Secretogranin 2 binds LILRB4 resulting in immunosuppression

Xing Yang; Ryan Huang; Meng Fang; Yubo He; Jingjing Xie; Xiaoye Liu; Chengcheng Zhang; Qi Lou; Mi Deng; Wei Xiong; Cheryl Lewis; Zade Sadek; Ankit Gupta; Lianqi Chen; Xuewu Zhang; Lei Guo; Lin Xu; Ningyan Zhang; Zhiqiang An; Cheng Cheng Zhang

doi:10.1038/s41590-025-02233-4

. Author manuscript; available in PMC: 2026 Jan 24.

Published in final edited form as: Nat Immunol. 2025 Jul 24;26(9):1567–1580. doi: 10.1038/s41590-025-02233-4

Secretogranin 2 binds LILRB4 resulting in immunosuppression

Xing Yang ¹, Ryan Huang ¹, Meng Fang ¹, Yubo He ¹, Jingjing Xie ¹, Xiaoye Liu ¹, Chengcheng Zhang ¹, Qi Lou ¹, Mi Deng ¹, Wei Xiong ², Cheryl Lewis ³, Zade Sadek ¹, Ankit Gupta ⁴, Lianqi Chen ⁵, Xuewu Zhang ⁵, Lei Guo ⁶, Lin Xu ⁶, Ningyan Zhang ², Zhiqiang An ², Cheng Cheng Zhang ^1,^✉

PMCID: PMC12414541 NIHMSID: NIHMS2106102 PMID: 40707822

Abstract

Immunosuppressive myeloid cells are important in a variety of physiological and pathological contexts, including tumor development, but how hormones might regulate their activity is unclear. Secretogranins, a family of secretory proteins in endocrine and neuronal cells, are proposed to function as prohormones or hormones, but their specific receptors are unknown. Here we show that secretogranin 2 (SCG2), a granin family member, functionally interacts with leukocyte immunoglobulin-like receptor B4 (LILRB4) on monocytic cells. Tumor-derived SCG2 promotes tumor growth in myeloid-specific LILRB4 transgenic mice in a T cell-dependent manner, whereas SCG2 deficiency in host mice impairs tumor progression and reduces infiltration of immunosuppressive monocytic cells. Blockade of LILRB4 abrogates SCG2-induced signaling, immunosuppression and tumor growth. Mechanistically, this SCG2–LILRB4 interaction triggers SHP recruitment and SHP-independent STAT3 activation. These findings define a function for SCG2 in regulating monocytic immunosuppression and suggest that the SCG2–LILRB4 axis might be a therapeutic target.

Breakthroughs in immune checkpoint blockade therapy have revolutionized cancer treatment. However, existing T cell-based immunotherapies, including PD-1/PD-L1 blockade, are only effective in approximately 10–30% of individuals with cancer^1,2. The mechanisms of immune surveillance repression in the tumor microenvironment (TME) of advanced human cancers are highly heterogeneous³. Heterogeneous immunosuppressive myeloid cells, including myeloid-derived suppressor cells (MDSCs), tumor-associated macrophages and a number of other types of myeloid cells, mediate immunosuppression in the TME. Reprogramming immunosuppressive myeloid cells is becoming an attractive strategy for cancer treatment and combating therapeutic resistance⁴.

Hormones coordinate physiological functions across tissues, but their influence on immunosuppressive myeloid cells remains poorly defined. Secretogranin 2 (SCG2; chromogranin C), a granin family protein produced by endocrine and neuronal cells, contributes to secretory granule formation and may act as a prohormone or hormone, although its receptor has been unknown^5–7. SCG2 was detected in full-length form or as cleaved peptides in different tissues⁸, and its fragments have been implicated in neuroinflammation, neurogenesis, immune response, energy homeostasis and reproduction^5,6,9. SCG2 is overexpressed in multiple cancer cells⁶ and was identified as a biomarker of breast cancer¹⁰, colorectal cancer^11,12, bladder cancer¹³, renal cell carcinoma¹⁴, non-small cell lung cancer¹⁵ and other cancers^6,16. Its expression is associated with the immunosuppressive myeloid polarization of the TME and poor survival^10,12,16. Deficiencies in SCG2 decrease proliferation, migration and invasion of tumor cells¹¹. Furthermore, SCG2 was suggested to be a candidate serum biomarker for multiple sclerosis, mild cognitive impairment, prestage Alzheimer’s disease and neurodevelopmental disorders⁶.

The leukocyte immunoglobulin-like receptor subfamily B (LILRB) proteins, which are expressed mainly on hematopoietic cells, are type 1 transmembrane receptors that contain immunoreceptor tyrosine-based inhibitory motifs (ITIMs)¹⁷. Numerous studies have contributed to our current understanding of LILRB biology^17–22, suggesting that LILRBs represent a new group of myeloid checkpoint targets. LILRB4, which is mainly expressed on human monocytic cells, including monocytic MDSCs (M-MDSCs)^23,24, is the most restrictively expressed member of the LILRB family. LILRB4 and gp49b, its mouse relative with a different expression pattern and certain distinct ligands²¹, were shown to play a role in support of the immunosuppressive TME and tumor development^24–27. Previously identified extracellular binding proteins of LILRB4, including ApoE²⁴, fibronectin^28,29 and ciliary neurotrophic factor receptor(CNTFR)³⁰, can also bind to other receptors. It was unclear whether there exists a specific functional ligand that induces LILRB4 signaling in the TME of solid tumors and how inhibitory receptor signaling regulates the immunosuppressive activity of myeloid cells.

Here, we identify SCG2 as a ligand of LILRB4 that promotes immunosuppressive signaling in monocytic cells. SCG2–LILRB4 engagement drives SHP recruitment and SHP-independent STAT3 activation, enhances monocytic cell-mediated suppression and supports tumor progression. These findings identify a hormone–checkpoint axis that orchestrates myeloid immunosuppression in the TME.

Results

SCG2 interacts specifically with LILRB4

Although SCG2 is associated with immunomodulatory effects¹², how it regulates immune activity is unknown.SCG2 was identified as a potential LILRB4 binding protein through a genome-wide screening of ~6,500 human membrane and secreted proteins (Extended Data Fig. 1a). Concordantly, we observed that SCG2 binds to human monocytes but not T cells, B cells, natural killer (NK) cells or neutrophils (Fig. 1a,b and Extended Data Fig. 1b). SCG2 no longer interacted with THP-1 cells, which are a monocyte line, after deletion of LILRB4 (Fig. 1c,d). Using LILRB chimeric reporter cells^31,32, we confirmed the SCG2–LILRB4 interaction as well as previously reported interactions between LILRB4 and ApoE²⁴, fibronectin^28,29 and CNTFR³⁰, but not CD166 (ref. 33) or galectin 8 (ref. 34; Fig. 1e). SCG2 is highly conserved (82% identity between human and mouse), and both human and mouse soluble SCG2 activated LILRB4 reporter cells, an effect blocked by a described antibody that blocks LILRB4 (ref. 24; Extended Data Fig. 1c,d).

Fig. 1 | — **a,b,** Flow cytometry assay of SCG2 binding to human monocytes. Human peripheral blood cells (5 × 10⁶) were incubated with recombinant SCG2–His protein (100 μg ml⁻¹) for 1 h at 4 °C. a, Representative flow cytometry histograms. b, Quantification of the percentage of SCG2 binding cells (n = 3 biological replicates); Ctrl, control; FMO, fluorescence minus one. **c,d**, Flow cytometry assay of SCG2 binding to THP-1 cells. WT or *LILRB4*-KO THP-1 cells (1 × 10⁶) were incubated with or without recombinant SCG2-His protein (100 μg ml⁻¹) for 1 h at 4 °C. c, Representative flow cytometry histograms. d, Quantification of the percentage of SCG2 binding cells (n = 3 biological replicates). e, LILRB4 reporter cells induced by the indicated proteins. LILRB4 reporter cells (5 × 10⁴) were seeded on 96-well plates precoated with serial dilutions of the indicated proteins (starting at 20 μg ml⁻¹) for 24 h. Green fluorescent protein-positive (GFP⁺) cells were quantified by flow cytometry (n = 3 biological replicates); EC₅₀, half-maximum effective concentration. f, ELISA-based binding assay. The indicated proteins were serially diluted and coated onto ELISA plates overnight. Binding was assessed using 1 μg ml⁻¹ purified LILRB4-ECD–Fc protein, followed by detection with anti-human Fc–horseradish peroxidase (Fc–HRP) secondary antibody (n = 3 biological replicates); OD₄₅₀, optical density at 450 nm. g, Co-IP assay of the interaction between LILRB4-ECD–Fc and SCG2–Flag in HEK293T cells; WCL, whole-cell lysate. h, Pulldown assay between SCG2 and LILRB4-ECD. i, BioLayer interferometry (Octet) measurement of the binding kinetics between SCG2–His and immobilized LILRB4-ECD, fit to a 1:1 model. j, HEK293T reverse binding assays. HEK293T cells were transfected with plasmids encoding membrane-anchored SCG2 fusion proteins for 48 h. Cells (2 × 10⁵) were collected and incubated with purified recombinant LILRB4–Fc (25 μg ml⁻¹) for 1 h at 4 °C. Binding was detected by flow cytometry; SN, secretoneurin (one of the bioactive peptides of SCG2); CT, C terminus of SCG2 (amino acids 283–617 of full-length (FL) SCG2); NT, N terminus of SCG2 (amino acids 31–179 of full-length (FL) SCG2). k, Mapping of LILRB4 domains required for SCG2 binding. HEK293T cells were co-transfected with LILRB4 domain deletion plasmids containing a Flag tag and SCG2 plasmids. Interactions were assessed by Co-IP; dD1, D1 deletion; dD2, D2 deletion. l,m, SCG2 selectively activates LILRB4. LILRB family or LAIR1 reporter cells (RC) were stimulated with plate-bound SCG2 or antibodies (10 μg ml⁻¹) for 24 h (l; n = 2 biological replicates). LILRB4 reporter cells were stimulated with the indicated plate-bound granin proteins (10 μg ml⁻¹) for 24 h (m; n = 3 biological replicates). GFP⁺ cells were quantified by flow cytometry. All data represent three independent experiments and are presented as mean ± s.d.

We characterized SCG2–LILRB4 binding affinity using multiple assays. In enzyme-linked immunosorbent assays (ELISAs), SCG2 showed a stronger affinity for LILRB4 (half-maximum effective concentration = 0.243 μg ml⁻¹) than ApoE (1.361 μg ml⁻¹), fibronectin (0.356 μg ml⁻¹) or CNTFR (0.848 μg ml⁻¹; Fig. 1f). Co-immunoprecipitation (Co-IP) in HEK293T cells and pulldown assays with purified proteins confirmed the interaction between SCG2 and the extracellular domain (ECD) of LILRB4 (Fig. 1g,h and Extended Data Fig. 1e). In a competition assay, SCG2 was able to displace ApoE but not CNTFR fromLILRB4 (Extended Data Fig. 1f,g), suggesting that SCG2 can outcompete ApoE binding to LILRB4. However, the addition of SCG2 further increased ApoE-induced activation of LILRB4 reporter cells (Extended Data Fig. 1h), suggesting that SCG2 and ApoE can additively activate LILRB4 under certain conditions. A biolayer interferometry assay confirmed the tight binding of both human and mouse SCG2 to LILRB4 (human SCG2 K_d of 44.52 ± 0.5181 nM, mouse SCG2 K_d of 15.32 ± 0.2974 nM; Fig. 1i and Extended Data Fig. 1i).

Molecular domain mapping by HEK293T reverse binding assays showed that the N terminus of SCG2 (amino acids 31–179 of full-length SCG2) was responsible for the interaction with LILRB4 (Fig. 1j and Extended Data Fig. 1j). The deletion of immunoglobulin-like domain 1 (D1), but not immunoglobulin-like domain 2 (D2), of LILRB4 abrogated SCG2 binding (Fig. 1k). Other LILRBs and LAIR1 reporter cells were not activated by SCG2 (Fig. 1l), and their ECDs did not bind SCG2 by flow cytometry (Extended Data Fig. 1k), consistent with co-IP analyses (Extended Data Fig. 1l). Of note, SCG2 did not activate LILRAs, mouse PIRB or gp49b reporter cells (Extended Data Fig. 1m,n). Of the five tested granin family members, SCG2 showed the highest activation of LILRB4 reporter cells (Fig. 1m and Extended Data Fig. 1o). These results indicate a specific interaction between SCG2 and LILRB4.

SCG2 induces LILRB4 activation and function in cultured cells

To determine whether SCG2 activates LILRB4 signaling, we reconstituted the signaling pathway in 293T cells by ectopically coexpressing LILRB4 with the Src kinase Lyn, SHP1 and SHP2. Under these conditions, a basal level of phosphorylation of LILRB4 was detected. The introduction of SCG2 increased LILRB4 phosphorylation and recruitment of SHP1 and SHP2, and these effects were blocked by anti-LILRB4 (Fig. 2a). To further confirm that LILRB4 activation is mediated by SCG2 binding, we screened LILRB4 mutants deficient in SCG2 interaction. The D1 domain of LILRB4 mediates its binding to SCG2 (Fig. 1k), suggesting that critical residues within this region are responsible for the interaction. To identify these residues, we constructed a series of LILRB4 truncation mutants to map the SCG2 binding region. Among these, deletion of amino acids 75–84, corresponding to the fourth β-sheet, abolished the interaction with SCG2. Based on this result, we generated single-amino-acid mutations within this region to further define key residues. Notably, we identified that a single F83A mutation in LILRB4 prevented the interaction with SCG2 (Extended Data Fig. 2a–c). This result was further confirmed using LILRB4-F83A reporter cells, which were not activated by SCG2 (Extended Data Fig. 2d). SCG2 did not induce LILRB4-F83A phosphorylation or the recruitment of SHP1 and SHP2 to the mutant receptor (Fig. 2b and Extended Data Fig. 2e). Mutation of the second ITIM of LILRB4 disrupted the ability of SCG2 to elicit downstream SHP1, SHP2 and LILRB4 phosphorylation (Fig. 2c). Previous work showed that coligation of LILRB4 suppresses FcγRI signaling³⁵. Using THP-1 cells expressing Flag–LILRB4, we confirmed that coated SCG2 or anti-LILRB4 suppressed CD64-induced interleukin-8 (IL-8) production, accompanied by increased LILRB4 and SHP phosphorylation (Fig. 2d,e). By contrast, SCG2 and ApoE failed to alter IL-8 secretion inLILRB4-knockout (LILRB4-KO) THP-1 cells, and anti-LILRB4 restored IL-8 in wild-type (WT) cells (Fig. 2e,f), confirming that SCG2–LILRB4 interaction suppresses FcγRI signaling.

Fig. 2 | — a, SCG2 induces phosphorylation of LILRB4 and recruitment of SHP1/SHP2 in a recombinant HEK293T system. HEK293T cells were co-transfected with LILRB4–Flag (2 μg), SHP1 (300 ng), SHP2 (300 ng), Lyn (100 ng) and SCG2 (300 ng), as indicated. IgG1 or anti-LILRB4 (10 μg ml⁻¹) was added as indicated. At 48 h after transfection, cells were lysed and subjected to IP and immunoblotting with the indicated antibodies. b, SCG2 fails to induce phosphorylation of the LILRB4–F83A mutant. HEK293T cells were co-transfected with the LILRB4–F83A–Flag mutant (2 μg) in place of LILRB4–Flag, along with the same components as in a. Cells were lysed 48 h post-transfection and analyzed by IP and immunoblotting. c, SCG2 induced phosphorylation of the ITIM of LILRB4 in HEK293T cells. HEK293T cells were co-transfected as in a and b, and lysates were analyzed by IP and immunoblotting. d, SCG2 activated endogenous LILRB4 in THP-1 cells when coligated with FcγRI. THP-1 cells stably expressing LILRB4–Flag were incubated in 96-well plates coated with SCG2 (10 μg ml⁻¹) or anti-LILRB4 (10 μg ml⁻¹), along with anti-CD64 (10 μg ml⁻¹), for 10 min. Lysates were analyzed by IP and immunoblotting. **e,f,** SCG2 inhibited IL-8 production downstream of FcγRI engagement. e, WT or *LILRB4*-KO THP-1 cells were incubated in 96-well plates coated with SCG2 or ApoE (10 μg ml⁻¹), together with anti-CD64 (10 μg ml⁻¹) for 24 h. f, WT THP-1 cells treated as in e, with the addition of anti-LILRB4 (10 μg ml⁻¹) as indicated. Supernatants were collected and analyzed by ELISA (n = 3 biological replicates). All data represent three independent experiments and are presented as mean ± s.d.

