While anti-inflammatory interventions show preclinical efficacy (5–8), critical questions remain unresolved: What molecular switches control microglia phenotypic transitions? How do neuron-glia interactions evolve across disease stages? Answering these questions is essential for developing stage-specific therapies.

The triggering receptor expressed on myeloid cells 2 (TREM2), a transmembrane receptor orchestrating immunometabolic regulation in microglia/myeloid cell, has emerged as a critical regulator of neuroimmune homeostasis in the central nervous system (CNS). Unlike receptors with narrower roles in isolated processes (e.g., cytokine sensing or phagocytic signaling), TREM2 integrates amyloid-β clearance, lipid metabolism, microglial survival, and disease-specific immune responses in the CNS (9–12). Critically, TREM2 loss-of-function variants are strongly linked to Alzheimer’s disease risk in human genetics, underscoring its nonredundant role in neuroprotection (13, 14). Mechanistic studies reveal that TREM2 dysfunction exacerbates neurodegenerative progression in both Alzheimer’s and Parkinson’s diseases through impaired clearance of apoptotic debris and dysregulated neuroimmune cross-talk (9–18). Emerging evidence extends TREM2’s regulatory significance to retinal pathophysiology, where it modulates microglial dynamics during degenerative processes. Our current understanding positions TREM2 as a master regulator governing three critical aspects of retinal microglial function: activation state transitions, chemotactic navigation toward injured photoreceptors, and production of paracrine survival factors (19–21). Notably, TREM2 deficiency exacerbates retinal pathology by impairing microglial capacity to contain immune cell infiltration post–retinal detachment (19) and compromises the neurosupportive phenotype of microglia necessary for photoreceptor preservation during degeneration (20, 21). These findings collectively establish TREM2 as a gatekeeper of retinal neural integrity through spatiotemporal immunomodulation. Despite these advances, critical knowledge gaps hinder therapeutic translation: (i) no systematic analysis of TREM2’s temporal regulation across RP stages, (ii) unknown mechanisms linking TREM2 signaling to photoreceptor-derived signals, and (iii) uncharacterized cell-cell communication networks in microglia/myeloid cell populations. Here, we resolve these questions through longitudinal investigation in rd10 RP models, revealing an evolutionary conserved TREM2 regulatory circuit with dual-phase functionality. We demonstrate that TREM2 deficiency elicits early-stage pathogenesis by signal transducers and activators of transcription 2 (STAT2)–mediated microglia/myeloid cell hyperactivation and neuroinflammation and late-stage protection by nuclear factor κB (NF-κB)/STAT2–driven microglia/myeloid cell apoptosis with phagocytic arrest. Mechanistically, we identify a photoreceptor-microglia/myeloid cell metabolic axis mediated by arachidonic acid (AA)/salicylic acid (SA)/creatinine and PSAP/GPR37-dependent intermicroglia/myeloid cell apoptosis propagation. These findings fundamentally redefine TREM2’s role in neurodegeneration while establishing temporal therapeutic windows for RP intervention—insights with direct relevance to Alzheimer’s disease and related CNS disorders.

RESULTS

Spatiotemporal dynamics of TREM2⁺ microglia/myeloid cell reveal stage-specific involvement in RP

To elucidate TREM2’s role in RP progression, we conducted longitudinal profiling of retinal microglia/myeloid cell in rd10 mice. Quantitative analysis revealed progressive TREM2 up-regulation in both transcriptomic and protein levels coinciding with degenerative progression in rd10 retinas compared with those in age-matched C57BL/6J mice (Fig. 1, A to D, and fig. S1, A and B).

Fig. 1. — (A) Quantitative polymerase chain reaction (qPCR) analysis of *Trem2* expression in the retinas of rd10 and C57BL/6J mice at P16, P19, P22, and P25 (n = 4 to 6 mice per group). (B to D) Western blot analysis [(B) and (C)] and quantification (D) of TREM2 expression in the retinas of rd10 and C57BL/6J mice (n = 4 to 5 mice per group). (E) Representative retinal sections from rd10 and C57BL/6J mice following staining with anti-TREM2 and anti–Iba-1 antibodies. The yellow arrows indicate rMG cells, the yellow arrowheads indicate TREM2⁻ microglia/myeloid cells, and the white arrowheads indicate TREM2⁺ microglia/myeloid cells. Scale bar, 20 μm. ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer. (F and G) Quantification of microglia/myeloid cells (F) and TREM2⁺ and TREM2⁻ microglia/myeloid cells (G) in the retinas of rd10 and C57BL/6J mice (n = 4 to 8 mice per group). The data are presented as the means ± SEM and were analyzed via unpaired two-tailed Student’s t tests (rd10 mice versus C57BL/6J mice: *P < 0.05, **P < 0.01, ***P < 0.001, and ***P < 0.0001).

In the healthy retina, microglia were normally distributed in the outer plexiform layer (OPL), inner plexiform layer (IPL) and ganglion cell layer (GCL) (Fig. 1E). In the retinas of rd10 mice, however, we observed that microglia/myeloid cell became activated, which migrated to the outer nuclear layer (ONL) from the OPL, especially at postnatal day 22 (P22) and P25 (Fig. 1E). Moreover, threefold microglia/myeloid cell density increase peaking at P25 in the retinas of rd10 mice (Fig. 1, E to F). Crucially, we identified a spatiotemporal pattern of TREM2-expressing microglia/myeloid cell (TREM2⁺) dynamics: (i) The number of TREM2⁺ microglia/myeloid cell was significantly increased with age in the retinas of rd10 mice (9.03 ± 9.13%, 20.1 ± 6.7%, 41.2 ± 13.1%, and 53.4 ± 5.18% at P19, P22, P25, and P30, respectively) compared with those in age-matched C57BL/6J mice (3.23 ± 2.32%, 2.85 ± 3.20%, 3.18 ± 4.22%, and 3.67 ± 3.22% at P19, P22, P25, and P30, respectively) (Fig. 1, E and G, and fig. S1, C to E), and (ii) TREM2⁺ microglia/myeloid cells were localized mainly in the OPL and ONL of the retinas of rd10 mice (Fig. 1E), where photoreceptors are located, indicating that TREM2⁺ microglia/myeloid cell may be associated with photoreceptor degeneration. This spatiotemporal correlation between TREM2⁺ microglia/myeloid cell expansion and photoreceptor loss suggests that TREM2 marks disease-associated microglia/myeloid cells (DAM) in retinal degeneration and that microglia/myeloid cell TREM2 induction parallels degenerative progression.

Stage-dependent immunomodulation and neuroprotection by TREM2 ablation in RP

To delineate TREM2’s temporal regulatory roles, we established TREM2^−/−/rd10 mice through genetic crossing and investigated the impact of TREM2 deficiency on microglia/myeloid cell activation and photoreceptor degeneration in the retinas of rd10 mice at P16 (early stage), P19 (initial stage), P22 and P25 (intermediate stage), and P30 (late stage). We found that TREM2 knockout triggered early microglia/myeloid cell activation and inflammation in the retinas of rd10 mice. In TREM2^−/−/rd10 mice, we observed that the number of CD68⁺ microglia/myeloid cell was significantly greater than age-matched rd10 mice at P16 and that the activated microglia/myeloid (aMG) cells extended dendrites into the ONL from the OPL (Fig. 2, A and B, and fig. S2, A and B), indicating that TREM2 deficiency induces early microgliosis. Afterward, the numbers of both microglia/myeloid cell and CD68⁺ microglia/myeloid cell were gradually decreased in the retinas of TREM2^−/−/rd10 mice from P22 (Fig. 2, A and B, and figs. S2, A and B, and fig. S3, A to C), at which point microglia/myeloid cells are overactivated in the retinas of rd10 mice, indicating that TREM2 deficiency inhibits microglia/myeloid cell activation in the intermediate and late stages of RP. Evidently, we found that the expression levels of proinflammatory cytokines (Tnfα, Il1b, and Il6) in the retinas of TREM2^−/−/rd10 mice were significantly increased at P16 (Fig. 2, C to E), but were significantly decreased from P19, when compared with their respective age-matched rd10 controls (Fig. 2, C to E). These results indicate that TREM2 deficiency presents a paradoxical stage-dependent duality in microglia/myeloid cell responses in RP: early phase pathogenesis (P16) and late phase protection (P22 to P30). Our longitudinal analyses (P16 to P30) specifically targeted degeneration-associated microglia/myeloid cell responses in rd10 retinas, bypassing developmental compensation mechanisms. Through multimodal analysis, we confirmed the cell-autonomous role of TREM2 in retinal microglia/myeloid cell regulation and ruled out the potential compensatory inflammation mechanisms (fig. S3, D to H).

Fig. 2. — (A) Retinal sections from TREM2^−/−/rd10, rd10, and C57BL/6J mice following staining with anti-CD68 and anti–Iba-1 antibodies. The white arrowheads indicate the migration of aMG cells into the ONL. Scale bar, 20 μm. (B) Quantification of microglia/myeloid cells in the ONL of the retinas of TREM2^−/−/rd10, rd10, and C57BL/6J mice. (C to E) qPCR analysis of *Tnf*α (C), *Il1b* (D), and *Il6* (E) expression in the retinas of TREM2^−/−/rd10, rd10, and C57BL/6J mice (n = 5 to 6 mice per group). (F) Representative retinal sections from TREM2^−/−/rd10, rd10, and C57BL/6J mice following staining with an anti-red/green opsin (R/G opsin) antibody. Scale bar, 20 μm. (G and H) Quantification of the cone OS/IS length (G) and ONL thickness (H) in retinal sections (n = 4 to 6 mice per group). (I and J) TUNEL staining (I) and quantification of TUNEL⁺ cells in the ONL (J) of the retinas of TREM2^−/−/rd10, rd10, and C57BL/6J mice (n = 4 to 5 mice per group). The white arrowheads indicate TUNEL⁺ cells. Scale bar, 20 μm. (K and L) Representative ERG recordings of TREM2^−/−/rd10, rd10, and C57BL/6J mice at P25 under scotopic (K) and photopic (L) conditions. The data are presented as the means ± SEM and were analyzed via one-way analysis of variance (ANOVA) with Tukey’s multiple comparison test (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ^#P < 0.05, ^##P < 0.01, and ^####P < 0.0001). OS, outer segment; IS, inner segment.

Concurrent with immune modulation, TREM2 deficiency conferred structural and functional preservation in rd10 mice. We observed that the length of the outer segment (OS) and inner segment (IS) of cone photoreceptors and the ONL thickness, where rod photoreceptors are dominated in the mouse retina, were significantly increased in the retinas of TREM2^−/−/rd10 mice compared with the retinas of age-matched rd10 mice at P25 and P30 (Fig. 2, F to H), indicating morphological preservation of rod and cone photoreceptor in the intermediate and late stages. We observed that the number of terminal deoxynucleotidyl transferase (TdT)–mediated deoxyuridine triphosphate (dUTP) nick end labeling (TUNEL⁺) cells in the ONL was decreased in the retinas of TREM2^−/−/rd10 mice compared with those of age-matched rd10 mice from P22 (Fig. 2, I and J), further indicating the preservation of photoreceptors. Functionally, electroretinogram (ERG) recordings showed that a- and b-wave amplitudes were significantly elevated under scotopic conditions (Fig. 2K and fig. S2, C and D) and that b-wave amplitudes were significantly increased under photopic conditions (Fig. 2L and fig. S2, E and F) in P25 TREM2^−/−/rd10 mice compared with age-matched rd10 mice, indicating vision improvement.

Together, these findings suggest that TREM2 exhibits temporal functional inversion—protective early versus detrimental late, and elimination of aMG correlates with neuroprotection in degenerative contexts.

Single-cell transcriptomics reveals TREM2-dependent microglia/myeloid cell reprogramming in RP

To investigate the molecular changes in TREM2⁺ microglia/myeloid cells in rd10 mice following TREM2 deficiency, we performed single-cell RNA sequencing (scRNA-seq) of microglia/myeloid cells from the retinas of TREM2^−/−/rd10 and rd10 mice at P16 (early stage), P25 (intermediate stage), and P30 (late stage). A total of 33,360 microglia/myeloid cells from the retinas of TREM2^−/−/rd10 and rd10 mice were sequenced (fig. S5A). We observed highly dynamic microglia/myeloid cell responses following TREM2 deficiency and defined 12 transcriptional states of microglia/myeloid cells on the basis of molecular signatures and functions (Fig. 3, A and B). We found that microglia/myeloid cells across all clusters strongly expressed the noncanonical microglia/myeloid cells marker genes Cst3, Sparc, and Tmsb4x (22, 23), whereas canonical microglia/myeloid cell marker genes, including Cx3cr1, P2ry12, Tmem119, and Trem2, were uniquely and robustly expressed in a few clusters (fig. S5, B to H).

Usually, most of microglia/myeloid cells in the retinas of rd10 mice at P16, representative of the early stage of RP, are resting microglia/myeloid cells (rMG) (7, 24). Here, we observed that more than 50% of the microglia/myeloid cells in the retinas of rd10 mice at P16 with expression of rMG cell marker genes were in cluster 4 (Fig. 3C and fig. S5, A to D) and defined cluster 4 as rMG_4. In addition, we identified three more rMG clusters, rMG_1, rMG_2, and rMG_3, whose transcriptomic profiles were similar to those of rMG_4 (Fig. 3D). Moreover, we identified one cluster of aMG cells with increased expression of inflammation-associated genes such as Fosb and Timp3 (25–27) (Fig. 3, A and B) and one cluster of proliferated microglia/myeloid (pMG) cells with increased expression of proliferation-associated genes, including Top2a and Pclaf (Fig. 3, A and B) (28–31).

Similarly, we detected six clusters of DAMs, DAM_1, DAM_2, DAM_3, DAM_4, DAM_5, and DAM_6 (Fig. 3, A and B), with high expression of well-known DAM marker genes, such as CD74, Il1a, Tyrobp, Gpnmb, Fabp5, Tnc, and Pf4 (32–35). Moreover, we identified that DAM-associated differentially expressed genes (DEGs), including Trem2, Ctsb, Ctsd, Fth1, Lyz2, Lpl, Spp1, Gpnmb, Cybb, and Fabp5 (32, 33), were significantly up-regulated in all six DAM clusters and pMG (DAMs) (Fig. 3 E). We found that Trem2 was uniquely expressed at a high level in DAMs (Fig. 3A and fig. S5H) but not in any of the four rMG clusters or aMG (non-DAMs) (Fig. 3A and fig. S5H), suggesting that TREM2⁺ microglia/myeloid cells are DAMs.

