The increasing incidence and prevalence of AF often leads to severe clinical outcomes. We integrated three pAF datasets, and a total of 77 CR samples were selected by brushing between pAF samples and normal controls (0.05, | log2FC |≥0.2), of which the expression of 48 genes was found to increase, including 29 immune-related genes. Based on the above-mentioned network, the top 10 genes with the greatest degree of connectivity were selected as hub genes. From these 10 hub genes, we selected six immune-related genes for modeling to predict the occurrence of pAF—RBBP4, KAT7, KANSL2, ACTB, TRRAP, and KAT2B. TBX5 plays a critical role in the development of the heart and the forelimbs; Holt–Oram syndrome is a genetic disorder affecting the heart and the upper limbs caused by mutations in TBX5. The transcriptional activity of TBX5 is enhanced by acetylation of lys339 by histone acetyltransferases KAT2A and KAT2B, which is essential for the nuclear retention of TBX5. Knockout of kat2a and kat2b transcripts severely interferes with heart and limb development [15]. A study conducted by DouwePons et al. found a reduction in the risk of coronary heart disease mortality and restenosis with mutations in the KAT2B gene promoter [16]. Because KAT2B is negatively regulated in cardiac hypoplasia and coronary heart disease, these conditions are strong risk factors for pAF. Indeed, KAT2B gene expression is negatively regulated in pAF patients. β-actin (the protein product of ACTB) is a protein that facilitates migration, division, growth, signaling, and cytoskeletal structure to support cell growth and division [17]. Thus, β-actin may play a role in vascular remodeling, thrombosis, hypertension, and stroke pathogenesis. The expression of ACTB is significantly downregulated in stroke [18]. Polycomb inhibits histone methyltransferase activity during Complex 2 (PRC2)-mediated differentiation of heart cells, thereby controlling their development; the histone-binding protein RBBP4 is therefore an important component of PRC2. Thus, when PRC2 is defective, Polycomb can prevent the development of heart cells, as evidenced by a fatal congenital heart malformation in mice lacking Polycomb [19].
The MYST family acetyltransferase MOF regulates oxidative phosphorylation by controlling the expression of mtDNA respiratory genes in aerobic respiratory cells; the combination of MOF and mtDNA requires the assistance of KANSL. The knockout of MOF genes leads to an imbalance of pathways related to mitochondrial metabolism and oxidative phosphorylation and to the occurrence of heart failure and hypertrophic cardiomyopathy in mice [20]. Among the top five genes intolerant of missense variations is TRRAP, an evolutionarily conserved gene. Missense mutations of TRRAP cause various deformities of the brain, heart, kidney, and urogenital system [21]. In a rat model studied by Hao et al., intracellular reactive oxygen species (ROS) levels are increased through the miR-134-5p/KAT7/MnSOD/catalase axis, thereby increasing the size of cardiomyocytes and activating cardiac fibroblasts. That study proved that miR-134-5p knockout effectively limited adverse myocardial remodeling in rats, which was manifested as hypertrophy and fibrosis of cardiomyocytes with less cardiac tissue. KAT7 reverses ROS accumulation in miR-134-5p cells and alleviates cardiac fibrosis by promoting the transcription of manganese superoxide dismutase (MnSOD) and catalase [22].