SCG2 supports tumor development through LILRB4

SCG2 expression was negatively correlated with cytotoxic T lymphocyte infiltration and survival, while showing a positive correlation with the infiltration of MDSCs within the same cohorts across multiple human tumor types (Extended Data Fig. 3a–c), suggesting an association with immunosuppression of the TME. To further study the functional interaction between SCG2 and LILRB4 in vivo, we generated LILRB4 transgenic mice (LILRB4-Tg) by crossing Rosa26-hLILRB4 knock-in mice with LysM-Cre mice, restricting human LILRB4 expression to myeloid cells (Extended Data Fig. 4a). First, we confirmed that transgenic LILRB4 was expressed on bone marrow (BM)-derived macrophages (BMDMs) and myeloid cells from LILRB4-Tg mice (Extended Data Fig. 4b–e). Second, the expression of cell markers on myeloid cells from LILRB4-Tg mice was comparable with that from WT mice (Extended Data Fig. 4f). Furthermore, the percentages of myeloid cells, T cells, regulatory T cells and NK cells were not changed (Extended Data Fig. 4g–j). This suggests that the expression of human LILRB4 in mice does not affect the immune system under physiological conditions. SCG2-overexpressing tumor cell lines(MC38, B16F10 and LLC) were subcutaneously injected into WT and LILRB4-Tg mice. Tumors developed faster in LILRB4-Tg mice than in WT mice, which indicates that human LILRB4 supports tumor development in mice. Importantly, SCG2-expressing tumors grew faster than control vector tumors in LILRB4-Tg mice but not in WT mice (Fig. 3a–c and Extended Data Fig. 5a,b). Ectopic expression of mouse SCG2 in MC38 cells also increased tumor growth in LILRB4-Tg mice (Extended Data Fig. 5c), consistent with the ability of mouse SCG2 to activate LILRB4 reporter cells (Extended Data Fig. 1c,d). Moreover, injection of recombinant SCG2 protein into palpable MC38 tumors in LILRB4-Tg mice enhanced tumor growth (Fig. 3d and Extended Data Fig. 5d). In a metastasis model, intravenous injection of SCG2-expressing B16F10 cells enhanced tumor metastasis in LILRB4-Tg mice but not WT mice (Fig. 3e,f). Finally, anti-LILRB4 blocked the ability of SCG2 to promote tumor growth in vivo (Fig. 3g and Extended Data Fig. 5e). Levels of LILRB4 expressed on MDSCs were decreased after anti-LILRB4 treatment (Extended Data Fig. 5f), which is consistent with a previous observation³⁶. These results indicated that SCG2 supports tumor development in an LILRB4-dependent manner.

Fig. 3 | — a–c, Tumor growth of SCG2-expressing or vector tumor cells in *LILRB4*-Tg mice subcutaneously injected with MC38 (a; n = 7 or 9), B16F10 (b; n = 6 or 8) or LLC (c; n = 7) cells. Tumor size was measured every 2 days (same as below). d, MC38 tumor growth in *LILRB4*-Tg mice intratumorally injected with recombinant SCG2 protein (10 μg in 50 μl of PBS) or PBS (n = 7). e,f, B16F10 lung metastasis model in WT and *LILRB4*-Tg mice. Representative lung images at the experimental endpoint (e; n = 7) and survival analysis (f; n = 6 or 7) are shown; i.v., intravenous. g, MC38-Vec and MC38-SCG2 tumor growth in *LILRB4*-Tg mice treated with anti-LILRB4 and isotype control (n = 8). h, CD4⁺/CD8⁺ T cell depletion in MC38-bearing *LILRB4*-Tg mice. Tumor-bearing mice were treated with isotype control or anti-CD4/CD8 (n = 6 or 8). i, Myeloid subset cell depletion in MC38-bearing *LILRB4*-Tg mice. For G-MDSC depletion, mice were treated with anti-Ly6G (n = 5). For M-MDSC depletion, mice were treated with anti-CSF1R (n = 5). **j,k**, Immune profiling of B16F10 tumors collected from WT or *LILRB4*-Tg mice on day 15. Percentages of MDSCs (j) and CD4⁺ and CD8⁺ T cells and regulatory T (T_reg) cells (k) were measured by flow cytometry (n = 6 or 7). Data on tumor size are presented as mean ± s.e.m.; all other data are shown as mean ± s.d. P values were determined by two-way analysis of variance (ANOVA) with a Tukey’s multiple comparisons test (a–c and g–i) or a Bonferroni’s multiple comparisons test (d), one-way ANOVA with a Holm–Sidak’s multiple comparisons test (e,j and k) or log-rank (Mantel–Cox) test for survival curves (f).

To investigate the potential role of T cells, we evaluated tumor growth in LILRB4-Tg mice depleted of CD4⁺ and CD8⁺ T cells. SCG2-expressing tumors showed no growth advantage under these conditions (Fig. 3h and Extended Data Fig. 5g,h), indicating that T cells are required for SCG2–LILRB4-mediated tumor promotion. To evaluate the role of myeloid cells, we selectively depleted myeloid subsets in LILRB4-Tg mice. Monocytes and macrophages were depleted using an antibody to CSF1R, whereas neutrophils were depleted with an antibody to Ly6G.SCG2 no longer promoted tumor growth following anti-CSF1R treatment, whereas the effect persisted in mice treated with anti-Ly6G (Fig. 3i and Extended Data Fig. 5i, j). These results suggest that the tumor-promoting effect of SCG2 is predominantly mediated by M-MDSCs, rather than granulocytic MDSCs (G-MDSCs, also known as PMN-MDSCs), in LILRB4-Tg mice.

Next, we used flow cytometry to quantify the frequencies of immune cells in tumors that did or did not express SCG2 in WT and LILRB4-Tg mice. Percentages of M-MDSCs were significantly higher in LILRB4-Tg mice implanted with tumors expressing SCG2 than in other groups, whereas no differences were observed in G-MDSCs, tumor-associated macrophages, NK cells or dendritic cells (Fig. 3j and Extended Data Fig. 5k). SCG2 expression decreased the percentage of CD4⁺ and CD8⁺ T cells only in LILRB4-Tg mice, without affecting T_reg cell frequencies (Fig. 3k). Moreover, the percentage of granzyme B⁺ cells among CD8⁺ T cells was substantially decreased in LILRB4-Tg mice implanted with tumors expressing SCG2 (Extended Data Fig. 5l). Treatment with anti-LILRB4 counteracted the effects of SCG2 (Extended Data Fig. 5m,n). Furthermore, using CD11c-Cre LILRB4-Tg mice, we found that SCG2–LILRB4 interaction supports tumor development by suppressing dendritic cell function (data not shown). These data indicate that the interaction between SCG2 and LILRB4 results in a monocytic cell-mediated immunosuppressive TME.

SCG2 deficiency impairs tumor progression in LILRB4-Tg mice

As SCG2 is detected in human serum⁹ and mouse SCG2 can induce activation of LILRB4 reporter cells, we asked whether host-derived SCG2 contributes to tumor progression in vivo. We generated LILRB4-Tg mice lackingSCG2 (LILRB4-Tg Scg2^−/− mice), which exhibit markedly reduced SCG2 levels (Extended Data Fig. 6a). The proportions of myeloid cells, T cells and NK cells were similar in the spleens of WT, Scg2^−/−, LILRB4-Tg and LILRB4-Tg Scg2^−/− mice (Extended Data Fig. 6b). SCG2 deficiency in LILRB4-Tg mice significantly diminished tumor growth, whereas SCG2 deficiency had no effect in WT mice (Fig. 4a–f and Extended Data Fig. 6c–e). This suggests that the tumor-supportive function of host-derived SCG2 depends on LILRB4 expression in myeloid cells. MC38-vector(MC38-Vec) tumors grew more slowly in LILRB4-Tg Scg2^−/− mice than in LILRB4-Tg mice, whereas SCG2-expressing MC38-SCG2 tumors exhibited accelerated growth in LILRB4-Tg mice relative to MC38-Vec tumors (Fig. 4g), indicating that both tumor-produced and host-derived SCG2 contribute to tumor development. Moreover, no significant difference was observed in the growth of MC38-SCG2 tumors between LILRB4-Tg mice and LILRB4-Tg Scg2^−/− mice (Fig. 4g), suggesting that high expression of SCG2 can sustain tumor-promoting activity even in the absence of host-derived SCG2. We observed that tumor growth in LILRB4-Tg Scg2^−/− mice was comparable to that in WT mice, suggesting that ApoE or other LILRB4 interactors are insufficient to compensate for SCG2 loss (Extended Data Fig. 6f). In the absence of LILRB4, tumors grew slower in Apoe^−/− mice than in WT mice (Extended Data Fig. 6g), indicating that ApoE has functional receptors in the TME other than LILRB4, and LILRB4 is not solely responsible for mediating the immunosuppressive function of ApoE within the TME. Treatment with anti-PD-1 reduced tumor growth in LILRB4-Tg mice, and anti-PD-1 combined with SCG2 deficiency significantly further suppressed tumor growth (Fig. 4h).

Fig. 4 | — a–e, Different tumor model growth in *LILRB4*-Tg and *LILRB4*-Tg *Scg2*^−/− mice. *LILRB4*-Tg or *LILRB4*-Tg *Scg2*^−/− mice were subcutaneously injected with MC38 (a; n = 9), B16F10 (b; n = 8), LLC (c; n = 9), EO771 (d; n = 8) and CT2A (e; n = 8) cells on the right flank. Tumor size was measured every 2 days, and tumor weights were measured on the final day (same below). f, B16F10 tumor growth in WT and *Scg2*^−/− mice. WT or *Scg2*^−/− mice were subcutaneously injected with B16F10 cells in the right flank (n = 5 or 6). g, MC38-Vec or MC38-SCG2 tumor growth in *LILRB4*-Tg and *LILRB4*-Tg *Scg2*^−/− mice.LILRB4-Tg mice or LILRB4-Tg *Scg2*^−/− mice were subcutaneously implanted with MC38-Vec or MC38-SCG2 cells in the right flank (n = 5). h, MC38 tumor growth in *LILRB4*-Tg and *LILRB4*-TG *Scg2*^−/− mice treated with anti-PD-1 or isotype control (n = 5 or 6). i–k, Flow cytometry analysis of tumor-infiltrating immune cells from MC38 tumors in *LILRB4*-Tg or *LILRB4*-Tg *Scg2*^−/− mice 21 days after implantation. MDSCs (i), CD4⁺ and CD8⁺ cells (j), CD44⁺CD62L⁺ and CD44⁺CD62L⁻ T cell subsets (k; n = 10) are shown. Tumor volume data are presented as mean ± s.e.m.; all other data are shown as mean ± s.d. P values were determined by two-way ANOVA with a Bonferroni’s multiple comparisons test (tumor volume; a–f) or Tukey’s multiple comparisons test (tumor volume; g), one-way ANOVA with a Holm–Sidak’s multiple comparisons test (tumor weight; g and h) or two-tailed unpaired Student’s t-test (tumor weight; a–f and i–k); cDC1, type 1 conventional dendritic cell; cDC2, type 2 conventional dendritic cell; T_EM, effector memory T cells; T_CM, central memory T cells; TAM, tumor-associated macrophage.

Tumors in LILRB4-Tg Scg2^−/− mice exhibited substantially decreased M-MDSC infiltration and a significant increase in the infiltration of CD4⁺ and CD8⁺ T cells compared to tumors in LILRB4-Tg mice (Fig. 4i,j), as further confirmed by immunofluorescence staining of tumor tissue (Extended Data Fig. 6h). We found that both central memory (CD44⁺CD62L⁺) and effector memory (CD44⁺CD62L⁻) T cell populations were increased in LILRB4-Tg Scg2^−/− mice (Fig. 4k). Similar results were obtained in B16F10, LLC, EO771 and CT2A syngeneic tumor models (Extended Data Fig. 6i). These results indicate that host-derived SCG2 induces an immunosuppressive TME in LILRB4-Tg mice and that SCG2 deficiency reduces the inhibitory role of LILRB4.

LILRB4 enhances immunosuppressive activity of MDSCs

To understand how SCG2 influences the activity of MDSCs, T cells were cocultured with BM-derived MDSCs (BM-MDSCs) isolated from WT mice or LILRB4-Tg mice in 96-well plates coated with SCG2 or bovine serum albumin (BSA; as a control). SCG2 suppressed T cell proliferation only in cocultures with MDSCs derived from LILRB4-Tg mice (Fig. 5a), and this effect was blocked by anti-LILRB4 (Fig. 5b). SCG2 inhibited T cell proliferation in an MDSC dose-dependent manner (Extended Data Fig. 7a), which indicates that SCG2 does not directly affect T cells. Furthermore, the production of interferon-γ (IFNγ) was consistent with T cell proliferation (Extended Data Fig. 7b). Myeloid cell-derived nitric oxide (NO) and reactive oxygen species (ROS) are essential for T cell inhibition³⁷, and the presence of SCG2 resulted in higher levels of NO and ROS in MDSCs from LILRB4-Tg mice but not in MDSCs from WT mice (Extended Data Fig. 7c,d), and NO inhibitor blocked SCG2-induced MDSC suppressive functions (Extended Data Fig. 7e). Additionally, SCG2 increased monocyte migration in vitro, which was blocked by anti-LILRB4 (Extended Data Fig. 7f). SCG2 also decreased the expression of activation markers CD80 and CD86 and increased the expression of the immunosuppressive marker CD163 on BMDMs from LILRB4-Tg mice but had no effect on the expression of these markers in WT mice (Extended Data Fig. 7g). These effects were blocked by treatment of macrophages from LILRB4-Tg mice with anti-LILRB4 (Extended Data Fig. 7g).

Fig. 5 | — a, T cell proliferation in BM-MDSC–T cell coculture assay. Carboxyfluorescein succinimidyl ester (CFSE)-labeled CD3⁺ T cells were cocultured with WT or LILRB4⁺ BM-MDSCs in 96-well plates coated with BSA or SCG2 (10 μg ml⁻¹). The positive control (PC) indicated only T cells with anti-CD3/CD28 beads (n = 3 biological replicates). T cell proliferation was analyzed by flow cytometry (same as below). b, T cell proliferation in BM-MDSC–T cell coculture assays treated with anti-LILRB4 or isotype control (n = 3 biological replicates). c, T cell proliferation in splenic MDSC–T cell coculture assays. Splenic MDSCs (5 × 10⁴) from MC38-Vec or MC38-SCG2 tumor-bearing *LILRB4*-Tg mice were cocultured with CFSE-labeled CD3+ T cells (either 2.5 × 10⁵ (1:5) or 5× 10⁵ (1:10)) for 2.5–3 days (n = 3 biological replicates); NC, negative control. d, T cell proliferation in splenic MDSC–T cell coculture assays. Splenic MDSCs (5 × 10⁴) from MC38 tumor-bearing *LILRB4*-Tg or *LILRB4*-Tg *Scg2*^−/− mice were cocultured with CFSE-labeled CD3⁺ T cells (either 2.5 × 10⁵ (1:5) or 5 × 10⁵ (1:10)) for 2.5–3 days (n = 3 biological replicates). e, IFNγ production in the coculture supernatant from d, measured by ELISA (n = 3 biological replicates). f, Volcano plot showing differentially expressed genes in M-MDSCs isolated from MC38 tumors in WT or *LILRB4*-Tg mice, based on RNA sequencing of M-MDSCs sorted by FACS. **g,h**, KEGG (g) and GO (h) pathway enrichment analysis of differentially expressed genes in M-MDSCs from *LILRB4*-Tg versus WT tumor-bearing mice; T_H17, IL-17-producing helper T cells; T_H1, type 1 helper T cells; T_H2, type 2 helper T cells. i, GSEA of the JAK–STAT signaling pathway comparing M-MDSCs from WT and *LILRB4*-Tg mice; NES, normalized enrichment score. j, Relative mRNA expression levels of the indicated genes in M-MDSCs from WT, *LILRB4*-Tg or *LILRB4*-Tg *Scg2*^−/− MC38 tumor-bearing mice (n = 8 or 9 biological replicates). k, Flow cytometry analysis of p-STAT3 levels in M-MDSCs from both tumors and spleens of MC38-Vec or MC38-SCG2 tumor-bearing *LILRB4*-Tg mice (n = 6 biological replicates). Data are presented as mean ± s.d. P values were determined by one-way ANOVA with Holm–Sidak’s multiple comparisons test (a–e and j), Wald’s test (f) or two-tailed unpaired Student’s t-test (k); MFI, mean fluorescence intensity.