To investigate the role of TREM2⁺ DAMs in the progression of RP, we performed the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of DEGs between TREM2⁺ DAMs and non-DAMs in rd10 or between TREM2⁻ DAMs and non-DAMs in TREM2^−/−/rd10 retinas, respectively. We found that 251 DEGs in TREM2⁺ DAMs were consistently up-regulated compared to non-DAMs in rd10 retinas (Fig. 3F), and these up-regulated DEGs were primarily associated with mitogen-activated protein kinase (MAPK) signaling pathway, phosphatidylinositol 3-kinase–Akt signaling pathway, chemokine signaling pathway, tumor necrosis factor (TNF) signaling pathway, NF- κB signaling pathway, and phagosome and apoptosis pathways (Fig. 3G). Meanwhile, we identified 564 constantly down-regulated DEGs that were principally involved in the regulation of fatty acid metabolism in TREM2⁺ DAMs in rd10 retinas (Fig. 3, H and I). These results indicate that TREM2⁺ DAMs might contribute to inflammation and photoreceptor apoptosis in RP. Moreover, we observed that the number of consistently up-regulated or down-regulated DEGs that were identified between TREM2⁻ DAMs and non-DAMs in TREM2^−/−/rd10 retinas from P16 to P30 was significantly lower (fig. S5, I and J) than those in the retinas of rd10 mice (Fig. 3, F and H). Inflammation- and phagocytosis-associated genes that were up-regulated in TREM2⁺ DAMs in rd10 retinas (fig. S4, A and B) were significantly down-regulated in TREM2⁻ DAMs in TREM2^−/−/rd10 retinas, especially at P25 and P30 (fig. S4, A and B). Conversely, apoptosis-associated genes that were down-regulated in TREM2⁺ DAMs in rd10 retinas (fig. S4B) were significantly up-regulated in TREM2⁻ DAMs in TREM2^−/−/rd10 retinas (fig. S4B). Together, these results suggest that TREM2 deficiency increases apoptosis but decreases inflammation and phagocytosis in microglia/myeloid cells at the intermediate and late stages of RP.

Furthermore, we identified 25 DEGs that were down-regulated in TREM2⁻ DAMs from TREM2^−/−/rd10 retinas at P16 but up-regulated from P25 to P30 compared with TREM2⁺ DAMs from rd10 retinas (fig. S4C). Meanwhile, we found that five DEGs were significantly up-regulated in TREM2⁻ DAMs from TREM2^−/−/rd10 retinas at P16 but down-regulated at P25 and P30 compared with TREM2⁺ DAMs from rd10 retinas (fig. S4D). KEGG pathway analysis revealed that these consistently up-regulated or down-regulated DEGs between TREM2⁻ DAMs and TREM2⁺ DAMs from P25 to P30 were associated with inflammation and apoptosis, including the chemokine pathway, NOD-like receptor pathway, Janus kinase (JAK)–STAT pathway, and necroptosis pathway (fig. S4E). In particular, we observed that inflammation-related genes were significantly up-regulated in TREM2⁻ DAMs from TREM2^−/−/rd10 retinas at P16 but down-regulated at P25 and P30 compared with TREM2⁺ DAMs from rd10 retinas (fig. S4F). Conversely, apoptosis-associated DEGs were significantly up-regulated in TREM2⁻ DAMs from TREM2^−/−/ rd10 retinas at P25 and P30 compared to those in TREM2⁺ DAMs from rd10 retinas (fig. S4F). These results further suggest that TREM2⁺ microglia/myeloid cells may be involved in regulating inflammation in the early stage of RP but mediating apoptosis in the intermediate and late stages of RP, indicating the potential stage-dependent duality of TREM2 in RP.

Similarly, we analyzed DEGs in non-DAMs between TREM2^−/−/rd10 and rd10 retinas. We observed that 13 DEGs were significantly down-regulated in non-DAMs from TREM2^−/−/rd10 retinas at P16 but up-regulated at P25 and P30 and that 7 DEGs were up-regulated in non-DAMs from TREM2^−/−/rd10 retinas at P16 but down-regulated at P25 and P30 compared with those non-DAMs from rd10 retinas (fig. S5, K and L). These consistently up-regulated or down-regulated DEGs in non-DAMs from P25 to P30 were associated mainly with inflammatory responses and cell apoptosis (fig. S5M). In particular, we found that inflammation-associated genes were significantly up-regulated in non-DAMs from TREM2^−/−/rd10 retinas at P16 but down-regulated at P25 and P30 and that apoptosis-related genes were significantly down-regulated in non-DAMs from TREM2^−/−/rd10 retinas at P16 but up-regulated at P25 and P30 compared to those in non-DAMs from rd10 mice (fig. S5N), indicating the potential impact of TREM2⁺ microglia/myeloid cells on its neighboring microglia/myeloid cells. Collectively, these findings uncovered a TREM2-mediated transcriptional continuum governing microglia/myeloid cell fate determination in RP progression.

TREM2 ablation orchestrates microglia/myeloid cell apoptosis and senescence clearance in RP

To validate the findings of the transcriptomic analysis of TREM2⁺ microglia/myeloid cells, we next investigated microglia/myeloid cell apoptosis in the retinas of rd10 mice following TREM2 deficiency. We found that the number of microglia/myeloid cells was significantly increased at P16 but decreased from P22 to P30 in the retinas of TREM2^−/−/rd10 mice compared to those in age-matched rd10 mice (Fig. 4, A and B). Moreover, we found that the percentages of caspase-3⁺ microglia/myeloid cells in the ONL, OPL, and IPL were significantly greater in the retinas of TREM2^−/−/rd10 mice than those in age-matched rd10 mice from P19 (Fig. 4, A and C to E), indicating that TREM2 deficiency triggers microglia/myeloid cell apoptosis at later time points. Consistently, we performed flow cytometry analysis of microglia/myeloid cell apoptosis and observed that the percentage of annexin-V⁺/PI⁺ cells among CD11b⁺/CD45⁺ microglia/myeloid cells was significantly greater in the retinas of TREM2^−/−/rd10 mice than those in the retinas of age-matched rd10 mice from P22 (Fig. 4, F to I), further indicating the acceleration of microglia/myeloid cell apoptosis following TREM2 deficiency in the intermediate and late stages of RP (Fig. 4J).

Fig. 4. — (A) Retinal whole mounts from TREM2^−/−/rd10, rd10, and C57BL/6J mice were stained with anti-Caspase-3 and anti-CD11b antibodies. The white arrowheads indicate Caspase-3⁺ microglia/myeloid cells. Scale bar, 20 μm. (B and C) Quantification of microglia/myeloid cells (B) and Caspase-3⁺ microglia/myeloid cells (C) in the OPL and IPL of the retinas of TREM2^−/−/rd10, rd10, and C57BL/6J mice (n = 4 to 9 mice per group). (D) Representative retinal sections from TREM2^−/−/rd10, rd10, and C57BL/6J mice stained with anti-Caspase-3 and anti–Iba-1 antibodies. The red arrowheads indicate Caspase-3⁺ microglia/myeloid cells, and the white arrowheads indicate Caspase-3⁻ microglia/myeloid cells in the ONL. Scale bar, 20 μm. (E) Quantification of Caspase-3⁺ microglia/myeloid cells in the ONL of the retinas of TREM2^−/−/rd10, rd10, and C57BL/6J mice (n = 4 to 5 mice per group). (F to I) Flow cytometry analysis (F) and quantification of the percentages of annexin-V⁺/PI⁻ (G), annexin-V⁺/PI⁺ (H), and annexin-V⁻/PI⁺ (I) cells among CD11b⁺/CD45⁻ and CD11b⁺/CD45⁺ microglia/myeloid cells in the retinas of TREM2^−/−/rd10, rd10, and C57BL/6J mice (n = 4 to 8 mice per group). (J) Schematic illustration showing the different alterations in microglia/myeloid cells upon photoreceptor degeneration in the retinas of TREM2^−/−/rd10 and rd10 mice. The data are presented as the means ± SEM and were analyzed via one-way ANOVA with Tukey’s multiple comparison test (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; ^#P < 0.05, ^###P < 0.001, and ^####P < 0.0001).

Moreover, we observed that TREM2⁺ DAMs strongly expressed senescence genes, including Cdkn1a/P21, Itgax, Lgals3, Spp1, and CCL2–5, compared to non-DAMs in the retinas of rd10 mice (fig. S5O), indicating the accumulation of senescent microglia/myeloid cells in the retinas of rd10 mice. Following TREM2 deficiency, however, these senescence genes were significantly down-regulated in TREM2⁻ DAMs in the retinas of TREM2^−/−/rd10 mice (fig. S5P), suggesting a reduction in the number of senescent microglia/myeloid cells in the retinas of rd10 mice following TREM2 deficiency that may result in the attenuation of inflammation in RP.

Collectively, our findings suggest that TREM2 maintains senescence-associated secretory phenotype in DAMs and TREM2-mediated apoptosis-senescence coupling drives inflammatory resolution in RP.

STAT2 emerges as a master regulator of microglia/myeloid cell apoptosis through NF-κB crosstalk following TREM2 ablation in RP

To understand the molecular mechanisms underlying microglia/myeloid cell apoptosis following TREM2 deficiency, we identified the top 20 transcription factors (TFs) among 6000 microglia/myeloid cells from the retinas of TREM2^−/−/rd10 mice and age-matched rd10 mice using SECNIC analysis (Fig. 5A and figs. S6A and S7, A to D). We observed that most of the TFs were associated with inflammation and cell apoptosis pathways, including the TNF signaling pathway, Toll-like receptor signaling pathway, the MAPK signaling pathway, the JAK-STAT signaling pathway, NF- κB signaling pathway, and apoptosis and necroptosis pathways (Fig. 5B).

Fig. 5. — (A) SCENIC analysis of top 20 TF activities in 1000 retinal microglial/myeloid cells from P25 rd10 mice. (B) KEGG pathway analysis of TFs in 6000 retinal microglial/myeloid cells from rd10 and TREM2^−/−/rd10 mice at P16, P25, and P30. (C) Retinal immunohistochemistry for p-STAT2 (red) and Iba-1 (green) in TREM2^−/−/rd10 and rd10 mice. White arrowheads, p-STAT2⁺ microglia/myeloid cells. Scale bar, 20 μm. (D) Quantification of p-STAT2⁺ microglial/myeloid cells in the ONL (n = 5 mice per group). (E) Retinal whole-mount staining for CD68 (red) and Iba-1 (green) in P25 TREM2^−/−/rd10 mice treated with siSTAT2 or siCTR. Scale bar, 20 μm. (F and G) Quantification of total (F) and CD68⁺ (G) microglial/myeloid cells in IPL/OPL of siSTAT2- versus siCTR-treated TREM2^−/−/rd10 mice. (H) Retinal sections stained for cleaved Caspase-3 (red) and Iba-1 (green) in siSTAT2- or siCTR-treated TREM2^−/−/rd10 mice. Scale bar, 20 μm. (I) Caspase-3⁺ microglial/myeloid cell counts (n = 5 mice per group). (J) R/G opsin⁺ cone photoreceptors (red) in retinal sections of treated mice. Insets, high-magnification views. Scale bar, 20 μm. (K and L) Cone OS/IS length (K) and ONL thickness (L) (n = 5 mice per group). (M and N) TUNEL⁺ cells in ONL [quantified in (M); representative images in (N)]. Scale bar, 20 μm. Data are means ± SEM; *P < 0.05, **P < 0.01, ***P < 0.001 (unpaired two-tailed t test; TREM2^−/−/rd10 versus rd10 or siSTAT2 versus siCTR).

Among the TFs, Stat2 was highly enriched in microglia/myeloid cells from both TREM2^−/−/rd10 and rd10 mice at P25 (fig. S6, B to G). In particular, Stat2 was highly expressed in TREM2⁺ DAMs (fig. S6, E to G), and it was significantly increased following TREM2 deficiency in TREM2^−/−/rd10 retinas compared to those in rd10 retinas (fig. S7E). Consistently, we observed that the percentage of phosphorylated STAT2–positive (p-STAT2⁺) microglia/myeloid cells was significantly greater in the retinas of TREM2^−/−/rd10 mice than in the retinas of age-matched rd10 mice (Fig. 5, C and D), indicating that STAT2 may be involved in TREM2-mediated inflammation and microglia/myeloid cell apoptosis in RP.

To confirm the role of STAT2, we knocked down microglia/myeloid cell STAT2 in the retinas of TREM2^−/−/rd10 mice via intravitreal injections of STAT2-specific small interfering RNA (siRNA) (siSTAT2) (fig. S7, F to H). We observed that, compared with control siRNA (siCTR) treatment, knockdown of microglia/myeloid cell STAT2 significantly increased the number of CD68⁺ microglia/myeloid cells (Fig. 5, E to G) and reduced the percentage of caspase-3⁺ microglia/myeloid cells in the retinas of TREM2^−/−/rd10 mice at P25 (Fig. 5, H and I), suggesting that STAT2 knockdown directly increases microglia/myeloid cell activation and inhibits microglia/myeloid cell apoptosis in the retinas of rd10 mice following TREM2 deficiency. Moreover, compared with siCTR treatment, STAT2 knockdown significantly decreased the OS/IS length of cone photoreceptors (Fig. 5, G to K) and ONL thickness (Fig. 5, G and L) and increased the number of TUNEL⁺ cells in the ONL in the retinas of P25 TREM2^−/−/rd10 mice (Fig. 5, M and N), indicating deterioration of photoreceptor cell death. Consequently, STAT2 knockdown significantly decreased b-wave amplitudes in P25 TREM2^−/−/rd10 mice under both scotopic and photopic conditions compared with siCTR treatment (figs. S6, H and I, and S7, I to L), suggesting enhancement of vision decline. Together, these data indicate that STAT2 knockdown aggravates neuroinflammation and subsequently deteriorates photoreceptor degeneration and vision loss in the retinas of rd10 mice following TREM2 deficiency.

To directly verify the role of microglia/myeloid cell STAT2, we transfected primary microglia/myeloid cells isolated from the brains of TREM2^−/− or C57BL/6J mice at 6 weeks of age with siSTAT2 or siCTR and then incubated them with the supernatant of rod cells isolated from the retinas of rd10 (rRCM25) or C57BL/6J (cRCM25) mice at P25. Compared with cRCM25 or phosphate-buffered saline (PBS) treatment, rRCM25 treatment significantly increased the expression of Stat2, Bax and Caspase-3 in microglia/myeloid cells from TREM2^−/− mice (fig. S7, M to O) and the numbers of p-STAT2⁺ microglia/myeloid cells and Caspase-3⁺ microglia/myeloid cells among the microglia/myeloid cells from TREM2^−/− mice (fig. S8, A to C), which is consistent with the findings from our scRNA-seq analysis and in vivo studies showing that TREM2 deficiency increased microglia/myeloid cell apoptosis in the retinas of rd10 mice (fig. S4, C to F, and Fig. 4). However, STAT2 knockdown significantly decreased the expression of apoptosis-associated genes and the number of Caspase-3⁺ microglia/myeloid cells among microglia/myeloid cells from TREM2^−/− mice after rRCM25 treatment (figs. S7, M to O, and S8, A to C), which is also consistent with our in vivo results (Fig. 5, H to I).

To directly evaluate the effect of STAT2 on microglia/myeloid cell phagocytosis following TREM2 deficiency, we transfected microglia/myeloid cells isolated from TREM2^−/− mice with siSTAT2 and cocultured them with rod cells isolated from the retinas of rd10 mice. We found that microglia/myeloid cells following TREM2 deficiency exhibited significantly lower phagocytic activity than microglia/myeloid cells from C57BL/6J mice (fig. S8, D and E), which is consistent with previous observations (9, 36, 37). However, the phagocytic activity of siSTAT2-treated microglia/myeloid cells isolated from TREM2^−/− mice was significantly greater than that of siCTR-treated microglia/myeloid cells (fig. S8, D and E), suggesting that STAT2 knockdown promotes phagocytosis in microglia/myeloid cells following TREM2 deficiency.