There is increasing evidence that immune and inflammatory responses play a role in the pathogenesis of AF. Therefore, the ssGSEA algorithm was used to study immune infiltration in AF. In this study, we found that AF patients had higher levels of TILs, mast cells, neutrophils, Tregs, Check − poin, parainflammation, T cell co-inhibition, T cell co-stimulation, HLA, cytolytic activity, and Type I IFN response. AF is often associated with an increased inflammatory response and is characterized by monocyte/macrophage infiltration. There is some evidence that intermediate CD14 + + CD16 + monocytes are closely linked to the pathogenesis of AF and represent functional remodeling of the left atrium [23]. CD3 + T cells are more common in patients with AF than in those with sinus rhythm. A small number of CD20 + B cells are found in AF, but not in sinus rhythm patients [24]. Post-ablative recurrent AF (RAF) occurs in ~ 50% of patients, and neutrophils constitute the majority of immune cells in these patients, with the ratio of neutrophils to lymphocytes increasing in peripheral circulation after surgery [25, 26]. Angioi-induced AF is largely a result of AngII activating neutrophils, instead of just directly affecting fibroblasts and cardiomyocytes [27]. Mast cells are also involved in inflammation associated with cardiovascular disease [28]. As a major component of cardiac fibrosis pathogenesis, mast cells release a multitude of fiber-forming mediators, thereby stimulating proliferating and differentiating cardiac fibroblasts [29]. AF is a condition characterized by platelet-derived growth factor-A (PDGF-A)-mediated fibrosis caused by mast cells [30]. Numerous studies have shown that T cells are closely associated with atherosclerosis, pericarditis, and other cardiovascular diseases [31]. T cells are speculated to increase hypertrophy-related gene expression and induce AF by activating nuclear calcineurin, leading to atrial hypertrophy [32]. Tregs produce IL-10, an anti-inflammatory cytokine [33]. IL-6 levels are significantly increased in AF patients and inhibit the function of Tregs, promoting the expression of α-SMA, collagen I, and collagen III, and leading to atrial fibrosis. Of note, the number of Tregs in peripheral blood and right atrial tissues of AF patients has been found to be significantly decreased [34]. In contrast, our findings suggest that Tregs are elevated in patients with AF, which may be related to our inclusion of patients with pAF. The tissues of the patients with atrial fibrillation included in the study were obtained from the left atrial ear, in which thrombi form easily when pAF occurs. Tregs regulate thrombolysis by controlling the recruitment and differentiation of monocytes and the activity of matrix metalloproteinase (MMP); increased Treg levels accelerate the absorption of thrombi [35]. The expression level of HLA in AF was also significantly increased [36]. To date, miRNAs are considered possible targets for AF gene therapy [37]. We also predicted related regulatory miRNAs based on six immHub genes (RBBP4, KAT7, KANSL2, ACTB, TRRAP, and KAT2B): miR-3177-5p, miR-4694-5p, miR-3681, miR-3713-5p, and miR-4259-5p. Wang et al. showed that circ-SIRT1 can upregulate its host gene SIRT1 through the post-transcriptional levels of miR-3681-3p/miR-5195-3p, and this SIRT1 activation can inhibit cardiac hypertrophy and promote autophagy [38]. Exosomal miR-3681-5p may be used as a biomarker in ST-segment elevation myocardial infarction [39]. However, for miR-3177-5p and other microRNAs, there are few regulatory channels related to cardiovascular diseases, which still need further investigation.
We constructed a nomogram prediction model based on six immHub genes. The AUC values were 0.861 in the training dataset and 0.83 in the validation dataset, suggesting that RBBP4, KAT7, KANSL2, ACTB, TRRAP, and KAT2B are potential biomarkers for predicting the risk of AF. The results of this study provide insights into the pathogenic mechanism of AF and can help the development of future targeted therapies. As a medical tool, nomograms are useful not only in predicting disease risk or survival outcomes but also in screening high-risk patients and ensuring appropriate treatment. The results of this study show good discriminative ability. The approach of using a prognostic model is simple and the results are easy to understand. First, adding all the "points" to obtain a "total score" can be achieved by drawing a vertical line from each factor to the "point" line. Finally, the risk of AF can be obtained by drawing a vertical line from the "total score" to the "risk of AF." Nevertheless, the occurrence of AF is associated with risk factors including age ≥ 65 years, hypertension, diabetes, previous history of stroke/transient ischemic attack/thromboembolism, vascular disease, heart failure, obesity, sleep apnea, coronary artery disease, heart failure, and alcohol or tobacco use [40, 41]. When using the level of gene expression to evaluate the risk of AF in clinical practice, the risk related to the occurrence of atrial fibrillation should still be considered comprehensively. There are some limitations to our study. Because our study is based on a bioinformatics analysis of transcriptome data from a public dataset, its findings may not be consistent with clinical reality. We studied AF-related CR genes and selected the six most relevant hub genes that would predict modeling in AF. However, because there are few reports on CRs associated with AF, the findings of our study must be validated by future in vitro and in vivo investigations.