Targeting Interactions Between Siglec-10 and α3β1 Integrin Enhances Macrophage-Mediated Phagocytosis of Pancreatic Cancer

Abstract

Tumor-associated macrophages (TAM) in the pancreatic ductal adenocarcinoma (PDAC) tumor microenvironment (TME) exhibit immunosuppressive phenotypes and impaired phagocytic activity, facilitating tumor progression and immune evasion. In this study, we identified integrin α3β1, composed of ITGA3 and ITGB1 subunits, as a sialylated glycoprotein ligand for Siglec-10, an inhibitory glyco-immune checkpoint receptor highly expressed on TAMs in PDAC. The interaction between Siglec-10 on TAMs and α3β1 on PDAC cells suppressed macrophage-mediated phagocytosis, thereby promoting immune evasion. Consistently, disrupting Siglec-10 interactions using mAbs significantly enhanced macrophage phagocytosis of PDAC cells and alleviated myeloid cell-mediated inhibition of T-cell proliferation and activation in vitro. In both a xenograft mouse model engrafted with human macrophages and a human Siglec-10 transgenic mouse model, targeting Siglec-10 with mAbs reduced PDAC growth. These findings suggest that Siglec-10 interactions are key mediators of TAM-driven immune evasion in PDAC and highlight the therapeutic potential of targeting these interactions to restore antitumor immunity.

Significance: Pancreatic tumor cells exploit integrin α3β1 to engage the immunosuppressive checkpoint receptor Siglec-10 on myeloid cells, driving immune evasion, which can be targeted with antibody-mediated blockade of Siglec-10 to restore antitumor immunity.

Figure 1.. Siglec-10 is predominantly expressed on myeloid cells within the PDAC TME and is associated with worse disease progression.

(a) UMAP visualization of immune cell populations within the PDAC TME. (b) UMAP of Siglec-10 expression and a pie chart illustrating the percentage of Siglec-10+ cells within each myeloid cell subset. (c) A comparison of the expression levels of various Siglecs in myeloid cells within the PDAC TME. Kruskal–Wallis test with Dunn’s multiple comparisons correction. Means with standard deviations (SD) are shown. Percentages shown represent the proportion of myeloid cells expressing each of these Siglecs. (d) A comparison of the gene expression of the indicated genes between Siglec-10+ and Siglec-10− macrophages, MDSCs, and DCs. Analysis was performed using the Bioturing’s Talk2Data V4 platform. (e) Siglec-10 expression in the PDAC TME across different disease states. Kruskal–Wallis test with Dunn’s multiple comparisons correction. Means with standard errors of the mean (SEM) are shown. (f) Comparison of Siglec-10 expression between normal tissues and pancreatic adenocarcinoma (PAAD) tissues from the TCGA dataset. Unpaired t test. (g) Survival analysis of pancreatic tumor patients in the Human Protein Atlas dataset showing the correlation between Siglec-10 expression and overall survival.

Figure 2.. PDAC cell lines express Siglec-10 ligands on glycoproteins other than CD24.

Flow cytometric analysis of Siglec-10 ligand and CD24 expression on (a–f) PDAC cell lines, (g–j) ovarian cancer cell lines, and (k–l) a breast cancer cell line. Siglec-10 ligands were detected using either (1) a recombinant Siglec-10-Fc chimeric protein and a secondary antibody (x-axis), with CD24 detected using a specific antibody (y-axis), or (2) a recombinant Siglec-10-Fc chimeric protein containing the inhibitory R119A mutation, which abrogates sialic acid binding (x-axis), with CD24 detected using a specific antibody (y-axis). Panels on the left (a, c, e, g, i, k) show representative flow plots. Panels on the right (b, d, f, h, j, l) show the percentage of positive cells and the median fluorescence intensity (MFI) for Siglec-10 ligands (using the non-mutated recombinant Siglec-10 protein), CD24, or cells expressing both. Means with SD are shown.

Figure 3.. ITGA3 and ITGB1 are sialylated glycoprotein ligands of Siglec-10 glycoproteins on PDAC cells.

