Integrated single-cell transcriptomics identifies γδ T-cell heterogeneity and a candidate HLA-E–NKG2A regulatory axis in pancreatic ductal adenocarcinoma

Background: Pancreatic ductal adenocarcinoma (PDAC) has a profoundly immunosuppressive tumour microenvironment (TME) that limits anti-tumour immunity and contributes to resistance to immunotherapy. Although γδ T-cells can integrate innate and adaptive immune signals, their abundance, transcriptional states and regulatory pathways in PDAC remain incompletely defined.

Methods: We performed integrated single-cell RNA sequencing analysis of 39 pancreatic tissue samples, comprising 33 PDAC tumours and 6 adjacent normal tissues. After dataset integration, immune cell annotation, and stringent per-cell gating, γδ T-cells were quantified and profiled for checkpoint-, ligand-, and chemokine-related programmes.

Results: γδ T-cells were detectable across PDAC samples but showed substantial inter-sample heterogeneity in abundance. Among candidate inhibitory pathways, PDCD1, CD274, and HAVCR2 expression in γδ T-cells did not differ significantly between tumour and adjacent tissues, whereas KLRC1 (encoding NKG2A) showed a tumour-associated difference at the single-cell level, with a consistent directional pattern in sample-level summaries. NKG2A expression was comparable between γδ T-cells and NK cells, suggesting a shared inhibitory programme. HLA-E, the ligand for NKG2A, showed higher epithelial-cell expression in tumour than adjacent tissue in sample-level summaries (median 1.06 vs. 0.57; BH-q = 0.035). Chemokine analysis identified enrichment of CCL2, CCL4, CCL5, CXCL8 and CXCL12, with limited CXCL9/10/11 signalling within the PDAC TME. Within γδ T-cells, CXCR4 was the trafficking receptor, followed by CCR6, CCR7 and CXCR6.

Conclusions: PDAC-infiltrating γδ T-cells show marked inter-sample heterogeneity and variable inhibitory and trafficking-related programmes. Integrated transcriptomic analysis nominates HLA-E–NKG2A as a candidate regulatory axis, with NK cells included as a biologically relevant comparator. Chemokine receptor patterns, particularly CXCR4, CCR6, CCR7 and CXCR6, suggest candidate trafficking features. These findings are hypothesis-generating and require spatial, protein-level and functional validation.