Markers of immunogenicity in pancreatic cell lines
Cytotoxicity by the selected chemotherapeutic agents against three pancreatic tumour cell lines PANC-1, Miapaca-2 and Bxcp-3, was assessed (Fig. 1; Table 1). Consistent with previous studies GEM was cytotoxic towards these tumour cell lines which in turn demonstrated different susceptibility to GEM-based killing (Fig. 1a; Table 1). Combinations of GEM with either ZA or OXP had an additive effect on tumour cell cytotoxicity, lowering GEM’s EC50 (Table 1). In contrast POM only demonstrated a discernible cytotoxic effect either alone or in combination with GEM against Miapaca-1 cells (Table 1) consistent with our previous work combining GEM with lenalidomide, an immune modulatory drug similar to POM .
The ability of GEM to increase HLA-class I expression on pancreatic tumour cell lines was assessed. GEM increased HLA expression after 48 h stimulation of each of the cell lines studied (Fig. 1b, c). Interestingly increases in HLA appeared to be independent of the ability of GEM to kill tumour cells (Fig. 1b) occurring at sub-optimal concentrations (10–100 nM) to induce cytotoxicity and at a time point before which cytotoxicity is observed. Incubation of GEM with each cell line increased the expression of HLA-class I, PDL-1 and CD47 on each tumour cell line (Fig. 1c). Concentrations of GEM (10 nM) unable to induce cytotoxicity were able to increase the expression of ULBP1, 2, 5 and 6 in addition to MIC A/B (Fig. 2a). In contrast GEM stimulation resulted in a decrease in ULBP3. The ability of GEM to alter the expression of immune checkpoints was also assessed, of five checkpoints measured GEM significantly increased the expression of PDL-1 and CD47 in each cell line studied but had no effect on the expression of CD39, Galectin 9 or HVEM (Figs. 1c, 2b). Neither OXP, ZA or POM could increase the expression of HLA-class I, ULBP’s, MIC A/B but didn’t significantly alter the ability of GEM to upregulate these receptors (Fig. 2). These data indicate that, of the agents studied, the ability to increase the expression of markers of tumour recognition was restricted to GEM.
Next, the ability of each chemotherapeutic agent to induce markers of ICD from pancreatic tumour cell lines was determined. GEM demonstrated an ability to induce the cell surface expression of Calreticulin on PANC-1 tumour cell lines (Fig. 3a). Similar effects were observed with Miapaca-2 and Bxcp-3 cells (data not shown). OXP, an established inducer of ICD was also capable of promoting CRT translocation whereas neither ZA or POM induced observable CRT translocation. Secretion of ATP and HMGB1 are also established markers of ICD . GEM, ZA and POM were unable to induce the expression ATP or HMGB1, in contrast to OXP (Fig. 3b, c). Combinations of chemotherapeutic agents had no additive effect on markers of ICD (data not shown).
Effect of chemotherapeutic agents on monocyte derived dendritic cells
PANC-1 cells were used to study the effect of GEM on functional responses from immune effector cells (Fig. 3e). First, CFSE-stained PANC-1 cells were stimulated with either single agent or GEM-based combinations for 24 and 48 h prior to co-culture with monocyte derived dendritic cells (MDDC) for 4 h (Fig. 3d, e). Stimulation of PANC-1 cells tended to increase their uptake into MDDC with statistically significant increases observed for GEM (Fig. 3d) and OXP but not ZA (Fig. 3f). Combinations of agents did not increase uptake of PANC-1 cells into MDDC (data not shown).
We sought to determine whether treatment of pancreatic tumour cells with chemotherapy alters the maturation of MDDC exposed to their supernatants and whether direct incubation of MDDC with chemotherapeutic agents effects their maturation. Incubation of MDDC with supernatant from PANC-1 cells significantly reduced the expression of HLA-class I, HLA-class II, PDL-1 and CD40 compared to untreated MDDC (Fig. 4a, c, d, f). CCR7 and CD86 expression was not significantly altered. The supernatant of PANC-1 cells treated with single agents had no effect on any marker with the exception of supernatant from POM treated PANC-1 cells which significantly increased the expression of CD40 compared to that of supernatant from untreated PANC-1 cells.