Furthermore, by analyzing protein-protein interactions (PPIs) among the apoptosis-associated DEGs between TREM2⁻ DAMs from TREM2^−/−/rd10 retinas and TREM2⁺ DAMs from rd10 retinas, we observed that Stat2 directly interacted with DEGs that were associated with the NF-κB signaling pathway, including Nfκb1, Nfκb2, Relb, Nfκbia, Socs3, Mapk14, and Irak1 (fig. S8F), indicating that the NF-κB signaling pathway may be associated with STAT2-mediated microglia/myeloid cells apoptosis. Strikingly, transcriptional profiling demonstrated marked up-regulation of NF-κB–associated genes in TREM2⁻ DAMs from TREM2^−/−/rd10 retinas at degenerative stages (P25/P30) compared to TREM2⁺ DAMs from rd10 retinas (fig. S8G), mechanistically linking NF-κB signaling to STAT2-mediated microglial apoptosis. This finding aligns with emerging evidence implicating NF-κB–STAT2 cross-talk in survival-apoptosis balance (38–40) but represents the first demonstration of this axis in retinal degeneration.

Quantitative analysis corroborated elevated Stat2 and NF-κB family members (Nfkb1/Nfkb2/Rela/Relb) in TREM2^−/− microglia/myeloid cells versus wild-type controls (fig. S8H). Crucially, pharmacological NF-κB inhibition (BAY 11–7082) suppressed Stat2 expression by 53.6% (P < 0.0001; fig. S8I), establishing NF-κB as a direct upstream regulator of STAT2 in this paradigm. Trem2 deficiency did not alter Jak1/Jak2/Jak3 expression or Ifnar levels (P > 0.05, versus controls; fig. S8H), excluding canonical JAK-interferon involvement in STAT2 activation.

TREM2/STAT2 Axis governs paracrine apoptosis through PSAP/GPR37 signaling in RP

Our scRNA-seq data demonstrated that TREM2 deficiency in DAMs induces apoptosis in non-DAMs microglia/myeloid cells (fig. S5N), indicating potential communications between TREM2⁺ microglia/myeloid cells and its neighboring microglia/myeloid cells. Therefore, we performed CellChat analysis to reveal the direct cell-cell interactions between DAMs (including all six DAM clusters and pMG) and non-DAMs (including all four rMG clusters and aMG) from the retinas of TREM2^−/−/rd10 and rd10 mice (fig. S9A). We identified multiple ligand-receptor pairs mediating the interactions between DAMs and non-DAMs in different clusters, including Psap/Gpr37, Ccl/Ccr, Spp1/Cd44, Tnf/Tnfrsf1a, and Lgals9/Cd45 (figs. S9, A and B; S10, A to D; and S11, A to B). Among them, we found that Psap that was expressed in DAMs strongly interacted with target non-DAMs (Fig. 6A). Moreover, Psap expression was greater in DAMs (Fig. 6B), especially higher in TREM2⁻ DAMs from TREM2^−/−/rd10 retinas than those in TREM2⁺ DAMs from rd10 retinas. On the other hand, we observed that Gpr37, one of classic receptors of Psap, was expressed primarily in non-DAMs (Fig. 6B). These data suggest that the Psap/Gpr37 pair may play a critical role in regulating communication between TREM2⁺ microglia/myeloid cells and its neighboring microglia/myeloid cells.

Fig. 6. — (A) Hierarchical network of PSAP signaling across microglia/myeloid clusters (solid circles, sources; open circles, targets). (B) *Psap* and *Gpr37* expression in microglia/myeloid clusters from TREM2^−/−/rd10 and rd10 retinas. (C and D) PSAP⁺ microglia/myeloid cells (white arrowheads) in retinal sections (C) and their quantification in the ONL [(D) n = 5 to 6 mice per group]. Scale bar, 20 μm. (E and F) PSAP⁺ microglia/myeloid cells in siSTAT2- versus siCTR-treated TREM2^−/−/ rd10 mice [(E) quantification, n = 5 mice per group; (F) representative images]. (G) qPCR of *Psap*, *Stat2*, *Bax*, and *Caspase-3* in TREM2^−/− and C57BL/6J microglia/myeloid cells transfected with siPSAP/siCTR and incubated with rRCM25. (H to M) PSAP⁺ [(H) and (K)], p-STAT2⁺ [(I) and (L)], and Caspase-3⁺ [(J) and (M)] microglia/myeloid cell quantification and localization (arrowheads). (N) qPCR of *Gpr37*, *Bax*, and *Caspase-3* in C57BL/6J microglia/myeloid cells treated with spRCM1/2/3. (O and P) GPR37⁺ cell localization (arrowheads) and quantification. Data are means ± SEM. The results shown represent three to four independent experiments. Statistics: unpaired t test or one-way ANOVA with Tukey’s test (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).

To validate the findings from scRNA-seq analysis, we performed immunohistochemistry on TREM2^−/−/rd10 retinas. We observed that the percentage of PSAP⁺ microglia/myeloid cells in the ONL was significantly greater in the retinas of TREM2^−/−/rd10 mice than in the retinas of age-matched rd10 mice, especially at P25 and P30 (Fig. 6, C and D), indicating the up-regulation of PSAP. However, we found that the number of PSAP⁺ microglia/myeloid cells was significantly reduced in the retinas of TREM2^−/−/rd10 mice that received intravitreal injections of siSTAT2 compared with the retinas of TREM2^−/−/rd10 mice that received intravitreal injections of siCTR (Fig. 6, E and F), indicating that STAT2 regulates PSAP expression in microglia/myeloid cells in the retinas of rd10 mice following TRM2 deficiency.

To confirm the role of PSAP in TREM2/STAT2-regulated microglia/myeloid cell apoptosis, we transfected primary microglia/myeloid cells isolated from TREM2^−/− mice with siPSAP and treated them with rRCM25 in vitro. We observed that PSAP knockdown did not affect Stat2 expression or the number of p-STAT2⁺ microglia/myeloid cells compared with siCTR or rRCM25 treatment (Fig. 6, G to I, K, and L), confirming that PSAP is a downstream molecule of STAT2. Moreover, we found that compared with siCTR or rRCM25 treatment, PSAP knockdown significantly reduced the expression of proapoptosis genes and the number of Caspase-3⁺ microglia/myeloid cells among rRCM25-treated microglia/myeloid cells isolated from TREM2^−/− mice (Fig. 6, G, H, J, K, and M), indicating that PSAP may be involved in regulating TREM2/STAT2-mediated apoptosis in microglia/myeloid cells. On the other hand, we observed that incubation with a mouse recombinant PSAP protein (rPSAP) in microglia/myeloid cells isolated from TREM2^−/− mice did not alter Stat2 expression (fig. S11, C to E) but significantly increased Bax and Caspase-3 expression and the number of Caspase-3⁺ microglia/myeloid cells compared with PBS treatment (fig. S11, C and F to G), further suggesting that PSAP is a downstream molecule of STAT2 and PSAP regulates microglia/myeloid cell apoptosis.

To investigate whether PSAP regulated apoptosis in microglia via Gpr37, we incubated primary microglia/myeloid cells isolated from C57BL/6J mice with the supernatant of cultured microglia/myeloid cells isolated from TREM2^−/− mice after treatment with rPSAP (pRCM2) or PBS (pRCM1). We observed that pRCM2 treatment significantly increased the expression of Gpr37, Bax, and Casspase-3 and the numbers of GPR37⁺ microglia/myeloid cells and Caspase-3⁺ microglia/myeloid cells (fig. S11, H to L). Conversely, when we cultured microglia/myeloid cells from C57BL/6J mice with the supernatants of cultured microglia/myeloid cells isolated from TREM2^−/− mice after treatment with siPSAP and rRCM25 (spRCM3), siCTR and rRCM25 (spRCM2), or rRCM25 (spRCM1) only, we observed that expression levels of Gpr37 and proapoptosis genes and the number of both GPR37⁺ microglia/myeloid cells and Caspase-3⁺ microglia/myeloid cells were significantly decreased in spRCM3-treated microglia/myeloid cells, when compared with spRCM2- or spRCM1-treated microglia/myeloid cells (Fig. 6, N to P, and fig. S9, C and D). Together, these data suggest that PSAP may regulate microglia/myeloid cells apoptosis via GPR37 signal. Moreover, we observed that PSAP knockdown significantly increased the phagocytic activity of microglia/myeloid cells following TREM2 deficiency (fig. S9, E and F), whereas rPSAP treatment significantly reduced the phagocytic activity in microglia/myeloid cells isolated from TREM2^−/− mice (fig. S11, M and N), suggesting the involvement of PSAP in regulating the phagocytic dysfunction of microglia/myeloid cells induced by TREM2 deficiency. Collectively, these results suggest that the TREM2/STAT2 axis governs paracrine apoptosis of microglia/myeloid cells through PSAP/GPR37 signaling in RP.

Photoreceptor-derived metabolites drive microglia/myeloid cell apoptosis via TREM2/STAT2 axis in RP

To elucidate the mechanisms underlying TREM2/STAT2 activity in the context of RP, we performed untargeted metabolomic analysis of the supernatants of cultured rod cells isolated from the retinas of rd10 mice at P16 (P16RCM) and P22 (P22RCM). Untargeted metabolomic analysis with a C18 column revealed 37 differentially enriched metabolites (DEMs), including 22 up-regulated and 15 down-regulated DEMs, in P22RCM versus P16RCM (Fig. 7A). Z-score analysis of the top 20 DEMs revealed that several DEMs, including AA, alpha-linolenic ethanolamide, n-(2-hydroxyethyl) icosa-2,4,6,8, lysoPE (P-16:0/0:0), oleoyl ethanolamide, and SA, were the most significantly enriched metabolites in P22RCM compared with P16RCM (Fig. 7B). Correlation analysis of the top 20 DEMs revealed that the levels of AA, SA, alpha-linolenoyl ethanolamide, n-(2-hydroxyethyl)icosa-2,4,6,8,10-pentaenamide, lysoPE (P-16:0/0:0), and nicotinic acid metabolites were strongly correlated with the levels of other DEMs (Fig. 7C). Differential abundance (DA) analysis of pathways revealed that AA and SA were strongly involved in regulating cell death, phagocytosis, and secretory pathways (Fig. 7D), further suggesting that AA and SA may be critically involved in regulating microglia/myeloid cell dynamics in RP.

Fig. 7. — (A) Heatmap of DEMs in rod-conditioned media from rd10 mouse retinas at P16 (P16RCM) versus P22 (P22RCM) identified by C18-based untargeted metabolomics (n = 6 samples per group). (B) Z-score plot of top 20 DEMs. (C) Correlation matrix of top DEMs (red, positive; blue, negative; intensity reflects correlation strength; *P < 0.05). (D) Pathway analysis showing DA scores for DEM-associated pathways. Dot size indicates metabolite count per pathway. (E to H) qPCR analysis of apoptosis-related genes in TREM2^−/− versus wild-type microglia transfected with siSTAT2/siCTR and treated with AA (50 μM). (I) p-STAT2 localization in microglia (white arrows, p-STAT2⁺/Iba1⁺ cells). Scale bar, 20 μm. (J) Quantification of p-STAT2⁺ microglia/myeloid cells. Data represent means ± SEM from three to four experiments. Statistical analysis: one-way ANOVA with Tukey’s test (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).

Similarly, by untargeted metabolomic analysis with an amide column, we identified 23 DEMs, including 16 up-regulated and 7 down-regulated DEMs, between P22RCM and P16RCM (fig. S13A). Z-score analysis of the top 20 DEMs revealed that creatinine and hypoxanthine were highly enriched in P22RCM (fig. S13B). Moreover, we found that the levels of the metabolites creatinine, d-(−)-glutamine, nicotinate, and tranexamic acid were strongly correlated with those of the other DEMs (fig. S13C). DA analysis of enriched pathways revealed that most metabolites were associated with cellular metabolic processes (fig. S13D). Considering the higher z-score value and the stronger correlation with other DEMs (fig. S13, B and C), creatinine, which was previously reported to regulate immune responses and cell apoptosis (41–45), could be another critical metabolite that is involved in regulating microglia/myeloid cell apoptosis.

To further confirm the roles of AA, SA, and creatinine in microglia/myeloid cell apoptosis in RP, we transfected microglia/myeloid cells isolated from TREM2^−/− or C57BL/6J mice with siSTAT2 or siCTR followed by treatment with AA, SA, or creatinine. Compared with dimethyl sulfoxide (DMSO) or PBS treatment, AA, SA, or creatinine treatment significantly up-regulated the expression of Stat2, Bax, Caspase-3, and Psap in microglia/myeloid cells isolated from both TREM2^−/− and C57BL/6J mice, especially in microglia/myeloid cells isolated from TREM2^−/− mice (Fig. 7, E to H, and figs. S12, A and D, and S13, E to H), suggesting that all of these three metabolites induce microglia/myeloid cell apoptosis. However, following STAT2 knockdown, AA, SA, or creatinine treatment failed to elicit apoptosis-associated gene expression in microglia/myeloid cells isolated from both TREM2^−/− and C57BL/6J mice (Fig. 7, E to H, and figs. S12, A and D, and S13, E to H), suggesting that microglia/myeloid cell apoptosis induced by AA, SA, and creatinine might be STAT2 dependent. Moreover, we observed that AA, SA, or creatinine treatment significantly increased the numbers of p-STAT2⁺ microglia/myeloid cells and Caspase-3⁺ microglia/myeloid cells in microglia/myeloid cells isolated from both TREM2^−/− and C57BL/6J mice, especially in microglia/myeloid cells isolated from TREM2^−/− mice (Fig. 7, I and J, and figs. S12, E to J, and S13, I to L), indicating that AA, SA, and creatinine induce microglia/myeloid cell apoptosis via TREM2/STAT2 signaling.

Intriguingly, photoreceptor-derived conditioned medium (rRCM25) recapitulated these effects, which were attenuated by targeted metabolite suppression via ONO-RS-082 (PLA2 inhibitor, blocking AA release), dorsomorphin (AMPK inhibitor, disrupting SA metabolism), or cyclocreatine (creatine kinase competitor): reduced p-STAT2⁺ microglia by 54.7 to 74.6% and Caspase-3⁺ microglia by 48.2 to 61.2% (fig. S14, A to C). In vivo validation through intravitreal administration of individual metabolite in rd10 mice induced STAT2 hyperphosphorylation in microglia/myeloid cell (0.87- to 1.52-fold increase versus DMSO controls, P < 0.01; fig. S14, D and E). These multimodal findings establish a unique TREM2-STAT2 signaling axis through which photoreceptor-derived metabolites drive microglial apoptosis in RP.

Furthermore, we observed that, compared with DMSO or PBS treatment, AA, SA, or creatinine treatment significantly reduced the phagocytic activity of microglia/myeloid cells isolated from TREM2^−/− mice (figs. S12, K to N, and S14, F and G), which was largely restored following STAT2 knockdown (figs. S12, K to N, and S14, F and G), suggesting the involvement of TREM2/STAT2 signaling in the dysregulation of microglia/myeloid cell phagocytosis induced by AA, SA and creatinine. Collectively, we demonstrate that the metabolites AA, SA, and creatinine released by apoptotic rod photoreceptors in rd10 retinas induce microglia/myeloid cell apoptosis via elevating STAT2 signal following TREM2 deficiency.

