Identification of novel predictors of cardiovascular disease risk in MDS patients: Results of a prospective observational single-centre cohort study

Abstract

Abstract Α shared pathobiology between MDS and cardiovascular disease(CVD) is postulated, but many MDS patients do not develop CVD and only about 10% die from CVD, whereas there is lack of MDS-specific predictors of CVD risk and death. We conducted a prospective observational single-centre cohort study in lower risk(LR) MDS patients to identify patients at risk for CVD development and death. Patients underwent evaluation for CVD every 6 months by ultrasound and coronary artery calcium(CAC) scan. The carotid plaque score(CPS) was calculated as reported previously (Stein J, 2008). MESA, Framingham(FRS), PREVENT, HELLENIC and SMART risk scores and serum markers of CVD were also assessed. Progression free survival (PFS) was defined as the time from first evaluation to CVD event, MDS progression or death. Of 37 patients recruited in the study 18 had preexisting CVD(pCVD) and 19 had no prior history of CVD(naïve, nCVD). No differences between pCVD and nCVD patients were found for age, sex, WHO subtype according to 2022 classification, IPSSM and mutational profile, but pCVD patients were more often transfusion dependent (p=0.005) and had higher NT-proBNP levels (p=0.037) at baseline. In sharp contrast to general population (Okwuosa, TM et al. JACC 2012), both CAC and CPS scores were not associated with FRS, PREVENT and HELLENIC scores and were increased in low and intermediate risk patients, suggesting that the current CVD risk prediction tools are imprecise in MDS. 27 patients (pCVD:14, nCVD:13) had >1 evaluation. The mean annual CAC increase was 60±31% in pCVD and 54.5±20% in nCVD (p=0.5), markedly higher than expected in the general population (McCullough PA et al. Arch Int Med 2009). We ranked patients from slow (SP) to fast (FP) progressors for accelerated subclinicalatherosclerosis based on weighted annual absolute and relative differences in CAC and CPS, adjusted for the corresponding CVD risk scores. Of note, 7 nCVD patients were ranked as faster progressors from 7 pCVD ones. No associations of SP/FP rankings with age, sex, WHO subtype, IPSSM and mutational profile were found. By contrast, there was a strong correlation with transfusion status (p=0.007) and NT-proBNP levels (p=0.002) at baseline. With a median F/U of 38.8(95% CI 19.6-28) months 8 patients suffered CVD events and 10 died. No difference in PFS was found in SP vs FP and pCVD vs nCVD comparisons. As peripheral blood monocytes are directly involved in the development and progression of CVD we assessed their immunophenotypic profile by multiparametric flow cytometry. In line with their proatherogenic role we observed higher levels of intermediate (p=0.002) and CCR2+ intermediate (p=0.002) monocytes in pCVD vs nCVD patients, whereas SP/FP ranking also correlated with CCR2+ intermediate monocytes (p=0.013). Bulk RNA-seq was then performed on purified intermediate monocytes from 4 SP and 5 FP patients. We identified 176 upregulated and 182 downregulated genes in FP compared to SP patients. The 2 top downregulated genes wereDUSP1 and JUNB, both of which have a protective role against atherosclerosis by modulating monocyte behavior and inflammation. Another downregulated gene, shared with the atherosclerosis monocyte signature of the MESA cohort (Liu et al, Nat Commun 2017), was ALDH1A1, which protects against vascular calcification. GO enrichment revealed upregulation of genes associated with mitochondrial function and enrichment of pathways related to ribosomal small subunit assembly in FP patients, whereas genes associated with the integrated stress response, responses to hypoxia and calcium were downregulated. GSEA against MSignDB and KEGG2 showed a positive correlation of the molecular signature of FP patients to that of TNF signaling and lipid and atherosclerosis pathways. Collectively, the transcriptomic profile of intermediate monocytes from FP patients indicates a state of increased cellular stress triggered by chronic inflammation with a parallel loss of compensatory defense mechanisms. To our knowledge this is the first study in LR-MDS patients assessing longitudinally essential clinical and laboratory factors associated with CVD and providing objective measurements of subclinical atherosclerosis. Our results suggest that the established CVD risk models are dysfunctional in LR-MDS patients. Transfusion status, NT-proBNP levels and the molecular signature of intermediate monocytes may act as novel indicators of CVD risk in MDS patients.