Identification of potential crucial genes and pathways associated with vein graft restenosis based on gene expression analysis in experimental rabbits

Introduction

Occlusive artery disease is a major cause of morbidity and mortality worldwide (Bansilal, Castellano & Fuster, 2015). Despite development of novel treatments in past decades, coronary artery bypass graft (CABG) surgery remains the standard of care for patients with left main coronary artery disease (CAD) and three-vessel CAD (Serruys et al., 2009). In contrast, most patients with late-stage peripheral artery occlusive disease are treated with peripheral artery bypass graft surgery (Weintraub et al., 2012). Due to their advantages in availability and length, veins are the most commonly used conduits in coronary and peripheral artery vascular surgeries (Goldman et al., 2004). However, one major issue that has dominated the field for many years is that autologous veins are especially prone to failure. Although data have shown that the patency rates of vein grafts diminish from 98% immediately after surgery to <88% within the first month post-surgery and to 60% at 10 years after surgery. To date, no unequivocally effective treatments are available for vein graft failure (Elmore et al., 2016). Thus, clarifying the cellular and molecular mechanisms involved in VGR and identifying potential novel therapeutic targets for the prevention of restenosis are significant goals.

At present, vein grafts are thought to undergo an adaptation process in new arterial environments during which intimal thickening and structural vessel wall remodeling appear (Buccheri et al., 2016; De Vries et al., 2016). However, the cellular and molecular mechanisms underlying intimal hyperplasia and the vascular remodeling process are largely unknown. Several studies have identified individual genes involved in restenosis; for instance, early growth response protein 1 (Egr-1) was suggested to promote endothelial cell(EC) proliferation and induce vein graft restenosis by up-regulating ICAM-1 expression (Zhang et al., 2013). Tumor necrosis factor alpha-stimulated gene 6 (TSG-6) was indicated to suppress restenosis of vein grafts in rats by inhibiting the inflammatory response (Zhang et al., 2017). Despite the rapid development of genome sequencing and microarray technologies, very few studies have characterized genes and pathways associated with VGR at the transcriptome level, which may be partly due to the challenge of obtaining appropriate samples and establishing animal models.

Kalish et al.(2004) identified the transcriptome responses of canine vein bypass grafts using transcriptional profiling and defined the individual contributions of ECs and VSMCs in a cell-specific vein graft transcriptome analysis (Bhasin et al., 2012). However, a main weakness of the former study was the unavailability of canine-specific gene arrays in 2004, as mentioned by the authors (Bhasin et al., 2012). The most important limitation of the latter study lies in the fact that only the role of ECs and VSMCs were taken into consideration. However, other cells and molecular components, such as inflammatory cells and immune cells, also play pivotal roles in mediating VGR (De Vries et al., 2016).

The purpose of the current study was to determine potential crucial genes and pathways involved in VGR. A bioinformatics analysis was performed with gene expression data from our previous study (Wang et al., 2016), which avoided the limitations of the aforementioned microarray data. After screening differentially expressed genes (DEGs) and performing a functional enrichment analysis, protein-protein interaction (PPI) networks of the DEGs were constructed and visualized. The present study makes notable contributions to our understanding of the molecular mechanisms involved in VGR and provides valuable biological information for further exploration of potential candidate biomarkers and therapeutic targets for the prevention of VGR.

Materials & Methods

Results

Identification of DEGs

The statistical analysis software R (version 3.4.1; R Core Team, 2017) was used to identify differentially expressed genes using an adjusted p-value <0.05 and —log2FC (fold change)—≥1 (Fig. 1) as criteria. A PCA plot was generated using normalized data. The PCA plot shows clear separation between the surgery and control group samples (Fig. 2). A total of 858 up-regulated and 817 down-regulated DEGs were obtained in the surgery vein graft group compared with the control group. Hierarchical clustering analysis of the top 100 DEGs is shown in Fig. 3.