To investigate whether SCG2–LILRB4 signaling enhances the suppressive activity of systemic MDSCs, we compared the immunosuppressive capacity of splenic MDSCs isolated from MC38-Vec or MC38-SCG2 tumor-bearing LILRB4-Tg mice. Splenic MDSCs from SCG2-expressing tumors more potently suppressed T cell proliferation (Fig. 5c). Conversely, splenic MDSCs isolated from tumor-bearing LILRB4-Tg Scg2^−/− mice exhibited reduced suppression of T cell proliferation and IFNγ production compared to those from LILRB4-Tg mice (Fig. 5d,e). These findings suggest that SCG2–LILRB4 signaling exerts systemic immunosuppressive effects in tumor-bearing hosts. To investigate underlying mechanisms, RNA sequencing (RNA-seq) was performed on tumor-infiltrating M-MDSCs isolated from WT, LILRB4-Tg and LILRB4-Tg Scg2^−/− mice. MDSC-associated genes, such as Il6, Il10, S10Oa8 and Stat3, and monocyte-related chemokines, such as Cxcl2 and Cxcl3, were upregulated in M-MDSCs isolated from LILRB4-Tg mice compared to those from WT mice (Fig. 5f). Moreover, KEGG signaling pathway analysis indicated that LILRB4 expression affected the JAK–STAT signaling pathway (Fig. 5g). Gene ontology (GO) enrichment analysis showed that expression of LILRB4 in mouse M-MDSCs influences the regulation of cytokine production (Fig. 5h). Gene set enrichment analysis (GSEA) also confirmed that LILRB4 positively regulated the JAK–STAT signaling pathway (Fig. 5i). We also compared RNA-seq data from M-MDSCs isolated from LILRB4-Tg mice and LILRB4-Tg Scg2^−/− mice. KEGG analysis showed that SCG2 deficiency downregulated JAK–STAT signaling (Extended Data Fig. 7h). GO analysis showed that SCG2 deficiency decreased IL-6 production and downregulated myeloid leukocyte migration (Extended Data Fig. 7i). GSEA confirmed that M-MDSCs derived from LILRB4-Tg Scg2^−/− mice showed decreased JAK–STAT signaling (Extended Data Fig. 7j).

We also measured the expression levels of effector molecules that contribute to the suppressive functions of M-MDSCs by quantitative PCR with reverse transcription. Nos2, Cox2, Arg1, Il1O and Il6 mRNA levels were significantly increased in M-MDSCs from LILRB4-Tg mice compared to those from WT mice. SCG2 deficiency rescued this phenotype. In LILRB4-Tg Scg2^−/− MDSCs, levels of S100a8, S100a9 and Mmp9 were downregulated compared to levels in LILRB4-Tg mice (Fig. 5j). Moreover, levels of phosphorylated STAT3 (p-STAT3) were significantly higher in tumor- and spleen-infiltrating M-MDSCs from MC38-SCG2 tumor-bearing LILRB4-Tg mice than in those from MC38-Vec tumor-bearing LILRB4-Tg mice (Fig. 5k). These results indicate that SCG2−LILRB4 interactions drive the immunosuppressive activity of M-MDSCs, potentially through the IL-6−JAK2−STAT3 signaling axis.

SCG2−LILRB4 interaction strengthens IL-6−STAT3 signaling

STAT3 signaling plays a central role in MDSC accumulation and immunosuppressive function³⁸. Our RNA-seq data and p-STAT3 levels in M-MDSCs from tumor-bearing mice indicated that SCG2−LILRB4 interaction upregulated IL-6−STAT3 signaling. To confirm this directly, BM-MDSCs from WT or LILRB4-Tg mice were stimulated with IL-6 for 10 min. SCG2 significantly increased the phosphorylation of STAT3 in MDSCs from LILRB4-Tg mice but not WT mice (Fig. 6a and Extended Data Fig. 8a−c). Phosphorylation of JAK1 (Y1034/1035) and JAK2 (Y1007/1008) was not impacted under these conditions (Fig. 6a and Extended Data Fig. 8a). Additionally, the upregulation of p-STAT3 induced by SCG2 was not affected in the absence of ApoE in MDSCs from LILRB4-Tg Apoe^−/− mice (Extended Data Fig. 8d,e). Because no antibody is available that specifically detects phosphorylated LILRB4 in primary cells, we could not directly evaluate LILRB4 phosphorylation. To determine whether LILRB4 was activated, we evaluated the status of SHP1. SCG2 increased levels of phosphorylated SHP1 in LILRB4-Tg MDSCs but not in WT MDSCs, and BSA did not change levels of SHP1 phosphorylation (Fig. 6a and Extended Data Fig. 8a). Therefore, under these conditions, SCG2 induces LILRB4 activation. As expected, the upregulation of phosphorylated forms of STAT3, SHP1 and SHP2 induced by SCG2 was blocked by anti-LILRB4 (Fig. 6b and Extended Data Fig. 8f). Moreover, the presence of SCG2 increased levels of Arg1 mRNA, which encodes a factor that acts downstream of STAT3 (ref. 39), in LILRB4-expressing MDSCs in a dose- and time-dependent manner (Fig. 6c and Extended Data Fig. 8g). These findings indicate that the SCG2−LILRB4 interaction directly activates STAT3-mediated signaling.

Fig. 6 | — a, Immunoblot of IL-6-STAT3 signaling in BM-MDSCs from WT or *LILRB4*-Tg mice stimulated with or without IL-6 (20 ng ml⁻¹, 10 min) in 96-well plates coated with BSA or SCG2. b, Immunoblot of p-STAT3, p-SHP1 and p-SHP2 in BM-MDSCs from *LILRB4*-Tg mice stimulated as in a with anti-LILRB4 or lgG1 (10 μg ml⁻¹). c, *Arg1* mRNA expression in BM-MDSCs after IL-6 (20 ng ml⁻¹, 24 h) stimulation in plates coated with BSA or SCG2 (n = 3 biological replicates). **d,e**, Co-IP of LILRB4 (d) or LILRB4-intracellular domain (LILRB4-ICD; e) with STAT3 in HEK293T cells. **f,g**, Schematic representation of STAT3 truncation mutants (f) and co-IP of STAT3 truncation mutants with LILRB4 in HEK293T cells (g). h, Co-IP of LILRB4 with STAT3 in the presence of Src kinases in HEK293T cells. i, SCG2 increased phosphorylation of STAT3 in HEK293T cells co-transfected with LILRB4–Flag (2 μg), STAT3 (1 μg), Lyn (100 ng) and SCG2 (300 ng). j, Immunoblot of p-STAT3, p-SHP1 and p-SHP2 in BM-MDSCs stimulated with cytokines or TLR ligands (50 ng ml⁻¹, except IL-6 at 20 ng ml⁻¹) for 10 min in plates coated with BSA or SCG2. k, Co-IP of Flag-tagged LILRB4, D1-deletion (dD1) or D2-deletion (dD2) mutants with hemagglutinin (HA)-tagged IL-6R in HEK293T cells. I, Immunoblot of p-STAT3 in BM-MDSCs stimulated with IL-6 (20 ng ml⁻¹, 10 min) in plates coated with BSA or SCG2. Inhibitors were added as indicated: TPI (1 μM), SHP099 (1 μM), saracatinib (0.5 μM) or stattic (1 μM); iSHP1, SHP1 inhibitor; iSHP2, SHP2 inhibitor; iSrc, Src inhibitor; iSTAT3, STAT3 inhibitor. m, T cell proliferation in MDSC–T cell coculture assay treated with inhibitors. CFSE-labeled CD3⁺ T cells were cocultured with LILRB4⁺ MDSCs in plates coated with BSA or SCG2. Saracatinib (0.5 μM) and stattic (1 μM) were added as indicated (n = 3 biological replicates). n, MC38-Vec or MC38-SCG2 tumor growth in *LILRB4*-Tg mice treated with a STAT3 inhibitor. Tumor-bearing mice were treated with stattic (20 mg per kg (body weight), intraperitoneally) every 2 days for 2 weeks. Control mice received an equal volume of vehicle solution. Tumor size was measured every 2 days (n = 5). Data are presented as mean ± s.d. (c and m) and mean ± s.e.m. (n). P values were determined by two-way ANOVA with Bonferroni’s multiple comparisons test (c and m) or with Tukey’s multiple comparisons test (n).

STAT3 is a direct substrate of the Src kinase family⁴⁰; we tested the hypothesis that STAT3 might associate with LILRB4 that can be phosphorylated by Src or a complex of STAT3 and Src might be recruited to the ITIMs of LILRB4. Co-IP experiments showed that the intracellular domain of LILRB4 interacts with STAT3 (Fig. 6d,e), and LILRB4 immunoprecipitates endogenous STAT3 (Extended Data Fig. 8h). Experiments with truncated STAT3 constructs showed that the C-terminal region of STAT3 mediates the interaction with LILRB4 (Fig. 6f,g). Deletion of the SH2 domain of STAT3 decreased binding to LILRB4, and deletion of the linker domain of STAT3 abolished binding to LILRB4 (Extended Data Fig. 8i). Furthermore, LILRB4 bound to different Src family members (Hck, Fgr, Lyn and Src), and both LILRB4 and STAT3 were phosphorylated by Src kinase. All these Src kinases enhanced LILRB4 binding to STAT3 (Fig. 6h). Coexpression of SCG2 and LILRB4 increased the phosphorylation of LILRB4, and SCG2 enhanced phosphorylation of STAT3 and LILRB4–STAT3 binding in our 293T recombination system (Fig. 6i). However, when the ITIM of LILRB4 was mutated to a nonphosphorylatable form (Y360F, Y412F and Y442F), the LILRB4 ITIM mutant substantially reduced the binding to Src–STAT3 complexes, and the phenotype was lost (Extended Data Fig. 8j). These data indicate that LILRB4 could form a complex with Src–STAT3.

We next evaluated whether other cytokines, which are important for the function of MDSCs, could induce the activation of LILRB4. Gran-ulocyte–macrophage colony-stimulating factor (GM-CSF) induced very strongSHP2 phosphorylation independently of SCG2, whereas IL-6 only induced phosphorylation of SHP2 in the presence of SCG2 (Fig. 6j and Extended Data Fig. 8k). IL-6 increased levels of SHP1 phosphorylation in a manner that depended on SCG2, whereas the presence of SCG2 decreased levels of SHP1 phosphorylation induced by IFNγ, indicating that only IL-6 could activate LILRB4 in the presence of SCG2 (Fig. 6j). GM-CSF, IL-6 and IL-10 could activate STAT3, but only the phosphorylation of STAT3 induced by IL-6 was increased by SCG2 (Fig. 6j).SCG2 did not influence phosphorylation of STAT1 or STAT6 under IFNγ or IL-4 stimulation (Extended Data Fig. 8l). LILRB1 binding to IL-6R was previously reported³⁰, and co-IP showed that LILRB4 interacted with IL-6R and that deletion of the extracellular LILRB4 D2 domain substantially decreased this interaction (Fig. 6k). Inhibition of SHP1 or SHP2 did not affect the upregulation of p-STAT3 induced by the SCG2–LILRB4 interaction, but a Src inhibitor decreased p-STAT3 levels, and inhibition of STAT3 completely inhibited p-STAT3 induced by IL-6 in the presence of SCG2 (Fig. 6l). SCG2 no longer suppressed T cell proliferation after Src and STAT3 inhibition (Fig. 6m and Extended Data Fig. 8m). STAT3 inhibition blocked SCG2-mediated enhancement of tumor growth in LILRB4-Tg mice (Fig. 6n). Together, these results indicate that the interaction between SCG2 and LILRB4 directly upregulates signaling mediated through IL-6 and STAT3.

LILRB4 supports the immunosuppressive function of human M-MDSCs

We investigated the effects of the SCG2–LILRB4 interaction in human M-MDSCs. T cell proliferation was assessed in cocultures of CD14⁺ MDSCs derived from the peripheral blood of individuals with cancer. MDSCs effectively inhibited the proliferation of autologous T cells (Fig. 7a and Extended Data Fig. 9a). SCG2 further decreased T cell proliferation, and anti-LILRB4 abrogated the effect (Fig. 7a and Extended Data Fig. 9a), indicating that SCG2–LILRB4 interaction increases the immunosuppressive function of human MDSCs. Moreover, SCG2 decreased the expression of CD86, HLA-DR and CD40 in primary monocytic cells in anLILRB4-dependent manner (Fig. 7b).SCG2 upregulated p-STAT3 after IL-6 stimulation of MDSCs derived from the blood of healthy human donors or individuals with cancer. SCG2 also enhanced the levels of p-SHP1 and p-SHP2 in these MDSCs; however, the phosphorylation status of STAT1, NF-κB, p38, Erk and AKT were not affected by SCG2 (Fig. 7c and Extended Data Fig. 9b, c). Consistent with the result of MDSCs from mice, phosphorylation of JAK1 and JAK2 was not impacted by SCG2–LILRB4 interaction in human M-MDSCs (Extended Data Fig. 9d,e). In humanized mice (Extended Data Fig. 9f), SCG2-expressing A375 tumors grew faster, and anti-LILRB4 suppressed tumor growth regardless of SCG2 expression (Fig. 7d and Extended Data Fig. 9g,h). We next analyzed single-cell RNA-seq data of myeloid cells from eight cancer types⁴¹ to confirm the relationship between LILRB4 and the IL-6–STAT3 signaling pathway. We characterized the subsets of myeloid cells using unsupervised graph-based clustering based on canonical cell markers (Fig. 7e and Extended Data Fig. 10a,b). Levels of LILRB4 mRNA were positively correlated with levels of mRNAs that encode factors involved in IL-6–STAT3 signaling and with the MDSC-like score (Fig. 7f,g). GSEA profiles showed a significant enrichment of IL-6–STAT3 and MDSC-like gene sets in monocytes that express high levels of LILRB4 (Fig. 7h). These data indicate that SCG2 enhances IL-6–STAT3 signaling in human MDSCs.

Fig. 7 | — a, T cell proliferation in a human MDSC–T cell coculture assay. CFSE-labeled CD3⁺ T cells were cocultured with CD14⁺ MDSCs isolated from samples from individuals with cancer in the presence of anti-CD3/CD28 activation beads in plates coated with BSA or SCG2. Blocking anti-LILRB4 (10 μg ml⁻¹) was added as indicated. After 5 days, T cell proliferation was assessed by flow cytometry (n = 3 biological replicates). b, SCG2 downregulated CD86, HLA-DR and CD40 expression on human monocytic cells. CD14⁺ monocytes were cultured in plates coated with SCG2 in the presence of GM-CSF (50 ng ml⁻¹) for 4 days, followed by lipopolysaccharide stimulation (20 ng ml⁻¹, 24 h). Surface marker expression was analyzed by flow cytometry (n = 3 biological replicates). c, Immunoblot of CD14⁺ monocytes from healthy donors stimulated with or without IL-6 (20 ng ml⁻¹, 10 min) in plates coated with BSA or SCG2. d, Tumor growth in humanized mice injected subcutaneously with A375 cells expressing SCG2 or vector control in the right flank. Anti-LILRB4 or human IgG1 control was injected intraperitoneally every other day starting from day 7 to the endpoint. Tumor size was measured every 2 days, and tumor weights were measured on the final day (n = 6). e, Uniform manifold approximation and projection (UMAP) plot of single-cell RNA-seq (scRNA-seq) data showing myeloid cell populations across eight cancer types; CDC3, type 3 conventional dendritic cell; pDC, plasmacytoid dendritic cells. f, UMAP visualization of IL-6–STAT3 signaling score, MDSC-like score and LILRB4 expression levels in monocyte clusters. g, Violin plots comparing IL-6–STAT3 signaling and MDSC-like scores between LILRB4-low and LILRB4-high monocyte populations. h, GSEA enrichment plots showing increased IL-6–STAT3 and MDSC-like gene signatures in LILRB4-high monocytes compared to LILRB4-low monocytes. Data are presented as mean ± s.e.m. (d, left) or mean ± s.d. (a, b and d, right). P values were calculated using unpaired two-tailed Student’s t-test (b), one-way ANOVA with Holm–Sidak’s multiple comparisons test (a and d right), two-way ANOVA with Holm–Sidak’s multiple comparisons test (d, left) or two-sided Wilcoxon rank-sum test (h).

