Abstract
Despite advances in the diagnosis and management of cardiovascular diseases (CVDs), they remain the leading cause of death globally and their prevalence continues to increase, partly due to the aging population. Atrial fibrillation (AF) and heart failure (HF) are two of the most common forms of CVD. AF is the most common type of abnormal heart rhythm in which the heart beats in an irregular and often rapid manner. In contrast, HF is a condition where the heart cannot pump enough blood to meet physiological demands. While often regarded as being distinct conditions, AF and HF frequently coexist, with each condition capable of promoting and exacerbating the other.
AF and HF are both associated with an impaired quality of life and poor outcomes including increased hospitalisation rates and greater risk of mortality. Due to the significant health and economic burden these conditions pose, there is a need to improve current risk prediction tools and discover new biomarkers that can help identify at-risk individuals earlier in the disease course to facilitate preventative treatment, ultimately reducing morbidity and mortality. Accordingly, this thesis aimed to identify novel biomarkers (both genetic and protein) to improve risk prediction and help gain a deeper understanding of the mechanisms underlying AF and HF.
First, whether the addition of genetics into clinical risk scores (GRACE and CHARGE-AF scores) improves the prediction of AF in patients with coronary artery disease (CAD) was investigated. The GRACE score predicts 6-month mortality following an acute coronary syndrome while the CHARGE-AF score predicts 5-year AF risk in the general population. While neither score was designed for predicting AF risk in a CAD cohort, both include clinical risk factors of relevance to AF (e.g., history of myocardial infarction and heart failure). Forty-eight published genetic risk scores (GRSs), which estimate the cumulative contribution of common genetic variants to disease susceptibility, were reviewed and eight representative scores were selected for analysis. Associations of each individual GRS with AF at five different timepoints in CAD patients (4 months up to 10 years following an acute coronary event) were tested. This was followed by evaluating the predictive performance of the clinical risk scores with and without a GRS added and adjusting for covariates (sex, ethnicity, and the cardiac biomarker NT-proBNP).
Second, to gain insights into the molecular processes driving AF development as well as identify candidate biomarkers, RNA sequencing was performed on right atrial appendage tissues from tachypaced sheep with sustained AF (n = 4), sheep that did not develop sustained AF ("resistant", n = 7) and unpaced healthy controls (n = 5). In this study, sheep that developed sustained AF represent a model of persistent AF in humans (defined as continuous AF for ≥7 days) while the resistant sheep likely reflects paroxysmal AF (self-terminating AF).
Given that persistently high lactate concentrations are associated with poor cardiovascular outcomes while effective lactate clearance confers better outcomes (e.g., decreased risk of mortality), the final two chapters of this thesis investigated CNDP2 in the context of HF. CNDP2 is a gene that encodes a protein that may be involved in lactate clearance via Lac-Phe formation. Aim 3 of this thesis investigated evaluated the association between CNDP2 and HF outcomes through genotyping and gene expression analyses in left ventricular tissues from heart-healthy donors (n = 108) and end-stage HF patients (n = 110). The genotyping analyses aimed to identify genetic variants that may alter its expression, whereas the gene expression analyses assessed whether cardiac CNDP2 expression differs between heart-healthy individuals and those with end-stage HF. Associations between genetic variants that influence CNDP2 gene expression and risk of developing HF were then assessed in a cohort of patients with CAD (n = 204 developed HF within 5 years). Finally, attempts were made to establish a method for quantification of CNDP2 protein in biological samples, first using a commercial enzyme-linked immunosorbent assay (ELISA) kit and subsequently using a targeted mass spectrometry (MS) method developed in this thesis.
The key results from this research include finding that the GRSs for AF, which were derived from non-CAD populations, were associated with AF in patients with CAD and their addition to the clinical risk scores led to moderate increases in predictive performance. These findings suggest that GRSs may capture risk beyond traditional clinical risk factors and may help refine risk stratification. Meanwhile, RNA sequencing detected over 12,000 genes in the right atrial appendage tissues of the sheep models, including 385 previously implicated in AF genome-wide association studies. Importantly, gene expression analyses revealed distinct molecular patterns between the sheep groups, with the resistant group exhibiting a larger number of altered genes than the AF group when compared to controls. These findings suggest that resistance to sustained AF may involve the engagement of a broad transcriptional response. Further, twenty-seven candidate protein biomarkers encoded by the differentially expressed genes were identified, two of which (DAPK1 and EMID1) have not been previously reported to be associated with AF.
The genotyping and gene expression analyses performed on CNDP2 showed that it is detectable in human cardiac tissue and its expression may be modulated by common genetic variants including rs17088884, rs2346057 and rs7235253. Cardiac CNDP2 expression was lower in patients with HF compared to the heart-healthy donors, suggesting that those with HF may have a lower capacity to clear lactate via Lac-Phe formation. The CC genotype for the genetic variant rs7235253 was associated with an increased risk of developing HF within 5 years in patients with CAD. Finally, we found that the commercial ELISA kit could not reliably detect CNDP2, highlighting the need for independent validation of commercial assays to ensure accurate target detection. Experiments using the developed MS method showed that CNDP2 is absent in human plasma but is present at low concentrations in buffy coat. Future work will focus on investigating CNDP2 in human heart tissues.
Collectively, these findings expand our current knowledge of CVDs and have the potential to contribute to reducing its burden. The insights gained in this thesis provide a strong foundation for future studies that could enable more precise and targeted approaches to prevention, diagnosis, and treatment of CVD.