Intra-Tumoral CD8+ T-Cell Infiltration and PD-L1 Positivity in Homologous Recombination Deficient Pancreatic Ductal Adenocarcinoma

Abstract

The immune contexture of pancreatic ductal adenocarcinoma (PDAC) is generally immunosuppressive. A role for immune checkpoint inhibitors (ICIs) in PDAC has only been demonstrated for the rare and hypermutated mismatch repair (MMR) deficient (MMR-d) subtype. Homologous recombination repair (HR) deficient (HR-d) PDAC is more prevalent and may encompass up to 20% of PDAC. Its genomic instability may promote a T-cell mediated anti-tumor response with therapeutic sensitivity to ICIs. To investigate the immunogenicity of HR-d PDAC, we used multiplex immunohistochemistry (IHC) to compare the density and spatial distribution of CD8+ cytotoxic T-cells, FOXP3+ regulatory T-cells (Tregs), and CD68+ tumor-associated macrophages (TAMs) in HR-d versus HR/MMR-intact PDAC. We also evaluated the IHC positivity of programmed death-ligand 1 (PD-L1) across the subgroups. 192 tumors were evaluated and classified as HR/MMR-intact (n=166), HR-d (n=25) or MMR-d (n=1) based on germline testing and tumor molecular hallmarks. Intra-tumoral CD8+ T-cell infiltration was higher in HR-d versus HR/MMR-intact PDAC (p<0.0001), while CD8+ T-cell densities in the peri-tumoral and stromal regions were similar in both groups. HR-d PDAC also displayed increased intra-tumoral FOXP3+ Tregs (p=0.049) and had a higher CD8+:FOXP3+ ratio (p=0.023). CD68+ TAM expression was similar in HR-d and HR/MMR-intact PDAC. Finally, 6 of the 25 HR-d cases showed a PD-L1 Combined Positive Score of >=1, whereas none of the HR/MMR-intact cases met this threshold (p<0.00001). These results provide immunohistochemical evidence for intra-tumoral CD8+ T-cell enrichment and PD-L1 positivity in HR-d PDAC, suggesting that HR-d PDAC may be amenable to ICI treatment strategies.

Keywords: PD-L1; T-cell inflammation; immunotherapy; pancreatic cancer; tumor microenvironment.

Figure 1

Distribution of CD8+ T-cells in PDAC. (A) Definitions of intra- tumoral, peri- tumoral and stromal regions. (B) Representative H&E images for HR/MMR-intact, HR-d and MMR-d PDAC with corresponding immunostaining for CD8 (brown) and Pan-cytokeratin (PanCK, teal). Black arrows show examples of CD8+ staining. (C) Comparison of CD8+ T-cell densities in HR/MMR-intact versus HR-d PDAC across the three tumor regions. The MMR-d case is shown as a reference for an immunogenic PDAC. NS, not significant.

Figure 2

FOXP3+ Treg and CD68+ TAM infiltration in PDAC. (A) Representative FOXP3 (brown), CD68 (purple) and PanCK (teal) immunostaining for HR/MMR-intact, HR-d and MMR-d PDAC. Black and red arrows show examples of FOXP3+ and CD68+ staining, respectively. (B, C) Comparison of FOXP3+ Treg (B) and CD68+ TAM (C) densities in HR/MMR-intact versus HR-d PDAC across the intra-tumoral, peri- tumoral and stromal regions. (D) Comparison of overall CD8+:FOXP3+ ratios between HR/MMR-intact versus HR-d PDAC. The MMR-d case is shown as a reference. NS, not significant.

Figure 3

PD-L1 positivity in PDAC. (A) Representative PD-L1 immunostaining for HR/MMR- intact, HR-d and MMR-d PDAC. The top row shows tumors stained with PD-L1 (brown), while the bottom row shows the same tumor sections stained with PanCK (teal) following PD-L1 staining (brown). Red arrows in top panel show examples of PD-L1 staining. (B) Comparison of the proportion of cases in the HR/MMR-intact versus HR-d groups meeting the Combined Positivity Score (CPS) threshold of ≥ 1. Six of 25 HR-d cases had a CPS of ≥1, whereas none of the 163 evaluable HR/MMR-intact cases met the PD-L1 positivity threshold of ≥1. The MMR-d case scored >1.