Page 123 - Haematologica April 2020

TM+ monocytes in myelodysplastic syndromes
Thrombomodulin-expressing monocytes are clonally involved and are associated with the features of low- risk myelodysplastic syndromes
In order to investigate the clonal involvement of TM+ monocytes in MDS, cells from three different patients were sorted and screened for the presence of a known cytogenet- ic aberration according to their karyotype. Almost all monocytes of these patients showed high expression of TM.
One patient had a deletion of chromosome 5q (del5q) in all cells (karyotype: 46,XY,del(5)(q22q33)[10]), one showed trisomy for chromosome 8 in 65% of the cells (karyotype: 47,XY,+8[13]/46,XY[7]) and one case had a monosomy 7 (karyotype: 45,XY,-7[10]). In all cases, TM+ monocytes were highly involved in the dysplastic clone and showed a high percentage of cells with the respective cytogenetic abnor- mality, suggesting that they did not come from healthy CD34+ cells. The isolated CD34+ progenitor cells showed a similar pattern, since the known cytogenetic aberration was present in most analyzed cells. As expected, B cells were not involved. Partial involvement was found in whole BM samples (Figure 2A).
To further correlate TM expression with different MDS subgroups, the patients’ clinical data were collected and their IPSS-R scores, reflecting survival and the risk of dis- ease progression, were calculated. Furthermore, patients were categorized using the 2016 WHO classification sys- tem which incorporates clinical characteristics, PB and BM findings and cytogenetic analysis. The percentage of mono- cytes displaying TM expression was higher for patients who had a very low- or low-risk score in the IPSS-R than for patients in higher-risk groups and healthy controls (very low/low 40.1% vs. intermediate 22.7% vs. high/very high 28.3% vs. NBM 11.3%) (Figure 2B). Additionally, TM expression was elevated in WHO categories related to lower-risk disease, such as MDS single and multiple lineage dysplasia with or without ring sideroblasts, compared to categories related to higher-risk MDS, i.e. MDS with excess blasts 1 and 2. The percentage of monocytes expressing TM was higher in all WHO subgroups than in NBM (single/multiple lineage dysplasia with/without ring sider- oblasts, 40.9% vs. MDS with excess blasts 1 and 2, 24.2% vs. NBM, 11.3%) (Figure 2B). Using the percentage of BM blast cells as a reflection of disease stage, patients with blast percentages below 5% harbored higher numbers of TM+ monocytes than patients with 5% or more blast cells (41.3% vs. 25.7%, respectively; P<0.0001) (Figure 2C). Finally, a relation between the percentage of monocytes expressing TM and the presence of ring sideroblasts (ery- throblasts with mitochondrial iron accumulation) was found (present 45.4% vs. absent 33.2%; P=0.003) (Figure 2C). As an “indirect” indication of the presence of a SF3B1 mutation, the percentage of TM+ monocytes was compared between subtypes with ring sideroblasts and other MDS subtypes (Online Supplementary Figure S4). Significantly higher percentages of TM+ monocytes were found in cases with ring sideroblasts than in those with MDS with excess blasts 1 or 2. A trend to higher frequencies was observed for the comparisons with other subtypes.
Myelodysplastic syndrome-derived thrombomodulin- positive monocytes polarize CD4+ T cells
to an immunosuppressive phenotype
The next research focus was to study the effect of TM+ monocytes on the phenotype of CD4+ T cells and to
determine whether they could induce an anti-inflamma- tory T-cell phenotype. Healthy donor-derived CD4+ T cells were co-cultured with sorted TM- or TM+ mono- cytes from two MDS patients (Online Supplementary Figure S5). After 5 days, T cells were harvested and labeled with a comprehensive panel of metal-tagged antibodies (Online Supplementary Table S1) for mass cytometry (CyTOF). Using our data analysis pipeline, T- cell subsets were identified (Online Supplementary Figure S6) and compared between the conditions (i.e. day 0, control stimulated and cultures with TM- or TM+ mono- cytes) (Figure 3A). Compared to day 0, all conditions showed expansion of specific T-cell subsets. To further characterize these cell islands, T cells were clustered by SPADE on tSNE (see Methods). Using selected T-cell markers various subsets could be characterized (Figure 3B). Then frequencies of clusters between the conditions were compared (Figure 3C). Nodes highlighted with red circles refer to the most frequent T cell clusters in the TM- condition whereas black circles indicate a higher percentage in the TM+ condition. For both conditions the top five of highest frequencies were selected. The expression profiles of T cells in the identified clusters were evaluated next, using MEM (see Methods). T-cell clusters were divided into two groups for comparison: clusters with highest frequency in the TM+ condition (called “up”) and clusters with highest frequency in the TM- condition (called “down”). MEM scores were calcu- lated for each marker in each group for further compari- son.44 Interestingly, T cells that were predominantly pres- ent upon culturing with TM+ monocytes (group called “up”) showed an anti-inflammatory profile (Figure 3C). They were polarized toward Th2, Treg and PD-1- expressing clusters of T cells, since they expressed high levels of FoxP3, GATA3 and CD279 (PD-1) and had ele- vated concentrations of intracellularly measured inter- leukin-4 and interleukin-10. In contrast, T cells cultured in the presence of TM- monocytes (group called “down”) were mainly positive for interferon-γ and hardly for immunosuppressivecytokines.
The presence of thrombomodulin-positive monocytes is related to a better overall and leukemia-free survival
In order to determine the clinical significance of the presence of TM+ monocytes, overall survival and leukemia-free survival were calculated. As a cut-off for TM expression monocytes from the healthy donor cohort were used. The mean TM percentage plus two standard deviations was calculated resulting in a cut-off of 25.53%. In total, overall survival data for 122 MDS patients and leukemia-free survival data for 102 patients were avail- able. Interestingly, the presence of TM on MDS mono- cytes was significantly associated with a better overall sur- vival (P=0.006) as well as a better leukemia-free survival (P=0.029) (Figure 4A). The median overall survival for patients with TM+ monocytes was 58 months, whereas that of patients without TM+ monocytes was 30 months. For a subgroup of patients, data were available for further risk stratification into IPSS and IPSS-R risk groups. Following the hypothesis that TM is mainly present in an inflammatory environment, survival and leukemia-free survival data were also analyzed in low-risk MDS groups (i.e. IPSS low and intermediate groups or IPSS-R very low, low and intermediate groups). The presence of TM+ monocytes resulted in better overall survival in this sub-
haematologica | 2020; 105(4)
965