Induced transcription of TREM2 predominantly situated at the infection sites
The single-cell resolution landscape of OPC has not been mapped to date. To address this gap, we utilized scRNA-seq in a murine model (Fig.1A), encompassing normal tongue tissue, tongue tissue post-C. albicans infection, and tongue tissue of mice treated with CA to induce OPC. Following sequencing and subsequent seurat pipeline analysis, a total of 30,810 cells were identified (Fig.1B, C). Predominantly, these cells comprised non-immune cells such as fibroblasts, endothelial cells, and smooth muscle cells (Fig.1B), with immune cells accounting for only 12.19%, aligning with the observed low proportion of immune cells in tongue tissue. Further subsetting of these immune cells (Fig.1D, E) revealed macrophages distinguishing into TREM2+ and TREM2- subpopulations. Conventionally, macrophage subpopulations with distinct functions are denoted as M1 and M2 macrophages. However, it is crucial to note that the M1 and M2 classification, derived mainly from in vitro polarization assays, is more apt for describing the activation state of macrophages in vitro [24]. Indeed, differentiating macrophage subpopulations in murine tongue tissue using M1 or M2 markers proved challenging (Extended Data Fig.1A). With the increasing application of single-cell studies, macrophage subtyping leans towards utilizing TREM2 for distinction. Within tongue tissue, the confirmed TREM2+ macrophages exhibited an elevated proportion among immune cells post-C. albicans infection, while CA-treated mice showed a significant reduction in the proportion of TREM2+ macrophages (Fig.1F, Supplementary Table 1).
To ascertain the presence of TREM2+ macrophages in the human oral mucosa, we conducted an analysis of publicly available scRNA-seq data [25] of the oral mucosa (Extended Data Fig.1B). After subsetting immune cell populations (Extended Data Fig.1C) and subsequent clustering (Extended Data Fig.1E, F), we identified the presence of TREM2+ macrophages in normal human tissue (Extended Data Fig.1D). These findings indicated that these cells may also represent a type of tissue-resident macrophage, akin to microglial cells found in the brain, Kupffer cells in the liver, and alveolar macrophages in lung tissue. To validate this hypothesis, we performed immunofluorescence and flow cytometric analyses, confirming the presence of TREM2+ macrophages in normal murine tongue tissue (Extended Data Fig.1G), predominantly localized within the internal regions of the tongue tissue (Extended Data Fig.1H). Upon C. albicans infection, the proportion of TREM2+ macrophages in murine tongue tissue significantly increased (Extended Data Fig.1G), with these macrophages predominantly situated at the infection sites (Extended Data Fig.1H).
Hematopoietic-derived TREM2 is indispensable for protection against OPC
Previous studies have suggested an inhibitory role for TREM2 in tumor immunity [26-28]. Findings from the OPC mouse model demonstrate the complete clearance of C. albicans by Day 5 [29]. Intriguingly, during the early stages of OPC (Day 1), we observed an accumulation of TREM2-expressing macrophages (Extended Data Fig.1G, H). To further understand the role of TREM2 in OPC and rule out any inhibitory effects (Considering that Trem2 deficiency leads to rapid recovery in OPC mice), we conducted an assessment of TREM2 impact during Days 1 and 2. The results revealed that Trem2 deficiency increased susceptibility to OPC in mice, manifested by lower body weight and higher fungal burden in Trem2-/- mice (Fig.2A, B). Pathological findings also demonstrated more severe epithelial fungal invasion in Trem2-/- mice (Fig.2C). By the fifth day, WT mice exhibited undetectable levels of C. albicans, while Trem2-/- mice still harbored persistent fungal infections (Extended Data Fig.1I, J).
