Gut Microbiome, Metabolome and Immune Changes in Myelodysplastic Syndromes

Introduction. Myelodysplastic syndromes (MDS) are a group of multi-factorial disorders, in which the clonal expansion of somatically mutated hematopoietic stem and progenitor cells plays a crucial role. The influence of the bone marrow microenvironment, including inflammatory and metabolic dysbalances, on disease propagation becomes increasingly recognized. Although changes in gut microbiome have been described in various cancer types, its role in MDS has not been sufficiently elucidated yet. Due to its interconnection with carcinogenesis in solid tumours, systemic inflammation and metabolism, the gut microbiome could possibly play a role in MDS development as well. The goal of our study is to describe the composition of the gut microbiome in MDS patients (pts) and its association with inflammatory-metabolic parameters.

Methods. We analysed faecal samples from 16 pts with untreated MDS or chronic myelomonocytic leukemia (median age: 66,5 years [41-81], m/f=13/3) and healthy donors (n=8, median age: 60,5 years [54-70], m/f=3/5). Exclusion criteria were antibiotic treatment within the last 4 weeks, chronic inflammatory bowel disease and other malignant disorders. Gut microbial taxonomic profiles were generated from the frozen faecal samples using 16S rRNA gene amplicon sequencing. For bioinformatic analysis, MapSeq was used. Statistical analysis included diversity analysis (observed microbial richness and Shannon index, differences assessed by Wilcoxon tests); differences in gut microbiome composition were visualized using Principal Coordinate Analysis (PCoA) in which differences were assessed by PERMANOVA.

We performed flow cytometry to assess the cellular immune status including B-, T- and NK-cell populations, T/NK-subpopulations and expression of checkpoint molecules. Each parameter was analysed using Mann-Whitney U-test. Metabolite concentrations in plasma and urine were measured using NMR spectroscopy. Multivariate analysis of metabolic parameters was performed as a principal component analysis (PCA) and partial least-square discriminant analysis (PLS-DA), to evaluate the possible discrimination of the two groups.

Results. Ten MDS pts were referred to the low-risk group (low and intermediate 1 according to IPSS), 5 pts - high-risk (intermediate 2 and high), in one patient information was not available. All but one patient carried at least one MDS-typical somatic mutation (median: 3 [1-6]). According to cytogenetic analysis 2 pts belong to "very good" prognostic risk, 7 - "good", 4 - "intermediate", 2 - "very bad", in 1 patient the information was not available. The median time from the diagnosis to faecal sampling was 19 months [1-205].

Microbiome composition showed statistically significant differences between MDS and healthy subjects (p<0.05, R²=0.079, PERMANOVA). Gut microbial diversity and richness (the number of genera observed per patient), was lower in MDS, although not reaching statistical significance in this small cohort. Time from MDS diagnosis until faecal sampling did not have an impact on the microbial diversity. Pts with high-risk MDS tended to have the lowest gut microbial diversity and richness compared to pts in the low-risk group, both according to IPSS and IPSS-R scores.

Flow cytometry revealed that pts with MDS had reduced numbers of B-lymphocytes (CD19+: 56 vs 184/µL in MDS and healthy, accordingly, p<0.01), increased number of T effector memory cells (0.06 vs 0/µL, p<0.05) and PD-L2 expression on T-lymphocytes (MFI: 150 vs 86, p<0.01), reflecting prolonged inflammatory stimulation. In addition, we observed reduced NK-cell activity, demonstrated by the reduction of the activating receptors on NK cells (NKp30: 92.9 vs 99.05 %CD56dim, p<0.05; DNAM-1: 86.8 vs 97 %CD56bright, p<0.01).

PCA applied to plasma and urine metabolites allowed us to separate MDS from healthy counterparts. Further separation was accomplished using a supervised approach with PLS-DA, which highlighted the separation of the groups due to high VIP scores of the parameters, mostly affecting lipid metabolism.

Conclusions. We have demonstrated broad, statistically significant changes in microbiome, metabolome and immunome of pts with MDS. Whether these are cause or consequence of the disease development remains to be elucidated.

No relevant conflicts of interest to declare.