Among the five inhibitory LILRBs, all except LILRB1 bound STAT3, and these interactions were enhanced by Lyn kinase (Extended Data Fig. 10c). When the intracellular domain of LILRB1 was replaced with the intracellular domain of LILRB4, the interaction was restored, suggesting that the intracellular domains of LILRBs mediate STAT3 binding (Extended Data Fig. 10d). To investigate whether other LILRBs play a role in IL-6–STAT3 signaling, we incubated MDSCs with coated anti-LILRB and evaluated levels of p-STAT3 after IL-6 stimulation. LILRB2, LILRB3, LILRB4 and LILRB5 increased STAT3 phosphorylation, butLILRB1 did not (Extended Data Fig. 10e,f), which is consistent with binding of STAT3 to all LILRBs except LILRB1.

Discussion

Here, we demonstrate that LILRB4 serves as the receptor of SCG2 on monocytic cells. SCG2 activates LILRB4, but not gp49b or other LILRBs, leading to LILRB4 phosphorylation and recruitment of SHP1 and SHP2. SCG2, produced either from tumor cells or host mice, supports the immunosuppressive activity of monocytic cells and tumor development. SCG2 deficiency decreased tumor development in LILRB4-Tg mice, and blocking anti-LILRB4 effectively inhibited SCG2-induced LILRB4 signaling and immunosuppressive activity. Mechanistically, the SCG2–LILRB4 interaction allows SHP1- and SHP2-dependent and -independent signaling. In the SHP-independent mechanism, LILRB4 induces STAT3 activation through an LILRB4–Src–STAT3 complex that amplifies IL-6-mediated signaling. This work provides a molecular mechanism by which a hormone regulates the activity of immunosuppressive myeloid cells.

Like other granins, SCG2 can undergo precursor processing to generate bioactive peptides, including secretoneurin, EM66 and manserin⁵. We showed that full-length SCG2 is present in mouse serum and in the supernatant of SCG2-expressing cells (Extended Data Figs. 5a and 6a). There is a lack of literature defining the exact functions of extracellular full-length SCG2 or cleaved N-terminal peptides^5,8,9. Secretoneurin, a processed peptide of SCG2 that is composed of the 30- to 43-amino-acid central domain of SCG2, was reported to play roles in the regulation of monocytic cell activities, immune response, neurodevelopment, food intake and reproduction⁵. The function of EM66 is less clear but was suggested to regulate neuroendocrine differentiation and energy homeostasis⁴². Here, we demonstrated the function of SCG2 as a ligand of LILRB4 and showed that its binding to LILRB4 occurs at the N terminus; neither the SCG2 central fragments (secretoneurin and EM66) nor other tested granin family members (including chromogranin A and chromogranin B) bound to LILRB4. Therefore, this specific activity of SCG2 is unique and can explain the reported correlation between SCG2 expression with immunosuppressive myeloid cell infiltration and cancer supportive capability. Considering the expression and function of LILRB4 in microglia²², our results also support the validity of SCG2 as a biomarker for certain neurological disorders, such as Alzheimer’s disease. A recent study demonstrated that SCG2 is highly expressed in melanoma cells and downregulated STAT1 expression and disrupted major histocompatibility complex class I assembly¹⁶. However, this finding differs mechanistically from ours, as we focused on STAT signaling in monocytic cells. This distinction underscores the context-specific effects of SCG2 and highlights its differential roles in tumor versus immune cells.

Previously, we identified ApoE as an extracellular activator of LILRB4 signaling in monocytic acute myeloid leukemia cells²⁴, which was supported by independent studies from other groups^22,28. However, the specific functional ligand for LILRB4 in the TME of solid cancers was unclear. Here, we demonstrated that SCG2 has a higher affinity for LILRB4 binding than do other previously identified binding partners (Fig. 1e,f and Extended Data Fig. 1f). In particular, different from ApoE that can be produced by LILRB4⁺ monocytic cells, SCG2 is from a trans source, such as tumor cells or neuroendocrine cells, and binds LILRB4 on monocytic cells to induce downstream signaling independent of ApoE (Extended Data Fig. 8d,e). Theoretically, multiple ligands may enable a more precise regulation of LILRB4 signaling than single ligands. LILRB4 may have emerged relatively late in evolution and thus is an unoptimized inhibitory receptor; such a regulation may also facilitate further evolution.

The functions of immunosuppressive myeloid cells are regulated by both inhibitory and activating receptors⁴³. However, how exactly these two types of receptors regulate the immunosuppressive activity of myeloid cells is unclear. We discovered that LILRB4-mediated signaling cross-talks with the IL-6–STAT3 signaling axis, which plays a pivotal role in the accumulation and activation of MDSCs during tumor development³⁸. Importantly, LILRB4 activation increased phosphorylation of STAT3 by Src, which represents a different signaling pathway from the SHP1-and SHP2-mediated classical downstream signaling of LILRB4. We further showed that all LILRBs with the exception of LILRB1 were able to mediateSTAT3 phosphorylation. This represents a mechanism by which inhibitory ITIM-containing receptors coordinate with activating receptor signaling to support the activity of immunosuppressive myeloid cells. It aligns with a previous report that SCG2 upregulates MMP-9 (ref. 44), a known downstream effector of STAT3 activation⁴⁵. MMP-9 plays multifaceted roles in cancer progression⁴⁶, angiogenesis⁴⁷, immune response and neuroinflammation⁴⁸, neurogenesis⁴⁹, energy homeostasis⁵⁰ and reproduction⁵¹, paralleling the known functions of SCG2 and its cleaved peptides^5,6,9.Moreover, although the SCG2–LILRB4 interaction mediates immunosuppression, inflammation has been shown to inhibit gonadotropin-releasing hormone secretion⁵², which is concordant with a reported role of SCG2 in stimulating luteinizing hormone production⁵. However, because the cleaved fragments secretoneurin and EM66 do not bind LILRB4, they are likely to act through mechanisms distinct from those of full-length or N-terminal SCG2.

We observed that MC38-Vec tumors exhibited slower in vivo growth than WT MC38 tumors, likely due to increased immunogenicity resulting from lentiviral transduction and puromycin selection. Additionally, the variability observed in CD4⁺ and CD8⁺ T cells across tumor models in Extended Data Fig. 6i reflects tumor-type-dependent differences rather than technical variability. We also acknowledge that the subcutaneous tumor model may not recapitulate the TME. Nevertheless, we consistently observed similar trends in immunosuppressive monocytic cell accumulation and T cell infiltration across multiple tumor models. These consistent patterns support our conclusion that SCG2 promotes monocytic cell-mediated immunosuppression via LILRB4 signaling, although the magnitude of immune modulation may vary with tumor context.

Together, we demonstrate that a hormone regulates the activity of immunosuppressive myeloid cells via an inhibitory receptor both locally and systemically and uncover a mechanism in which inhibitory and activating receptor signaling cooperatively sustains immunosuppression. These findings may guide therapeutic strategies for cancer and inflammatory and neurological diseases.

Online content

Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41590-025-02233-4.

Methods

Cells

HEK293T (ATCC, CRL-11268), B16F10 (ATCC, CRL-6475), MC38 (a gift from Y. Fu, University of Texas Southwestern), LLC (ATCC, CRL-1642), EO771 (ATCC, CRL-3461), CT2A (Sigma-Aldrich, SCC194) and A375 (ATCC, CRL-1619) cells were maintained in DMEM with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin in a humidified atmosphere of 5% CO₂ at 37 °C. THP-1 (ATCC, TIB-202) and LILR reporter cells (generated and maintained in our laboratory^24,31,53,54), including parental 2B4 cells (a gift from H. Arase, Osaka University), were cultured in RPMI with 10% FBS and 1% penicillin/streptomycin. Cell lines were cultured for a maximum of 20 passages before they were discarded. All cell lines were routinely tested using a mycoplasma contamination kit (R&D Systems).

Mice

Lysm-Cre C57BL/6J mice, Apoe^−/− C57BL/6J and NOD-scid-IL2rg-null (NSG) mice were obtained from The Jackson Laboratory. C57BL/6N-Scg2^{tmLIIKOMPVIcg}/JMmucd (RRID:MMRRC_049562-UCD) mice were obtained from the Mutant Mouse Resource and Research Center (MMRRC) at the University of California, Davis, a National Institutes of Health-funded strain repository, and were donated to the MMRRC by The KOMP Repository, University of California, Davis (originating from S. Murray, The Jackson Laboratory).LILRB4-Tg mice were generated by CRISPR and backcrossed with C57BL/6J mice for at least six generations. LILRB4-Tg Scg2^−/− C57BL/6J mice were produced by crossing LILRB4-Tg mice with Scg2^−/− C57BL/6J mice. All mice were maintained under specific pathogen-free conditions, housed in individually ventilated cages with autoclaved bedding and fed ad libitum with a standard irradiated chow diet (LabDiet 5P76, PicoLab Rodent Diet 20, 0007688). Housing conditions consisted of a 12-h light/12-h dark cycle at 22 ± 2 °C with 40–60% humidity. All animal experiments were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee of University of Texas Southwestern.

Flow cytometry of tumor immune cells

Tumor-bearing mice were killed after inoculation on relevant days. Tumors were collected, dissociated with surgical scissors and digested with 1 mg ml⁻¹ collagenase IV and 200 μg ml⁻¹ DNase in PBS for 60 min in a 37 °C shaking incubator (110 rpm). Cell suspensions were filtered through a 70-μm cell strainer and centrifuged at 300g for 5 min. Cells were washed with PBS once and stained with Zombie viability dye in PBS at room temperature for 30 min. Cell suspensions were washed with FACS buffer (PBS containing 0.2% BSA, 2 mM EDTA and 1% penicillin/streptomycin) and blocked with Fc blocking antibody for 30 min. Finally, cell suspensions were stained with the indicated antibodies for 1 h on ice. Flow data were collected using FACS Calibur, FACS Melody and Cytek Northern Lights and analyzed by FlowJo V10 software. A detailed list of antibodies, including fluorochrome, clone, vendor and catalog number, is provided in the Reporting Summary.

Chimeric receptor reporter

LILRB4 chimeric receptor reporter cells were developed as described previously²⁴. The indicated proteins were precoated onto 96-well plates by adding 50 μl of the desired protein concentration to each well and incubating at 4 °C overnight. The next day, the plates were washed twice with PBS, and 5 × 10⁴ indicated reporter cells were seeded into each well. For soluble proteins, recombinant proteins were mixed with reporter cells before seeding. The indicated antibodies were added at a concentration of 10 μg ml⁻¹. After incubation for 24 h, the percentages of GFP⁺ reporter cells were measured by flow cytometry. Recombinant proteins included SCG2 (Sino Biological, 13441-H08H), CHGA (R&D, 10422-CH), CHGB (myBioSource, MBS143960), SCG3 (Sino Biological, 16012-H08H), SCG5 (myBioSource, MBS7609203), APOE (Sino Biological, 10817-H30E), fibronectin (Sino Biological, 10314-H08H), CNTFR (Sino Biological, 11012-H08B), MAG (Sino Biological, 13186-H08H), Eln (LSBio, 208234) and galectin-8 (Sino Biological, 10301-HNAE).

ELISA

Recombinant SCG2–His, ApoE–His, CNTFR-His and fibronectin proteins were coated onto 96-well ELISA plates overnight at 4 °C. The next day, the plate was washed with phosphate-buffered saline with 0.05% Tween-20 and blocked with 5% BSA for 1 h at room temperature. After washing three times, LILRB4–Fc recombinant protein (Sino Biological, 16742-HO2H) was added and incubated for 2 h. After washing four times, the plate was incubated with anti-Fc conjugated with horseradish peroxidase (Jackson Laboratories, 109–035-008) for 60 min. After washing five times, 50 μl of TMB buffer was added, and the reaction was quenched with 2 N H₂SO₄. The results were read using a plate reader, and absorbance was measured at 450 nm.

IP and immunoblotting

HEK293T cells were transfected with the indicated plasmids at 90% confluency with Polyjet (Signagen, SL100688). The cells were washed twice with cold PBS and lysed in whole-cell lysis buffer (50 mM Tris-HCl (pH7.4), 150 mM NaCl, 1% NP⋅40, 1 mM EDTA, 10% glycerol and protease inhibitor cocktail (Roche)) for 20 min on ice at 48 h after transfection. The cell lysates were then centrifuged at 11,000g for 15 min, and the clear supernatants were subjected to IP with anti-Flag M2 agarose resin (Sigma, F2426) following the manufacturer’s instructions. After 4 h of incubation at 4 °C, the beads were washed three times with whole-cell lysis buffer and twice with PBS and then boiled with 2× loading buffer for 10 min. The immunoprecipitants were subjected to standard immunoblotting analyses with the indicated specific antibodies. For IL-6 signaling stimulation, 4 × 10⁵ human CD14⁺ cells isolated from PBMCs or 4 × 10⁵ BM-MDSCs were incubated in 96-well plates coated with BSA or SCG2 for 4 h. These cells were then stimulated with 20 μg ml⁻¹) human IL-6 (Sino Biological, 10395-HNAE) or mouse IL-6 (Sino Biological, 50136-MNAE) for the indicated times, respectively. The cells were washed twice with cold PBS, lysed in whole-cell lysis buffer and boiled with 2× loading buffer for 10 min. The immunoprecipitants were subjected to standard immunoblotting analyses with the indicated specific antibodies. A detailed list of antibodies, including fluorochrome, clone, vendor and catalog number, is provided in the Reporting Summary.

Pulldown

Protein A/G agarose beads were washed twice with PBS and blocked with 5% BSA for 30 min at room temperature.LILRB4–Fc protein (2 μg) or IgG was diluted in 500 μl, and the indicated amounts of SCG2-His or ApoE-His and CNTFR-His proteins were mixed. The mixture was incubated with protein A/G beads at 4 °C for 4 h. After washing five times with phosphate-buffered saline with 0.05% Tween-20, the beads were boiled with 2× loading buffer for 10 min. The immunoprecipitants were subjected to standard immunoblotting analyses with the indicated specific antibodies.

Surface interferometry

The kinetics of association and dissociation between SCG2 and LILRB4 were measured by surface plasmon resonance (Octet Red384, FortéBio) as described previously⁵⁵. To study SCG2 binding to LILRB4, we first loaded pre-equilibrated ProA biosensor with 100 nM LILRB4–Fc protein in 1× kinetics buffer (Sartorius, 18–5010) for 120 s before a 300-s association step and a 300-s dissociation step. We then tested binding of serially diluted SCG2–His protein to LILRB4–Fc. Finally, we used Octet System Data Analysis Software (FortéBio) to fit the kinetic curves to equations based on 1:1 kinetics and calculate the K_d, k_a (on) and k_d (off). We conducted double referencing (sample and biosensor) for all measurements to subtract background noise and eliminate nonspecific binding.

HEK293T reverse binding assay

Secretory SCG2 was anchored onto cell membranes by fusing the cytoplasmic domain and transmembrane domain of TFR1 at its N terminus. To detect membrane expression of SCG2 by FACS, a Flag tag was fused at its C terminus. The indicated fragments of pLVX-TFRSA2-SCG2-Flag plasmids were transfected into HEK293T cells for 48 h. In total, 2 × 10⁵ transfected cells were incubated with 50 μl of the indicated LILRB–Fc protein (LILRB1–Fc, Sino Biological, 16014-H02H; LILRB2–Fc, Sino Biological, 14132-H02H; LILRB3–Fc, Sino Biological, 11978-H02H; LILRB4–Fc, Sino Biological, 16742-H02H; LILRB5–Fc, Sino Biological, 17221-HO2H; 20 μg ml⁻¹) for 1 h on ice. Cells were washed with FACS buffer twice and blocked with human IgG for 15 min on ice. These cells were stained with anti-Fc APC for 1 h, followed by anti-Flag PE for 30 min. After propidium iodide (Sigma, P4864) staining, the results were collected using a Cytek Northern Lights flow cytometer. Flow data were analyzed using FlowJo v10.

HEK293T reconstituted signaling assay

Detection of LILRB4 phosphorylation and SHP1/SHP2 recruitment was conducted as described previously⁵³. Briefly, 2 μg of LILRB4–Flag or LILRB4–Flag mutants, 300 ng of SHP1, 300 ng of SHP2 and 100 ng of Lyn with or without 300 ng of SCG2 were co-transfected into HEK293T cells. Four to six hours after transfection, the indicated antibodies were added, and fresh medium was replaced. Cells were collected after 48 h. Following co-IP with Flag M2 agarose resin, the immunoprecipitants were subjected to standard immunoblotting analyses with the indicated specific antibodies.