Over the past decade, the classification of Ly6C+ and Ly6C- macrophages has become a prevalent tool for studying monocyte-derived macrophages [30]. This classification helps in accurately defining the state of macrophages within complex internal microenvironments [30]. Ly6C+ macrophages undergo significant enrichment and display a pro-inflammatory phenotype, actively participating in acute inflammatory responses. Our investigation further unveiled a pronounced infiltration of TREM2+ macrophages at the infection site (Extended Data Fig.1H). We postulated that these TREM2+ macrophages might belong to the Ly6C+ subset. To validate this hypothesis, flow cytometric analysis was conducted, revealing differential presence of Ly6C+ TREM2+ macrophages before and after infection, while Ly6C- TREM2+ macrophages exhibited no such difference (Fig.2D). Additionally, we observed that TREM2+ macrophages at the infection site were also CX3CR1+ (Extended Data Fig.2A), a marker for infiltrating monocyte-derived macrophages. scRNA-seq data further indicated a high expression of CX3CR1 in both human and murine TREM2+ macrophages (Extended Data Fig.2B, C). Remarkably, the administration of the lymphocyte recirculation inhibitor FTY720 effectively suppressed the infiltration of Ly6C+ TREM2+ macrophages, with no impact on Ly6C-TREM2+ macrophage infiltration (Extended Data Fig.2D). This inhibition significantly heightened susceptibility to OPC in mice (Extended Data Fig.2E-H). These findings strongly support a protective role of hematopoietic-origin TREM2 in OPC. To pinpoint the specific source of TREM2 in the anti-fungal effect, we created chimeric mice by replacing the bone marrow (BM) of lethally irradiated wild-type (WT) mice with a reconstituting of WT BM and Trem2-/- BM. The results demonstrated that chimeric mice exhibited increased susceptibility to OPC (Extended Data Fig.2I, J). Therefore, we can confidently affirm that hematopoietic-origin TREM2 plays a pivotal protective role in OPC. Previous studies have firmly established that Trem2 expression is confined to macrophages within the myeloid cell compartment [19, 31], underscoring that TREM2 primarily exerts its function through macrophages of hematopoietic origin.
Recent investigations indicated an elevation in TREM2 expression within macrophages across various liver diseases and relevant mouse models [32-37]. However, whether TREM2 undergoes upregulation during fungal infections remains unexplored. To address this knowledge gap, we stimulated mouse bone marrow-derived macrophages (BMDMs) with C. albicans, revealing an augmentation in TREM2 protein (Extended Data Fig.2K). Intriguingly, IL-17 was also identified as a potential inducer of TREM2 protein upregulation (Extended Data Fig.2K). These findings suggested that, subsequent to macrophage infiltration into the OPC infection site, both C. albicans and IL-17 contributed to TREM2 protein upregulation, thereby facilitating the clearance of C. albicans.
Pro-inflammatory role of TREM2+ macrophages in OPC
To elucidate the molecular mechanisms through which TREM2 contributes to protection in OPC, we further analyzed scRNA-seq data. We observed that TREM2+ macrophages upregulated inflammatory chemokines (Cxcl1, Cxcl16, Ccl2, Ccl4, and Ccl9), antigen presentation (H2-DMb1, H2-Aa, H2-Eb1, H2-DMa, and Cd74), interferon response (Ifitm3), complement pathway (C1qa, C1qb, and C1qc), protein degradation (Lamp1, Ctss, Ctsz, and Lgmn), inflammation markers (Rgs10, Mpeg1, and Tnf), costimulatory molecules (Cd86), chemokine receptors (Ccr2 and Ccr5), pattern recognition receptors (Tlr2), and surface adhesion receptor (Cd44) (Fig.2E, Supplementary Table 2). In contrast, TREM2- macrophages primarily upregulated ribosomal proteins (Rps2, Rps6, Rps12, and Rpsa), the inflammatory chemokine (Ccl3), and the complement pathway component (C11bp) (Fig.2E, Supplementary Table 2). Pathway analysis revealed that TREM2+ macrophages primarily enhanced pathways related to antigen processing and presentation, phagosome formation, chemokine signaling, NF-κB activation, Th17 cell differentiation, cytokine-cytokine receptor interaction, and TNF signaling—all crucial for antifungal immunity (Fig.2E). On the other hand, TREM2- macrophages upregulated pathways associated with disease states, not involving antifungal immune responses (Fig.2E). Other antifungal immune cells, such as neutrophils, were enriched with pathways like phagosome formation, neutrophil extracellular trap formation, IL-17 signaling, C-type lectin receptor signaling, TNF signaling, cytokine-cytokine receptor interaction, and NF-κB signaling (Fig.2E). T cells exhibited enrichment in pathways vital to their function, including Th1 and Th2 cell differentiation, Th17 cell differentiation, and T cell receptor signaling (Fig.2E), reflecting the reliability of our cell clustering and functional analysis.