Treatment of PANC-1 cells with combinations of chemotherapeutic agents significantly increased the expression of MDDC markers of maturation. HLA-class I, class II, CD86 and CCR7, but not CD40, expression could be significantly increased by incubation of MDDC with supernatants from PANC-1 cells treated with GEM-based combinations containing ZA, POM and/or OXP (Fig. 4a, b, d–f) compared to supernatant from untreated PANC-1 cells. Next, MDDC were directly stimulated with chemotherapeutic agents and markers of maturation were assessed. Each agent, alone or in combination, significantly increased the expression of HLA-class I compared to the untreated control. In addition, combinations of GEM including ZA or POM significantly increased the expression of CD86 and CD40 respectively (data not shown). However, CCR7, HLA-class II, and PDL-1 were unchanged.
Stimulation of MDDC with 100 ng/ml of the TLR4 agonist LPS significantly increased the expression of each marker with the exception of CCR7 when compared to untreated or PANC-1 supernatant incubated MDDC. To assess the effect of TLR ligation in combination with chemotherapeutic agents Poly IC, a TLR3 agonist, was combined with single agent GEM, which had demonstrated no effect on DC maturation or POM which had demonstrated limited effects on DC maturation (Fig. 4e). Poly IC plus either GEM or POM resulted in increases in markers of MDDC maturation compared to the no stimulation control (Fig. 4g).
Effect of chemotherapeutic agents on T-cell responses
Next, the ability of treated MDDC to stimulate antigen specific CD8+ T-cell responses was assessed (Fig. 5). MDDC incubated with supernatant from treated PANC-1 tumour cells or directly stimulated with chemotherapeutic agents for 24 h were co-cultured with a peptide pool containing immunodominant epitopes from cytomegalovirus, Epstein Barr virus and influenza virus (CEF) and co-cultured with PBMC for a further 24 h. Intracellular cytokine staining was used to assess the antigen specific expression of IFN-γ from CD8+ T-cells (Fig. 5). Consistent with the ability of combination agents to induce MDDC maturation, combinations including GEM and POM with either ZA or OXP significantly increased the expression of IFN-γ from CD8+ T-cells. Single agent POM was able to stimulate increased CEF dependent expression of IFN-γ from CD8+ T-cells whereas single agent GEM reduced IFN-γ expression (Fig. 5b, c). To assess the direct effect of POM on T-cell activation, isolated T-cells were incubated with POM (1–100 nM) and CD69 expression was assessed (Fig. 5g). POM significantly increased the expression of CD69 on CD8+ T-cells (Fig. 6f). Combining POM with GEM showed that whilst GEM had no significant effect on CD69 expression it could partially block the effect of POM. Taken together these data suggest an inhibitory effect of GEM on T-cell activation.
CD8+, CD101+ T-cells expressing CD38 and PD-1 have been associated with poor prognosis in pancreatic cancer . CD101 is a marker of T-cell exhaustion  and CD38 and PD-1 co-expressing T-cells are thought to be dysfunctional . We sought to determine whether GEM pre-stimulation followed by T-cell activation could cause increases in CD38 and PD-1 expression on T-cells. Activation of T-cells with anti-CD3 plus anti-CD28 antibodies increased the expression of CD38 and PD-1; however, preincubation with gemcitabine resulted in no additional expression either CD4+ or CD8+ T-cell subsets. POM significantly increased the expression of IFN-γ from effector and central memory CD8+ T-cell populations, consistent with its ability to increase IFN-γ expression from CEF stimulated T-cells (Fig. 6) but did not significantly increase the expression of CD38 and PD-1 (data not shown). The frequency of CD101 on T-cell subsets was not changed by preincubation with POM or GEM and the expression of CD38 and PD-1 on CD101+ cells was also unchanged.
Finally, checkpoint blockade with anti-PD-1 antibodies has been implicated in the onset of dysfunction and apoptosis of T-cells associated with poor prognosis in PDAC . Consistent with these findings we found that preincubation of anti-PD-1 antibody (10 μg/ml) followed by T-cell activation resulted in increases in Annexin V staining and reductions in IFN-γ expression in CD8+ T-cells (Fig. 6). The ability of GEM or POM to co-stimulate T-cells in the presence of anti-PD-1 antibody was measured. Incubation of T-cells with GEM + anti-PD-1 antibody prior to T-cell activation resulted in reductions in both IFN-γ and Annexin V staining (Fig. 6b, e). In contrast incubation of T-cells with POM + anti-PD-1 antibody reduced the expression of Annexin V staining compared to anti-PD-1 incubation alone whilst increasing the expression of IFN-γ in CD8+ T-cells (Fig. 6c, f) but not CD4+ T-cells. Incubation of Pom with GEM + anti-PD-1 antibody was able to recapitulate IFN-γ expression in CD8+ T-cells but was unable to reduce the expression of Annexin V (Fig. 6d, g).