For the detection of STAT3 phosphorylation in HEK293T cells, 1 μg of LILRB4–Flag or LILRB4–Flag mutants, 400 ng of STAT3, 300 ng of SCG2 and 100 ng of Scr kinases were transfected as indicated. Cells were collected after 48 h and analyzed by co-IP and western blotting.

THP-1 LILRB4–FcR coligation assay

For LILRB4 phosphorylation detection, we referenced a previous study³⁵. THP-1 LILRB4–Flag cells were serum starved overnight. Ninety-six-well plates (nontreated flat-bottom plates) were coated with 10 μg ml⁻¹ anti-CD64 (R&D, MAB1257), anti-LILRB4 or SCG2 overnight. The next day, starved cells were collected and resuspended in medium without FBS at a concentration of 2 × 10⁶ cells per ml. One hundred microliters of cell suspension was added per well into coated plates. Plates were then placed into a tissue culture incubator for 10 min. Afterward, the plates were centrifuged at 300g for 5 min at 4 °C. The cells were lysed and analyzed by co-IP and western blotting.

For IL-8 detection, THP-1 WT cells or THP-1 LILRB4-KO cells were directly added into 96-well plates (coated with 10 μg ml⁻¹ lgGl along with SCG2 or ApoE). At the same time, the indicated antibodies were added, and the plates were incubated in a tissue culture incubator for 24 h. Cell-free supernatants were collected and analyzed for IL-8 by ELISA (Biolegend, 431504).

Generation of stable SCG2-expressing cell lines

Codon-optimized cDNA sequences encoding human SCG2 and mouse SCG2 were synthesized by a commercial vendor. The human SCG2 sequence was cloned into the pLVX-Puro lentiviral expression vector, which contains a puromycin resistance cassette for antibiotic selection. The mouse Scg2 sequence was cloned into the pLenti vector lacking a selectable marker. For lentivirus packaging, pLVX-hSCG2, pLVX-Vec, pLenti-mSCG2 or pLenti-Vec was co-transfected with psPAX2 and pMD2.G into HEK293T cells at a ratio of 6:4:10 using Lipofectamine 2000 (Invitrogen). Viral supernatants were collected 48–72 h after transfection and used for infection. Indicated tumor cell lines were infected with virus supernatants by centrifugation at 1,800rpm at 37 °C for 2 h. For human SCG2-expressing and corresponding control (Vec) cell lines, successfully transduced cells were selected with puromycin (2 μg ml⁻¹) for 3–5 days. MC38 cells transduced with mouse SCG2-expressing virus were used directly without antibiotic selection. SCG2 expression was confirmed by western blotting.

In vivo tumor growth and treatment

In total, 5 × 10⁵ MC38, 3 × 10⁵ B16F10, 7 × 10⁵ LLC or 2 × 10⁶ CT2 A vector or SCG2 stable cells were subcutaneously injected into LILRB4-Tg mice or Lysm-Cre mice (as WT control) in the right flank on day 0. The metastatic B16F10 tumor model was constructed by intravenous injection of 5 × 10⁵ B16F10 cells on day 0. For anti-LILRB4 treatment, MC38 tumor-bearing mice were intraperitoneally injected with 200 μg of anti-LILRB4 or 200 μg of human IgG1 control on days 7, 9, 11 and 13. For CD4⁺ and CD8⁺ T cell depletion, MC38 tumor-bearing mice were intraperitoneally injected with 200 μg of anti-CD4 (BioXCell, YTS177, BE-0003–3) and 200 μg of anti-CD8 (BioXCell, YTS169.4, BE0117) or 400 μg of isotype control (BioXCell, LTF-2, BE0090) on days 5, 7, 9 and 11 (from 13.8 to 0.11% for CD8⁺ T depletion and from 18.5 to 0.06% for CD4⁺ T depletion). For neutrophil depletion, MC38 tumor-bearing mice were intraperitoneally injected with 200 μg of anti-Ly6G (BioXCell, 1A8, BE-0075–1) on days 5, 8, 12 and 15. For monocyte and macrophage depletion, MC38 tumor-bearing mice were intraperitoneally injected with 400 μg of anti-CSF1R (BioXCell, AFS98, BE0213) on days 0, 3, 6, 9 and 12. For STAT3 inhibition, the STAT3 inhibitor stattic (MCE, HY-13818) was freshly dissolved for each injection in a vehicle containing 5% DMSO, 30% PEG300 and 65% sterile saline. Mice were injected intraperitoneally with stattic (20 mg per kg (body weight)) or an equal volume of vehicle control every 2 days starting from day 3 after tumor implantation. For recombinant SCG2 protein treatment, MC38 tumor-bearing mice were intratumorally injected with 10 μg of SCG2 (50 μl per mouse) or PBS on days 7, 9, 11 and 13.

We cannot obtain enough Scg2^−/− mice in both LILRB4-Tg and WT backgrounds at the same time because SCG2 deficiency influences reproduction 56. Therefore, we compared tumor growth in LILRB4-Tg mice to that in LILRB4-Tg Scg2^−/− mice and tumor growth in WT mice to that in Scg2^−/− mice, respectively. Similarly, 5 × 10⁵ MC38, 3 × 10⁵ B16F10, 7 × 10⁵, 5 × 10⁵ EO771 or 2 × 10⁶ CT2A cells were subcutaneously injected into the indicated mice in the right flank on day 0. Meanwhile, 200 μg of anti-PD-1 (BioXCell, 29F.1A12, BE0273) and isotype control (BioXCell, 2A3, BE0089) were intraperitoneally injected into MC38 tumor-bearing mice on days 7,9,11 and 13. The mice used in the in vivo experiment above were 8- to 12-week-old female mice.

The humanized mouse model was generated as previously reported⁵⁷. Briefly, 4- to 12-week-old NSG female mice (Jackson Laboratories, 005577) were irradiated at 250 cGy for 98 s. After 4 h of rest, the mice were retro-orbitally injected with 30,000 CD34⁺ cord blood cells resuspended in 200 μl of StemSpan (STEMCELL, 09650) with 25 ng ml⁻¹ macrophage colony-stimulating factor. The mice were then treated with trimethoprim sulfamethoxazole water for at least 14 days. The percentage of engrafted human cells was detected at 6 weeks. At 8 weeks, 2 × 10⁶ A375 vector or SCG2 stable skin melanoma cells were subcutaneously injected into mice in the right flank on day 0. Two hundred micrograms of anti-LILRB4 or 200 μg of human IgG1 control was injected intraperitoneally every other day starting from day 7 to the experimental endpoint.

Tumor volumes were measured using calipers to measure the length (L) and width (W) and calculated as V = L × W × W/2. Tumor weights were measured after mice were killed. The maximal tumor size permitted by the University of Texas Southwestern Institutional Animal Care and Use Committee was 2.0 cm in any dimension, and this limit was not exceeded in any experiment.

Mouse MDSC–T cell coculture

Mouse MDSC differentiation was reported previously^58,59. BM was collected from LILRB4-Tg mice or WT mice by flushing femurs and tibias. BM cells were strained through a 70-μm filter and centrifuged following resuspension in red blood cell lysis buffer (150 mM NH₄Cl, 10 mM NaHCO₃ and 1 mM EDTA) for 3 min. Cells were centrifuged for 5 min at 300g, the cell pellet was resuspended in 10 ml of medium, and the cells were counted. A total of 3 × 10⁶ cells were seeded per 10-cm dish in 10 ml of RPMI supplemented with 10% FBS and 1% penicillin/streptomycin, and 50 ng ml⁻¹) recombinant mouse GM-CSF was added. The cells were then cultured at 37 °C and 5% CO₂ in a humidified incubator for 3 days. Ly6G⁺ cells were removed using Ly6G beads (Miltenyi, 130–120-337). MDSCs were also isolated from the spleens of tumor-bearing mice. Spleens from tumor-bearing mice were placed onto a cell strainer and mashed with a sterile syringe plunger into the dish. The splenic cells were centrifuged and resuspended in red blood cell lysis buffer for 5 min. Lysis was stopped by the addition of 5 ml of complete medium. MDSCs were isolated following the instructions for the EasySep Mouse MDSC Isolation kit (STEMCELL, 19867). T cells were isolated from the spleens of WT non-tumor-bearing mice. T cells were isolated following the instructions for the Mojosort Mouse CD3 T cell Isolation kit (Biolegend, 480024). T cells were stained with CFSE for 10 min. Differentiated BM-MDSCs were collected and cocultured with CFSE-stained T cells (T cells:MDSCs = 1:1) in Immunocult-XF T cell Expansion medium (STEMCELL, 10981) containing 10 ng ml⁻¹ human IL-7 (Peprotech, 200–07), 5ng ml⁻¹ human IL-15 (Peprotech, 200–15) and Dynabeads Mouse T-activator (Thermo Fisher Scientific, 11456D). For MDSCs from tumor-bearing mice, MDSCs were cocultured with CFSE-labeled CD3⁺ T cells at MDSC:T cell ratios of 1:5 or 1:10. Antibodies or inhibitors were added as indicated. Cocultured cells were incubated for 3 days, and the percentage of proliferating T cells was measured by flow cytometry. The supernatants were collected for further detection.

Human MDSC–T cell coculture

Human blood from University of Texas Southwestern Tissue Management was carefully transferred to 50-ml tubes containing 15 ml of Ficoll (GE Healthcare, 45001752) without disturbing the Ficoll interface. The sample was centrifuged at 3,000rpm for 30 min at 20 °C without braking. The top phase was removed, and the interphase was transferred to a new 50-ml tube. The cells were washed twice by adding DMEM to 50 ml and then centrifuging at 1,500rpm for 10 min. The pellet was resuspended and counted. CD14⁺ MDSCs were isolated using CD14 Microbeads (Miltenyi, 130–050-201), and T cells were isolated using CD3 Microbeads (Miltenyi, 130–0597-043). CD14⁺ MDSCs were cocultured with CFSE-labeled T cells (MDSCs:T cells = 1:1) in the presence of 6.25 μl ml⁻¹ ImmunoCult Human CD3/CD28 T Cell Activator (STEMCELL, 10971) for 5 days in BSA- or SCG2-coated plates, and the indicated antibodies were added. Percent T cell proliferation was measured by flow cytometry. Primary human samples were collected by the University of Texas Southwestern Tissue Management Shared Resource. Informed consent was obtained under a protocol approved by the Institutional Review Board at University of Texas Southwestern.

Mouse M-MDSCRNA-seq

MC38 tumors from WT mice, LILRB4-Tg mice and LILRB4-Tg Scg2^−/− mice were placed onto a 12-well plate and mashed with a sterile syringe plunger in 1 ml of FACS buffer. Cell suspensions were filtered through a 70-μm cell strainer, followed by Fc blocker treatment. Before staining with CD11b–APC, Ly6C–FITC and Ly6G–PE for 1 h on ice, cell suspensions were washed once with FACS buffer. M-MDSCs (CD11b⁺Ly6C^hiLy6G⁻) were sorted after propidium iodide staining. Total RNA from the sorted M-MDSCs was purified using a Purelink RNA Mini kit (Thermo Fisher Scientific, 12183025) and sent for sequencing (Novogene, X202SC24031184-Z01-F001) or for quantitative PCR detection. Differentially expressed genes were analyzed using DEseq2 on the iDEP.96 platform⁶⁰ (http://bioinformatics.sdstate.edu/idep96/). Primers for quantitative PCR with reverse transcription are listed in Supplementary Table 1.

Human scRNA-seq

Single-cell transcriptome sequencing data were downloaded from the NCBI database GEO (GSE154763)⁴¹, which includes myeloid cells from multiple tumor types. A total of 65,698 cells were obtained from this dataset for subsequent analysis in the project. The Seurat function FindIntegrationAnchors was used to remove batch effects from different batches and integrate single-cell data (normalization method = ‘Log-Normalize’, dims = 1:20, reduction = ‘cca’)⁶¹. The FindVariableFeatures package was used to perform highly variable gene analysis using default parameters, and the highly variable genes identified were then subjected to principal component analysis. Cell clustering information was obtained using the FindClusters software package (dims.use = 1:30, resolution = 1), which resulted in the visualization of single-cell data using the UMAP algorithm⁶². To identify signature genes in each cell type, the Seurat functions FindMarkers and FindAllMarkers were used (min.pct = 0.1, logfc.threshold = 0.25). Cell types were annotated using the CellMarker database⁶³, and the marker genes for cell types were demonstrated using dot plots. GO enrichment analysis of differentially expressed genes was implemented by the clusterProfiler (version 4.0.2) package⁶⁴. GSEA was performed to show the enriched gene sets based on the expression of each gene. We used GSVA (R packages) to perform the GSVA analysis. Wilcoxon tests were conducted using R language to compare the differences between two samples or cell types. P values of <0.05 were considered to be significant. The scores of specific gene sets were generated by UCell (UCell is an R package for evaluating gene signatures in single-cell datasets)⁶⁵. R language was applied to plot heat maps (by pheatmap package), box plots, bar plots (by ggplot2 package) and so on.

Quantification and statistical analysis

Representative data from two or three independent experiments or indicated independent samples were expressed as mean ± s.e.m. for tumor growth curves, and all other data were expressed as mean ± s.d. Statistical analyses were performed using GraphPad Prism8 software (GraphPad Software). Statistical significance for two-sample comparisons was calculated by two-tailed Student’s t-test. Statistical significance for multiple-sample comparisons was calculated by one-way ANOVA or two-way ANOVA. Statistical significance for survival was calculated by log-rank test. No statistical methods were used to predetermine sample sizes, but our sample sizes are similar to those reported in previous publications^24,32,53. Data distribution was assumed to be normal, but this was not formally tested. Data collection and analysis were not performed blind to the experimental groups. Animals or samples were randomly assigned to experimental and control groups. No animals or data points were excluded from the analyses unless animals died before the experimental endpoint due to unrelated causes, such as injection errors or unrelated illness.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The following publicly available datasets were accessed for this study. TCGA CESC and TCGA MESO survival data were accessed via the GEPIA2 online tool. Raw RNA-seq data for TCGA CESC and MESO were retrieved from TCGA. RNA-seq and survival data for GSE65858, GSE54236 and GSE1993 were obtained from the Tumor Immune Dysfunction and Exclusion (TIDE) database. scRNA-seq data from multiple human tumors were downloaded from GEO under accession code GSE154763. Bulk RNA-seq data generated in this study have been deposited in the GEO database under accession code GSE285725. All other data supporting the findings of this study are provided within the article and its Supplementary Information files. Additional raw data and unique/stable reagents generated during this study are available from the corresponding author upon reasonable request. Source data are provided with this paper.

Extended Data

Extended Data Fig. 3 | — a, Kaplan–Meier survival curves comparing overall survival of patients stratified by high (red) versus low (blue) *SCG2* expression in TCGA CESC, TCGA MESO, GSE65858, GSE54236, and GSE1993 datasets. Survival analyses for TCGA CESC and TCGA MESO datasets were performed using GEPIA2 online tool (http://gepia2.cancer-pku.cn/#index) with median *SCG2* expression as the cutoff. Survival data and analyses for GSE65858, GSE54236, and GSE1993 were obtained from the Tumor Immune Dysfunction and Exclusion (TIDE) database (http://tide.dfci.harvard.edu/), using optimal cut-off values provided by TIDE. Statistical significance was determined using the log-rank (Mantel–Cox) test. **b-c**, Correlation analyses between *SCG2* expression and infiltration of myeloid-derived suppressor cell (MDSC) or cytotoxic T lymphocyte (CTL) across multiple tumor datasets. RNA-seq data were obtained from GEO and TCGA. Gene expression values in each dataset were log⁡2-transformed and normalized using the sample average for TIDE analysis. Cytotoxic T lymphocyte (CTL) infiltration scores were calculated based on the average expression of *CD8A*, *CD8B*, *GZMA*, *GZMB*, and *PRF1* within each dataset. Pearson correlation analyses were conducted between *SCG2* and MDSC infiltration, as well as between *SCG2* and CTL infiltration. Scatter plots for these correlations were generated using the ggplot2 package in R.