Further analysis revealed that under normal physiological conditions, TREM2+ macrophages showed gene oncology (GO) functional enrichment in inflammatory response, MAPK signaling, and antigen processing and presentation (Extended Data Fig.3A, Supplementary Table 3). Upon C. albicans infection, TREM2+ macrophages predominantly upregulated GO functions related to the cellular response to interleukin-1, endocytosis, cytokine-mediated signaling, and IL-17 signaling (Extended Data Fig.3B, Supplementary Table 3), indicating a pro-inflammatory role even prior to infection and promoting fungal clearance through activation of IL-17 signaling after C. albicans infection.
The upregulation of various chemokines, inflammatory mediators, and receptors in TREM2+ macrophages, coupled with the enhanced cytokine-cytokine receptor interaction pathway (Fig. 2E), suggested extensive interactions with other immune cells. Indeed, strong intercellular interactions were observed between TREM2+ and TREM2- macrophages, T cells, and neutrophils, with TREM2+ macrophages also engaging in self-regulation (Extended Data Fig.3C). These interactions predominantly involved signals through CCL, APP, CXCL, TNF, GRN, and VISFATIN (Extended Data Fig.3D). Notably, the TNF signaling was exclusively generated by TREM2+ macrophages and impacted neutrophils, T cells, TREM2+ and TREM2- macrophages (Extended Data Fig.3D). GSEA analysis further identified a significant upregulation of TNF signaling in TREM2+ macrophages compared to other immune cells (Extended Data Fig.3E), with only TREM2+ macrophages showing high expression of Tnf, and the TNF receptor being expressed in neutrophils, T cells, and both TREM2+ and TREM2- macrophages (Extended Data Fig.3F), aligning with the cell interaction analysis of TNF signaling regulation in these immune cells. Correlation analysis substantiated a significant positive association of TREM2 with TNF expression and TNF signaling pathways in both humans and mice (Extended Data Fig.3G), preliminarily confirming the unique role of TREM2+ macrophages in regulating TNF signaling.
Regulatory influence of TREM2 on macrophage TNF-α secretion
Our bioinformatics analysis has confirmed TREM2 as a potential regulator of TNF (Extended Data Fig.3G). To determine the relationship between TREM2 and TNF-α, we first examined the co-localization of TREM2+ macrophages with TNF-α. Our findings revealed that TREM2+ macrophages were capable of secreting TNF-α and were localized to the site of infection in OPC (Extended Data Fig.4 A). Flow cytometry analysis also indicated a significant infiltration of TREM2+ TNF-α+ macrophages in the infected tongue tissue, whereas the infiltration of TREM2- TNF-α+ macrophages showed no difference (Extended Data Fig.4B). Consistently, Ly6C+ TREM2+ macrophages were significantly more prevalent in the infected tongue tissue, with an increase in Ly6C+ TREM2+ TNF-α+ macrophages at the infection site, while there was no disparity observed with Ly6C- TREM2+ TNF-α+ and TREM2- TNF-α+ macrophages (Extended Data Fig.4C, D). Further analysis of TNF-α in tongue tissues revealed that the absence of Trem2 led to a significant decrease in TNF-α in OPC tongue tissues (Extended Data Fig.4E). TNF-α is primarily secreted by macrophages, T cells, and B cells, which also inhabit the immunological microenvironment of the mouse tongue (Fig.1D). To ascertain which cells were affected by the Trem2 deficiency leading to reduced TNF-α in OPC tongue tissues, we analyzed the secretion of TNF-α by these cells before and after Trem2 knockout. The results indicated that the absence of Trem2 mainly affected the secretion of TNF-α by macrophages (Extended Data Fig.4F), significantly reducing the infiltration of TNF-α-secreting macrophages at the infection site (Extended Data Fig.4G), while B cells, CD4+ T cells, and CD8+ T cells showed no difference in TNF-α secretion (Extended Data Fig.5A).
To investigate how TREM2 affects the secretion of TNF-α by macrophages, we first studied whether there were differences in macrophage infiltration levels. We found that Trem2 knockout significantly reduced the number of macrophages at the infection site (Extended Data Fig.5B), corresponding with a notable decrease in macrophages at the infection site (Extended Data Fig.4G). Previous research suggests that TREM2 is critical for promoting microglia proliferation and inhibiting apoptosis [12, 38-40]. We hypothesized that the significant reduction in macrophages following Trem2 knockout could be due to inhibited proliferation and activated apoptosis. To test this hypothesis, we examined the impact of TREM2 on macrophage proliferation and apoptosis, finding that Trem2 knockout significantly inhibited macrophage proliferation in OPC (Extended Data Fig.5C) and promoted apoptosis (Extended Data Fig.5D). Immunofluorescence results also showed a propensity for promoted apoptosis and inhibited proliferation in the OPC infection sites lacking Trem2 (Extended Data Fig.5E). In summary, our study has identified that TREM2 influences the proliferation and apoptosis of macrophages in OPC tissues, which significantly reduces the number of TNF-α-secreting macrophages, thereby decreasing the levels of TNF-α within the tissue.