Our study uncovers a previously unrecognized biphasic regulatory property of TREM2 in retinal degeneration, challenging conventional neurocentric perspectives on microglial dynamics. Integrated multi-omics analyses demonstrate that TREM2 transitions from a STAT2-mediated anti-inflammatory modulator during early RP to an NF-κB/STAT2–driven proapoptotic accelerator in late stages. This phase-dependent functional inversion contrasts sharply with Alzheimer’s disease paradigms, where TREM2⁺ DAMs exhibit sustained neuroprotective phagocytosis (12, 36), highlighting retina-specific immunoregulatory networks.

DISCUSSION

Our findings delineate a previously unrecognized temporal duality in TREM2 functionality during retinal degeneration, establishing its stage-specific roles as both neuroinflammatory modulator and neurodegenerative accelerator. Through integrated multi-omics and functional validation, we reveal that TREM2 ablation induces early STAT2-mediated microglia/myeloid cell priming followed by NF-κB/STAT2–coordinated apoptotic clearance, ultimately preserving photoreceptor architecture and visual function (Fig. 8). Crucially, we identify photoreceptor-derived metabolites (AA/SA/creatinine) as advanced paracrine regulators that amplify STAT2 signaling through TREM2-dependent mechanisms, creating a self-reinforcing loop driving microglia/myeloid cell dynamics (Fig. 8). The discovery of PSAP/GPR37-mediated intermicroglia/myeloid cell communication further unveils an evolutionary conserved mechanism synchronizing inflammatory resolution across neural tissues (Fig. 8). These paradigm-shifting insights redefine therapeutic strategies for RP, proposing: (i) time-sensitive TREM2 modulation (early enhancement versus late inhibition), (ii) STAT2-targeted neuroprotection, and (iii) metabolic intervention disrupting neurodegeneration-immunometabolic cross-talk. The identified temporal therapeutic window extends beyond ophthalmology, offering mechanistic frameworks for Alzheimer’s and Parkinson’s diseases where microglia/myeloid cell senescence shares conserved regulatory pathways. Our work establishes metabolic-immune axis manipulation as a groundbreaking approach in neurodegenerative therapeutics, bridging molecular mechanisms with clinical-stage intervention strategies.

Fig. 8. — Our study uncovers a temporally bifurcated regulatory axis in which TREM2/STAT2 signaling orchestrates microglia/myeloid cell dynamics and photoreceptor survival during RP progression. In early RP phases, TREM2 deficiency triggers STAT2-driven microglia/myeloid cell hyperactivation, exacerbating neuroinflammation. During intermediate-late stages, amplified STAT2/NF-κB cross-talk induces microglia/myeloid cell apoptosis and phagocytic arrest, paradoxically preserving photoreceptors. Mechanistically, TREM2⁻ microglia/myeloid cells synchronize apoptosis through PSAP/GPR37-mediated intercellular communication, while photoreceptor-derived metabolites (AA/SA/creatinine) fuel STAT2 activation, establishing a self-perpetuated degenerative loop. This work provides three transformative insights: (i) Phase-dependent TREM2 functional inversion redefines therapeutic targeting strategies; (ii) PSAP/GPR37 emerges as a glial apoptotic synchronizer; and (iii) metabolic-immune cross-talk drives neurodegenerative feedback circuits, positioning retinal degeneration as a paradigm for compartmentalized disease progression.

We identify the PSAP/GPR37-mediated paracrine apoptotic cascade as a central coordinator of microglia/myeloid cell homeostasis. Spatial transcriptomics revealed that STAT2 activation triggers PSAP secretion in TREM2⁻ microglia/myeloid cells, inducing GPR37-dependent apoptosis in neighboring cells. This mechanism extends PSAP’s known role in neuronal apoptosis (46, 47) to glial networks, establishing a microenvironmental synchronization framework for neurodegeneration.

Metabolomic profiling identified AA/SA/creatinine as potent STAT2 activators, creating a feedback loop between photoreceptor apoptosis and immune clearance. While these metabolites are established apoptosis regulators in oncology (48–50), their discovery as glial modulators represents a conceptual breakthrough. The STAT2 dependence reveals actionable nodes for therapeutic targeting.

We delineate TREM2⁺ senescent microglia/myeloid cell clearance mechanisms, uncovering retina-specific senescence propagation: (i) photoreceptor metabolite dependency and (ii) PSAP-mediated spatial cascades. These findings position retinal degeneration as an ideal model for studying compartmentalized senescence spread, offering advanced insights into systemic neurodegenerative processes.

Our data advocate a stage-specific therapeutic strategy: early TREM2 enhancement to suppress neuroinflammation versus late-phase inhibition to promote senescence clearance. Repurposing STAT2 inhibitors from oncology trials (51) shows translational promise. Current limitations include (i) unresolved spatiotemporal regulation of STAT2/NF-κB cross-talk, (ii) uncharacterized metabolic flux underlying AA/SA production, and (iii) incomplete human tissue validation. Future priorities include (i) phase-optimized drug delivery systems, (ii) CRISPR-Cas9–mediated spatiotemporal TREM2 modulation, and (iii) metabolic pathway engineering.

Our study unveils stage-specific therapeutic interventions for RP by decoding TREM2’s dual-phase functionality and its interplay with metabolic-immune networks. Time-dependent TREM2 manipulation—early enhancement to suppress STAT2-driven neuroinflammation and late inhibition to disrupt NF-κB/STAT2–mediated neurodegeneration—exploits a critical therapeutic window. This biphasic approach overcomes limitations of static targeting, aligning with disease progression dynamics. Targeting photoreceptor-derived metabolites (AA/SA/creatinine) or microglial receptors (e.g., GPR37) interrupts the neurodegeneration-immunometabolic axis. Small molecules blocking PSAP-GPR37 signaling could prevent intermicroglial senescence and apoptosis propagation, offering a first-in-class strategy to contain spatially degenerative foci. These strategies redefine RP therapeutics by integrating temporal, spatial, and metabolic dimensions. The PSAP/GPR37 pathway—an evolutionarily conserved coordinator of microglial apoptosis—provides a blueprint for Alzheimer’s and Parkinson’s therapies, where analogous microglial senescence mechanisms operate. By prioritizing time-resolved immune modulation and microenvironment-targeted metabolic intervention, our framework bridges mechanistic discovery with clinical translatability, establishing a paradigm shift for neurodegenerative diseases with unmet therapeutic needs.

MATERIALS AND METHODS

Mice

All experimental procedures were approved by the Animal Subjects Ethics Sub-committee (ASESC) of The Hong Kong Polytechnic University. C57BL/6J (stock no: 000664), rd10 (stock no: 004297), and TREM2 knockout (TREM2^−/−; stock no: 027197) mice were acquired from The Jackson Laboratory. Our prior work (52) confirmed that homozygous TREM2^−/− mice exhibit (i) complete absence of TREM2 transcript via RNA sequencing, (ii) no overt photoreceptor degeneration compared to age-matched C57BL/6J controls, and (iii) exacerbated neuroinflammation, as evidenced by elevated proinflammatory cytokine levels and microglial activation. To establish the dual mutant line, rd10 mice carrying the Pde6b mutation were backcrossed with TREM2^−/− mice for >six generations to generate TREM2^−/−/rd10 mice. All animals were maintained under specific pathogen–free conditions at the Centralised Animal Facilities with controlled temperature (22° ± 1°C), humidity (55 ± 5%), and 12:12-hour light-dark cycles, with ad libitum access to autoclaved food and water.

In vivo experiments

Intravitreal siRNA injections

To study the role of STAT2 on microglia/myeloid cell apoptosis and photoreceptor degeneration in rd10 retinas following TREM2 deficiency, we injected intravitreally the STAT2-specific siRNA (Thermo Fisher Scientific, HK) into TREM2^−/−/rd10 mouse eyes as previously described (52, 53). In brief, the mouse was put under a dissecting microscope after anesthesia with a mixture of ketamine hydrochloride (100 mg/kg) and xylazine (50 mg/kg), and an incision in the superior nasal sclera was made using a sterile and sharp 31G needle. One microliter of siRNA solution carried by a glass pipette was slowly injected into the vitreous along the same incision followed by treating the antibiotic eye ointment on the eye surface to prevent infection. One hundred micromolar STAT2 siRNA or negative control siRNA was intravitreally injected to TREM2^−/−/rd10 mouse eyes for three times since P16, 2 days apart. The retina tissues were collected for further analysis at P25. The knockdown efficiency of STAT2 was evaluated by immunofluorescence.

Immunocytochemistry and confocal imaging

After anesthesia with a mixture of ketamine hydrochloride (100 mg/kg) and xylazine (50 mg/kg), the mouse’s eye was enucleated with curved tweezers and the retina was isolated from the vitreous and sclera in PBS followed by fixing the tissue in 4% paraformaldehyde (PFA) for 1 hour before dehydrating in 30% sucrose overnight at 4°C. The retinal sections were made by a cryostat microtome with a serial frozen sectioning at the thickness of 14 μm. Following three washes in PBS, the retinal sections were incubated with rat TREM2 (R&D systems, 1:500), rabbit Iba-1 (Wako, 1:500), goat Iba-1 (Wako, 1:500), rabbit Caspase-3 (Cell Signaling Technology, 1:250), rabbit red/green (R/G) opsin (Millipore, 1:250), rabbit p-STAT2 (ABclonal, 1:250; Invitrogen, 1:250), rabbit PSAP (Proteintech, 1:250), and rat CD68 (Bio-Rad, 1:500) primary antibodies in blocking buffer containing 3% normal donkey serum (NDS), 1% bovine serum albumin (BSA), and 0.3% Triton X-100 in PBS (pH 7.4) overnight. After incubating the retinal sections with secondary antibodies including donkey anti-rabbit Alexa Flour 488 (Invitrogen, 1:500), donkey anti-rabbit Alexa Flour 594 (Invitrogen, 1:500), donkey anti-rat Alexa Fluor 555 (Abcam, 1:500), and anti-goat Alexa Flour 488 (Molecular Probes, 1:500) for 2 hours at room temperature (RT), the slides were mounted with Dako fluorescence mounting medium for further observation. For whole-mounted retinas, rat TREM2 (R&D systems, 1:500), rabbit Iba-1 (Wako, 1:500), goat Iba-1 (Wako, 1:500), rat CD11b (Invitrogen, 1:100), rabbit Caspase-3 (Cell Signaling Technology, 1:250), rabbit glial fibrillary acidic protein (Dako, 1:500), mouse glutamine synthetase (Sigma-Aldrich, 1:500), and rat CD68 (Bio-Rad, 1:500) primary antibodies were incubated for 24 hours followed by incubating the tissues with above secondary antibodies for 2 hours at RT before mounting the tissues in above mounting medium.

Fluorescence images of retinal sections and whole-mounted retinas were captured by a Zeiss LSM 800 Upright Confocal Microscope (Zeiss, USA) with a pixel resolution of 1024 by 1024 and Plan-Apochromat 20×/0.8 objective. Z-stack images with an interval of 0.8 μm was acquired. For the quantification of microglia/myeloid cell density in the ONL of the retina and OS/IS length of R/G opsin–positive cone photoreceptors, three areas at 100 μm (central), 1 mm (middle), and 1.8 mm (peripheral) from the optical nerve head in each retinal section were captured. The OS/IS length of R/G opsin–positive cone photoreceptors in each image was measured by ImageJ software as previously described (5, 54). The microglia/myeloid cell density in the ONL of the retina in each image were manually counted. For quantification of microglia/myeloid cell number in OPL and IPL in the image from the whole-mounted retina, four sampling areas with 638.9 μm–by–638.9 μm squares along the dorsal-ventral axis of retinal whole-mounts at 200 μm and 1 mm from the optic nerve head on both sides were photographed, and the numbers of Iba-1⁺ or CD11b⁺ and of CD68⁺Iba-1⁺, TREM2⁺/Iba-1⁺, or Caspase-3⁺/CD11b⁺ microglia/myeloid cells were manually counted.

TUNEL assay

TUNEL assay was performed as our previously described (5, 52). In brief, the retinal section was fixed in 4% PFA for 15 min at 37°C followed by permeabilizing with proteinase K for 15 min at RT. After second fixation with 4% PFA for 5 min at 37°C and two washes with PBS, the retinal section was incubated with 100 μl TdT reaction buffer for 10 min at 37°C. Afterward, the retinal section was incubated with TdT reaction mixture that was prepared with EdUTP and TdT enzyme in TdT reaction buffer for 1 hours at 37°C, followed by incubating the TUNEL reaction cocktails containing Alexa Fluor 594 picolyl azide and 10X Click-iT Plus TUNEL reaction buffer additive (Invitrogen) for 30 min at 37°C. After several washes in PBS, the retinal section was incubated with 4′,6-diamidino-2-phenylindole for 15 to 20 min and mounted with Dako fluorescence mounting medium. Fluorescence images of retinal sections after TUNEL assay were captured by a Zeiss LSM 800 Upright Confocal Microscope (Zeiss, USA). For quantification of TUNEL-positive cells in the ONL of the retina in each image, three views in each retinal section at 100 μm (central), 1 mm (middle), and 1.8 mm (peripheral) from the optical nerve head were captured, and TUNEL-positive cells were manually counted.

Immunoblotting

The total proteins in fresh retina tissue were extracted with radioimmunoprecipitation assay buffer (Abcam) containing proteinase and phosphatase inhibitor cocktails (Roche) on ice for 30 min followed by centrifugating at 10,000g for 15 min. The supernatant was collected, and the protein concentration was quantified with Pierce rapid gold BCA protein assay kit (Invitrogen). Twenty micrograms of protein was loaded into 10% SDS–polyacrylamide gel electrophoresis gel before running gel at 80 voltages for 90 min. Afterward, the proteins in gel were transferred to a polyvinylidene difluoride membrane. Following blocking with 5% BSA for 1 hour, the membrane containing the proteins was incubated with primary antibodies including rabbit TREM2 (ABclonal, 1:500) and mouse β-actin (Invitrogen, 1:2000) in blocking buffer overnight at 4°C. After three washes with TBST buffer, the membrane was incubated with goat anti-rabbit immunoglobulin G (IgG) and goat anti-mouse IgG (Invitrogen, 1:1000) conjugated to horseradish peroxidase for 2 hours. The membrane with protein was evaluated by ChemiDoc Imaging Systems (Bio-Rad, California, USA) after incubating with SuperSignal West Pico PLUS Chemiluminescent Substrate (Invitrogen) or ECL Select Western blotting detection reagent (AmershamTM). The optical density value of each band was measured using ImageJ software.

Quantitative real-time PCR

The total RNA from fresh retina tissues was extracted with TransZol Up Plus RNA Kit (TransGen, China) or TRIzol Reagent (Invitrogen), and cDNA was synthetized using the PrimeScript RT Master Mix (TaKaRa). Quantitative polymerase chain reaction (qPCR) was performed with TB Green Premix Ex Taq (Tli RNase H Plus) (TaKaRa) using QuantStudio 7 Flex Real-Time PCR System (Applied Biosystems) or QuantStudio 5 Flex Real-Time PCR System (Applied Biosystems). Glyceraldehyde-3-phosphate dehydrogenase was used as a control, and the data were analyzed using 2^-ΔΔCt method. Gene-specific primer used for qPCR are listed in table S1.

