Abstract
Background: The formation of thrombi inside arteries and veins is a process that underlies the pathogenesis of the most important causes of morbidity and mortality in the modern world: namely cardiovascular diseases (CVD), a term encompassing acute myocardial infarction (AMI), ischemic stroke (IS) and peripheral arterial occlusive disease (PAOD); and venous thromboembolism (VTE). So far, the discussion on how much CVD and VTE share and differ in terms of pathological mechanisms has been based on morphological studies about the composition of venous and arterial thrombi, epidemiological studies about the association and common risk factors of CVD and VTE, and inferences upon the cellular and molecular mechanisms of each of these conditions. Objective: Here we introduce a new strategy to obtain insights on the association of CVD and VTE based on the comparison of transcriptomic signatures associated with these conditions. Methods: publicly available datasets (from Gene Expression Omnibus Database) of gene expression data from patients with AMI, IS, PAOD, cardioembolic stroke (CS), and VTE were identified and retrieved. Stringent selection criteria regarding patient selection, microarray methods and RNA sources were applied to limit heterogeneity as much as possible. A meta-analysis was performed using a robust statistical method, allowing the identification of genes that are commonly expressed in the same direction in all five conditions. We also identified a list of genes that were divergently expressed between VTE and CVD. Both lists were then used for functional and correlation analyses aimed to gain additional insights on the molecular pathogenesis of arterial and venous thrombosis. Results: Five datasets fulfilled the inclusion criteria and were included in our analyses: the dataset of VTE patients originally compared gene expression levels in patients with single or recurrent VTE with healthy controls (GSE19151); three datasets involving CVD compared gene expression levels of patients with PAOD (GSE 27034), AMI (GSE 48060) and IS (GSE 22255) with study-specific healthy controls. A dataset comparing patients with CS (GSE58294) and healthy controls was also included. In total, data from 163 adult patients and 145 healthy controls were analyzed. The meta-analysis generated a list of 473 genes whose expression was changed in the same direction in all five conditions (168 up-regulated and 305 down-regulated). The top 20 up- and down-regulated genes are shown in figure 1a. The functional analysis of the 168 up-regulated genes identified pathways associated with hemostasis, innate immunity and neutrophil activation and degranulation as overrepresented in patients compared to controls. When gene expression fold-change levels were analyzed across all five diseases, a lower correlation coefficient was observed between VTE and CVD, when compared to correlations observed within CVD such as AMI and PAOD (fig.1b). Moreover, an unsupervised cluster analysis showed that the expression pattern of these 473 genes in AMI, PAOD and IS are more closely related to each other than to VTE and CS (fig. 1a). When the 124 genes that were divergently expressed between VTE and CVD were functionally analyzed, genes that were upregulated only in CVD were associated with activation of innate immunity, whereas genes that were upregulated in VTE were associated with cell maintenance pathways. Discussion and conclusion: our comparative analyses provide new data on how venous and arterial thrombosis differ at a transcriptional level. Specifically, we identified a set of 473 genes whose expression levels were consistently changed in both CVD and VTE, and 124 genes with divergent expression levels between these conditions. The prominence of pathways associated with hemostasis, innate immunity and neutrophil function in both CVD and VTE are in line with the concept of immunothrombosis in the pathogenesis of thrombotic disorders. The clustering of MI, IS and PAOD compared to VTE and CS is coherent with the concept that CS is caused by thrombi formed in a different microenvironment than IS. Additional studies are warranted on the role of these genes and pathways in the pathogenesis of CVD and VTE, as well as on the biological relevance of these thrombosis-specific gene expression signatures in other thrombotic disorders.
Disclosures
De Paula: Hematology and Transfusion Medicine Center, University of Campinas: Employment.
Is It Possible to Find Needles in a Haystack? Meta-Analysis of 1000+ MS/MS Files Provided by the Russian Proteomic Consortium for Mining Missing Proteins