Regulation of TNF-α secretion by TREM2 confers protection in OPC
TNF-α plays a pivotal role in systemic fungal infections, and studies suggest that deleting Tnf in vivo increases susceptibility to OPC in mice [41]. We administered TNF-α monoclonal antibodies to WT mice, which did not affect their body weight (Extended Data Fig.6A), but resulted in significantly lower body weight on days 1 and 2 post-C. albicans infection (Extended Data Fig.6B), along with increased fungal burden (Extended Data Fig.6C), and more severe epithelial invasion (Extended Data Fig.6D). These findings implied that the decrease in TNF-α levels in tissues due to Trem2 deficiency may be a contributing factor to increased susceptibility to OPC. To test this hypothesis, we restored TNF-α in the Trem2-deficient OPC model and observed that although this restoration did not significantly reverse the weight loss caused by Trem2 deficiency (Extended Data Fig.6E), it completely reversed the susceptibility to OPC (Extended Data Fig.6F, G). This evidence corroborated that TREM2 exerted a protective role in OPC by modulating the secretion of TNF-α.
Modulation of macrophage inflammatory response by TREM2-mediated TNF-α secretion
Inflammatory cytokines, chemokines, antimicrobial peptides, and defensins are critical for the clearance of fungi, recruitment of immune cells, and regulation of their functions. TNF-α serves as a potential regulator of inflammatory cytokines, and its deficiency leads to a marked suppression of inflammatory activity. In our in vitro co-culture of BMDMs with C. albicans, we discovered that the absence of Trem2 significantly inhibited the expression of Tnf, Il6, Il23, S100a8, S100a9, Defb1, Cxcl1, Cxcl2, Ccl2, Ccl7, and S100a7a (Extended Data Fig.6H and Extended Data Fig.7A). Exogenous supplementation of TNF-α was able to reverse these differences (Extended Data Fig.6H and Extended Data Fig.7A). The expression of these genes is primarily regulated by inflammatory pathways, and we found that Trem2 deficiency notably inhibited the activation of NF-κB's P65 subunit and the degradation of IκBα (Extended Data Fig.7B), as well as the activation of Syk (Extended Data Fig.7B), and suppressed the activity of the MAPK pathway's P38 (Extended Data Fig.7C). The supplementation of exogenous TNF-α reversed these deficiencies. Although PI3k/Akt is a downstream signal of TREM2, in the context of C. albicans infection, TREM2 did not rely on this pathway (Extended Data Fig.7D). In conclusion, we have elucidated that TREM2 is capable of modulating the secretion of TNF-α, thereby amplifying the inflammatory signaling response in the context of C. albicans infection.
Regulation of Th17 cell differentiation by TREM2 via TNF-α
IL-17 is essential for combating OPC, predominantly produced by Th17 cells, innate lymphoid cells (ILCs), and γδT cells. The absence of Trem2 suppressed the expression of Tnf, Il6, and Il23, cytokines that induce Th17 differentiation [42-44], leading us to postulate that Trem2 deficiency may also affect Th17 cells in OPC. To verify this hypothesis, we assessed the immune cells secreting IL-17, IL-17F, and IL-22 in the tongues of mice. The results showed that Trem2 deficiency significantly inhibited the proportion of Th17 cells without affecting other IL-17-secreting immune cells (Extended Data Fig.8A), indicating a specific impact of TREM2 on Th17 cells. Trem2 deficiency also significantly reduced the proportions of IL-17F-producing Th17, CD4- αβT cells, and ILCs (Extended Data Fig.8B), and even the secretion of IL-22 by Th22 and CD4- αβT cells was notably suppressed (Extended Data Fig.8C). Also, in contrast to WT, Trem2-/- mice demonstrated a markedly diminished secretion of IL-17 in their lingual supernatant (Extended Data Fig.7E). Interestingly, Trem2 deficiency unexpectedly resulted in increased neutrophil infiltration (Extended Data Fig.8D), possibly due to the increased fungal burden, as systemic infection studies have found that more C. albicans can lead to increased neutrophil infiltration, with neutrophil levels seemingly scaling with fungal load [45, 46]. Trem2 deficiency also significantly suppressed Th1 infiltration (Extended Data Fig.8E), while IFN-γ is not essential in OPC since mice lacking IFN-γ or the IL-12 subunit IL-12p35 do not exhibit increased susceptibility to OPC [47, 48]. Trem2 deficiency did not affect the infiltration of dendritic cells (DCs), eosinophils, or mast cells at the site of infection (Extended Data Fig.8F).