(a) Experimental design for the pull-down of Siglec-10 ligands on PDAC cells. Recombinant Siglec-10 Fc (as well as a no-protein control or a Siglec-5 control) was allowed to bind its physiological ligands on the surface of PDAC cells, followed by an HRP-conjugated anti-Fc secondary antibody. In the presence of H₂O₂, HRP generated short-lived radicals that facilitated the transfer of biotin to proximal Siglec-10 ligands. Biotinylated Siglec-10 ligands were pulled down using streptavidin beads and identified by mass spectrometry. Created in BioRender. Saini, P. (2025) https://BioRender.com/lhnibjl. (b) A total of 4,044 proteins were identified, with enriched binding compared to a control using the anti-Fc antibody only without Siglec-10 protein. Of these, 110 proteins showed enrichment relative to a Siglec-5 control. Six proteins, CD47, CD59, CD73, ITGB6, ITGA3, and ITGB1, were significantly overexpressed in PAAD tissues compared to normal tissues in the TCGA dataset. (c) Response curves showing interactions between Siglec-10 and the six glycoproteins measured by surface plasmon resonance (SPR). Two concentrations (1000 nM, green; 100 nM, red) were tested for all glycoproteins, while ITGA3 was also tested at 300 nM (green) and 30 nM (red). (d) Binding of the SNA lectin (specific for sialic acid) to ITGA3 and ITGB1 recombinant glycoproteins was measured by a lectin array. Sialidase-treated glycoproteins (blue bars) showed significantly reduced binding compared to untreated glycoproteins (red bars). Unpaired t-tests. Means with SEM are shown. (e) SPR response curves comparing the binding of intact (sialylated) ITGA3 and desialylated ITGA3 to immobilized Siglec-10. (f) SPR response curves comparing the binding of intact (sialylated) ITGB1 and desialylated ITGB1 to immobilized Siglec-10.

Figure 4.. ITGA3 and ITGB1 are overexpressed in pancreatic cancer tissues and associated with poor survival.

(a–c) Comparison of CD24 (a), ITGA3 (b), and ITGB1 (c) expression between normal tissues and PAAD tissues in the TCGA dataset. Unpaired t tests. (d–f) Survival analyses of pancreatic tumor patients in the TCGA dataset showing the correlation between CD24 (d), ITGA3 (e), and ITGB1 (f) expression and overall survival. (g–i) Expression of CD24 (g), ITGA3 (h), and ITGB1 (i) in the PDAC TME across different disease states. Kruskal–Wallis test with Dunn’s multiple comparisons correction. Means with SEM are shown.

Figure 5.. ITGA3-expressing PDAC cells evade macrophage-mediated phagocytosis.

(a) ITGA3 knock-down in PANC-1 PDAC cells was achieved using CRISPR-Cas9. Flow cytometry confirmed the reduction of ITGA3 expression in knockout cells (ITGA3-KO) compared to non-targeting gRNA-treated cells and untreated controls. Data from three independent replicates show the percentage of ITGA3-expressing cells and the mean fluorescence intensity (MFI). Means and SEM are shown. Unpaired t-tests. (b) Experimental schematic illustrating co-culture of red-labeled PDAC cells (ITGA3-KO or non-targeting gRNA-treated) with monocyte-derived macrophages. Phagocytosis was monitored over time using live-cell imaging on the Incucyte Live-Cell Analysis System. (c) (left) Phagocytosis over time and area under the curve (AUC). Data show macrophage-mediated phagocytosis of PDAC cells, normalized to the 0-hour time point. Paired t-tests. N = 3 donors, with experiments performed in triplicate for each donor. (right) Representative images showing increased red fluorescence as an indicator of enhanced phagocytosis. (d) Schematic of ITGA3^low PDAC cell enrichment. ITGA3^high PDAC cells were separated using anti-ITGA3 antibody-coated columns, and the flow-through, containing ITGA3^low cells, was collected. (e) Flow cytometry analysis of ITGA3 expression in MIA PaCa-2 cells after enrichment for ITGA3^low cells and comparison with non-enriched controls (Ctrl cells). (f) Experimental schematic illustrating the phagocytic assay setup, where monocyte-derived macrophages were co-cultured with either ITGA3^high (control) or ITGA3^low PDAC cells. (g-h) Phagocytosis analysis of ITGA3^high (control) versus ITGA3^low MIA PaCa-2 (g) or PANC1 (h) cells by macrophages derived from different donors. N = 5–6, with experiments performed in triplicate for each donor. Top: Representative images showing increased red fluorescence as an indicator of enhanced phagocytosis. Bottom left: Live imaging data from individual donors (each symbol represents one donor’s data done in triplicate per donor). Bottom right: AUC data from multiple donors, showing significantly higher phagocytic activity with ITGA3^low cells. Paired t-tests. Some images in panels B, D, and F were created in BioRender. Saini, P. (2025) https://BioRender.com/lbo67eu.

Figure 6.. Development of Siglec-10 blocking antibodies to enhance macrophage-mediated phagocytosis of PDAC cells and prevent myeloid cell–mediated inhibition of T cell proliferation and activation in vitro.