PPI network analysis

Based on the information in the STRING database, the PPI network consists of 695 nodes and 2,098 interactions. The top 34 hub nodes with more than 20 degrees were screened and defined as the “PPI hub genes” gene list. Moreover, nine hub genes obtained from the intersection of the “KEGG genes” and “PPI hub genes” gene lists (Fig. 4) are listed in Table 4. In addition, a comprehensive PPI network was constructed by merging the 9 PPI networks of the nine hub genes and visualized using the Cytoscape software (Fig. 5). A total of 695 nodes and 2,098 edges were analyzed using the plug-in MCODE, and the top three significant modules were screened (Fig. 6).

Table 4:

9 hub genes appearing higher degrees and involved in key pathways associated with VGR.

Gene	Degree	KEGG pathway	logFC	P Value
FGFR1	21	MAPK signaling pathway, PI3K-Akt signaling pathway	−1.59	5.13E−07
PIK3CG	24	TNF signaling pathway, Toll-like receptor signaling pathway, Chemokine signaling pathway	2.52	1.42E−13
CDK1	29	Cell cycle	4.34	9.07E−12
YWHAE	34	Cell cycle	1.36	2.57E−12
YWHAG	35	Cell cycle	1.89	1.41E−10
CCNB2	21	Cell cycle	3.70	2.87E−10
MAD2L1	21	Cell cycle	2.25	2.08E−13
PCNA	30	Cell cycle, DNA replication	1.16	2.51E−12
GNB4	20	Chemokine signaling pathway	1.91	5.52E−16

DOI: 10.7717/peerj.4704/table-4

To further lend some more validity to the analysis results, the day 7 versus day 1 comparison was served as a control relative to day 7 versus sham surgery comparison. The expression of these nine hub genes in three groups (day 1, day 7 and control) is shown in Fig. 7. As expected, most hub DEGs identified in day 7 versus sham surgery group, are also differentially expressed in day 7 versus day 1 group, which may due to the influence of surgery hit, and all hub DEGs are also differentially expressed in day 1 versus sham surgery group.

Discussion

Improving understanding of the cellular and molecular mechanisms underlying VGR is critically significant for reducing the rates of vein graft failure. In the present investigation, an integrated bioinformatics analysis was performed to identify potential key genes and pathways involved in regulating VGR. In summary, a total of 858 up-regulated and 817 down-regulated DEGs were screened. A total of 14 potential crucial pathways were selected from the functional annotation enrichment analysis. In the PPI network analysis, a comprehensive PPI network of the 9 hub genes was constructed. Furthermore, three significant modules were screened from the PPI network by module analysis.