Primary microglia/myeloid cell isolation by MACS

Fresh brain tissues from TREM2^−/− and age-matched C57BL/6J mice at 6-week-old were collected in precooled Hanks’ balanced salt solution (HBSS). After several washes in sterile PBS buffer, the brain tissue was dissociated with 500 μl of papain enzyme and deoxyribonuclease (DNase) at 37°C for 30 min followed by centrifugating at 1500 rpm for 5 min. Following one wash with PBS, cell pellets were resuspended with 30% Percoll in HBSS solution and centrifugated at 500g for 15 min at 18°C. The supernatant containing myeline was discarded and cell pellets were resuspended in precooled fluorescence-activated cell sorting (FACS) buffer before incubating cells with anti-CD11b magnetic microbeads (Miltenyi Biotec) for 25 to 30 min at 4°C. After one-time wash, the cells were resuspended in magnetic-activated cell sorting (MACS) buffer and loaded into prebalanced LS column (Miltenyi Biotec), which was fixed in magnetic separator (Miltenyi Biotec). The unlabeled flow-through containing unlabeled cells were discarded, and the LS column containing labeled cells was washed with 6 to 8 ml of MACS buffer. LS column carrying labeled cells were removed from magnetic separator, and magnetically labeled cells were quickly flushed out with 5 to 7 ml of MACS buffer in a 15-ml falcon tube. After centrifugation at 1500 rpm for 5 min, the labeled cells were collected for further analysis. Isolation of primary microglia/myeloid cells from TREM2^−/− and C57BL/6J retinas at P16, P25, and P30 were performed with similar approaches.

Flow cytometry

After enucleation, eyeballs were quickly isolated with curved tweezers and the retina was separated from the vitreous and sclera in precooled PBS. The retina was dissociated with 150 μl of papain enzyme in EBSS solution and 12.5 μl of DNase (Worthington Biochemical Corporation) at 37°C for 30 min. After washing one time, cell samples were incubated with Fc blocking antibody for 10 min on ice before incubating with allophycocyanin-conjugated anti-CD45 (Invitrogen) and phycoerythrin (PE)–conjugated anti-CD11b (Invitrogen) in staining buffer for 30 min on ice. Following rinsing with FACS buffer for one time, the cell samples were incubated with fluorescent isothiocyanate (FITC)-conjugated anti-annexin V (Abcam or MCE) and propidium iodide (PI) (Abcam or MCE) in 1× binding buffer for 10 min before analyzing with BD FACSAria III Cell Sorter (BD Bioscience). All data were analyzed with FowJo V10 software.

Isolation of rod cells

Fresh retinas from rd10 and age-matched C57BL/6J mice at P25 were collected in precooled PBS and digested with papain enzyme and DNase in 37°C for 30 min. After one time wash with PBS, cell pellets were resuspended in FACS buffer and analyzed with BD FACSAria III Cell Sorter (BD Bioscience). The isolation of rod cells was performed as previously described (55). An 80-μm nozzle was used to isolate rod cells, and the standard forward (FSC) and sideward scatter (SSC) settings were used for particle detection. A distinct subpopulation with low FSC and SSC in flow cytometry dot plot could be observed, and which was sorted as pure rod cells. The cell purity of rod cells was verified by immunohistochemistry. Cell viability of cultured rod photoreceptors was evaluated by CCK-8 assay (fig. S15).

ERG analysis

ERGs were performed as previously described by us (52, 56). Briefly, mouse was put on a platform heater at 37°C after anesthesia with a mixture of ketamine hydrochloride (100 mg/kg) and xylazine (50 mg/kg) after dark adaption overnight. The pupils were dilated with 1% mydriacyl (Alcon), and the cornea was kept moist by treating 3% hypromellose lubricating gel solution. ERG recording was performed using a Celeris ERG system (Diagnosys, USA) with TOUCH/TOUCH protocol. A scotopic ERG was detected at 0.01, 0.1, 1, and 3 cd·s/m² light intensities, followed by photopic ERG detection at 3 and 10 cd.s/m² light intensities after 10-min light adaptation with background light intensity at 30 cd·s/m². Ten sweeps were acquired with each light stimulus. The distance between the baseline and the negative peak were measured as the amplitude of ERG a-wave, and the amplitude of b-wave was calculated between the bottom of the a-wave and the top of the tallest curve.

Single-cell RNA sequencing

Sample preparation, library construction, and reads filtering for scRNA-seq

For scRNA-seq on microglia/myeloid cells, retinas from TREM2^−/−/rd10 and age-matched Rd10 mice at P16, P25 and P30 were dissociated individually with papain dissociation system (Worthington Biochemical Corporation) for 30 min at 37°C. After centrifugation at 400g, 4°C for 5 min, the cell pellets were resuspended in 2% BSA for one wash followed by blocking cell samples with Fc blocking antibody (Invitrogen) on ice for 10 to 15 min. Afterward, cell samples were incubated with FITC-conjugated anti-CD45 (Invitrogen) and PE-conjugated anti-CD11b (Invitrogen) in staining buffer for 30 min on ice. After rinsing for one time, the cell samples were analyzed with BD FACSAria III Cell Sorter (BD Biosccence). CD11b⁺/CD45⁺ cells (31) were sorted as microglia/myeloid cells, and cell viability >85% microglia/myeloid cells were prepared for scRNA-seq analysis.

The single-cell libraries were constructed with 10x Genomics platform, and the sequencing was performed using DNBSEQ sequencing platform. The sequencing depth for microglia/myeloid cells isolated from P16, P25, and P30 Rd10 retinas were at 42.8, 42.1, and 31.2 k mean reads per cell, respectively. The sequencing depth for microglia/myeloid cells isolated from P16, P25, and P30 TREM2^−/−/rd10 retinas were at 93.0, 35.5, and 56.5 k mean reads per cell, respectively. The sequencing saturation rates for microglia/myeloid cells isolated from P16, P25, and P30 Rd10 retinas were 79.6, 86.6, and 63.9%, respectively, and the sequencing saturation rates for microglia/myeloid cells isolated from P16, P25, and P30 TREM2^−/−/rd10 retinas were 94.5, 54.4, and 62.7%, respectively. Sample demultiplexing, barcode processing, and single-cell counting were performed using the Cell Ranger (version 5.0.1). RNA reads were aligned with the mouse reference genome (refdata-gex-mm10-2020-A) using STAR alignment before calculating UMI counts. The raw output data were processed and analyzed using the Seurat packages (version 5.0.1) in R software (version 4.3.1) (57). The quality control (QC) was performed by filtering out cells identified with (i) a gene count less than 200, (ii) a gene count greater than the maximum gene count * 90%, and (iii) the top 15% of cells with the highest proportion of mitochondria reads. The correction of cell cycle influences and removal of potential doublets were conducted by DoubletFinder package (version 2.0.2) (58). After QC, a total of 4392, 5942, and 8639 microglia/myeloid cells from P16, P25, and P30 Rd10 retinas, and a total of 1826, 7508, and 5053 microglia/myeloid cells from P16, P25, and P30 TREM2^−/−/rd10 retinas were used for further analysis, respectively.

Data integration, dimensionality reduction, and cell clustering

After normalizing the filtered gene-barcode matrices with the “NormalizedData” function in R using Seurat packages, the top 2000 highly variable genes were identified with “FindVariableFeatures” function. Furthermore, gene expression matrices were scaled and centered using “ScaleData” function followed by performing principal components analysis (PCA) and uniform manifold approximation and projection dimension reduction using the top 20 principal components (PCs). To remove the batch effects across all samples, all data were processed using “FindIntergrationAnchors” and “IntegrateData” functions in Seurat with default parameters. For microglia/myeloid cells scRNA-seq dataset, clustering analysis was performed with “FindNeighbors” (first 20 PCs) and “FindClusters” (resolution = 0.4) functions, and which ultimately yielded 12 cell clusters. The cell annotation was conducted on the basis of the top marker gene expressions in each cluster, which were calculated by the “FindAllMarkers” function.

DEGs identification and pathway enrichment analysis

DEGs analysis was conducted using FindAllMarkers or “FindMarkers” (min.pct = 0.1 and thresh.use = 0.25) functions with Wilcoxon rank sum test. DEGs were identified as |log₂FC| > 0.25 and adjust P value <0.05. The cluster-specific DEGs were used for KEGG pathway enrichment analysis using ShinyGO0.77 (http://bioinformatics.sdstate.edu/go/) (59). The PPI among DEGs was analyzed using STRING (version 12.0) and Cytoscape (version 3.10.1) (60).

Analysis of cell-cell interactions

Cell-cell communications were performed with CellChat (Version 1.6.1) (61). CellChat objects were generated from the Seurat object, and the data were preprocessed using the “subsetData,” “identifyOverExpressedGenes,” and “identifyOverExpressedInterations” functions with default parameters. The “computeCommunProb” function was then used to calculate communication probability and generate a communication network. The analysis was performed using the CellChatDB mouse database. To identify and visualize the cell-cell interaction among different microglia/myeloid cell clusters, we used “netVisual_chord_gene,” “netVisual_individual,” “netVisual_aggregate,” and “netVisual_bubble” functions.

TF module analysis

We used the SCENIC package (version 1.3.1) in R software to infer TFs and gene regulatory networks for microglia/myeloid cell clusters (62). To achieve this, we constructed a gene expression matrix from 6000 cells (1000 microglia/myeloid cells randomly selected from each sample) based on raw counts. The RcisTarget databases “mm9-500 bp-upstream-7species.mc9nr.feather” and “mm9-tss-centered-10 kb-7species.mc9nr.feather,” containing TF motif scores for gene promoters and around transcription start sites for the mouse reference genome, were used for reference. We used “runCorrelation” and “GENIE3” to calculate the correlation of candidate regulators. The downstream pipeline for constructing gene regulatory networks included “runSCENIC_1_coexNetwork2modules,” “runSCENIC_2_createRegulons,” “runSCENIC_3_scoreCells,” and “runSCENIC_4_aucell_binarize.” The projection of the area under the curve (AUC) and TF expression onto t-SNE plots was visualized by “AUCell_plotTSNE.” Cell type–specific regulators were determined on the basis of the Regulon Specificity Score (62).

Untargeted metabolomics analysis

Rod cells from rd10 retinas at P16 and P22 were isolated as described above and seeded into a sterile 12-well plate. Rod cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin in a 37°C incubator for 2 hours followed by changing medium with serum-free culture medium to incubate for 10 to 12 hours. Afterward, the cell supernatants were collected in a sterile EP tube and centrifuged at 1000g, 4°C for 10 min. Next, the supernatants were centrifuged at 25,000g, 4°C for 15 min followed by freezing dry. The dry sample was reconstituted in 50% methanol and transferred to a split-new EP tube for untargeted metabolomics assay. Ten microliters of the reconstituted sample was injected on a Waters ACQUITY UPLC BEH C18 column (1.7 μm, 2.1 mm by 100 mm, Waters, USA) or a BEH Amide column (1.7 μm 2.1 mm by100 mm, Waters, USA). The Waters UPLC I-Class Plus (Waters, USA) random Q Exactive high-resolution mass spectrometer (Thermo Fisher Scientific, USA) was used for separation and detection of metabolites. The metabolite ion peak extraction and metabolite identification were performed by Compound Discoverer 3.3 (Thermo Fisher Scientific, USA) software and analyzed by a combination of bmdb (BGI metabolome database), mzcloud database, and chemspider online database. Afterward, all data files from Compound Discoverer were inputted into MetaX for data preprocessing and further analysis. QC of data was performed by PCA using log transformation and Paretoscaling, and the metabolite functional annotation was performed by the Human Metabolome Database and the KEGG PATHWAY database. The differential metabolites between two groups were identified by partial least squares–discriminant analysis (PLS-DA), orthogonal PLS-DA (OPLS-DA), and variable important for the projection (VIP). The differential metabolites screening criteria was set as 1. VIP of OPLS-DA model ≥1; 2. Fold change ≥1.2 or ≤0.83; 3. P value <0.05. The top 20 differential metabolites were used to do z-score analysis and correlation analysis. The z-score was calculated as: z = (x − μ)/σ, where x is the specific expression of a metabolite in a certain sample, μ as the mean expression of the control group, and σ for the SD of the control group. The correlation between two metabolites was calculated by the Spearman correlation coefficient among all metabolites.

In vitro experiments

siRNA transfection

STAT2- and PSAP-specific siRNAs were purchased from Invitrogen and Synbio Technologies (Suzhou, China). For siRNA transfection, primary microglia/myeloid cells that were isolated from TREM2^−/− and age-matched C57BL/6J mice at 6-week-old as described above were seeded in a sterile 12-well plate and cultured with DMEM containing 10% FBS, 1% penicillin-streptomycin and GlutaMAX Supplement for 24 hours before changing the medium for continuous cell culture for another 6 days (63, 64). After changing the culture medium, microglia/myeloid cells were transfected with 10 μM siRNA and 9 μl of Lipofectamine RNAiMAX Reagent (Life Technologies) for 6 to 8 hours followed by changing the medium to culture another 36 hours before collection for further analysis.

Coculture of primary microglia/myeloid cells and rod cells

Primary microglia/myeloid cells isolated from TREM2^−/− mice were seeded in a sterile 12-well plate and cultured for 7 days as described above. After changing the culture medium, microglia/myeloid cells were transfected with siSTAT2 or siCTR for 30 to 36 hours followed by coculturing with rod cells isolated from rd10 (rRCM25) and age-matched C57BL/6J (cRCM25) retinas at P25 in a sterile culture-insert (Millipore) for 10 to 12 hours before collection for further analysis.

Similarly, primary microglia/myeloid cells from TREM2^−/− mice were transfected with siPSAP and siCTR for 30 to 36 hours followed by coculturing with rRCM25 for 10 to 12 hours before collecting cells for further analysis. Meanwhile, cell supernatants of primary microglia/myeloid cells from TREM2^−/− mice treated with rRCM25 +siPSAP (spRCM3), rRCM25 + siCTR (spRCM2), or rRCM25 (spRCM1) were added into microglia/myeloid cells from C57BL/6J mice for culturing another 12 hours before collecting microglia/myeloid cells for further analysis.

To evaluate the role of microglia/myeloid cells phagocytizing rod photoreceptors, microglia/myeloid cells from TREM2^−/− mice were transfected with siSTAT2, siPSAP, or siCTR for 30 to 36 hours followed by coculturing with rod cells isolated from rd10 retinas at P25 for another 10 to 12 hours before collecting cells for further analysis.

rPSAP protein, NF-κB inhibitor, and metabolites stimulation

To explore the role of PSAP in TREM2/STAT2 signaling–mediated microglia/myeloid cells apoptosis, microglia/myeloid cells from TREM2^−/− mice were seeded in a sterile 12-well plate and cultured for 7 days as described above. After changing the culture medium, microglia/myeloid cells were incubated with mouse rPSAP (200 ng/ml) or PBS for 12 hours before collecting cells for further analysis. Meanwhile, the cell supernatants of microglia/myeloid cells from TREM2^−/− mice treated with rPSAP (pRCM2) and PBS (pRCM1) were collected and added into microglia/myeloid cells isolated from C57BL/6J mice for culturing another 12 hours before collecting cells for further analysis. To investigate the role of NF-κB pathway in regulating STAT2 activation, microglia/myeloid cells isolated from TREM2^−/− mice were incubated with NF-κB inhibitor (BAY 11-7082, 20 μM) for 12 hours before collecting cells for further analysis.

