Cross Talk between the Immune Compartment and the Tumor Cells in Myelodysplastic Syndromes (MDS)

Background: Abnormalities of the immune system and innate immune signaling have been described in patients (pts) with MDS, however the interaction between the MDS tumor cells and the immune compartment is poorly defined. Immune modulatory therapies are being explored in MDS and a better understanding of the interaction and potential cross-talk with immune components and MDS tumor cells is critical in developing new therapeutic strategies.

Methods: We developed a multiscale immune profiling strategy to map the immune environment of treatment naïve pts with lower- and higher-risk MDS, using mass cytometry by time of-flight (CyTOF) combined with Olink proteomics analyses. We obtained bone marrow (BM) and peripheral blood (PB) from 33 pts with MDS, (17 higher-risk; 16 lower-) and 10 healthy donors. Pts were representative of the MDS distribution across age, gender, mutational status and histologic subtype. PB and BM were freshly processed and stained with an antibody panel for myeloid cells. For the lymphoid panel, all the pts' samples were stained simultaneously to avoid batch effect. We processed the BM and the PB plasma for Olink.

Results: We profiled myeloid cells, NK cells and T cells in pts with MDS to determine differences in their constituents. Phenograph clustering across all pts revealed distinct T lymphoid metaclusters, that corresponded to known immune cell populations. We observed a robust immune response in the BM from MDS patients. T cells and myeloid cells were found in equal proportion in the BM from MDS and healthy patients. However, the B cell population were present at significantly lower frequency in the MDS BM (p<0.05), whereas gamma delta (gd)T cells and NK cells were at significantly higher frequency in the MDS pts (p <0.05). Paired mass cytometric analysis revealed a distinct composition and phenotype of T cell subsets. Among the T cell compartment, regulatory T cells (Treg) (p<0.001), cytolytic CD8 T cells (expressing perforin (p<0.01) or granzyme (p<0.001)) and exhausted CD8 T cells (expressing PD-1 (p= 0.09), LAG-3 (p<0.001) or TIM-3) seemed to accumulate in the MDS BM. Compared to healthy BM Treg in MDS BM expressed higher levels of TIGIT, TIM3, CD27, CD25, CD28 and OX40. Cytolytic NK cells accumulate in MDS BM (p<0.05). In addition to measuring adaptive lymphocyte responses to tumors, we also analyzed the distribution of innate lymphocytes and in particular NK cells in MDS BM. In the NK cell population, the cytotoxic NK cells (CD56dim CD16+) were found in a higher proportion (p<0.05) than the cytokinic counterpart (CD56bright) in the MDS BM. NK cells that accumulate in the MDS BM expressed higher level of CD45RA, CD38, granzyme B, CD56 suggesting their superior cytotoxic potential at a critical developmental stage of retaining enhanced sensitivity to pro-inflammatory cytokines derived from myeloid cells. We also analyzed a panel of 92 unique cytokines, chemokines and inflammatory markers produced in the tumor milieu using OLink Proteomics proximity extension assay. Many chemokines and cytokines were differently expressed in the MDS BM compared to the healthy BM, suggesting an important inflammatory reaction.

Conclusion: These data demonstrate new findings suggesting interactions and cross-talk between Tregs, NK cells, CD8 T cells and gdT cells in the BM of pts with untreated MDS. The data also suggest that there is a strong immune response in the MDS BM, particularly an intense cytotoxic response with the accumulation of cytotoxic CD8 T cells, NK cells and gdT cells. These results are are contrary to those observed in most solid tumors. This cytotoxic response seemed to be counteracted by the presence of Treg. Each of the subgroups of T cells (as well as NK cells) have been implicated as possibly promoting immune tolerance and tumor growth in different model systems and other tumor types. NK cells and gdT cells are known to respond to virus-infected and transformed cells rapidly and without the need for prior sensitization. However, anecdotal evidence from numerous independent studies may suggest that a strong but prolonged NK cell- and gdT cell-mediated antiviral response may result in dysregulation and promote tumorigenesis and/or aggravate pathology. Additional studies are underway to further characterize the immune landscape in those untreated and treated and to study the interaction of the immune components with tumor cells.