Subsequent studies revealed that administration of TNF-α monoclonal antibodies markedly reduced the populations of IL-17-producing Th17 cells (Extended Data Fig.8G), Th17 cells secreting IL-17F (Extended Data Fig.8H), as well as Th22 and CD4- αβT cells that produce IL-22 in OPC (Extended Data Fig.8I). To conclusively validate that TREM2 influences Th17 cell differentiation through TNF-α, we undertook a co-culture experiment. Naïve CD4+ T cells were incubated with supernatant derived from BMDMs stimulated with C. albicans. We found that Trem2 knockout significantly inhibited Th17 cell differentiation (Extended Data Fig.9A) and the secretion of IL-17F by CD4+ T cells (Extended Data Fig.9B), which could be reversed with exogenous TNF-α. In summary, the data confirmed that TREM2 modulated the secretion of TNF-α, impacting Th17 differentiation to regulate IL-17 signaling.
TREM2 regulates macrophage and neutrophil killing abilities through TNF-α
Macrophages contribute to the clearance of C. albicans through phagocytosis, killing, and ROS production, while neutrophils also produce NETs to kill fungi. TNF-α regulates these functions of macrophages and neutrophils [49, 50]. The single-cell analysis also indicated that TNF-α secreted by TREM2+ macrophages acted upon both macrophages and neutrophils (Extended Data Fig.3C, D). Indeed, Trem2 deficiency significantly inhibited macrophage phagocytosis (Extended Data Fig.9C), killing ability (Extended Data Fig.9D), and ROS production (Extended Data Fig.9G), with exogenous TNF-α reversing these functional deficits. To clarify if Trem2 deficiency attenuated the killing ability of neutrophils due to decreased TNF-α, we co-cultured neutrophils with supernatant obtained from BMDMs that had been stimulated with C. albicans. Consistent with the regulation of macrophage functions by TNF-α, we found that Trem2 knockout inhibited neutrophil phagocytosis of C. albicans (Extended Data Fig.9E), killing (Extended Data Fig.9F), and ROS production (Extended Data Fig.9G). Trem2 knockout also significantly inhibited NETs at the site of infection (Extended Data Fig.9H), a deficit also observed in tongue tissues treated with TNF-α monoclonal antibodies (Extended Data Fig.9I). The NETs deficiency caused by Trem2 knockout could be reversed by supplementing TNF-α (Extended Data Fig.9J). In the in vitro setting, neutrophils co-cultured with supernatant from BMDMs demonstrated a significant inhibition in the production of NETs following Trem2 knockout. This reduction in NETs production could be counteracted by the addition of exogenous TNF-α (Extended Data Fig.9K). In conclusion, we demonstrated that TREM2 influenced the fungicidal functions of macrophages and neutrophils by regulating TNF-α.