To study the role of metabolites generated from apoptotic rod photoreceptors, microglia/myeloid cells isolated from TREM2^−/− and age-matched C57BL/6J mice were transfected with siSTAT2 and siCTR for 30 to 36 hours followed by incubation with AA (50 μM), SA (5 mM), or creatinine (200 μM) for 12 hours before collecting cells for further analysis. To evaluate the role of metabolites in regulating microglia/myeloid cells phagocytizing rod photoreceptors, microglia/myeloid cells isolated from TREM2^−/− and C57BL/6J mice were transfected with siSTAT2 or siCTR for 30 to 36 hours followed by incubation with AA (50 μM), SA (5 mM), or creatinine (200 μM) and by coculturing with rod cells isolated from rd10 retinas at P25 for another 10 to 12 hours before collecting the cells for further analysis. To validate the causative role of photoreceptor-derived metabolites in activating STAT2 signaling in microglia, primary microglia/myeloid cells isolated from C57BL/6J mice were exposed for 12 hours to rd10 retinal photoreceptor-conditioned medium (rRCM25, P25) pretreated with specific metabolic inhibitors: ONO-RS-082 (20 μM, PLA2 inhibitor suppressing AA release), dorsomorphin (100 μM, AMPK inhibitor disrupting SA metabolism), or cyclocreatine (20 μM, creatine kinase antagonist). Postincubation cellular responses were analyzed to assess STAT2 pathway activation.

Quantitative real-time PCR

The microglia/myeloid cells were collected after different treatments, and the total RNA were extracted with TransZol Up Plus RNA Kit (TransGen, China) or TRIzol Reagent (Invitrogen), and cDNA synthesis and qPCR assay were performed as described above. Gene-specific primer used for qPCR are listed in table S1.

Immunocytochemistry and confocal imaging

After different treatments, the microglia/myeloid cells were fixed with 4% PFA for 15 min followed by rinsing with PBS for three times before incubating with primary antibodies including rabbit p-STAT2 (ABclonal, 1:250; Invitrogen, 1:250), rabbit Caspase-3 (Cell Signaling Technology, 1:250), rabbit PSAP (Proteintech, 1:250), rabbit GPR37 (Invitrogen, 1:250) and goat Iba-1 (Wako, 1:500) in blocking buffer containing 3% NDS, 1% BSA and 0.3% Triton X-100, overnight at 4°C. After several washes, microglia/myeloid cells were incubated with donkey anti-goat Alexa Fluor 488 (Invitrogen, 1:500) and donkey anti-rabbit Alexa Flour 594 (Invitrogen, 1:500) for 2 h at RT, followed by mounting slides with Dako fluorescence mounting medium. Imagines were captured by a Zeiss LSM 800 Upright Confocal Microscope (Zeiss, USA). Z-stack images were acquired at 0.5 μm intervals using a 20× objective. An orthogonal projection was made to compose a final image.

Statistical analysis

All experiments involving imaging and quantification were repeated at least three times with similar results. Animal numbers used in each group were indicated in figure legends. Data plotting and statistical tests were performed using GraphPad Prism software version 8.0. Data are represented as means ± SEM and analyzed with unpaired two-tailed Student’s t test or one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test. In all graphs, statistical significance was described as *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; #P < 0.05, ##P < 0.01, ###P < 0.001, ####P < 0.0001.

Acknowledgments

Funding: We declare that we received no funding in support of this research.

Author contributions: This study was designed by R.L. and B.L. The experiments were performed by J.-M.W., J.-Q.F., and R.L. The scRNA data were analyzed by R.L. and J.Z. The experimental data were analyzed by R.L., J.-M.W., and J.-Q.F. The main manuscript text was written by R.L., and B.L.

Competing interests: The authors declare that they have no competing interests.

Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials.

Supplementary Materials

This PDF file includes:

Figs. S1 to S15

Table S1

sciadv.adw9299_sm.pdf^{(7.1MB, pdf)}

REFERENCES AND NOTES

1.Eid A. T., Eid K. T., Odom J. V., Hinkle D., Leys M., Autosomal dominant retinitis pigmentosa secondary to TOPORS mutations: A report of a novel mutation and clinical findings. J. Clin. Med. 13, 1498 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Hartong D. T., Berson E. L., Dryja T. P., Retinitis pigmentosa. Lancet 368, 1795–1809 (2006). [DOI] [PubMed] [Google Scholar]
3.Wright A. F., Chakarova C. F., Abd El-Aziz M. M., Bhattacharya S. S., Photoreceptor degeneration: Genetic and mechanistic dissection of a complex trait. Nat. Rev. Genet. 11, 273–284 (2010). [DOI] [PubMed] [Google Scholar]
4.Zhao L., Hou C., Yan N., Neuroinflammation in retinitis pigmentosa: Therapies targeting the innate immune system. Front. Immunol. 13, 1059947 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Peng B., Xiao J., Wang K., So K. F., Tipoe G. L., Lin B., Suppression of microglial activation is neuroprotective in a mouse model of human retinitis pigmentosa. J. Neurosci. 34, 8139–8150 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Yang W., Xiong G., Lin B., Cyclooxygenase-1 mediates neuroinflammation and neurotoxicity in a mouse model of retinitis pigmentosa. J. Neuroinflammation 17, 306 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Zhao L., Zabel M. K., Wang X., Ma W., Shah P., Fariss R. N., Qian H., Parkhurst C. N., Gan W. B., Wong W. T., Microglial phagocytosis of living photoreceptors contributes to inherited retinal degeneration. EMBO Mol. Med. 7, 1179–1197 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Blank T., Goldmann T., Koch M., Amann L., Schön C., Bonin M., Pang S., Prinz M., Burnet M., Wagner J. E., Biel M., Michalakis S., Early microglia activation precedes photoreceptor degeneration in a mouse model of CNGB1-linked retinitis pigmentosa. Front. Immunol. 8, 1930 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Li R.-Y., Qin Q., Yang H.-C., Wang Y.-Y., Mi Y.-X., Yin Y.-S., Wang M., Yu C.-J., Tang Y., TREM2 in the pathogenesis of AD: A lipid metabolism regulator and potential metabolic therapeutic target. Mol. Neurodegener. 17, 40 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Li Y., Xu H., Wang H., Yang K., Luan J., Wang S., TREM2: Potential therapeutic targeting of microglia for Alzheimer’s disease. Biomed. Pharmacother. 165, 115218 (2023). [DOI] [PubMed] [Google Scholar]
11.Wang Y., Cella M., Mallinson K., Ulrich J. D., Young K. L., Robinette M. L., Gilfillan S., Krishnan G. M., Sudhakar S., Zinselmeyer B. H., Holtzman D. M., Cirrito J. R., Colonna M., TREM2 lipid sensing sustains the microglial response in an Alzheimer’s disease model. Cell 160, 1061–1071 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Wei W., Zhang L., Xin W., Pan Y., Tatenhorst L., Hao Z., Gerner S. T., Huber S., Juenemann M., Butz M., Huttner H. B., Bähr M., Fitzner D., Jia F., Doeppner T. R., TREM2 regulates microglial lipid droplet formation and represses post-ischemic brain injury. Biomed. Pharmacother. 170, 115962 (2024). [DOI] [PubMed] [Google Scholar]
13.Jay T. R., Miller C. M., Cheng P. J., Graham L. C., Bemiller S., Broihier M. L., Xu G., Margevicius D., Karlo J. C., Sousa G. L., Cotleur A. C., Butovsky O., Bekris L., Staugaitis S. M., Leverenz J. B., Pimplikar S. W., Landreth G. E., Howell G. R., Ransohoff R. M., Lamb B. T., TREM2 deficiency eliminates TREM2+ inflammatory macrophages and ameliorates pathology in Alzheimer’s disease mouse models. J. Exp. Med. 212, 287–295 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Jay T. R., von Saucken V. E., Landreth G. E., TREM2 in neurodegenerative diseases. Mol. Neurodegener. 12, 56 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.You Y. F., Chen M., Tang Y., Yu W. X., Pang X. W., Chu Y. H., Zhang H., Shang K., Deng G., Zhou L. Q., Yang S., Wang W., Xiao J., Tian D. S., Qin C., TREM2 deficiency inhibits microglial activation and aggravates demyelinating injury in neuromyelitis optica spectrum disorder. J. Neuroinflammation 20, 89 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Carmona S., Zahs K., Wu E., Dakin K., Bras J., Guerreiro R., The role of TREM2 in Alzheimer’s disease and other neurodegenerative disorders. Lancet Neurol. 17, 721–730 (2018). [DOI] [PubMed] [Google Scholar]
17.Mecca C., Giambanco I., Donato R., Arcuri C., Microglia and aging: The role of the TREM2-DAP12 and CX3CL1-CX3CR1 axes. Int. J. Mol. Sci. 19, 318 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Kleinberger G., Yamanishi Y., Suárez-Calvet M., Czirr E., Lohmann E., Cuyvers E., Struyfs H., Pettkus N., Wenninger-Weinzierl A., Mazaheri F., Tahirovic S., Lleó A., Alcolea D., Fortea J., Willem M., Lammich S., Molinuevo J. L., Sánchez-Valle R., Antonell A., Ramirez A., Heneka M. T., Sleegers K., van der Zee J., Martin J. J., Engelborghs S., Demirtas-Tatlidede A., Zetterberg H., Van Broeckhoven C., Gurvit H., Wyss-Coray T., Hardy J., Colonna M., Haass C., TREM2 mutations implicated in neurodegeneration impair cell surface transport and phagocytosis. Sci. Transl. Med. 6, 243ra286 (2014). [DOI] [PubMed] [Google Scholar]
19.Zhou W., Zhou Y., He J., Rao Y., Fei P., Li J., TREM2 deficiency in microglia accelerates photoreceptor cell death and immune cell infiltration following retinal detachment. Cell Death Dis. 14, 219 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Zhou W., He J., Shen G., Liu Y., Zhao P., Li J., TREM2-dependent activation of microglial cell protects photoreceptor cell during retinal degeneration via PPARγ and CD36. Cell Death Dis. 15, 623 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
21.C. Yu, E. M. Lad, R. Mathew, S. Littleton, Y. Chen, K. Schlepckow, S. Degan, L. Chew, J. Amason, J. Kalnitsky, C. B. Rickman, A. D. Proia, M. Colonna, C. Haass, D. R. Saban, Microglia at Sites of Atrophy Restrict the Progression of Retinal Degeneration via Galectin-3 and Trem2 Interactions. bioRxiv [Preprint] (2023). 10.1101/2023.07.19.549403. [DOI] [PMC free article] [PubMed]
22.Vecchiarelli H. A., Tremblay M.-È., Microglial transcriptional signatures in the central nervous system: Toward a future of unraveling their function in health and disease. Annu. Rev. Genet. 57, 65–86 (2023). [DOI] [PubMed] [Google Scholar]
23.Spiteri A. G., Wishart C. L., Pamphlett R., Locatelli G., King N. J. C., Microglia and monocytes in inflammatory CNS disease: Integrating phenotype and function. Acta Neuropathol. 143, 179–224 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Zabel M. K., Zhao L., Zhang Y., Gonzalez S. R., Ma W., Wang X., Fariss R. N., Wong W. T., Microglial phagocytosis and activation underlying photoreceptor degeneration is regulated by CX3CL1-CX3CR1 signaling in a mouse model of retinitis pigmentosa. Glia 64, 1479–1491 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Cervantes-Madrid D. L., Nagi S., Asting Gustafsson A., FosB transcription factor regulates COX-2 expression in colorectal cancer cells without affecting PGE2 expression. Oncol. Lett. 13, 1411–1416 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Liao Y., Wang Z., Wang L., Lin Y., Ye Z., Zeng X., Wei F., MicroRNA-27a-3p directly targets FosB to regulate cell proliferation, apoptosis, and inflammation responses in immunoglobulin a nephropathy. Biochem. Biophys. Res. Commun. 529, 1124–1130 (2020). [DOI] [PubMed] [Google Scholar]
27.Nuttall R. K., Silva C., Hader W., Bar-Or A., Patel K. D., Edwards D. R., Yong V. W., Metalloproteinases are enriched in microglia compared with leukocytes and they regulate cytokine levels in activated microglia. Glia 55, 516–526 (2007). [DOI] [PubMed] [Google Scholar]
28.Zhao H., Lai X., Zhang W., Zhu H., Zhang S., Wu W., Wang S., Tang M., Deng Z., Tan J., MiR-30a-5p frequently downregulated in prostate cancer inhibits cell proliferation via targeting PCLAF. Artif. Cells Nanomed. Biotechnol. 47, 278–289 (2019). [DOI] [PubMed] [Google Scholar]
29.Ma F., Zhi C., Wang M., Li T., Khan S. A., Ma Z., Jing Z., Bo C., Zhou Q., Xia S., Huang S., Huang S., Zhang Z., Jia H., Cui X., Yao M., Ji T., Dysregulated NF-κB signal promotes the hub gene PCLAF expression to facilitate nasopharyngeal carcinoma proliferation and metastasis. Biomed. Pharmacother. 125, 109905 (2020). [DOI] [PubMed] [Google Scholar]
30.Dräger N. M., Sattler S. M., Huang C. T., Teter O. M., Leng K., Hashemi S. H., Hong J., Aviles G., Clelland C. D., Zhan L., Udeochu J. C., Kodama L., Singleton A. B., Nalls M. A., Ichida J., Ward M. E., Faghri F., Gan L., Kampmann M., A CRISPRi/a platform in human iPSC-derived microglia uncovers regulators of disease states. Nat. Neurosci. 25, 1149–1162 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Mathys H., Adaikkan C., Gao F., Young J. Z., Manet E., Hemberg M., De Jager P. L., Ransohoff R. M., Regev A., Tsai L. H., Temporal tracking of microglia activation in neurodegeneration at single-cell resolution. Cell Rep. 21, 366–380 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Deczkowska A., Keren-Shaul H., Weiner A., Colonna M., Schwartz M., Amit I., Disease-associated microglia: A universal immune sensor of neurodegeneration. Cell 173, 1073–1081 (2018). [DOI] [PubMed] [Google Scholar]
33.Keren-Shaul H., Spinrad A., Weiner A., Matcovitch-Natan O., Dvir-Szternfeld R., Ulland T. K., David E., Baruch K., Lara-Astaiso D., Toth B., Itzkovitz S., Colonna M., Schwartz M., Amit I., A unique microglia type associated with restricting development of Alzheimer’s disease. Cell 169, 1276–1290.e17 (2017). [DOI] [PubMed] [Google Scholar]
34.Manrique-Castano D., Dzyubenko E., Borbor M., Vasileiadou P., Kleinschnitz C., Roll L., Faissner A., Hermann D. M., Tenascin-C preserves microglia surveillance and restricts leukocyte and, more specifically, T cell infiltration of the ischemic brain. Brain Behav. Immun. 91, 639–648 (2021). [DOI] [PubMed] [Google Scholar]
35.Xie K., Liu Y., Hao W., Walter S., Penke B., Hartmann T., Schachner M., Fassbender K., Tenascin-C deficiency ameliorates Alzheimer’s disease-related pathology in mice. Neurobiol. Aging 34, 2389–2398 (2013). [DOI] [PubMed] [Google Scholar]
36.Hwang M., Savarin C., Kim J., Powers J., Towne N., Oh H., Bergmann C. C., Trem2 deficiency impairs recovery and phagocytosis and dysregulates myeloid gene expression during virus-induced demyelination. J. Neuroinflammation 19, 267 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Pang X. W., Chu Y. H., Zhou L. Q., Chen M., You Y. F., Tang Y., Yang S., Zhang H., Xiao J., Deng G., Wang W., Shang K., Qin C., Tian D. S., Trem2 deficiency attenuates microglial phagocytosis and autophagic-lysosomal activation in white matter hypoperfusion. J. Neurochem. 167, 489–504 (2023). [DOI] [PubMed] [Google Scholar]
38.Khorooshi R., Babcock A. A., Owens T., NF-kappaB-driven STAT2 and CCL2 expression in astrocytes in response to brain injury. J. Immunol. 181, 7284–7291 (2008). [DOI] [PubMed] [Google Scholar]
39.Zhang J., Ma F., Li Z., Li Y., Sun X., Song M., Yang F., Wu E., Wei X., Wang Z., Yang L., NFKB2 mediates colorectal cancer cell immune escape and metastasis in a STAT2/PD-L1-dependent manner. MedComm 5, e521 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Zhang J., Wu Y.-J., Hu X.-X., Wei W., New insights into the Lck-NF-κB signaling pathway. Front. Cell Dev. Biol. 11, 1120747 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Kim S., Simon E., Myers L., Hamm L. L., Jazwinski S. M., Programmed cell death genes are linked to elevated creatine kinase levels in unhealthy male nonagenarians. Gerontology 62, 519–529 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Li B., Yang L., Creatine in T cell antitumor immunity and cancer immunotherapy. Nutrients 13, 1633 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Biasibetti-Brendler H., Schmitz F., Pierozan P., Zanotto B. S., Prezzi C. A., de Andrade R. B., Wannmacher C. M. D., Wyse A. T. S., Hypoxanthine induces neuroenergetic impairment and cell death in striatum of young adult Wistar rats. Mol. Neurobiol. 55, 4098–4106 (2018). [DOI] [PubMed] [Google Scholar]
44.Kim Y. J., Ryu H. M., Choi J. Y., Cho J. H., Kim C. D., Park S. H., Kim Y. L., Hypoxanthine causes endothelial dysfunction through oxidative stress-induced apoptosis. Biochem. Biophys. Res. Commun. 482, 821–827 (2017). [DOI] [PubMed] [Google Scholar]
45.Yoo H. C., Yu Y. C., Sung Y., Han J. M., Glutamine reliance in cell metabolism. Exp. Mol. Med. 52, 1496–1516 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Li T., Zeng L., Gao W., Cui M. Z., Fu X., Xu X., PSAP induces a unique Apaf-1 and Smac-dependent mitochondrial apoptotic pathway independent of Bcl-2 family proteins. Biochim. Biophys. Acta 1832, 453–474 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Zhou X., Sun L., Bracko O., Choi J. W., Jia Y., Nana A. L., Brady O. A., Hernandez J. C. C., Nishimura N., Seeley W. W., Hu F., Impaired prosaposin lysosomal trafficking in frontotemporal lobar degeneration due to progranulin mutations. Nat. Commun. 8, 15277 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Wang B., Wu L., Chen J., Dong L., Chen C., Wen Z., Hu J., Fleming I., Wang D. W., Metabolism pathways of arachidonic acids: Mechanisms and potential therapeutic targets. Signal Transduct. Target. Ther. 6, 94 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Zhang Y., Liu Y., Sun J., Zhang W., Guo Z., Ma Q., Arachidonic acid metabolism in health and disease. MedComm 4, e363 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Lu J., Cong T., Wen X., Li X., Du D., He G., Jiang X., Salicylic acid treats acne vulgaris by suppressing AMPK/SREBP1 pathway in sebocytes. Exp. Dermatol. 28, 786–794 (2019). [DOI] [PubMed] [Google Scholar]
51.Gamero A. M., Young M. R., Mentor-Marcel R., Bobe G., Scarzello A. J., Wise J., Colburn N. H., STAT2 contributes to promotion of colorectal and skin carcinogenesis. Cancer Prev. Res. 3, 495–504 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Li R., Zhang J., Wang Q., Cheng M., Lin B., TPM1 mediates inflammation downstream of TREM2 via the PKA/CREB signaling pathway. J. Neuroinflammation 19, 257 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Ahmed Z., Kalinski H., Berry M., Almasieh M., Ashush H., Slager N., Brafman A., Spivak I., Prasad N., Mett I., Shalom E., Alpert E., Di Polo A., Feinstein E., Logan A., Ocular neuroprotection by siRNA targeting caspase-2. Cell Death Dis. 2, e173 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Wang K., Xiao J., Peng B., Xing F., So K. F., Tipoe G. L., Lin B., Retinal structure and function preservation by polysaccharides of wolfberry in a mouse model of retinal degeneration. Sci. Rep. 4, 7601 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Feodorova Y., Koch M., Bultman S., Michalakis S., Solovei I., Quick and reliable method for retina dissociation and separation of rod photoreceptor perikarya from adult mice. MethodsX 2, 39–46 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Li R., Liang Y., Lin B., Accumulation of systematic TPM1 mediates inflammation and neuronal remodeling by phosphorylating PKA and regulating the FABP5/NF-κB signaling pathway in the retina of aged mice. Aging Cell 21, e13566 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Butler A., Hoffman P., Smibert P., Papalexi E., Satija R., Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat. Biotechnol. 36, 411–420 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
58.McGinnis C. S., Murrow L. M., Gartner Z. J., DoubletFinder: Doublet detection in single-cell RNA sequencing data using artificial nearest neighbors. Cell Syst. 8, 329–337.e4 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Ge S. X., Jung D., Yao R., ShinyGO: A graphical gene-set enrichment tool for animals and plants. Bioinformatics 36, 2628–2629 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Szklarczyk D., Gable A. L., Nastou K. C., Lyon D., Kirsch R., Pyysalo S., Doncheva N. T., Legeay M., Fang T., Bork P., Jensen L. J., von Mering C., The STRING database in 2021: Customizable protein-protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res. 49, D605–D612 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Jin S., Guerrero-Juarez C. F., Zhang L., Chang I., Ramos R., Kuan C. H., Myung P., Plikus M. V., Nie Q., Inference and analysis of cell-cell communication using CellChat. Nat. Commun. 12, 1088 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Aibar S., González-Blas C. B., Moerman T., Huynh-Thu V. A., Imrichova H., Hulselmans G., Rambow F., Marine J. C., Geurts P., Aerts J., van den Oord J., Atak Z. K., Wouters J., Aerts S., SCENIC: Single-cell regulatory network inference and clustering. Nat. Methods 14, 1083–1086 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Vijaya A. K., Iešmantaitė M., Mela V., Baltriukienė D., Burokas A., Microglia isolation from aging mice for cell culture: A beginner’s guide. Front. Cell. Neurosci. 17, 1082180 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Bohlen C. J., Bennett F. C., Tucker A. F., Collins H. Y., Mulinyawe S. B., Barres B. A., Diverse requirements for microglial survival, specification, and function revealed by defined-medium cultures. Neuron 94, 759–773.e8 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figs. S1 to S15

