A Specific Host/Microbial Signature of Plasma-Derived Extracellular Vesicles Is Associated to Thrombosis and Marrow Fibrosis in Polycythemia Vera

Abstract

Polycythemia vera is a myeloproliferative neoplasm with increased risk of thrombosis and progression to myelofibrosis. However, no disease-specific risk factors have been identified so far. Circulating extracellular vesicles (EVs) are mostly of megakaryocyte (MK-EVs) and platelet (PLT-EVs) origin and, along with phosphatidylethanolamine (PE)-EVs, play a role in cancer and thrombosis. Interestingly, circulating microbial components/microbes have been recently indicated as potential modifiers of inflammation and coagulation. Here, we investigated phenotype and microbial DNA cargo of EVs after isolation from the plasma of 38 patients with polycythemia vera. Increased proportion of MK-EVs and reduced proportion of PLT-EVs identify patients with thrombosis history. Interestingly, EVs from patients with thrombosis history were depleted in Staphylococcus DNA but enriched in DNA from Actinobacteria members as well as Anaerococcus. In addition, patients with thrombosis history had also lower levels of lipopolysaccharide-associated EVs. In regard to fibrosis, along with increased proportion of PE-EVs, the EVs of patients with marrow fibrosis were enriched in DNA from Collinsella and Flavobacterium. Here, we identified a polycythemia-vera-specific host/microbial EV-based signature associated to thrombosis history and marrow fibrosis. These data may contribute to refining PV prognosis and to identifying novel druggable targets.

Keywords: biomarker; extracellular vesicles; marrow fibrosis; microbial DNA cargo; polycythemia vera; thrombosis.

Figure 1

Proportion of circulating MK-, PLT-, and TF-EVs of PV patients. (A,B) Frequency of MK- and PLT-EVs in PV patients with (n = 14) or without thrombosis history (n = 24). (C) Frequency of TF-EVs of PV patients with (n = 24) or without (n = 14) hypertension. (A–C) Data are expressed as percentage of MK-, PLT-, and TF-EVs and presented as min to max with median (Mann–Whitney test; * p < 0.05). (D,E) TF-EV percentages of PV patients positively correlate with age and JAK2V617F-variant allele frequency (Spearman’s correlation test).

Figure 2

Proportion of circulating PE-EVs of PV patients. (A) Frequency of PE-EVs in PV patients with low (LR-PV; n = 11) vs. high (HR-PV; n = 27) risk. (B) Proportion of PE-EVs of PV patients with (n = 24) or without (n = 14) hypertension. (C) PE-EVs percentages of PV patients positively correlate with age (Spearman’s correlation test). (D) Frequency of PE-EVs in PV patients with (n = 10) or without (n = 28) marrow fibrosis. (A,B,D) Data are expressed as percentage of PE-EVs and presented as min to max with median (Mann–Whitney test; * p < 0.05).

Figure 3

Proportion of LPS-associated EVs after isolation from the plasma of PV patients. (A) Representative dot-plots of LPS-associated EVs after isolation from the platelet-poor plasma of 1 PV patients with thrombosis history and 1 PV patient without thrombosis history. (B) Proportion of isolated LPS-associated EVs of PV patients with (n = 14) or without thrombosis (n = 24). Data are expressed as percentages and presented as min to max with median (Mann–Whitney test; * p < 0.05).

Figure 4

Microbial DNA signatures of thrombosis history and marrow fibrosis in EVs isolated from the plasma of PV patients. (A) Alpha diversity, estimated according to the inverse Simpson and Shannon indices, of the microbial DNA cargo of EVs isolated from PV patients stratified by the occurrence of thrombosis or marrow fibrosis. No significant differences were found (p ≥ 0.06, Wilcoxon rank-sum test). (B) Principal coordinates analysis (PCoA) based on Jaccard similarity between the genus-level profiles of EVs isolated from the plasma of PV patients, stratified as above. No significant segregation was found between study groups (p ≥ 0.3, PERMANOVA). Boxplots showing the relative abundance distribution of families (C) and genera (D) that were significantly differentially represented between PV patients with and without history of thrombosis or marrow fibrosis (* p < 0.05; ** p < 0.01; Wilcoxon rank-sum test).

Figure 5

The gut microbiome of PV patients. (A) Alpha diversity estimated according to inverse Simpson (top) and Shannon (bottom) indices of PV patients with or without thrombosis history and marrow fibrosis. (B) Principal coordinates analysis (PCoA) based on Jaccard similarity between the genus-level microbiota profiles of PV patients with or without thrombosis history and marrow fibrosis. (C) Boxplot showing the relative abundance distribution of the genera that were differentially represented between PV patients with or without marrow fibrosis (# p < 0.05; ** p < 0.01; Wilcoxon rank-sum test).