TREM2 recognition of candidalysin initiates TNF Signaling
TREM2, as a receptor, is capable of recognizing a variety of ligands, including a wide range of anionic molecules that are free, can bind to cell membranes, and include components from both Gram-positive and Gram-negative bacteria (such as Neisseria Gonorrhoeae, Escherichia coli, and Staphylococcus aureus), DNA, lipoproteins, and phospholipids [12, 51]. Building on our findings, we observed that Trem2 deficiency presented similar impairments to PRR deficiencies, leading to compromised antifungal inflammatory responses, including signal pathway deficiencies (Extended Data Fig.6H and Extended Data Fig.7A-C). This led us to propose that TREM2 might be involved in the recognition of C. albicans. To test this hypothesis, we co-incubated the extracellular segment of TREM2 with both spore and hyphal forms of C. albicans and found that both human and mouse TREM2 recognized only the hyphal form of the fungus, not the yeast form (Fig.3A, B), suggesting the presence of PAMPs on the hyphae recognized by TREM2. To determine the specific PAMPs recognized by TREM2, we incubated its extracellular domain with various PAMPs. Previous research has identified that TREM2 can recognize sphingolipid [52], which are abundantly present on the cell membrane of fungi. Here, we have also purified C. albicans sphingolipid to investigate whether TREM2 is involved in the recognition of fungal surface sphingolipids. Finally, we discovered that human and mouse TREM2 could engage with multiple PAMPs (LPS, a known ligand of TREM2, served as a positive control in the study) but mainly recognized candidalysin (Fig.3C, D), the first acknowledged classical virulence factor of C. albicans [53, 54]. After infection, C. albicans forms hyphae, which induce the expression of the ECE1 gene, coding for Ece1p [55, 56]. Ece1p is then processed by Kex2p to produce immature candidalysin, which is finally secreted by the hyphae after Kex1p removes the terminal R93 to generate mature candidalysin (SIIGIIMGILGNIPQVIQIIMSIVKAFKGNK) [54, 57, 58]. This revealed why TREM2 only recognized the hyphal form of C. albicans and not the yeast form. Immunofluorescence also showed that candidalysin could co-localize with TREM2 on THP-1 cells (Fig.3E), and overexpression of TREM2 in 293T cells with candidalysin confirmed that TREM2 could bind to candidalysin (Fig.3F). In the absence of Trem2, the recognition of candidalysin by BMDMs was significantly inhibited (Extended Data Fig.10A). These multifaceted results affirmed that TREM2 could directly bind and recognize candidalysin.
When PAMPs bind to their receptors, they induce an inflammatory signal response. Does the recognition of candidalysin by TREM2 also trigger such signaling activation? To address this question, we stimulated BMDMs with candidalysin and found that it could induce the activation of P65, but this activation was suppressed in the absence of Trem2 (Extended Data Fig.10B, C), with the regulation of Tnf expression consistent with P65 activity (Extended Data Fig.10D). Other PAMPs, only zymosan had a similar effect (Extended Data Fig.10E), indicating that TREM2 could recognize candidalysin and zymosan to provoke an inflammatory response.
Given the significant role of candidalysin and the primary recognition of it by TREM2 (Fig. 3A-D), we investigated the specific binding sites of TREM2 for candidalysin. Employing AlphaFold, DeepMind's AI-powered tool for predicting protein 3D structures from amino acid sequences, we modeled the interaction between the extracellular domain of human TREM2 and the fungal toxin candidalysin. Following this, protein docking with haddock 2.4 identified that the highest scoring interactions (Select the intersection of the top-scoring results from two rounds of molecular docking) showed TREM2 binding to candidalysin's G65, N73, and N91-N92 (located on a loop structure) (Extended Data Fig.10F). To confirm the specificity of TREM2 for these sites, we mutated or deleted them (Extended Data Fig.10G), and found that mutations and deleting the loop site significantly inhibited the recognition of candidalysin by human and mouse TREM2 (Extended Data Fig.10H). The mutations at these amino acid sites do not alter the early activity of the early NF-kB signaling pathway (Extended Data Fig.10I); however, as time progresses, there is a significant suppression of the NF-kB signal (Extended Data Fig.10J), suggesting that these site mutations lead to an inability of TREM2 to form a stable, long-term binding structure with candidalysin. Candidalysin's identified sites correspond to TREM2's D131, R136, and P169 (Extended Data Fig.10F). Mutating these sites in TREM2 and then overexpressing it in 293T cells with candidalysin showed that the mutations in TREM2's D131, R136, and P169 significantly inhibited the recognition of candidalysin by TREM2 (Fig.4A-C). It is noteworthy that mutations at the N73, G65, and N91-N92 residues of candidalysin markedly impaired its interaction with TREM2 (Fig.4A-C). This finding underscores the critical role these residues play in the molecular recognition processes of TREM2. Together these findings ascertain that the residues D131, R136, and P169 of TREM2 are involved in recognizing the candidalysin sites G65, N73, and N91-K92.