Extended Data Fig. 4 | — a, Schematic diagram illustrating the generation of myeloid-specific LILRB4-Tg mice. The human LILRB4 gene was introduced downstream of a loxP-flanked STOP cassette containing a puromycin resistance gene (Pur/Stop cassette), driven by the constitutive CAG promoter. These mice were crossed with LysM-Cre mice expressing Cre recombinase under the control of the myeloid-specific LysM promoter. Upon Cre-mediated recombination, the STOP cassette was excised specifically in myeloid lineage cells, resulting in cell-specific expression of the human LILRB4 transgene. **b-c**, Immunoblot analysis (b) and Flow cytometry analysis (c) validating LILRB4 expression on bone marrow-derived macrophages (BMDMs) isolated from wild-type WT and LILRB4-Tg mice (n = 3 biological replicates). d, Representative flow cytometry gating strategy for dendritic cells (DCs, CD11c⁺MHC-II⁺), macrophages (CD11b⁺F4/80⁺), monocytes (CD11b⁺Ly6C^hi), and neutrophils (CD11b⁺Ly6G⁺) in spleen single-cell suspensions from LILRB4-Tg mice. e, Quantification of LILRB4 expression levels (MFI) and percentages of LILRB4-positive cells among DCs, macrophages, monocytes, and neutrophils in spleens from LILRB4-Tg mice (n = 5 mice per group). f, Comparative analysis of cell surface marker expression profiles (CD40, CD80, CD86, and CD206) on monocytes, macrophages, and DCs isolated from spleens of WT and LILRB4-Tg mice (n = 5 mice per group), assessed by flow cytometry. **g-j**, Detailed gating strategies employed for the comprehensive immunophenotyping of spleen immune cell populations from WT and LILRB4-Tg mice, including analysis of monocytes (CD11b⁺Ly6^hiLy6G⁻), neutrophils (CD11b⁺Ly6C⁻Ly6G⁺), cDC1 (CD11c⁺MCHII⁺XCR1⁺), cDC2(CD11c⁺MCHII⁺CD172a⁺), T cells (CD3⁺CD4⁺ or CD3⁺CD8⁺), NK cells (CD3⁻NK1.1⁺), and Treg cells (CD3⁺CD4⁺CD25⁺CD127⁻). Quantification of percentages of these immune cell subsets in the spleen is shown (n = 3 biological replicates). All data are present as mean ± s.d. p values were determined by two-tailed unpaired Student’s t-test.

Extended Data Fig. 5 | — a, Immunoblot of SCG2 expression in MC38, B16F10 and LLC cells infected with lentiviruses encoding pLVX-SCG2-puro or empty vector (pLVX-puro). b, Tumor weights from WT or LILRB4-Tg mice subcutaneously injected with indicated SCG2-expressing or vector control MC38 (n = 7 or 9), B16F10 (n = 6 or 8), or LLC (n = 7) cells. Tumors were harvested and weighed at the endpoint. c, Tumor growth curve and tumor weights in LILRB4-Tg mice and WT mice injected subcutaneously with WT MC38 cells or mouse SCG2-expressing MC38 cells (n = 7 or 9). d, Tumor weights of MC38 tumor-bearing LILRB4-Tg mice intratumorally injected with recombinant SCG2 or PBS (n = 7). e, Tumor weights and representative images from MC38-Vec and MC38-SCG2 tumors in LILRB4-Tg mice intraperitoneally injected with anti-LILRB4 antibody or control IgG (n = 8). f, Flow cytometric analysis of LILRB4 expression on tumor-infiltrating M-MDSCs and G-MDSCs from tumor-bearing mice treated with anti-LILRB4 antibody (n = 8). g, Frequency of CD4⁺ and CD8⁺ T cells in peripheral blood of mice treated intraperitoneally with anti-CD4/CD8 antibodies or control IgG. h, Tumor weights and representative images from MC38-Vec and MC38-SCG2 tumor-bearing mice treated with anti-CD4/CD8 antibodies or control IgG (n = 6 or 8). i, Percentages of peripheral blood M-MDSCs, G-MDSCs, and macrophages after intraperitoneal treatment with antiLy6G or anti-CSF1R antibody (n = 5). j. Tumor weights from MC38 tumor-bearing mice treated with anti-Ly6G or anti-CSF1R antibodies (n = 5). k, Immune cell infiltration analysis of B16F10-Vec and B16F10-SCG2 tumors from WT and LILRB4-Tg mice (n = 7). I, Frequency of GZMB⁺ cells among tumor-infiltrating CD8⁺ T cells in MC38-Vec and MC38-SCG2 tumors (n = 9). m-n, Analysis of tumor-infiltrating immune cell populations in MC38-Vec and MC38-SCG2 tumors from mice treated with anti-LILRB4 antibody or control IgG (n = 7). Tumor growth curves are present as mean ± s.e.m.; other data are present as mean ± s.d. p values were determined by two-tailed unpaired Student’st-test (d, i, and I), one-way ANOVA with Holm-Sidak’s multiple comparisons test (b, c (right), e, f, h, j, k, m, n), or two-way ANOVA with Tukey’s multiple comparisons test (c, left).

Extended Data Fig. 6 | — a, Immunoblot analysis of SCG2 expression levels in brain tissues and serum from LILRB4-Tg and LILRB4-Tg SCG2^−/− mice (n = 5). b, Percentages of immune cell populations in spleens from WT, SCG2^−/−, LILRB4-Tg, and LILRB4-Tg SCG2^−/−, mice (n = 6). c, Representative images of excised tumors (MC38, EO771, LLC, and CT2A models) from indicated mouse groups at study endpoint. **d-e**, Tumor growth curves and final tumor weights showing no significant difference between SCG2^−/− mice and WT mice bearing MC38 (n = 6) or LLC (n = 6) tumors. Tumor volumes were measured every two days; tumors were excised and weighed at endpoint. f, MC38 tumor growth in WT, LILRB4-Tg, mice and LILRB4-Tg SCG2^−/− mice. Tumor size was measured every two days, and tumor weights were measured on the final day (n = 9 or 10). g, Tumor growth in WT or ApoE^−/− mice. Mice were subcutaneously implanted with LLC cells (5×10⁵) on the right flank. Tumor size was measured every two days, and tumor weights were measured on the final day (n = 7 or 8). h, Representative immunofluorescence staining and quantification of CD8⁺ T cells and M-MDSCs in MC38 tumors from LILRB4-Tg mice and LILRB4-Tg SCG2^−/− mice (3 independent experiments, n = 10). Scale bars, 100 μm. i, Quantification of immune cell infiltration (M-MDSCs, G-MDSCs, macrophages, CDC1, CDC2, CD4⁺ and CD8⁺ T cells) in tumor tissues (B16F10 (n = 8), LLC (n = 9), EO771 (n = 8) and CT2A (n = 8)) isolated from LILRB4-Tg mice and LILRB4-Tg SCG2^−/− mice at endpoint. Tumor growth data are present as mean ± s.e.m.; all other data are present as mean ± s.d. p values were determined by two-tailed unpaired Student’s t-test or two-way ANOVA with Bonferroni’s multiple comparisons test (tumor volume (d-g)).

Extended Data Fig. 7 | — a, T cell proliferation assay in BM-MDSC-T cell co-culture assay. CFSE-labeled CD3⁺ T cells were co-cultured with LILRB4⁺ MDSCs at indicated MDSC-to-T cell ratios. T cell proliferation was analyzed by flow cytometry (n = 3 biological replicates). b, ELISA quantification of IFN-γ in culture supernatants from co-cultures described in Fig. 5a (n = 3 biological replicates). c, Nitric oxide (NO) concentrations in culture supernatants from co-cultures described in Fig. 5a, using Griess reagent assay (n = 3 biological replicates). d, ROS production in BM-MDSCs from WT or LILRB4-Tg mice, cultured for 24 h on plates coated with BSA or SCG2 (10 μg/ml). ROS levels were quantified by flow cytometry using DCFH-DA staining (n = 3 biological replicates). e, T cell proliferation of CFSE-labeled CD3⁺ T cells co-cultured with BM-MDSCs (1:1 ratio) from LILRB4-Tg mice. Co-cultures were stimulated with anti-CD3/CD28 beads in plates coated with BSA or SCG2, and treated with the NOS inhibitor L-NMMA (500 μM, MCE) or vehicle (DMSO). T cell proliferation was analyzed by flow cytometry (n = 3 biological replicates). f, Transwell migration assay of purified human CD14⁺ monocytes or T cells toward SCG2 (10 μg/ml) or BSA (control). Cells were placed in the upper chamber, and after 16 h incubation, migrated cells in the lower chamber were quantified by flow cytometry (n = 3 biological replicates). g. Flow cytometry analysis of CD80, CD86 and CD163 expression in BMDMs from WT and LILRB4-Tg mice. Cells were cultured for 24 h with LPS (100 ng/ml) in 96-well plates coated with BSA or SCG2, and treated with the anti-LILRB4 or IgG1 antibodies (n = 3 biological replicates). **h-i**, KEGG pathway (h) and Gene Ontology (GO) (i) enrichment analyses of genes differentially expressed in M-MDSCs isolated from LILRB4-Tg mice compared to LILRB4-Tg SCG2^−/− mice. Significantly enriched pathways are indicated. Statistical significance was calculated by hypergeometric test with Benjamini-Hochberg correction for multiple testing. j, Gene set enrichment analysis (GSEA) demonstrating enrichment of JAK-STAT signaling pathway genes in M-MDSCs from LILRB4-Tg mice compared with those from LILRB4-Tg SCG2^−/− mice. Data are present as mean ± s.d. p values were determined by one-way ANOVA with Holm-Sidak’s multiple comparisons test (b-g).

Extended Data Fig. 8 | — a, Quantification and statistical analysis of Western blot results shown in Fig. 6a (n = 3 biological replicates). **b-c**, Immunoblot (b) and quantification (c) of p-STAT3 levels in BM-MDSCs from LILRB4-Tg mice stimulated with or without IL-6 for 10,20,30 min in plates coated with BSA, SCG2 or ApoE (n = 3 biological replicates). **d-e**, Immunoblot (d) and quantification (e) of p-STAT3 levels in BM-MDSCs from WT, LILRB4-Tg, or LILRB4-Tg ApoE^−/− mice. Cells were stimulated with or without IL-6 for 10 min in plates coated with BSA, SCG2 or ApoE (n = 3 biological replicates). f, Quantification and statistical analysis of Western blot results shown in Fig. 6b (n = 3 biological replicates). g, mRNA level of Arg1 in BM-MDSCs stimulated with IL-6 under different conditions. Cells were incubated with varying concentration of IL-6 for 24 h or with IL-6 for indicated times in plates coated with BSA or SCG2 (n = 3 biological replicates). h, IP Flag-tagged LILRB4 and endogenous STAT3 in THP-1 cells. THP-1 LILRB4-Flag stable cells were subjected to IP with anti-Flag beads followed by IB with anti-STAT3 antibody. i, Co-IP of Flag-tagged STAT3 domain deletion and HA-tagged LILRB4 in HEK293T cells. j, SCG2 induced phosphorylation of STAT3 in a recombinant HEK293T system. HEK293T cells were co-transfected with LILRB4-Flag, ITIM-mutant LILRB4-Flag (2 μg), STAT3 (1 μg), Lyn (100 ng), and SCG2 (300 ng) as indicated. At 48 h post-transfection, cell lysates were collected and subjected to IP/IB analysis. k, Quantification and statistical analysis of Western blot results shown in Fig. 6j (n = 3 biological replicates). I, Immunoblot of p-STAT1 and p-STAT6 in BM-MDSCs from LILRB4-Tg mice stimulated with or without IFN-γ or IL-4 (20 ng/mL) for 10 min in plates coated with BSA or SCG2. m, Quantification and statistical analysis of Western blot results shown in Fig. 61 (n = 3 biological replicates). Data are presented as mean ± s.d. p values were determined by one-way ANOVA with Tukey’s multiple comparisons test (a, f, k, and m), two-way ANOVA with Tukey’s multiple comparisons test (c and e) or two-way ANOVA with Bonferroni’s multiple comparisons test (g).

Extended Data Fig. 9 | — a, T cell proliferation assay of CFSE-labeled CD3⁺ T cells co-cultured with CD14⁺ MDSCs isolated from cancer patient peripheral blood. Cells were cultured in plates coated with BSA or SCG2 in the presence of anti-CD3/CD28 activation beads. Anti-LILRB4 blocking antibodies were added as indicated. T cell proliferation was assessed by flow cytometry (n = 3 biological replicates). b, Quantification and statistical analysis of Western blot results shown in Fig. 7c (n = 3 biological replicates). c. Immunoblot of CD14⁺ cells from cancer patient blood stimulated with or without IL-6 for 10 min in 96-well plates coated with BSA or SCG2. Lysates were analyzed using the indicated antibodies. **d-e**, Immunoblot (d) and quantification (e) of IL-6-STAT3 signaling activity in CD14⁺ cells from health blood stimulated with or without IL-6 for 5,10 min in plates coated with BSA or SCG2 (n = 3 biological replicates). f, Percentage of human CD45⁺ cells in peripheral blood of humanized mice 8 weeks after CD34⁺ cells injection into NSG mice. Human cell engraftment was evaluated by flow cytometry. g. Immunoblot analysis confirming SCG2 ectopic expression in A375 melanoma cells transduced with pLVX-SCG2 lentivirus. h, Representative tumor images of the A375 tumor model in humanized mice intraperitoneally injected with anti-LILRB4 antibody or control IgG (n = 9 or 10). Data are presented as mean ± s.d. p values were determined by one-way ANOVA with Holm-Sidak’s multiple comparisons test (a), one-way ANOVA with Tukey’s multiple comparisons test (b), or two-way ANOVA with Bonferroni’s multiple comparisons test (e).

Extended Data Fig. 10 | — a, Proportions of myeloid cell subsets (monocytes, macrophages, DCs, and mast cells) across eight tumor types, analyzed from publicly available single-cell RNA-seq datasets. b, Bubble heatmap showing expression of selected signature genes across myeloid subsets. Dot size represents the percentage of cells expressing each gene; colored intensity indicates normalized average expression levels. c, Co-IP of Flag-tagged LILRB1–5 and STAT3 in the presence or absence of Lyn in HEK293T cells. HEK293T cells were co-transfected with Flag-LILRB1–5 (2 μg), Lyn (500 ng), and STAT3 (2 μg) plasmids. After 48 h, lysates were subjected to IP with anti-Flag beads and immunoblotting. d, Co-IP analysis of STAT3 binding to Flag-tagged LILRB1, LILRB4, or achimeric receptor combining the ECD of LILRB1 and ICD of LILRB4 (ECD^LLRBI-ICD^LLRB4), co-transfected in HEK293T cells. After 48 h, lysates were subjected to IP with anti-Flag beads and immunoblotting. **e-f**, Immunoblot (e) and quantification (f) of p-STAT3, p-SHP1, and p-SHP2 in human CD14⁺ monocytes from healthy donors stimulated with or without IL-6 (20 ng/mL) for 10 min in 96-well plates coated with BSA or indicated antibodies (n = 3 biological replicates). Data are presented as mean ± s.d. p values were determined by one-way ANOVA with Tukey’s multiple comparisons test.

Supplementary Material

Source Data Extended Fig 1

NIHMS2106102-supplement-Source_Data_Extended_Fig_1.xlsx^{(11.1KB, xlsx)}

Source Data Extended Fig 2

NIHMS2106102-supplement-Source_Data_Extended_Fig_2.xlsx^{(9.2KB, xlsx)}

Source Data Extended Fig 4

NIHMS2106102-supplement-Source_Data_Extended_Fig_4.xlsx^{(14.2KB, xlsx)}

Source Data Extended Fig 5

NIHMS2106102-supplement-Source_Data_Extended_Fig_5.xlsx^{(27.8KB, xlsx)}

Source Data Extended Fig 6

NIHMS2106102-supplement-Source_Data_Extended_Fig_6.xlsx^{(29.2KB, xlsx)}

Peer Review File

NIHMS2106102-supplement-Peer_Review_File.pdf^{(3MB, pdf)}

Source Data Extended Fig 7

NIHMS2106102-supplement-Source_Data_Extended_Fig_7.xlsx^{(15.3KB, xlsx)}

Source Data Extended Fig 8

NIHMS2106102-supplement-Source_Data_Extended_Fig_8.xlsx^{(20.1KB, xlsx)}

Source Data Extended Fig 9

NIHMS2106102-supplement-Source_Data_Extended_Fig_9.xlsx^{(15.9KB, xlsx)}

Source Data Extended Fig 10

NIHMS2106102-supplement-Source_Data_Extended_Fig_10.xlsx^{(11.9KB, xlsx)}

Source Data Fig 1

NIHMS2106102-supplement-Source_Data_Fig_1.xlsx^{(13KB, xlsx)}

Source Data Fig 2

NIHMS2106102-supplement-Source_Data_Fig_2.xlsx^{(9.5KB, xlsx)}

Source Data Fig 3

NIHMS2106102-supplement-Source_Data_Fig_3.xlsx^{(33.9KB, xlsx)}

Source Data Fig 4

NIHMS2106102-supplement-Source_Data_Fig_4.xlsx^{(33.3KB, xlsx)}

Source Data Fig 5

NIHMS2106102-supplement-Source_Data_Fig_5.xlsx^{(19.2KB, xlsx)}

Source Data Fig 6

NIHMS2106102-supplement-Source_Data_Fig_6.xlsx^{(13.7KB, xlsx)}

Supplementary Table 1

NIHMS2106102-supplement-Supplementary_Table_1.xlsx^{(9.4KB, xlsx)}

Source Data Fig 7

NIHMS2106102-supplement-Source_Data_Fig_7.xlsx^{(16.1KB, xlsx)}

Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41590-025-02233-4.

