Abstract
Introduction
Triple-negative breast cancer (TNBC) is characterized by an immunosuppressive, myeloid-rich tumor microenvironment. Sialic acid-binding immunoglobulin-like lectins (Siglecs), specifically the inhibitory receptors Siglec-7, -9, -10, and -15 on myeloid cells, promote immune evasion by binding tumor-surface sialoglycans. Although cell-surface sialoglycoproteins like CD24, MUC1, and GPNMB have been identified for several of these Siglecs, the dominant glyco-immune checkpoints driving local immunosuppression remains largely unexplored. CDCP1 is an oncogenic transmembrane glycoprotein overexpressed in TNBC and associated with TNBC migration, metastasis, and adverse patient outcomes. This study aims to evaluate Siglec-binding heterogeneity across TNBC lines and the functional role of CDCP1 as a Siglec ligand.
Methods
Because Siglec recognition is influenced by sialic acid linkage and glycan context, CDCP1 glycosylation was characterized by PNGase F and neuraminidase digestion followed by immunoblotting. Sialylation was further assessed by metabolic labelling with Ac4ManNAz followed by copper-free click chemistry and streptavidin enrichment. To investigate linkage-associated sialylation, SNA (α2,6) and MAL-II (α2,3) lectin affinity pulldown was performed, with enriched CDCP1 proteoforms identified by immunoblotting. To determine whether Siglec-binding profiles differ between TNBC cell lines, baseline Siglec-7, -9, -10 and -15 binding was assessed by flow cytometry using Siglec-Fc chimeras in malignant (BT-549, HCC1143, MDA-MB-231) and non-malignant (MCF10A) cell lines. CRISPR-Cas9 ribonucleoproteins were utilized to knock out CDCP1 in MDA-MB-231 cells. Clones were screened by Western blotting for total CDCP1 expression and flow cytometry for surface CDCP1 expression relative to a CRISPR-SHAM control. Verified KO clones with near-background surface expression were evaluated for changes in Siglec-Fc binding relative to CRISPR-SHAM controls. Direct physical engagement between CDCP1 and Siglecs was determined via Siglec pulldown followed by Western blotting for CDCP1 detection.
Results
Enzymatic deglycosylation revealed extensive N-glycosylation and terminal sialylation of CDCP1, while metabolic labelling and copper-free click chemistry independently demonstrated the incorporation of sialic acid into CDCP1 proteoforms. Notably, both full-length (~135 kDa) and cleaved (~70 kDa) CDCP1 were recovered in SNA- and MAL-II-enriched fractions, supporting the presence of α2,6- and α2,3-linked sialylated glycans on both proteoforms. The relative enrichment of these proteoforms varied across cell lines, revealing heterogeneity in CDCP1 sialylation that provides a biochemical foundation for its subsequent investigation as a candidate Siglec ligand. Siglec-Fc binding profiles were highly heterogeneous. MDA-MB-231 exhibited significantly higher baseline Siglec-9 and Siglec-15 binding compared to other lines. Western blot screening of CDCP1-disrupted MDA-MB-231 clones identified five clones with altered total CDCP1 expression. Subsequent flow cytometry confirmed near-background cell-surface CDCP1 levels in three of these clones (E4, B3, and D6). Preliminary Siglec-binding analysis showed CRISPR-mediated CDCP1 deletion significantly reduced Siglec-9 binding in two KO clones compared to SHAM controls, while one clone exhibited variable binding. Siglec-7 binding remained unaltered. Paradoxically, Siglec-15 binding was consistently increased across all three CDCP1-KO clones relative to SHAM controls.
Conclusion
These results demonstrate Siglec-binding profiles are heterogeneous in TNBC and identified CDCP1 as a potentially functional ligand for the inhibitory receptor Siglec-9. Furthermore, the unexpected upregulation of Siglec-15 binding upon CDCP1 loss may suggest a compensatory glycan-remodeling or receptor-compensatory mechanism. These findings highlight the complex, dynamic nature of the tumor glycoproteome and glyco-immune