Table S1

sciadv.adw9299_sm.pdf^{(7.1MB, pdf)}

[R1] 1.Eid A. T., Eid K. T., Odom J. V., Hinkle D., Leys M., Autosomal dominant retinitis pigmentosa secondary to TOPORS mutations: A report of a novel mutation and clinical findings. J. Clin. Med. 13, 1498 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R2] 2.Hartong D. T., Berson E. L., Dryja T. P., Retinitis pigmentosa. Lancet 368, 1795–1809 (2006). [DOI] [PubMed] [Google Scholar]

[R3] 3.Wright A. F., Chakarova C. F., Abd El-Aziz M. M., Bhattacharya S. S., Photoreceptor degeneration: Genetic and mechanistic dissection of a complex trait. Nat. Rev. Genet. 11, 273–284 (2010). [DOI] [PubMed] [Google Scholar]

[R4] 4.Zhao L., Hou C., Yan N., Neuroinflammation in retinitis pigmentosa: Therapies targeting the innate immune system. Front. Immunol. 13, 1059947 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R5] 5.Peng B., Xiao J., Wang K., So K. F., Tipoe G. L., Lin B., Suppression of microglial activation is neuroprotective in a mouse model of human retinitis pigmentosa. J. Neurosci. 34, 8139–8150 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] 6.Yang W., Xiong G., Lin B., Cyclooxygenase-1 mediates neuroinflammation and neurotoxicity in a mouse model of retinitis pigmentosa. J. Neuroinflammation 17, 306 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R7] 7.Zhao L., Zabel M. K., Wang X., Ma W., Shah P., Fariss R. N., Qian H., Parkhurst C. N., Gan W. B., Wong W. T., Microglial phagocytosis of living photoreceptors contributes to inherited retinal degeneration. EMBO Mol. Med. 7, 1179–1197 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R8] 8.Blank T., Goldmann T., Koch M., Amann L., Schön C., Bonin M., Pang S., Prinz M., Burnet M., Wagner J. E., Biel M., Michalakis S., Early microglia activation precedes photoreceptor degeneration in a mouse model of CNGB1-linked retinitis pigmentosa. Front. Immunol. 8, 1930 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] 9.Li R.-Y., Qin Q., Yang H.-C., Wang Y.-Y., Mi Y.-X., Yin Y.-S., Wang M., Yu C.-J., Tang Y., TREM2 in the pathogenesis of AD: A lipid metabolism regulator and potential metabolic therapeutic target. Mol. Neurodegener. 17, 40 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R10] 10.Li Y., Xu H., Wang H., Yang K., Luan J., Wang S., TREM2: Potential therapeutic targeting of microglia for Alzheimer’s disease. Biomed. Pharmacother. 165, 115218 (2023). [DOI] [PubMed] [Google Scholar]

[R11] 11.Wang Y., Cella M., Mallinson K., Ulrich J. D., Young K. L., Robinette M. L., Gilfillan S., Krishnan G. M., Sudhakar S., Zinselmeyer B. H., Holtzman D. M., Cirrito J. R., Colonna M., TREM2 lipid sensing sustains the microglial response in an Alzheimer’s disease model. Cell 160, 1061–1071 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R12] 12.Wei W., Zhang L., Xin W., Pan Y., Tatenhorst L., Hao Z., Gerner S. T., Huber S., Juenemann M., Butz M., Huttner H. B., Bähr M., Fitzner D., Jia F., Doeppner T. R., TREM2 regulates microglial lipid droplet formation and represses post-ischemic brain injury. Biomed. Pharmacother. 170, 115962 (2024). [DOI] [PubMed] [Google Scholar]

[R13] 13.Jay T. R., Miller C. M., Cheng P. J., Graham L. C., Bemiller S., Broihier M. L., Xu G., Margevicius D., Karlo J. C., Sousa G. L., Cotleur A. C., Butovsky O., Bekris L., Staugaitis S. M., Leverenz J. B., Pimplikar S. W., Landreth G. E., Howell G. R., Ransohoff R. M., Lamb B. T., TREM2 deficiency eliminates TREM2+ inflammatory macrophages and ameliorates pathology in Alzheimer’s disease mouse models. J. Exp. Med. 212, 287–295 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R14] 14.Jay T. R., von Saucken V. E., Landreth G. E., TREM2 in neurodegenerative diseases. Mol. Neurodegener. 12, 56 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R15] 15.You Y. F., Chen M., Tang Y., Yu W. X., Pang X. W., Chu Y. H., Zhang H., Shang K., Deng G., Zhou L. Q., Yang S., Wang W., Xiao J., Tian D. S., Qin C., TREM2 deficiency inhibits microglial activation and aggravates demyelinating injury in neuromyelitis optica spectrum disorder. J. Neuroinflammation 20, 89 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R16] 16.Carmona S., Zahs K., Wu E., Dakin K., Bras J., Guerreiro R., The role of TREM2 in Alzheimer’s disease and other neurodegenerative disorders. Lancet Neurol. 17, 721–730 (2018). [DOI] [PubMed] [Google Scholar]

[R17] 17.Mecca C., Giambanco I., Donato R., Arcuri C., Microglia and aging: The role of the TREM2-DAP12 and CX3CL1-CX3CR1 axes. Int. J. Mol. Sci. 19, 318 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R18] 18.Kleinberger G., Yamanishi Y., Suárez-Calvet M., Czirr E., Lohmann E., Cuyvers E., Struyfs H., Pettkus N., Wenninger-Weinzierl A., Mazaheri F., Tahirovic S., Lleó A., Alcolea D., Fortea J., Willem M., Lammich S., Molinuevo J. L., Sánchez-Valle R., Antonell A., Ramirez A., Heneka M. T., Sleegers K., van der Zee J., Martin J. J., Engelborghs S., Demirtas-Tatlidede A., Zetterberg H., Van Broeckhoven C., Gurvit H., Wyss-Coray T., Hardy J., Colonna M., Haass C., TREM2 mutations implicated in neurodegeneration impair cell surface transport and phagocytosis. Sci. Transl. Med. 6, 243ra286 (2014). [DOI] [PubMed] [Google Scholar]

[R19] 19.Zhou W., Zhou Y., He J., Rao Y., Fei P., Li J., TREM2 deficiency in microglia accelerates photoreceptor cell death and immune cell infiltration following retinal detachment. Cell Death Dis. 14, 219 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R20] 20.Zhou W., He J., Shen G., Liu Y., Zhao P., Li J., TREM2-dependent activation of microglial cell protects photoreceptor cell during retinal degeneration via PPARγ and CD36. Cell Death Dis. 15, 623 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R21] 21.C. Yu, E. M. Lad, R. Mathew, S. Littleton, Y. Chen, K. Schlepckow, S. Degan, L. Chew, J. Amason, J. Kalnitsky, C. B. Rickman, A. D. Proia, M. Colonna, C. Haass, D. R. Saban, Microglia at Sites of Atrophy Restrict the Progression of Retinal Degeneration via Galectin-3 and Trem2 Interactions. bioRxiv [Preprint] (2023). 10.1101/2023.07.19.549403. [DOI] [PMC free article] [PubMed]

[R22] 22.Vecchiarelli H. A., Tremblay M.-È., Microglial transcriptional signatures in the central nervous system: Toward a future of unraveling their function in health and disease. Annu. Rev. Genet. 57, 65–86 (2023). [DOI] [PubMed] [Google Scholar]