TREM2 activates NF-κB via the DAP12/Syk axis to regulate the secretion of TNF-α
TREM2 recognized its ligand, leading to the recruitment and activation of the adaptor proteins DAP10 and DAP12 [15]. This process activated downstream signaling pathways, notably Syk and PI3k/Akt [15]. Previous research had established that DAP12 primarily activates the Syk pathway, while DAP10 is associated with activating PI3k/Akt [12, 59]. In our study, we found that TREM2's recognition of C. albicans predominantly stimulated the Syk pathway (Extended Data Fig.7B), but not the PI3k/Akt pathway (Extended Data Fig.7D), suggesting an interaction between TREM2 and candidalysin via DAP12. Our experimental findings supported this hypothesis. Upon stimulation with candidalysin, there was a significant aggregation of DAP12 in the vicinity of TREM2, unlike DAP10, which did not show a notable increase (Fig.4D, E). This indicated that TREM2's engagement with candidalysin primarily involved DAP12, leading to downstream signaling activation. Moreover, selectively diminishing Dap12 expression, rather than Dap10, significantly decreased the Syk activation triggered by candidalysin (Extended Data Fig.11A). Given that Syk acts upstream of NF-κB [60, 61], reducing Dap12 expression or directly inhibiting Syk significantly hindered the activation of the NF-κB subunit P65 (Extended Data Fig.11A, B). NF-κB is known for its critical role in the regulation of inflammatory mediators including TNF-α, IL-6, chemokines, and antimicrobial proteins during infection [62]. The knockdown of Dap12 or the inhibition of Syk both significantly suppressed the expression of Tnf (Extended Data Fig.11C, D). Thus, we concluded that TREM2's recognition of candidalysin prompted an immune response through the DAP12/Syk axis, which then activated the NF-κB signaling pathway. This activation regulated TNF-α secretion, orchestrating an effective inflammatory and immune response in the context of OPC.
TREM2 agonist Hsp60 exhibits therapeutic effects in CA-induced OPC
CA-induced OPC is commonly seen in clinical settings, and our scRNA-seq data also indicated a significant reduction in the proportion of TREM2+ macrophages in the CA-induced OPC model (Fig.1F), suggesting that CA might suppress TREM2+ macrophages, contributing to OPC susceptibility. Indeed, in the CA-induced OPC model, we observed fewer TREM2+ TNF-α+ macrophages at the infection site on day 1 (Extended Data Fig.12A), and by day 3, these cells were completely undetectable (Extended Data Fig.12B). Flow cytometric analysis also showed that CA inhibited Ly6C+ TREM2+ macrophages (Extended Data Fig.12C) and Ly6C+ TREM2+ TNF-α+ macrophages (Extended Data Fig.12D).
Hsp60, as a specific TREM2 ligand agonist used to stimulate TREM2 activity [63-65], may have therapeutic potential in OPC. Although CA significantly reduced the number of TREM2+ macrophages (Extended Data Fig.12A-D), a subset of TREM2+ macrophages remained in the lingual tissue (Extended Data Fig.12C, D), providing a basis for using Hsp60 to stimulate TREM2 as a treatment for CA-induced OPC. We explored the therapeutic effects of Hsp60, TNF-α, and soluble TREM2 (sTREM2) in a CA-induced OPC model. The results showed no difference in body weight between the mice on days 1 and 2 of the model (Extended Data Fig.12E), but on day 3, mice treated with Hsp60 had higher body weight compared to controls (Extended Data Fig.12F), suggesting potential phenotypic differences at this time point. On day 3, the Hsp60 and TNF-α treatment groups exhibited lower fungal burdens and milder C. albicans epithelial invasion and damage (Extended Data Fig.12G, H). However, we did not observe a therapeutic effect from sTREM2 (Extended Data Fig.12E-H), and in vitro stimulation of BMDMs with sTREM2 did not reverse the suppression of the NF-kB pathway caused by Trem2 deficiency (Extended Data Fig.13A, B), reflecting that TREM2 mainly exerts its function through receptor-mediated inflammation rather than in a soluble form. To confirm that Hsp60 mediated its therapeutic effects through TREM2, we treated Trem2-deficient BMDMs with Hsp60 in vitro and found that Hsp60 could enhance the expression of Tnf in WT BMDMs, but this enhancement was ineffective in the absence of Trem2 (Extended Data Fig.13C). Similarly, we found that Hsp60 induced Th17 differentiation (Extended Data Fig.13D) and enhanced macrophage phagocytosis and killing of C. albicans (Extended Data Fig.13E, F) through TREM2. In conclusion, we have determined that the TREM2 agonist Hsp60 can have therapeutic effects in CA-induced OPC by activating TREM2 function.