(a) Schematic model illustrating Siglec-10-mediated suppression of macrophage phagocytosis. In the left panel, Siglec-10 on macrophages binds to glycan ligands on PDAC cells, including ITGA3, ITGB1, and CD24, triggering inhibitory signaling and suppressing phagocytosis. In the right panel, blocking Siglec-10 with an antibody prevents inhibitory signaling and enhances macrophage phagocytic capacity. (b) ELISA screening of recombinant antibodies from the top hybridoma clones for Siglec-10 binding. Binding to immobilized Siglec-10 (blue) and Siglec-5 (gray) proteins is shown. (c) Flow cytometric analysis of antibody selectivity, showing binding to CHO-K1 cells expressing Siglec-10 (blue) but not Siglec-5 (gray). (d) AUC analysis of in vitro phagocytosis assays screening various Siglec-10 antibody clones, along with commercially available anti-CD24 and anti-Siglec-10 antibodies, for their ability to enhance macrophage-mediated phagocytosis of AsPC-1 PDAC cells. Means with SEM are shown. (e) AUC analysis of the in vitro phagocytosis assay using the top-performing Siglec-10 antibody clone with macrophages differentiated from monocytes of four healthy donors. Statistical significance was determined using Friedman’s ANOVA test. Means with SEM are shown. (f) Time-course analysis of the in vitro phagocytosis assay comparing the top Siglec-10 blocking antibody clone (68A11A1, blue) with the isotype control (gray). Data represent n = 4 independent experiments. (g) ELISA-based binding analysis of the recombinant 68A11A1 antibody to immobilized recombinant Siglec-10 and Siglec-5 proteins across different dilutions. (h) Evaluation of the recombinant Siglec-10 antibody (clone 68A11A1) and anti-CD24 antibody in enhancing macrophage-mediated phagocytosis of multiple PDAC cell lines (AsPC-1, MIA PaCa-2, and PANC-1). Phagocytosis was normalized to the isotype control for each cell line and conducted using macrophages derived from monocytes of 5–8 healthy donors. Each symbol represents an individual donor; statistical significance was assessed using ratio paired t-tests compared to isotype control. Means with SEM are shown. (i) Triple co-culture assay involving cancer-associated fibroblasts (CAFs), PANC-1 PDAC cells, and monocyte-derived macrophages, showing phagocytosis kinetics, AUC quantification, and representative images. Statistical significance assessed using paired t-tests. (j-k) Flow cytometry analysis of CellTrace Violet (CTV)-labeled human CD8⁺ T cells co-cultured with human monocytes ± anti-Siglec-10 antibody in the presence of anti-CD3/CD28 beads for 5 days. (j) T cell proliferation; (k) granzyme B expression. ANOVA with post hoc comparisons.

Figure 7.. Blocking Siglec-10 interactions enhances macrophage-mediated phagocytosis and reduces PDAC tumor growth in vivo.

(a) Experimental design for subcutaneous tumor development and Siglec-10 antibody treatment. AsPC-1 PDAC cells were subcutaneously injected into immunodeficient NSG mice. After 7 days, when tumors reached ~40 mm³, mice were randomized into treatment groups. On the indicated days, ~3 × 10⁶ monocyte-derived macrophages were intravenously injected along with 200 μg of either isotype control or anti-Siglec-10 antibody. Tumor size was measured on specified days using a vernier caliper. (b) Tumor volume over time (left) and at the last time point (right) in mice treated with isotype control or anti-Siglec-10 antibody. n = 10 mice per group. Means with SEM are shown for the longitudinal analysis. For the last time point analysis, box and whiskers represent the minimum to maximum range, with all individual data points shown. Mann–Whitney U test. (c–e) Heatmaps showing differential gene expression in FACS-sorted human macrophages from tumors of NSG mice treated with anti-Siglec-10 antibody, highlighting: (c) immune activation pathways, (d) phagocytosis-related genes, and (e) metabolism-related genes. Red indicates higher expression; blue indicates lower expression. (f) Flow cytometry analysis of peritoneal cells from wild-type B6 mice and transgenic B-hSIGLEC10 mice. B6 mice express Siglec-G but not Siglec-10, whereas B-hSIGLEC10 mice express Siglec-10 but not Siglec-G on myeloid cells. Means with SEM are shown. (g) Experimental design for tumor challenge in B-hSIGLEC10 transgenic mice. KPC-derived murine PDAC cells (2838c3) were subcutaneously injected. After 5 days, mice were randomized and treated intravenously with either isotype control or anti–Siglec-10 antibody. Some images in panel g were created in BioRender. Saini, P. (2025) https://BioRender.com/ut55ra9. (h) Tumor volume in B-hSIGLEC10 mice treated with isotype control or anti–Siglec-10 antibody. Tumor volume over time (left) and at the last time point (right) in mice treated with isotype control or anti-Siglec-10 antibody. Means with SEM are shown for the longitudinal analysis. N=4–5 per group. For the last time point analysis, box and whiskers represent the minimum to maximum range, with all individual data points shown. Unpaired t test. (i) Mouse body weight during treatment.