[R23] 23.Spiteri A. G., Wishart C. L., Pamphlett R., Locatelli G., King N. J. C., Microglia and monocytes in inflammatory CNS disease: Integrating phenotype and function. Acta Neuropathol. 143, 179–224 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R24] 24.Zabel M. K., Zhao L., Zhang Y., Gonzalez S. R., Ma W., Wang X., Fariss R. N., Wong W. T., Microglial phagocytosis and activation underlying photoreceptor degeneration is regulated by CX3CL1-CX3CR1 signaling in a mouse model of retinitis pigmentosa. Glia 64, 1479–1491 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R25] 25.Cervantes-Madrid D. L., Nagi S., Asting Gustafsson A., FosB transcription factor regulates COX-2 expression in colorectal cancer cells without affecting PGE2 expression. Oncol. Lett. 13, 1411–1416 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R26] 26.Liao Y., Wang Z., Wang L., Lin Y., Ye Z., Zeng X., Wei F., MicroRNA-27a-3p directly targets FosB to regulate cell proliferation, apoptosis, and inflammation responses in immunoglobulin a nephropathy. Biochem. Biophys. Res. Commun. 529, 1124–1130 (2020). [DOI] [PubMed] [Google Scholar]

[R27] 27.Nuttall R. K., Silva C., Hader W., Bar-Or A., Patel K. D., Edwards D. R., Yong V. W., Metalloproteinases are enriched in microglia compared with leukocytes and they regulate cytokine levels in activated microglia. Glia 55, 516–526 (2007). [DOI] [PubMed] [Google Scholar]

[R28] 28.Zhao H., Lai X., Zhang W., Zhu H., Zhang S., Wu W., Wang S., Tang M., Deng Z., Tan J., MiR-30a-5p frequently downregulated in prostate cancer inhibits cell proliferation via targeting PCLAF. Artif. Cells Nanomed. Biotechnol. 47, 278–289 (2019). [DOI] [PubMed] [Google Scholar]

[R29] 29.Ma F., Zhi C., Wang M., Li T., Khan S. A., Ma Z., Jing Z., Bo C., Zhou Q., Xia S., Huang S., Huang S., Zhang Z., Jia H., Cui X., Yao M., Ji T., Dysregulated NF-κB signal promotes the hub gene PCLAF expression to facilitate nasopharyngeal carcinoma proliferation and metastasis. Biomed. Pharmacother. 125, 109905 (2020). [DOI] [PubMed] [Google Scholar]

[R30] 30.Dräger N. M., Sattler S. M., Huang C. T., Teter O. M., Leng K., Hashemi S. H., Hong J., Aviles G., Clelland C. D., Zhan L., Udeochu J. C., Kodama L., Singleton A. B., Nalls M. A., Ichida J., Ward M. E., Faghri F., Gan L., Kampmann M., A CRISPRi/a platform in human iPSC-derived microglia uncovers regulators of disease states. Nat. Neurosci. 25, 1149–1162 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R31] 31.Mathys H., Adaikkan C., Gao F., Young J. Z., Manet E., Hemberg M., De Jager P. L., Ransohoff R. M., Regev A., Tsai L. H., Temporal tracking of microglia activation in neurodegeneration at single-cell resolution. Cell Rep. 21, 366–380 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R32] 32.Deczkowska A., Keren-Shaul H., Weiner A., Colonna M., Schwartz M., Amit I., Disease-associated microglia: A universal immune sensor of neurodegeneration. Cell 173, 1073–1081 (2018). [DOI] [PubMed] [Google Scholar]

[R33] 33.Keren-Shaul H., Spinrad A., Weiner A., Matcovitch-Natan O., Dvir-Szternfeld R., Ulland T. K., David E., Baruch K., Lara-Astaiso D., Toth B., Itzkovitz S., Colonna M., Schwartz M., Amit I., A unique microglia type associated with restricting development of Alzheimer’s disease. Cell 169, 1276–1290.e17 (2017). [DOI] [PubMed] [Google Scholar]

[R34] 34.Manrique-Castano D., Dzyubenko E., Borbor M., Vasileiadou P., Kleinschnitz C., Roll L., Faissner A., Hermann D. M., Tenascin-C preserves microglia surveillance and restricts leukocyte and, more specifically, T cell infiltration of the ischemic brain. Brain Behav. Immun. 91, 639–648 (2021). [DOI] [PubMed] [Google Scholar]

[R35] 35.Xie K., Liu Y., Hao W., Walter S., Penke B., Hartmann T., Schachner M., Fassbender K., Tenascin-C deficiency ameliorates Alzheimer’s disease-related pathology in mice. Neurobiol. Aging 34, 2389–2398 (2013). [DOI] [PubMed] [Google Scholar]

[R36] 36.Hwang M., Savarin C., Kim J., Powers J., Towne N., Oh H., Bergmann C. C., Trem2 deficiency impairs recovery and phagocytosis and dysregulates myeloid gene expression during virus-induced demyelination. J. Neuroinflammation 19, 267 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R37] 37.Pang X. W., Chu Y. H., Zhou L. Q., Chen M., You Y. F., Tang Y., Yang S., Zhang H., Xiao J., Deng G., Wang W., Shang K., Qin C., Tian D. S., Trem2 deficiency attenuates microglial phagocytosis and autophagic-lysosomal activation in white matter hypoperfusion. J. Neurochem. 167, 489–504 (2023). [DOI] [PubMed] [Google Scholar]

[R38] 38.Khorooshi R., Babcock A. A., Owens T., NF-kappaB-driven STAT2 and CCL2 expression in astrocytes in response to brain injury. J. Immunol. 181, 7284–7291 (2008). [DOI] [PubMed] [Google Scholar]

[R39] 39.Zhang J., Ma F., Li Z., Li Y., Sun X., Song M., Yang F., Wu E., Wei X., Wang Z., Yang L., NFKB2 mediates colorectal cancer cell immune escape and metastasis in a STAT2/PD-L1-dependent manner. MedComm 5, e521 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R40] 40.Zhang J., Wu Y.-J., Hu X.-X., Wei W., New insights into the Lck-NF-κB signaling pathway. Front. Cell Dev. Biol. 11, 1120747 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R41] 41.Kim S., Simon E., Myers L., Hamm L. L., Jazwinski S. M., Programmed cell death genes are linked to elevated creatine kinase levels in unhealthy male nonagenarians. Gerontology 62, 519–529 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R42] 42.Li B., Yang L., Creatine in T cell antitumor immunity and cancer immunotherapy. Nutrients 13, 1633 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R43] 43.Biasibetti-Brendler H., Schmitz F., Pierozan P., Zanotto B. S., Prezzi C. A., de Andrade R. B., Wannmacher C. M. D., Wyse A. T. S., Hypoxanthine induces neuroenergetic impairment and cell death in striatum of young adult Wistar rats. Mol. Neurobiol. 55, 4098–4106 (2018). [DOI] [PubMed] [Google Scholar]

[R44] 44.Kim Y. J., Ryu H. M., Choi J. Y., Cho J. H., Kim C. D., Park S. H., Kim Y. L., Hypoxanthine causes endothelial dysfunction through oxidative stress-induced apoptosis. Biochem. Biophys. Res. Commun. 482, 821–827 (2017). [DOI] [PubMed] [Google Scholar]

[R45] 45.Yoo H. C., Yu Y. C., Sung Y., Han J. M., Glutamine reliance in cell metabolism. Exp. Mol. Med. 52, 1496–1516 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R46] 46.Li T., Zeng L., Gao W., Cui M. Z., Fu X., Xu X., PSAP induces a unique Apaf-1 and Smac-dependent mitochondrial apoptotic pathway independent of Bcl-2 family proteins. Biochim. Biophys. Acta 1832, 453–474 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R47] 47.Zhou X., Sun L., Bracko O., Choi J. W., Jia Y., Nana A. L., Brady O. A., Hernandez J. C. C., Nishimura N., Seeley W. W., Hu F., Impaired prosaposin lysosomal trafficking in frontotemporal lobar degeneration due to progranulin mutations. Nat. Commun. 8, 15277 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R48] 48.Wang B., Wu L., Chen J., Dong L., Chen C., Wen Z., Hu J., Fleming I., Wang D. W., Metabolism pathways of arachidonic acids: Mechanisms and potential therapeutic targets. Signal Transduct. Target. Ther. 6, 94 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R49] 49.Zhang Y., Liu Y., Sun J., Zhang W., Guo Z., Ma Q., Arachidonic acid metabolism in health and disease. MedComm 4, e363 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R50] 50.Lu J., Cong T., Wen X., Li X., Du D., He G., Jiang X., Salicylic acid treats acne vulgaris by suppressing AMPK/SREBP1 pathway in sebocytes. Exp. Dermatol. 28, 786–794 (2019). [DOI] [PubMed] [Google Scholar]

[R51] 51.Gamero A. M., Young M. R., Mentor-Marcel R., Bobe G., Scarzello A. J., Wise J., Colburn N. H., STAT2 contributes to promotion of colorectal and skin carcinogenesis. Cancer Prev. Res. 3, 495–504 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R52] 52.Li R., Zhang J., Wang Q., Cheng M., Lin B., TPM1 mediates inflammation downstream of TREM2 via the PKA/CREB signaling pathway. J. Neuroinflammation 19, 257 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R53] 53.Ahmed Z., Kalinski H., Berry M., Almasieh M., Ashush H., Slager N., Brafman A., Spivak I., Prasad N., Mett I., Shalom E., Alpert E., Di Polo A., Feinstein E., Logan A., Ocular neuroprotection by siRNA targeting caspase-2. Cell Death Dis. 2, e173 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R54] 54.Wang K., Xiao J., Peng B., Xing F., So K. F., Tipoe G. L., Lin B., Retinal structure and function preservation by polysaccharides of wolfberry in a mouse model of retinal degeneration. Sci. Rep. 4, 7601 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R55] 55.Feodorova Y., Koch M., Bultman S., Michalakis S., Solovei I., Quick and reliable method for retina dissociation and separation of rod photoreceptor perikarya from adult mice. MethodsX 2, 39–46 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R56] 56.Li R., Liang Y., Lin B., Accumulation of systematic TPM1 mediates inflammation and neuronal remodeling by phosphorylating PKA and regulating the FABP5/NF-κB signaling pathway in the retina of aged mice. Aging Cell 21, e13566 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R57] 57.Butler A., Hoffman P., Smibert P., Papalexi E., Satija R., Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat. Biotechnol. 36, 411–420 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R58] 58.McGinnis C. S., Murrow L. M., Gartner Z. J., DoubletFinder: Doublet detection in single-cell RNA sequencing data using artificial nearest neighbors. Cell Syst. 8, 329–337.e4 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R59] 59.Ge S. X., Jung D., Yao R., ShinyGO: A graphical gene-set enrichment tool for animals and plants. Bioinformatics 36, 2628–2629 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R60] 60.Szklarczyk D., Gable A. L., Nastou K. C., Lyon D., Kirsch R., Pyysalo S., Doncheva N. T., Legeay M., Fang T., Bork P., Jensen L. J., von Mering C., The STRING database in 2021: Customizable protein-protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res. 49, D605–D612 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R61] 61.Jin S., Guerrero-Juarez C. F., Zhang L., Chang I., Ramos R., Kuan C. H., Myung P., Plikus M. V., Nie Q., Inference and analysis of cell-cell communication using CellChat. Nat. Commun. 12, 1088 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R62] 62.Aibar S., González-Blas C. B., Moerman T., Huynh-Thu V. A., Imrichova H., Hulselmans G., Rambow F., Marine J. C., Geurts P., Aerts J., van den Oord J., Atak Z. K., Wouters J., Aerts S., SCENIC: Single-cell regulatory network inference and clustering. Nat. Methods 14, 1083–1086 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R63] 63.Vijaya A. K., Iešmantaitė M., Mela V., Baltriukienė D., Burokas A., Microglia isolation from aging mice for cell culture: A beginner’s guide. Front. Cell. Neurosci. 17, 1082180 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R64] 64.Bohlen C. J., Bennett F. C., Tucker A. F., Collins H. Y., Mulinyawe S. B., Barres B. A., Diverse requirements for microglial survival, specification, and function revealed by defined-medium cultures. Neuron 94, 759–773.e8 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Temporal biphasic regulation of photoreceptor degeneration by microglial TREM2: A metabolic-immune nexus in retinitis pigmentosa

Rong Li

Jing Zhang

Jiang-Mei Wu

Jun-Qi Fan

Bin Lin

Roles

Abstract

INTRODUCTION

RESULTS

Spatiotemporal dynamics of TREM2+ microglia/myeloid cell reveal stage-specific involvement in RP

Fig. 1. Spatiotemporal dynamics of TREM2+ microglia/myeloid cell reveal stage-specific involvement in RP.

Stage-dependent immunomodulation and neuroprotection by TREM2 ablation in RP

Fig. 2. Stage-dependent immunomodulation and neuroprotection by TREM2 ablation in RP.

Single-cell transcriptomics reveals TREM2-dependent microglia/myeloid cell reprogramming in RP

Fig. 3. Single-cell transcriptomics reveals TREM2-dependent microglia/myeloid cell reprogramming in RP.

TREM2 ablation orchestrates microglia/myeloid cell apoptosis and senescence clearance in RP

Fig. 4. TREM2 ablation orchestrates microglia/myeloid cell apoptosis and senescence clearance in RP.

STAT2 emerges as a master regulator of microglia/myeloid cell apoptosis through NF-κB crosstalk following TREM2 ablation in RP

Fig. 5. STAT2 emerges as a master regulator of microglia/myeloid cell apoptosis through NF-κB cross-talk following TREM2 ablation in RP.

TREM2/STAT2 Axis governs paracrine apoptosis through PSAP/GPR37 signaling in RP

Fig. 6. TREM2/STAT2 axis governs paracrine apoptosis through PSAP/GPR37 signaling in RP.

Photoreceptor-derived metabolites drive microglia/myeloid cell apoptosis via TREM2/STAT2 axis in RP

Fig. 7. Photoreceptor-derived metabolites drive microglia/myeloid cell apoptosis via TREM2/STAT2 axis in RP.

DISCUSSION

Fig. 8. Schematic illustration showing the role of TREM2/STAT2 signaling in mediating microglia/myeloid cell apoptosis and photoreceptor degeneration in RP.

MATERIALS AND METHODS

Mice

In vivo experiments

Intravitreal siRNA injections

Immunocytochemistry and confocal imaging

TUNEL assay

Immunoblotting

Quantitative real-time PCR

Primary microglia/myeloid cell isolation by MACS

Flow cytometry

Isolation of rod cells

ERG analysis

Single-cell RNA sequencing

Sample preparation, library construction, and reads filtering for scRNA-seq

Data integration, dimensionality reduction, and cell clustering

DEGs identification and pathway enrichment analysis

Analysis of cell-cell interactions

TF module analysis

Untargeted metabolomics analysis

In vitro experiments

siRNA transfection

Coculture of primary microglia/myeloid cells and rod cells

rPSAP protein, NF-κB inhibitor, and metabolites stimulation

Quantitative real-time PCR

Immunocytochemistry and confocal imaging

Statistical analysis

Acknowledgments

Supplementary Materials

This PDF file includes:

REFERENCES AND NOTES

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases

Spatiotemporal dynamics of TREM2⁺ microglia/myeloid cell reveal stage-specific involvement in RP

Fig. 1. Spatiotemporal dynamics of TREM2⁺ microglia/myeloid cell reveal stage-specific involvement in RP.