Acknowledgements

This work was supported by the National Cancer Institute (R01CA248736, R01CA263079 and Lung Cancer SPORE Development Research Program; C.C.Z.), the Cancer Prevention and Research Institute of Texas (RP220032, C.C.Z.; RP15150551 and RP190561, Z.A.), the Welch Foundation (AU-0042–20030616, Z.A.; I-1702, X.Z.), Immune-Onc Therapeutics (Sponsored Research Grant 111077, C.C.Z.) and National Institutes of Health (R35GM130289, X.Z.), the CPRIT training grant (RP210041, X.Y.; RP210041, L.C.). This work received support from University of Texas Southwestern Simmons Comprehensive Cancer Center’s Tissue Management Shared Resource and was supported by the National Cancer Institute of the National Institutes of Health (P30CA142543, C.L.).

Footnotes

Competing interests

Authors R.H., Y.H., M.D., W.X., N.Z., Z.A. and C.C.Z. are listed as inventors on relevant patent applications that were exclusively licensed to Immune-Onc Therapeutics by the Board of Regents of the University of Texas System. Authors M.D., Z.A., N.Z. and C.C.Z. hold equity in and had Sponsored Research Agreements with Immune-Onc Therapeutics. The University of Texas has a financial interest in Immune-Onc in the form of equity and licensing. The other authors declare no competing interests.

Additional information

Extended data is available for this paper at https://doi.org/10.1038/s41590-025-02233-4.

References

1.Sharma P & Allison JP The future of immune checkpoint therapy. Science 348, 56–61 (2015). [DOI] [PubMed] [Google Scholar]
2.Sharma P et al. Immune checkpoint therapy—current perspectives and future directions. Cell 186, 1652–1669 (2023). [DOI] [PubMed] [Google Scholar]
3.Zebley CC, Zehn D, Gottschalk S & Chi H T cell dysfunction and therapeutic intervention in cancer. Nat. Immunol. 25, 1344–1354 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Barry ST, Gabrilovich DI, Sansom OJ, Campbell AD & Morton JP Therapeutic targeting of tumour myeloid cells. Nat. Rev. Cancer 23, 216–237 (2023). [DOI] [PubMed] [Google Scholar]
5.Mitchell K et al. Secretoneurin is a secretogranin-2 derived hormonal peptide in vertebrate neuroendocrine systems. Gen. Comp. Endocrinol. 299, 113588 (2020). [DOI] [PubMed] [Google Scholar]
6.Bartolomucci A et al. The extended granin family: structure, function, and biomedical implications. Endocr. Rev. 32, 755–797 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Troger J et al. Granin-derived peptides. Prog. Neurobiol. 154, 37–61 (2017). [DOI] [PubMed] [Google Scholar]
8.Kirchmair R, Hogue-Angeletti R, Gutierrez J, Fischer-Colbrie R & Winkler H Secretoneurin—a neuropeptide generated in brain, adrenal medulla and other endocrine tissues by proteolytic processing of secretogranin II (chromogranin C). Neuroscience 53, 359–365 (1993). [DOI] [PubMed] [Google Scholar]
9.Lim SH et al. Nanoplasmonic immunosensor for the detection of SCG2, a candidate serum biomarker for the early diagnosis of neurodevelopmental disorder. Sci. Rep. 11, 22764 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Tao D et al. Identification of angiogenesis-related prognostic biomarkers associated with immune cell infiltration in breast cancer. Front. Cell Dev. Biol. 10, 853324 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Weng S et al. SCG2: a prognostic marker that pinpoints chemotherapy and immunotherapy in colorectal cancer. Front. Immunol. 13, 873871 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Wang H et al. SCG2 is a prognostic biomarker associated with immune infiltration and macrophage polarization in colorectal cancer. Front. Cell Dev. Biol. 9, 795133 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Zhou J et al. Promoting effect and immunologic role of secretogranin II on bladder cancer progression via regulating MAPK and NF-κB pathways. Apoptosis 29, 121–141 (2024). [DOI] [PubMed] [Google Scholar]
14.Fukumoto W et al. Potential therapeutic target secretogranin II might cooperate with hypoxia-inducible factor 1α in sunitinib-resistant renal cell carcinoma. Cancer Sci. 114, 3946–3956 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Cury SS et al. Tumor transcriptome reveals high expression of IL-8 in non-small cell lung cancer patients with low pectoralis muscle area and reduced survival. Cancers 11, 1251 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Steinfass T et al. Secretogranin II influences the assembly and function of MHC class I in melanoma. Exp. Hematol. Oncol. 12, 29 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Hou J et al. Antibody-mediated targeting of human microglial leukocyte Ig-like receptor B4 attenuates amyloid pathology in a mouse model. Sci. Transl. Med. 16, eadj9052 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Hirayasu K & Arase H Functional and genetic diversity of leukocyte immunoglobulin-like receptor and implication for disease associations. J. Hum. Genet. 60, 703–708 (2015). [DOI] [PubMed] [Google Scholar]
19.Van der Touw W, Chen HM, Pan PY & Chen SH LILRB receptor-mediated regulation of myeloid cell maturation and function. Cancer Immunol. Immunother. 66, 1079–1087 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Thomas R, Matthias T & Witte T Leukocyte immunoglobulin-like receptors as new players in autoimmunity. Clin. Rev. Allergy Immunol. 38, 159–162 (2010). [DOI] [PubMed] [Google Scholar]
21.Deng M et al. Leukocyte immunoglobulin-like receptor subfamily B (LILRB): therapeutic targets in cancer. Antib. Ther. 4, 16–33 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Redondo-García S et al. Human leukocyte immunoglobulin-like receptors in health and disease. Front. Immunol. 14, 1282874 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Dobrowolska H et al. Expression of immune inhibitory receptor ILT3 in acute myeloid leukemia with monocytic differentiation. Cytom. B 84, 21–29 (2013). [DOI] [PubMed] [Google Scholar]
24.Deng M et al. LILRB4 signalling in leukaemia cells mediates T cell suppression and tumour infiltration. Nature 562, 605–609 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Sharma N, Atolagbe OT, Ge Z & Allison JP LILRB4 suppresses immunity in solid tumors and is a potential target for immunotherapy. J. Exp. Med. 218, e20201811 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Sultan H et al. Neoantigen-specific cytotoxic TR1 CD4 T cells suppress cancer immunotherapy. Nature 632, 182–191 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Singh L et al. ILT3 (LILRB4) promotes the immunosuppressive function of tumor-educated human monocytic myeloid-derived suppressor cells. Mol. Cancer Res. 19, 702–716 (2021). [DOI] [PubMed] [Google Scholar]
28.Paavola KJ et al. The fibronectin–ILT3 interaction functions as a stromal checkpoint that suppresses myeloid cells. Cancer Immunol. Res. 9, 1283–1297 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Su MT et al. Blockade of checkpoint ILT3/LILRB4/gp49B binding to fibronectin ameliorates autoimmune disease in BXSB/Yaa mice. Int. Immunol. 33, 447–458 (2021). [DOI] [PubMed] [Google Scholar]
30.Verschueren E et al. The immunoglobulin superfamily receptome defines cancer-relevant networks associated with clinical outcome. Cell 182, 329–344 (2020). [DOI] [PubMed] [Google Scholar]
31.Deng M et al. A motif in LILRB2 critical for ANGPTL2 binding and activation. Blood 124, 924–935 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Huang R et al. LILRB3 supports immunosuppressive activity of myeloid cells and tumor development. Cancer Immunol. Res. 12, 350–362 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Xu Z et al. ILT3.Fc–CD166 interaction induces inactivation of p70 S6 kinase and inhibits tumor cell growth. J. Immunol. 200, 1207–1219 (2017). [DOI] [PubMed] [Google Scholar]
34.Wang Y et al. Discovery of galectin-8 as an LILRB4 ligand driving M-MDSCs defines a class of antibodies to fight solid tumors. Cell Rep. Med. 5, 101374 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Lu HK et al. Leukocyte Ig-like receptor B4 (LILRB4) is a potent inhibitor of FcγRI-mediated monocyte activation via dephosphorylation of multiple kinases. J. Biol. Chem. 284, 34839–34848 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Morse JW et al. Fcγ receptors promote antibody-induced LILRB4 internalization and immune regulation of monocytic AML. Antib. Ther. 7, 13–27 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Veglia F, Sanseviero E & Gabrilovich DI Myeloid-derived suppressor cells in the era of increasing myeloid cell diversity. Nat. Rev. Immunol. 21, 485–498 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Johnson DE, O’Keefe RA & Grandis JR Targeting the IL-6/JAK/STAT3 signalling axis in cancer. Nat. Rev. Clin. Oncol. 15, 234–248 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Vasquez-Dunddel D et al. STAT3 regulates arginase-I in myeloid-derived suppressor cells from cancer patients. J. Clin. Invest. 123, 1580–1589 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Schreiner SJ, Schiavone AP & Smithgall TE Activation of STAT3 by the Src family kinase Hck requires a functional SH3 domain. J. Biol. Chem. 277, 45680–45687 (2002). [DOI] [PubMed] [Google Scholar]
41.Cheng S et al. A pan-cancer single-cell transcriptional atlas of tumor infiltrating myeloid cells. Cell 184, 792–809 (2021). [DOI] [PubMed] [Google Scholar]
42.Trebak F et al. A potential role for the secretogranin II-derived peptide EM66 in the hypothalamic regulation of feeding behaviour. J. Neuroendocrinol. 10.1111/jne.12459 (2017). [DOI] [PubMed] [Google Scholar]
43.Christofides A et al. SHP-2 and PD-1–SHP-2 signaling regulate myeloid cell differentiation and antitumor responses. Nat. Immunol. 24, 55–68 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Lin C, Zhao P, Sun G, Liu N & Ji J SCG2 mediates blood–brain barrier dysfunction and schizophrenia-like behaviors after traumatic brain injury. FASEB J. 38, e70016 (2024). [DOI] [PubMed] [Google Scholar]
45.Song Y et al. FRA-1 and STAT3 synergistically regulate activation of human MMP-9 gene. Mol. Immunol. 45, 137–143 (2008). [DOI] [PubMed] [Google Scholar]
46.Huang H Matrix metalloproteinase-9 (MMP-9) as a cancer biomarker and MMP-9 biosensors: recent advances. Sensors 18, 3249 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Rundhaug JE Matrix metalloproteinases and angiogenesis. J. Cell. Mol. Med. 9, 267–285 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Vafadari B, Salamian A & Kaczmarek L MMP-9 in translation: from molecule to brain physiology, pathology, and therapy. J. Neurochem. 139, 91–114 (2016). [DOI] [PubMed] [Google Scholar]
49.Reinhard SM, Razak K & Ethell IM A delicate balance: role of MMP-9 in brain development and pathophysiology of neurodevelopmental disorders. Front. Cell. Neurosci. 9, 280 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Kang L, Mayes WH, James FD, Bracy DP & Wasserman DH Matrix metalloproteinase 9 opposes diet-induced muscle insulin resistance in mice. Diabetologia 57, 603–613 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Zhu Y Metalloproteases in gonad formation and ovulation. Gen. Comp. Endocrinol. 314, 113924 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Daniel JA et al. Endotoxin inhibition of luteinizing hormone in sheep. Domest. Anim. Endocrinol. 25, 13–19 (2003). [DOI] [PubMed] [Google Scholar]
53.Wu G et al. LILRB3 supports acute myeloid leukemia development and regulates T-cell antitumor immune responses through the TRAF2–cFLIP–NF-κB signaling axis. Nat. Cancer 2, 1170–1184 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Chen H et al. Antagonistic anti-LILRB1 monoclonal antibody regulates antitumor functions of natural killer cells. J. Immunother. Cancer 8, e000515 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Hitt BD et al. Anti-tau antibodies targeting a conformation-dependent epitope selectively bind seeds. J. Biol. Chem. 299, 105252 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Mitchell K et al. Targeted mutation of secretogranin-2 disrupts sexual behavior and reproduction in zebrafish. Proc. Natl Acad. Sci. USA 117, 12772–12783 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Frecha C, Fusil F, Cosset FL & Verhoeyen E In vivo gene delivery into hCD34⁺ cells in a humanized mouse model. Methods Mol. Biol. 737, 367–390 (2011). [DOI] [PubMed] [Google Scholar]
58.Alattar H, Xu H, Zenke M & Lutz MB Fully functional monocytic MDSC generation from the murine HoxB8 cell line. Eur. J. Immunol. 53, e2350466 (2023). [DOI] [PubMed] [Google Scholar]
59.Eckert I, Ribechini E & Lutz MB In vitro generation of murine myeloid-derived suppressor cells, analysis of markers, developmental commitment, and function. Methods Mol. Biol. 2236, 99–114 (2021). [DOI] [PubMed] [Google Scholar]
60.Ge SX, Son EW & Yao R iDEP: an integrated web application for differential expression and pathway analysis of RNA-seq data. BMC Bioinformatics 19, 534 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Hao Y et al. Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Narayan A, Berger B & Cho H Assessing single-cell transcriptomic variability through density-preserving data visualization. Nat. Biotechnol. 39, 765–774 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Hu C et al. CellMarker 2.0: an updated database of manually curated cell markers in human/mouse and web tools based on scRNA-seq data. Nucleic Acids Res. 51, D870–D876 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Wu T et al. clusterProfiler 4.0: a universal enrichment tool for interpreting omics data. Innovation 2, 100141 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Andreatta M & Carmona SJ UCell: robust and scalable single-cell gene signature scoring. Comput. Struct. Biotechnol. J. 19, 3796–3798 (2021). [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

Source Data Extended Fig 1

NIHMS2106102-supplement-Source_Data_Extended_Fig_1.xlsx^{(11.1KB, xlsx)}

Source Data Extended Fig 2

NIHMS2106102-supplement-Source_Data_Extended_Fig_2.xlsx^{(9.2KB, xlsx)}

Source Data Extended Fig 4

NIHMS2106102-supplement-Source_Data_Extended_Fig_4.xlsx^{(14.2KB, xlsx)}

Source Data Extended Fig 5

NIHMS2106102-supplement-Source_Data_Extended_Fig_5.xlsx^{(27.8KB, xlsx)}

Source Data Extended Fig 6

NIHMS2106102-supplement-Source_Data_Extended_Fig_6.xlsx^{(29.2KB, xlsx)}

Peer Review File

NIHMS2106102-supplement-Peer_Review_File.pdf^{(3MB, pdf)}

Source Data Extended Fig 7

NIHMS2106102-supplement-Source_Data_Extended_Fig_7.xlsx^{(15.3KB, xlsx)}

Source Data Extended Fig 8

NIHMS2106102-supplement-Source_Data_Extended_Fig_8.xlsx^{(20.1KB, xlsx)}

Source Data Extended Fig 9

NIHMS2106102-supplement-Source_Data_Extended_Fig_9.xlsx^{(15.9KB, xlsx)}

Source Data Extended Fig 10

NIHMS2106102-supplement-Source_Data_Extended_Fig_10.xlsx^{(11.9KB, xlsx)}

Source Data Fig 1

NIHMS2106102-supplement-Source_Data_Fig_1.xlsx^{(13KB, xlsx)}

Source Data Fig 2

NIHMS2106102-supplement-Source_Data_Fig_2.xlsx^{(9.5KB, xlsx)}

Source Data Fig 3

NIHMS2106102-supplement-Source_Data_Fig_3.xlsx^{(33.9KB, xlsx)}

Source Data Fig 4

NIHMS2106102-supplement-Source_Data_Fig_4.xlsx^{(33.3KB, xlsx)}

Source Data Fig 5

NIHMS2106102-supplement-Source_Data_Fig_5.xlsx^{(19.2KB, xlsx)}

Source Data Fig 6

NIHMS2106102-supplement-Source_Data_Fig_6.xlsx^{(13.7KB, xlsx)}

Supplementary Table 1

NIHMS2106102-supplement-Supplementary_Table_1.xlsx^{(9.4KB, xlsx)}

Source Data Fig 7

NIHMS2106102-supplement-Source_Data_Fig_7.xlsx^{(16.1KB, xlsx)}

Data Availability Statement

[R1] 1.Sharma P & Allison JP The future of immune checkpoint therapy. Science 348, 56–61 (2015). [DOI] [PubMed] [Google Scholar]