All nine hub genes showed higher degrees in the PPI network analysis and were involved in potential key pathways. Several previous studies demonstrated that fibroblast growth factor receptor 1 (FGFR1) activated the Akt/mTOR pathway (Chen & Friesel, 2009) and regulated phenotypic modulation of vascular smooth muscle cells (VSMCs) (Chen & Friesel, 2009; Chen, Simons & Friesel, 2009). As described in our previous study, the mTOR signaling pathway was indicated to play a major role in the arterialization of vein grafts (Wang et al., 2016). Moreover, a fibroblast growth factor receptor antagonist was validated to block growth factor-mediated VSMC proliferation by inhibiting binding of fibroblast growth factor 2 (FGF2) to VSMCs and soluble FGFR1 (Segev et al., 2002). VSMC proliferation was also thought to play a critical role in the pathophysiology of restenosis. The second hub gene PIK3CG (phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit gamma) was shown to be involved in three key pathways: the TNF signaling pathway, Toll-like receptor signaling pathway, and chemokine signaling pathway (Table 4). CDK1 (cyclin-dependent kinase 1) is a protein-coding gene that plays a key role in regulating the eukaryotic cell cycle by modulating the centrosome cycle and mitotic onset and thus cell proliferation. Cell proliferation requires the expression of CDKs, and CDKs are considered a proliferation signaling cascade of VSMCs (Schad et al., 2011). Evidence indicated that radiation inhibited VSMC proliferation via cell cycle arrest by enhancing p21 expression and suppressing CDK1 and 2 (Kim et al., 2004). Consistently, a derivative of the CDK inhibitor roscovitine was validated to potently induce G1 phase arrest and therefore inhibit VSMC proliferation (Sroka et al., 2010). YWHAE (tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein epsilon) has 34 degrees in the PPI network and is also included in the cell cycle pathway. Moreover, YWHAE belongs to the 14-3-3 family of proteins that induce signal transduction by binding to phosphoserine-containing proteins. Intriguingly, a recent study provided evidence that up-regulation of YWHAB in endothelial cells played a pivotal role in intimal hyperplasia following carotid artery injury via enhancing endothelial cell proliferation and migration (Feng et al., 2017). YWHAG knockdown in zebrafish was reported to reduce the brain size and increase the diameter of the heart tube (Komoike et al., 2010). CCNB2 (cyclin B2) is a member of the cyclin family, specifically the B-type cyclins. Evidence indicated that IL-18 promoted cell proliferation via NF-κB and the p38/ATF2 pathway by targeting CCNB2 (Zhang et al., 2015). Down-regulation of MAD2L1 was demonstrated to be involved in suppressing the proliferation, migration, invasion, apoptosis induction and cell cycle arrest of cancer cells (Li, Bai & Zhang, 2017). PCNA (proliferating cell nuclear antigen) is a molecular marker for proliferation due to its role in replication and is involved in both the cell cycle pathway and DNA replication; PCNA had 30 degrees in the PPI network. Previous studies showed that PCNA expression was correlated with stent-induced in-stent restenosis (Gao et al., 2013) and proposed that PCNA might be a potential future target (Wang, 2014). Guanine nucleotide-binding protein subunit beta-4 (GNB4) was involved in the chemokine signaling pathway and showed 20 degrees in the PPI network. Surprisingly, we noted that six of the nine hub genes (CDK1, YWHAE, YWHAG, CCNB2, MAD2L1, and PCNA) were involved in the cell cycle pathway, which further supported the enormous significance of this pathway in regulating VGR (Table 4). The key role of the cell cycle pathway in regulating VGR has been established (Sriram & Patterson, 2001); Hierarchical clustering analysis of DEGs in cell cycle pathways was shown in Fig. 8. However, the underlying molecular mechanisms remain largely unknown. Based on the abovementioned evidence, these hub genes may potentially interact with each other and regulate the VGR program by modulating the cell cycle pathway. Future studies focusing on revealing the specific cellular and molecular mechanisms involved in regulating the VGR program may provide novel therapeutic targets for VGR inhibition.

Over the past few decades, many powerful research tools in genomic technologies have advanced rapidly, such as genome sequencing and microarray technologies. Microarray technology has enhanced our ability to screen samples for thousands of genes at once, which may provide novel insights into potential molecular mechanisms involved in regulating the VGR program. Unfortunately, high throughput data associated with VGR are extremely scarce at present. “Restenosis” was used as a search term to search in the GEO (Gene Expression Omnibus) database (https://www.ncbi.nlm.nih.gov/geo/), and only 7 datasets were returned; this lack may be partly due to the challenge of obtaining appropriate samples in vivo and establishing a VGR animal model. Hence, we suggest that collaborations between clinicians and basic researchers may accelerate our understanding of the cellular and molecular mechanisms involved in VGR and therefore improve the clinical efficacy of occlusive artery disease.

Due to the extreme lack of high throughput data associated with VGR, more research is required to screen key genes and pathways involved in this process. Further experimental investigations are needed to estimate the role of these candidate genes and pathways. However, there are certain limitations in the present study, the aforementioned results, including the gene expression levels and their functions, were needed to be validated by experiments, which would be carried out in our further studies.

Conclusions

In conclusion, this study provides a set of potential therapeutic targets for future investigations into the molecular mechanisms involved in VGR. However, we acknowledge that there were certain limitations in this study. The aforementioned results, including the expression level of genes and the definite functions of potential hub genes and key pathways, were needed to be validated by experiments, which would be conducted in our further studies.

Supplemental Information