Lymphocyte Co-Expression of CD38/HLA-DR Is a Marker for Anti-Tumor Activity in Hodgkin Lymphoma in Vitro: Evidence for a PD1-Independent Mechanism

Introduction

The immunosuppressive tumor microenvironment has become of increasing interest in Hodgkin lymphoma (HL), in particular due to the recent success of checkpoint inhibitors (CI) as part of therapeutic strategies. The mechanism of these agents action in HL, however, remains elusive. Some studies have shown that cytotoxic T-lymphocytes may not be responsible for clinical efficacy, and that tumor-associated macrophages may also be targeted by these agents. We recently described a positive association between increasing proportion of CD38/HLA-DR co-positive lymphocyte in affected nodal tissue and clinical outcome in children with HL. We developed an in vitro model to further evaluate our findings, in this study.

Methods

Peripheral blood mononuclear cells collected from healthy volunteers were used to generate effector cytoxic T lymphocytes (CTL), natural killer (NK) and CD4 positive T (CD4+T) cells by incubating with IL-15 and IL-2. Both a short-term (4-day) incubation and a longer incubation was used to generate lymphocytes with low-level and higher CD38/HLA-DR co-positive cells, respectively. CD3/CD8 co-positive CTL, NK (CD56-positive/CD3-negative) and CD4-positive T cells were isolated using MACS system. In addition to CD38/HLA-DR expression, isolated cells were evaluated for expression of CD279 (PD-1) and CD274 (PDL-1) by flow cytometry. Two HL cell lines, HDLM-2 (nodular sclerosis HL) and KMH-2 (mixed cellular HL) were used as targets in effector cell-mediated cytotoxicity experiments. Cells were incubated at various effector:target ratios, and HL cell death was measured with a flow cytometric cell-mediated cytotoxicity assay. The resuts were corrected for alloreactive cell elimination. Blocking antibodies against PD-1 and PDL-1 were used for cytoxicity experiments, using CTL and CD4-positive T cells as effector cells, as well.

Results

Higher CD38/HLA-DR co-expression was seen in CTL after longer incubation (day 11) with IL-2 and IL-15, while peak expression was reached earlier in NK cells (day 4). Both CTL and NK cells demonstrate cytotoxicity against HDLM-2 and KMH-2 cell lines. Cytotoxicity was increased, as evidenced by lower (lethal unit 20%) LU20 effector:target ratio levels, at day 11 of incubation (compared to day 5) for CTL for both cell lines: 0.5 (day 5) vs 0.29 (day 11) for KMH-2 (p=0.02) and 1.1 (day 5) vs. 0.4 (day 11) for HDLM-2 (p=0.15) cells. There was no difference between cytotoxicity with CTL compared to NK cells for either cell line at day 11.

CTL, NK cells, and CD4-positive T cells all expressed both PD-1 and PDL-1, with no difference between cell types in percent positivity or mean channel intensity after cytokine incubation. PD-1 expression increased with incubation time in CTL, peaking at 50.5% on day 10, as opposed to NK cells, where it peaked at day 5-7. The co-expression of CD38/HLA-DR was higher in CTL compared to CD4-positive T cells (79% vs. 37% at day 7). PD-1 blockade did not inhibit CTL or CD4-positive T cell-mediated cytotoxicity in either cell line. This was also the case after PDL-1 blockade on tumor cells, indicating PD-1/PDL-1 pathway-independent HL cell elimination by CD38/HLA-DR co-positive CTL and CD4-positive T cells.

Discussion

Our results indicate that higher CD38/HLA-DR co-expression in CTL was associated with superior elimination of HL cells in vitro supporting our recent in vivo findings. Induced co-expression of CD38/HLA-DR was higher in CTL compared with NK cells and reached a peak level earlier in NK cells. Increasing expression of PD-1 and PDL-1 was observed for all three effectors cells with longer incubation time. Interestingly, there was no change in CTL or CD4-positive T cell-mediated cytoxicity of HL cells following PD-1 and PDL-1 blockade in vitro. In conclusion, both CTL and NK cells are effective against HL cells. The anti-tumor activity of CTL correlated with increasing levels of CD38/HLA-DR expression in our experimental model. Cytotoxicity was enhanced despite increased expression of PD-1, and, therefore, appears to be independent of the PD-1/PDL-1 pathway, suggesting involvement of other operational mechanisms. This model could be useful in further elucidating the interactions between immune effectors and HL cells.