Background: Myelodysplastic syndromes (MDS) are caused by a stem cell failure and often include a dysfunction of the immune system. However, the relationship between spatial immune cell distribution within the bone marrow (BM), in relation to genetic features and the course of disease has not been analyzed in detail. Methods: Histotopography of immune cell subpopulations and their spatial distribution to CD34⁺ hematopoietic cells was determined by multispectral imaging (MSI) in 147 BM biopsies (BMB) from patients with MDS, secondary acute myeloid leukemia (sAML), and controls. Results: In MDS and sAML samples, a high inter-tumoral immune cell heterogeneity in spatial proximity to CD34⁺ blasts was found that was independent of genetic alterations, but correlated to blast counts. In controls, no CD8⁺ and FOXP3⁺ T cells and only single MUM1p⁺ B/plasma cells were detected in an area of ≤10 μm to CD34⁺ HSPC. Conclusions: CD8⁺ and FOXP3⁺ T cells are regularly seen in the 10 μm area around CD34⁺ blasts in MDS/sAML regardless of the course of the disease but lack in the surrounding of CD34⁺ HSPC in control samples. In addition, the frequencies of immune cell subsets in MDS and sAML BMB differ when compared to control BMB providing novel insights in immune deregulation.

Myeloproliferative neoplasms (MPN) and myelodysplastic/myeloproliferative neoplasms (MDS/MPN) both harbor the potential to undergo myelodysplastic progression or acceleration and can transform into blast-phase MPN or MDS/MPN, a form of secondary acute myeloid leukemia (AML). Although the initiating transforming events are yet to be determined, current concepts suggest a stepwise acquisition of (additional) somatic mutations—apart from the initial driver mutations—that trigger disease evolution. In this study we molecularly analyzed paired bone marrow samples of MPN and MDS/MPN patients with known progression and compared them to a control cohort of patients with stable disease course. Cases with progression displayed from the very beginning a higher number of mutations compared to stable ones, of which mutations in five (ASXL1, DNMT3A, NRAS, SRSF2 and TP53) strongly correlated with progression and/or transformation, even if only one of these genes was mutated, and this particularly applied to MPN. TET2 mutations were found to have a higher allelic frequency than the putative driver mutation in three progressing cases (“TET2-first”), whereas two stable cases displayed a TET2-positive subclone (“TET2-second”), supporting the hypothesis that not only the sum of mutations but also their order of appearance matters in the course of disease. Our data emphasize the importance of genetic testing in MPN and MDS/MPN patients in terms of risk stratification and identification of imminent disease progression.

Purpose: Myelodysplastic syndromes (MDS) are caused by a stem cell failure, but the relationship between immune dysregulation and the course of disease has not yet been analyzed in detail. Experimental design: To get insights into the pathophysiologic and clinical relevance of the histotopography of immune cell subpopulations in this process, the immune cell infiltrate with focus on its spatial distribution was determined by multispectral imaging (MSI) in 147 bone marrow biopsies from MDS or secondary acute myeloid leukemia (sAML) patients and healthy controls (HC). In addition, the data were correlated to genetic alterations and clinical features of these patients including therapy response. Results: A high inter-tumoral heterogeneity in the frequency and spatial distribution of CD3⁺CD8⁺, CD3⁺CD8^-, CD3⁺FOXP3⁺ T cell subsets, MUM1p⁺CD3^- post-germinal B/plasma cells and CD34⁺ blasts was found in MDS and sAML samples. In HC only few B cells/plasma cells, but no T cell subpopulations were detected in the proximity to CD34⁺ blasts. In contrast, the frequency of these lymphocytes was increased in proximity to CD34⁺ blasts in both MDS and sAML independent of the karyotype, genetic alterations frequently detected in MDS, clinical risk stratification systems or treatment response to hypomethylating agents. Furthermore, an increased frequency of CD3⁺CD8⁺ T cells and MUM1p⁺ CD3^- B cells was found in responders to epigenetic drugs. Conclusions: Thus, we conclude that (i) T cell subsets do not belong to the normal stem cell niche, (ii) the presence of T and B cell subpopulations not directly affect the course of MDS, (iii) lymphocytes in the proximity to CD34⁺ blasts might indicate defective stem cell properties and (iv) the number of lymphocytes is a predictor of therapy response to hypomethylating agents.