[R2] 2.Sharma P et al. Immune checkpoint therapy—current perspectives and future directions. Cell 186, 1652–1669 (2023). [DOI] [PubMed] [Google Scholar]

[R3] 3.Zebley CC, Zehn D, Gottschalk S & Chi H T cell dysfunction and therapeutic intervention in cancer. Nat. Immunol. 25, 1344–1354 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R4] 4.Barry ST, Gabrilovich DI, Sansom OJ, Campbell AD & Morton JP Therapeutic targeting of tumour myeloid cells. Nat. Rev. Cancer 23, 216–237 (2023). [DOI] [PubMed] [Google Scholar]

[R5] 5.Mitchell K et al. Secretoneurin is a secretogranin-2 derived hormonal peptide in vertebrate neuroendocrine systems. Gen. Comp. Endocrinol. 299, 113588 (2020). [DOI] [PubMed] [Google Scholar]

[R6] 6.Bartolomucci A et al. The extended granin family: structure, function, and biomedical implications. Endocr. Rev. 32, 755–797 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R7] 7.Troger J et al. Granin-derived peptides. Prog. Neurobiol. 154, 37–61 (2017). [DOI] [PubMed] [Google Scholar]

[R8] 8.Kirchmair R, Hogue-Angeletti R, Gutierrez J, Fischer-Colbrie R & Winkler H Secretoneurin—a neuropeptide generated in brain, adrenal medulla and other endocrine tissues by proteolytic processing of secretogranin II (chromogranin C). Neuroscience 53, 359–365 (1993). [DOI] [PubMed] [Google Scholar]

[R9] 9.Lim SH et al. Nanoplasmonic immunosensor for the detection of SCG2, a candidate serum biomarker for the early diagnosis of neurodevelopmental disorder. Sci. Rep. 11, 22764 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R10] 10.Tao D et al. Identification of angiogenesis-related prognostic biomarkers associated with immune cell infiltration in breast cancer. Front. Cell Dev. Biol. 10, 853324 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R11] 11.Weng S et al. SCG2: a prognostic marker that pinpoints chemotherapy and immunotherapy in colorectal cancer. Front. Immunol. 13, 873871 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R12] 12.Wang H et al. SCG2 is a prognostic biomarker associated with immune infiltration and macrophage polarization in colorectal cancer. Front. Cell Dev. Biol. 9, 795133 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R13] 13.Zhou J et al. Promoting effect and immunologic role of secretogranin II on bladder cancer progression via regulating MAPK and NF-κB pathways. Apoptosis 29, 121–141 (2024). [DOI] [PubMed] [Google Scholar]

[R14] 14.Fukumoto W et al. Potential therapeutic target secretogranin II might cooperate with hypoxia-inducible factor 1α in sunitinib-resistant renal cell carcinoma. Cancer Sci. 114, 3946–3956 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R15] 15.Cury SS et al. Tumor transcriptome reveals high expression of IL-8 in non-small cell lung cancer patients with low pectoralis muscle area and reduced survival. Cancers 11, 1251 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R16] 16.Steinfass T et al. Secretogranin II influences the assembly and function of MHC class I in melanoma. Exp. Hematol. Oncol. 12, 29 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R17] 17.Hou J et al. Antibody-mediated targeting of human microglial leukocyte Ig-like receptor B4 attenuates amyloid pathology in a mouse model. Sci. Transl. Med. 16, eadj9052 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R18] 18.Hirayasu K & Arase H Functional and genetic diversity of leukocyte immunoglobulin-like receptor and implication for disease associations. J. Hum. Genet. 60, 703–708 (2015). [DOI] [PubMed] [Google Scholar]

[R19] 19.Van der Touw W, Chen HM, Pan PY & Chen SH LILRB receptor-mediated regulation of myeloid cell maturation and function. Cancer Immunol. Immunother. 66, 1079–1087 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R20] 20.Thomas R, Matthias T & Witte T Leukocyte immunoglobulin-like receptors as new players in autoimmunity. Clin. Rev. Allergy Immunol. 38, 159–162 (2010). [DOI] [PubMed] [Google Scholar]

[R21] 21.Deng M et al. Leukocyte immunoglobulin-like receptor subfamily B (LILRB): therapeutic targets in cancer. Antib. Ther. 4, 16–33 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R22] 22.Redondo-García S et al. Human leukocyte immunoglobulin-like receptors in health and disease. Front. Immunol. 14, 1282874 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R23] 23.Dobrowolska H et al. Expression of immune inhibitory receptor ILT3 in acute myeloid leukemia with monocytic differentiation. Cytom. B 84, 21–29 (2013). [DOI] [PubMed] [Google Scholar]

[R24] 24.Deng M et al. LILRB4 signalling in leukaemia cells mediates T cell suppression and tumour infiltration. Nature 562, 605–609 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R25] 25.Sharma N, Atolagbe OT, Ge Z & Allison JP LILRB4 suppresses immunity in solid tumors and is a potential target for immunotherapy. J. Exp. Med. 218, e20201811 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R26] 26.Sultan H et al. Neoantigen-specific cytotoxic TR1 CD4 T cells suppress cancer immunotherapy. Nature 632, 182–191 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R27] 27.Singh L et al. ILT3 (LILRB4) promotes the immunosuppressive function of tumor-educated human monocytic myeloid-derived suppressor cells. Mol. Cancer Res. 19, 702–716 (2021). [DOI] [PubMed] [Google Scholar]

[R28] 28.Paavola KJ et al. The fibronectin–ILT3 interaction functions as a stromal checkpoint that suppresses myeloid cells. Cancer Immunol. Res. 9, 1283–1297 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R29] 29.Su MT et al. Blockade of checkpoint ILT3/LILRB4/gp49B binding to fibronectin ameliorates autoimmune disease in BXSB/Yaa mice. Int. Immunol. 33, 447–458 (2021). [DOI] [PubMed] [Google Scholar]

[R30] 30.Verschueren E et al. The immunoglobulin superfamily receptome defines cancer-relevant networks associated with clinical outcome. Cell 182, 329–344 (2020). [DOI] [PubMed] [Google Scholar]

[R31] 31.Deng M et al. A motif in LILRB2 critical for ANGPTL2 binding and activation. Blood 124, 924–935 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R32] 32.Huang R et al. LILRB3 supports immunosuppressive activity of myeloid cells and tumor development. Cancer Immunol. Res. 12, 350–362 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R33] 33.Xu Z et al. ILT3.Fc–CD166 interaction induces inactivation of p70 S6 kinase and inhibits tumor cell growth. J. Immunol. 200, 1207–1219 (2017). [DOI] [PubMed] [Google Scholar]

[R34] 34.Wang Y et al. Discovery of galectin-8 as an LILRB4 ligand driving M-MDSCs defines a class of antibodies to fight solid tumors. Cell Rep. Med. 5, 101374 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R35] 35.Lu HK et al. Leukocyte Ig-like receptor B4 (LILRB4) is a potent inhibitor of FcγRI-mediated monocyte activation via dephosphorylation of multiple kinases. J. Biol. Chem. 284, 34839–34848 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R36] 36.Morse JW et al. Fcγ receptors promote antibody-induced LILRB4 internalization and immune regulation of monocytic AML. Antib. Ther. 7, 13–27 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R37] 37.Veglia F, Sanseviero E & Gabrilovich DI Myeloid-derived suppressor cells in the era of increasing myeloid cell diversity. Nat. Rev. Immunol. 21, 485–498 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R38] 38.Johnson DE, O’Keefe RA & Grandis JR Targeting the IL-6/JAK/STAT3 signalling axis in cancer. Nat. Rev. Clin. Oncol. 15, 234–248 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R39] 39.Vasquez-Dunddel D et al. STAT3 regulates arginase-I in myeloid-derived suppressor cells from cancer patients. J. Clin. Invest. 123, 1580–1589 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R40] 40.Schreiner SJ, Schiavone AP & Smithgall TE Activation of STAT3 by the Src family kinase Hck requires a functional SH3 domain. J. Biol. Chem. 277, 45680–45687 (2002). [DOI] [PubMed] [Google Scholar]

[R41] 41.Cheng S et al. A pan-cancer single-cell transcriptional atlas of tumor infiltrating myeloid cells. Cell 184, 792–809 (2021). [DOI] [PubMed] [Google Scholar]

[R42] 42.Trebak F et al. A potential role for the secretogranin II-derived peptide EM66 in the hypothalamic regulation of feeding behaviour. J. Neuroendocrinol. 10.1111/jne.12459 (2017). [DOI] [PubMed] [Google Scholar]

[R43] 43.Christofides A et al. SHP-2 and PD-1–SHP-2 signaling regulate myeloid cell differentiation and antitumor responses. Nat. Immunol. 24, 55–68 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R44] 44.Lin C, Zhao P, Sun G, Liu N & Ji J SCG2 mediates blood–brain barrier dysfunction and schizophrenia-like behaviors after traumatic brain injury. FASEB J. 38, e70016 (2024). [DOI] [PubMed] [Google Scholar]

[R45] 45.Song Y et al. FRA-1 and STAT3 synergistically regulate activation of human MMP-9 gene. Mol. Immunol. 45, 137–143 (2008). [DOI] [PubMed] [Google Scholar]

[R46] 46.Huang H Matrix metalloproteinase-9 (MMP-9) as a cancer biomarker and MMP-9 biosensors: recent advances. Sensors 18, 3249 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R47] 47.Rundhaug JE Matrix metalloproteinases and angiogenesis. J. Cell. Mol. Med. 9, 267–285 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R48] 48.Vafadari B, Salamian A & Kaczmarek L MMP-9 in translation: from molecule to brain physiology, pathology, and therapy. J. Neurochem. 139, 91–114 (2016). [DOI] [PubMed] [Google Scholar]

[R49] 49.Reinhard SM, Razak K & Ethell IM A delicate balance: role of MMP-9 in brain development and pathophysiology of neurodevelopmental disorders. Front. Cell. Neurosci. 9, 280 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R50] 50.Kang L, Mayes WH, James FD, Bracy DP & Wasserman DH Matrix metalloproteinase 9 opposes diet-induced muscle insulin resistance in mice. Diabetologia 57, 603–613 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R51] 51.Zhu Y Metalloproteases in gonad formation and ovulation. Gen. Comp. Endocrinol. 314, 113924 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R52] 52.Daniel JA et al. Endotoxin inhibition of luteinizing hormone in sheep. Domest. Anim. Endocrinol. 25, 13–19 (2003). [DOI] [PubMed] [Google Scholar]

[R53] 53.Wu G et al. LILRB3 supports acute myeloid leukemia development and regulates T-cell antitumor immune responses through the TRAF2–cFLIP–NF-κB signaling axis. Nat. Cancer 2, 1170–1184 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R54] 54.Chen H et al. Antagonistic anti-LILRB1 monoclonal antibody regulates antitumor functions of natural killer cells. J. Immunother. Cancer 8, e000515 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R55] 55.Hitt BD et al. Anti-tau antibodies targeting a conformation-dependent epitope selectively bind seeds. J. Biol. Chem. 299, 105252 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R56] 56.Mitchell K et al. Targeted mutation of secretogranin-2 disrupts sexual behavior and reproduction in zebrafish. Proc. Natl Acad. Sci. USA 117, 12772–12783 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R57] 57.Frecha C, Fusil F, Cosset FL & Verhoeyen E In vivo gene delivery into hCD34⁺ cells in a humanized mouse model. Methods Mol. Biol. 737, 367–390 (2011). [DOI] [PubMed] [Google Scholar]

[R58] 58.Alattar H, Xu H, Zenke M & Lutz MB Fully functional monocytic MDSC generation from the murine HoxB8 cell line. Eur. J. Immunol. 53, e2350466 (2023). [DOI] [PubMed] [Google Scholar]

[R59] 59.Eckert I, Ribechini E & Lutz MB In vitro generation of murine myeloid-derived suppressor cells, analysis of markers, developmental commitment, and function. Methods Mol. Biol. 2236, 99–114 (2021). [DOI] [PubMed] [Google Scholar]

[R60] 60.Ge SX, Son EW & Yao R iDEP: an integrated web application for differential expression and pathway analysis of RNA-seq data. BMC Bioinformatics 19, 534 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R61] 61.Hao Y et al. Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R62] 62.Narayan A, Berger B & Cho H Assessing single-cell transcriptomic variability through density-preserving data visualization. Nat. Biotechnol. 39, 765–774 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R63] 63.Hu C et al. CellMarker 2.0: an updated database of manually curated cell markers in human/mouse and web tools based on scRNA-seq data. Nucleic Acids Res. 51, D870–D876 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R64] 64.Wu T et al. clusterProfiler 4.0: a universal enrichment tool for interpreting omics data. Innovation 2, 100141 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R65] 65.Andreatta M & Carmona SJ UCell: robust and scalable single-cell gene signature scoring. Comput. Struct. Biotechnol. J. 19, 3796–3798 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Secretogranin 2 binds LILRB4 resulting in immunosuppression

Xing Yang

Ryan Huang

Meng Fang

Yubo He

Jingjing Xie

Xiaoye Liu

Chengcheng Zhang

Qi Lou

Mi Deng

Wei Xiong

Cheryl Lewis

Zade Sadek

Ankit Gupta

Lianqi Chen

Xuewu Zhang

Lei Guo

Lin Xu

Ningyan Zhang

Zhiqiang An

Cheng Cheng Zhang

Abstract

Results

SCG2 interacts specifically with LILRB4

Fig. 1 |. SCG2 interacts specifically with LILRB4.

SCG2 induces LILRB4 activation and function in cultured cells

Fig. 2 |. SCG2 induces LILRB4 activation and function in cultured cells.

SCG2 supports tumor development through LILRB4

Fig. 3 |. SCG2 supports tumor development through LILRB4.

SCG2 deficiency impairs tumor progression in LILRB4-Tg mice

Fig. 4 |. SCG2 deficiency impairs tumor progression in LILRB4-Tg mice.

LILRB4 enhances immunosuppressive activity of MDSCs

Fig. 5 |. LILRB4 enhances immunosuppressive activity of MDSCs.

SCG2−LILRB4 interaction strengthens IL-6−STAT3 signaling

Fig. 6 |. SCG2–LILRB4 interaction strengthens IL-6–STAT3 signaling.

LILRB4 supports the immunosuppressive function of human M-MDSCs

Fig. 7 |. LILRB4 supports immunosuppressive function of human M-MDSCs.

Discussion

Online content

Methods

Cells

Mice

Flow cytometry of tumor immune cells

Chimeric receptor reporter

ELISA

IP and immunoblotting

Pulldown

Surface interferometry

HEK293T reverse binding assay

HEK293T reconstituted signaling assay

THP-1 LILRB4–FcR coligation assay

Generation of stable SCG2-expressing cell lines

In vivo tumor growth and treatment

Mouse MDSC–T cell coculture

Human MDSC–T cell coculture

Mouse M-MDSCRNA-seq

Human scRNA-seq

Quantification and statistical analysis

Reporting summary

Data availability

Extended Data

Extended Data Fig. 1 |. SCG2 binds to LILRB4 specifically.

Extended Data Fig. 2 |. SCG2 does not bind LILRB4 F83A mutant.

Extended Data Fig. 3 |. Correlation of SCG2 with MDSCs, CTLs, or patient survival.

Extended Data Fig. 4 |. LILRB4 transgene does not change immune profile.

Extended Data Fig. 5 |. SCG2-LILRB4 interaction promotes tumor growth.

Extended Data Fig. 6 |. SCG2 loss restricts tumor progression in LILRB4-Tg mice.

Extended Data Fig. 7 |. SCG2-LILRB4 axis strengthens MDSC suppression.

Extended Data Fig. 8 |. SCG2-LILRB4 interaction upregulates IL-6-STAT3 activation.

Extended Data Fig. 9 |. LILRB4 promotes human M-MDSC suppressive function.

Extended Data Fig. 10 |. Conserved function of LILRBs in M-MDSC activation.

Supplementary Material

Acknowledgements

Footnotes

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS