Identification and multiparametric ranking of candidate genes involved in hepatitis C virus entry and host immune response using bioinformatics methods

Full Text

Introduction

The hepatitis C virus (HCV) remains a global public health problem, affecting more than 71 million people worldwide [1]. Approximately 75% of hepatitis C cases become chronic, and in severe cases, the disease can lead to liver cirrhosis or hepatocellular carcinoma [2]. Although direct-acting antiviral drugs exist that demonstrate efficacy exceeding 95%, their widespread use is limited by a number of factors, such as the high cost of therapy, side effects, the emergence of resistant viral variants, and the possibility of reinfection [3, 4]. Moreover, HCV infection is mostly asymptomatic and often does not induce sterilizing immunity, which contributes to reinfection or further disease progression [5]. Collectively, these challenges have led to a continuing rise in the number of HCV infections. The development of an effective vaccine against HCV has not yet been achieved, due to the significant genetic variability of the virus, its structural plasticity, and a lack of data on the precise spatial organization of the virion [6].

The immune system is capable of eliminating the virus during acute HCV infection in approximately 25% of cases; moreover, in individuals who have spontaneously recovered from a primary infection, the probability of clearance upon reinfection reaches 80% [7]. This indicates the formation of sterilizing immune memory, which is based on the induction of broadly neutralizing antibodies against the E1 and E2 envelope glycoproteins, as well as the development of a polyspecific T-cell response [8]. The E1 and E2 glycoproteins mediate viral entry into the cell through interaction with host cell receptors. HCV entry requires coordinated interaction with a complex of cellular receptors, including tetraspanin CD81, tight junction proteins claudin-1 (CLDN1) and occludin (OCLN), and the SR-BI receptor (scavenger receptor class B member 1, product of the SCARB1 gene) [9]. As an alternative to the SR-BI receptor, the virus can also use the low-density lipoprotein receptor (LDLR) for attachment [10].

Effective HCV clearance and the development of immune memory depend on the coordinated action of the innate and adaptive immune systems. Dendritic cells, natural killer cells, T-helper cells, and cytotoxic T-lymphocytes specific to viral antigens play a key role in this process [11]. In turn, HCV has developed complex mechanisms to evade the immune response, including inhibition of the interferon signaling cascade, disruption of antigen-presenting cell functions, and T-cell exhaustion [12, 13]. Host genetic characteristics that influence the effectiveness of these immune mechanisms may determine the outcome of the infection—ranging from spontaneous resolution to chronic infection with the risk of developing cirrhosis and hepatocellular carcinoma. However, the experimental identification of significant genetic determinants and their polymorphic variants among all human genes is a complex and resource-intensive task. In this regard, bioinformatics methods play a paramount role, enabling the ranking of candidate genes (CGs) based on the analysis of large datasets. Modern computational approaches make it possible to integrate heterogeneous information, model interactions in the virus–host system, and identify new molecular patterns associated with the development and progression of HCV infection.

Thus, identifying not only host genes encoding entry receptors but also genes regulating key stages of the antiviral immune response is an important step toward understanding the pathogenesis of HCV infection.

The aim of the study is to comprehensively assess the potential involvement of CGs and their products in the pathogenesis of HCV infection during the stages of viral entry into the cell and the formation of the host immune response, based on an integrative analysis of gene expression, the localization of their products at the cellular level, and their participation in molecular pathways and biological processes.

Materials and methods

Initial data

Previously, using comprehensive bioinformatics analysis and subsequent ranking, 35 CGs were identified that are associated with proteins involved in the stages of HCV attachment and entry into the cell and that potentially contribute to the pathogenesis of HCV infection. A group of genes was used as background genes (BGs), whose products, according to the literature, are involved in interacting with the viral glycoproteins E1 and E2 during the HCV cell entry stage: CD81, CLDN1, LDLR, OCLN, SCARB1 [14].

To gain insight into the putative biological mechanisms of the priority genes, we used the FUMA GWAS program in the functional mapping and GENE2FUNC gene annotation mode [15]. This mode allowed us to obtain biological information for each gene. FUMA GWAS was applied under the following conditions: ensemble version 102, GTEx v8 expression dataset: 30 major tissue types, the Benjamini–Hochberg multiple correction method for gene set enrichment testing, maximum adjusted p-value for gene set association < 0.05, minimum overlapping genes with gene sets ≥ 2.

Thus, the methodology for identifying genes potentially associated with the pathogenesis of hepatitis C consisted of several stages. The initial search for gene-protein interactions was performed using the HumanNet, STRING, and GeneMANIA protein-protein interaction databases [14]. The reliability of the predicted interactions was assessed using ROC analysis (AUROC metric). For subsequent functional annotation of the selected genes, the FUMA GWAS tool (GENE2FUNC mode) was used, which enabled a comprehensive characterization, including an assessment of the level and tissue specificity of gene expression, the subcellular localization of products, as well as their involvement in key biological processes and molecular pathways.

Methodology for evaluating candidate genes

To objectively analyze the functional significance of the CG, a ranking system was developed based on a point-based ranking, following the principles described previously [16]. The system involves a sequential evaluation based on several independent criteria, the results of which are combined into a single score.

The expression profile of the BGs was analyzed in various tissues of the body. The inclusion criterion for a particular tissue in the analysis was a gene expression level of log2 ≥ 2.51, corresponding to the average expression level. Tissues with low BG expression levels were excluded from the analysis. For each CG, the proportion of cases of its co-expression with CGs in the selected tissues was assessed.

CG scores (CGSExp) were calculated using the formula:

CGSExp = f_eCD81 + f_eOCLN + f_eCLDN1 + max (f_eSCARB1, f_eLDLR),

where f is the frequency of overlap between the expression profiles of the candidate genes and the corresponding background genes. Since the entry of HCV into the host cell requires the involvement of several key human receptors, this model treats all of them as equally important. Given that the SR-BI and LDLR receptors perform interchangeable functions in the virus attachment process, the operator max (f_SCARB1, f_LDLR) is used in the formula to assess the contribution of this stage, accounting for the maximum frequency of CG expression with one of these BGs.

To assess the functional and spatial proximity of CGs to BGs, a comparative analysis was performed across 3 categories:

• subcellular localization of proteins;
• involvement in molecular pathways;
• participation in biological processes.

Based on annotations from the FUMA GWAS database (GENE2FUNC mode), binary matrices were constructed for each category, where the presence of a common trait in a pair of genes was denoted as 1, and the absence as 0.

In the subsequent analysis, only statistically significant correlations with a p-value < 0.05 were considered. The total score for each individual analysis was calculated using the following formula:

CGSLoc = k_CD81 + k_OCLN + k_CLDN1 + max (k_SCARB1, k_LDLR),

where k is the correlation level between the CG and the BG (if p > 0.05, the value is set to 0). By analogy with the previous step, the operator max(k_SCARB1, k_LDLR) accounts for the maximum frequency of CG expression with one of these BGs.

Given the equal contribution of each analyzed aspect (expression, subcellular localization, involvement in molecular pathways and biological processes), the final score for the CG was determined as the sum of the scores obtained at each stage. The maximum score at each stage was 4, and the maximum total score was 16. To enhance the significance of CGs demonstrating associations across multiple characteristics, additional points were assigned: +2 points for a match across 2 characteristics, +3 for 3, and +4 for 4. The threshold for the total score was set at 4; CGs that scored this or higher were assessed as significant for the pathogenesis of hepatitis C.

Statistical analysis

To assess the degree of association between CGs and BGs across categories of subcellular localization, involvement in molecular pathways, and biological processes, the correlation coefficient φ was used. The choice of method was dictated by the binary nature of the analyzed data: based on annotations obtained from FUMA GWAS (GENE2FUNC mode), a binary matrix was constructed for each gene pair and each trait (1 — common annotation present, 0 — absent). For each such matrix, the value of φ and the achieved significance level were calculated. Correlations were considered significant at p < 0.05. The analysis was performed using the Prizm 10.2.3 software (GraphPad Software Inc.).

Results

Previously, using bioinformatics analysis methods — specifically, the construction of genetic and protein-protein interaction networks — CGs potentially involved in the pathogenesis of hepatitis C were identified [14]. The identified CGs were grouped into functional sets (Table 1).

 

Table 1. Functional groups of the CGs

Functional group	CGs	Role in HCV infection
Cell barriers and intercellular contacts	TJP1, TJP2, CLDN2, CLDN3, CLDN5, CLDN6, CLDN9, CLDN11, CLDN12, CLDN17, ESAM, F11R, IGSF8, GJB1, PDZK1, DAB1, MMP2	Maintenance of the barrier, viral penetration, cell migration, matrix remodeling
Lipid metabolism and lipoproteins	APOA1, APOA2, APOE, APOB, LIPC, LDLRAP1, MYLIP, PCSK9, LRPAP1	Lipovirus particle formation, lipid metabolism, viral entry
Immune response and interaction with the virus	CD19, CD9, IFITM1, C3, PTGFRN	Immune response, opsonization, viral inhibition, B-cell activation
Signaling pathways, proteolysis, and the cytoskeleton	CSNK1E, ADAM10, ITGB1, NEDD4L	Regulation of cellular processes, adhesion, proteolysis, viral interaction

 

Assessment of the expression levels of candidate genes

For further analysis, the selected genes were uploaded to FUMA GWAS using the GENE2FUNC mode. This resulted in a heat map showing the expression of the genes of interest across 30 different organ and tissue types (Fig. 1).

 

Fig. 1. Heat map of gene expression in tissues and organs. Genes involved in infection and/or the development of hepatitis C have been identified: CD81, CLDN1, LDLR, OCLN, SCARB1.

 

The following scale was used to assess expression levels: maximum expression (5.672), high (3.510–5.671), medium (2.51–3.50), low (1.51–2.50), and minimal (0–1.5).

Based on the analysis of the expression heatmap, it can be concluded that a fairly high level of expression is observed for most of the genes studied in the presented set of tissues and organs.

For CD81, the maximum expression level (5.672) is observed in 24 tissues, and a high level (5.030–5.671) in the remaining 6 tissues.

The SCARB1 gene demonstrates moderate and high expression levels in all analyzed organs and tissues. The maximum expression level is observed in the adrenal glands (5.672). High expression levels are observed in 22 tissues, and moderate levels in 7.

For LDLR, a wide range of tissues with varying expression levels is observed. High expression levels were detected in 23 tissues, and moderate levels in 6. The lowest expression level is observed in muscle (1.543).

For the CLDN1 gene, expression levels vary significantly among different tissue types. High expression levels are shown in 10 tissues, moderate levels in 11, and minimal levels in 8. Low expression levels are shown in blood vessels (1,572).

For OCLN, high expression levels are observed in thyroid tissues (3,614), and moderate levels in lung tissues (2,955). The gene exhibits low expression levels in 7 tissues. In other tissues, the expression level of this gene is minimal.

Based on the evaluation of expression profiles, the following CGs with the highest co-expression with BGs were identified:

CD81: CSNK1E (100%), IFITM1 (100%), ITGB1 (100%), LRPAP1 (100%), CD9 (97%), IGSF8 (97%), APOE (93%), C3, (93%), ESAM (93%), TJP1 (93%), F11R (90%), MMP2 (90%), MYLIP (90%), TJP2 (90%), CLDN5 (83%), LDLRAP1 (83%), CLDN12 (67%), PTGFRN (67%), ADAM10 (53%);

SCARB1: CD9 (80%), CLDN12 (80%), IGSF8 (80%), LDLRAP1 (80%), MYLIP (80%), APOE (77%), CSNK1E (77%), ESAM (77%) F11R (77%), IFITM1 (77%), ITGB1 (77%), LDLRAP1 (77%), LRPAP1 (77%), TJP2 (77%), C3 (73%), TJP1 (73%), MMP2 (70%), CLDN5 (67%), ADAM10 (60%), PTGFRN (60%);

LDLR: LDLRAP1 (90%), MYLIP (86%), APOE (83%), CD9 (83%), CSNK1E (83%), ESAM (83%), F11R (83%), IFITM1 (83%), IGSF8 (83%), ITGB1 (83%), LRPAP1 (83%), C3 (79%), CLDN5 (79%), TJP1 (79%), TJP2 (79%), CLDN12 (76%), MMP2 (76%), PTGFRN (72%), ADAM10 (66%);

CLDN1: LDLRAP1 (57%), ADAM10 (48%), APOE (48%), C3 (48%), CD9 (48%), CLDN12 (48%), CSNK1E (48%), F11R (48%), IFITM1 (48%), IGSF8 (48%), ITGB1 (48%), LRPAP1 (48%), TJP2 (48%);

OCLN: ADAM10, APOE, C3, CD9, CLDN12, CLDN3, CLDN5, CSNK1E, ESAM, F11R, GJB1, IFITM1, IGSF8, ITGB1, LDLRAP1, LRPAP1, MMP2, MYLIP, PTGFRN, TJP1, TJP2 (все 100%), NEDD4L (50%).

Based on an analysis of expression profiles for the CGs, the following ranking metrics were calculated: CSNK1E (3,31), IFITM1 (3,31), ITGB1 (3,31), LRPAP1 (3,31), LDLRAP1 (3,30), CD9 (3,28), IGSF8 (3,28), APOE (3,24), F11R (3,21), C3 (3,20), TJP2 (3,17), CLDN12 (2,95), ESAM (2,76), TJP1 (2,72), ADAM10 (2,67), MMP2 (2,66), CLDN5 (2,62), PTGFRN (2,39), MYLIP (1,76), CLDN3 (1,00), GJB1 (1,00), NEDD4L (1,00), PCSK9 (0,50).

Assessment of the cellular localization of candidate gene products

In addition to analyzing the tissue-specific expression of CGs, an important aspect is the assessment of the subcellular localization of their protein products. The products of these genes may directly interact with receptors involved in HCV entry and initiate intracellular signaling cascades that influence the efficiency of viral entry. Furthermore, it should be noted that proteins encoded by CGs do not necessarily function in the same cells where they are expressed; they may be secreted or transported to other cellular compartments or even to neighboring cells, modulating susceptibility to infection in adjacent tissue microenvironments. The subcellular localization of the products of the studied CGs is visualized in Fig. 2.

 

Fig. 2. Localization of CG products. Genes are plotted on the horizontal axis, while cell types and/or their structures are plotted on the vertical axis. The following genes are marked in red: CD81, CLDN1, LDLR, OCLN, SCARB1.

 

For a significant number of gene families (APOB, APOE, LDLRAP1, PCSK9, etc.), multiple localization sites of their products have been identified at the cellular level. Most of the products of the gene families under study are localized in the following types of junctional structures: tight junctions, the apical junctional complex, intercellular junctions, and adhesive junctions. To rank the significance of CG products, a correlation analysis was performed between gene pairs.

Based on the analysis of the subcellular localization of CG products with BGs, the following CGs with the highest correlation levels were identified for each BG at p < 0.05:

CD81: TJP1 (0,35), ADAM10 (0,29), CLDN6 (0,28), CLDN11 (0,28), PDZK1 (0,27);

SCARB1: IFITM1 (0,46), PDZK1 (0,46), ITGB1 (0,35), PTGFRN (0,32), CD19 (0,26), C3 (0,26);

LDLR: CD9 (0,49), LDLRAP1 (0,47), PCSK9 (0,47), APOB (0,26), APOE (0,26);

CLDN1: TJP1 (0,78), CLDN2 (0,67), CLDN6 (0,67), CLDN9 (0,67), CLDN11 (0,67), CLDN17 (0,67), F11R (0,67), CLDN3 (0,60), CLDN5 (0,60), CLDN12 (0,57), ESAM (0,57), TJP2 (0,57), GJB1 (0,26);

OCLN: CLDN3 (0,75), CLDN5 (0,75), CLDN6 (0,69), CLDN12 (0,62), CLDN2 (0,55), CLDN9 (0,55), CLDN17 (0,55), F11R (0,55), TJP1 (0,49), ESAM (0,45), TJP2 (0,45), IFITM1 (0,38), CLDN11 (0,38), GJB1 (0,37).

The analysis of subcellular localization yielded the following CG ranking scores: CLDN6 (1,63), TJP1 (1,62), CLDN3 (1,35), CLDN5 (1,35), CLDN11 (1,33), CLDN2 (1,22), CLDN9 (1,22), CLDN17 (1,22), CLDN12 (1,19), TJP2 (1,03), F11R (1,2), ESAM (1,00), IFITM1 (0,84), PDZK1 (0,73), GJB1 (0,60), CD9 (0,49), LDLRAP1 (0,47), PCSK9 (0,47), ITGB1 (0,35), PTGFRN (0,32), ADAM10 (0,29), APOB (0,26), APOE (0,26), CD19 (0,26), C3 (0,26).

Assessment of the involvement of candidate genes in molecular pathways

The results of functional annotation, illustrating the involvement of the identified CGs, CGs and their products in common molecular pathways, are shown in Fig. 3.

 

Fig. 3. Annotation of biological pathways for selected genes, their proteins, and metabolites at p < 0.05, according to WikiPathways. Genes involved in infection and/or the development of hepatitis C are highlighted: CD81, CLDN1, LDLR, OCLN, SCARB1.

 

The analysis revealed a significant enrichment of CGs in pathways associated with lipoprotein and cholesterol metabolism, as well as cellular plasticity. The three most significant pathways (with the lowest p-values):

1) WP_METABOLIC_PATHWAY_OF_LDL_HDL_AND_TG_INCLUDING_DISEASES (the metabolic pathway of low- and high-density lipoproteins (LDL, HDL) and triglycerides, including diseases). This pathway is directly related to the pathogenesis of HCV, since the virus forms lipoprotein particles, and its entry into hepatocytes critically depends on lipid metabolism receptors: SR-BI and LDLR. Enrichment in this pathway confirms that the identified genes are involved in establishing the metabolic and membrane microenvironment necessary for efficient viral uptake and internalization;

2) WP_CHOLESTEROL_METABOLISM (cholesterol metabolism). Cholesterol is a vital component of cell membranes and HCV lipoprotein envelopes. Its homeostasis directly influences membrane fluidity, the organization of lipid rafts where entry receptors (CD81, CLDN1) are localized, and the efficiency of viral fusion. Genes associated with this pathway can modulate a cell’s susceptibility to infection by regulating the availability of cholesterol for virion assembly and the functioning of the viral entry complex;

3) WP_EPITHELIAL_TO_MESENCHYMAL_TRANSITION_IN_COLORECTAL_CANCER (epithelial mesenchymal transition in colorectal cancer). Although this pathway is described in the context of oncology, its activation is highly relevant to HCV infection. The epithelial mesenchymal transition is associated with the remodeling of intercellular junctions, loss of cellular polarity and changes in the composition of the extracellular matrix – processes that directly affect the integrity of tight junctions in hepatocytes. Since tight junction proteins (CLDN1, OCLN) are essential HCV co-receptors, genes regulating the epithelial mesenchymal transition may influence the availability of these receptors and, consequently, the efficiency of viral entry, as well as contribute to fibrosis and the progression of chronic liver disease.

Thus, the results of the functional analysis indicate that the identified CGs are concentrated around two central axes of HCV pathogenesis:

1) modulation of host lipid metabolism, which is essential for the early stages of the viral cycle;

2) regulation of the integrity and dynamics of intercellular contacts, which determine receptor availability for entry and contribute to long-term infection outcomes.

Based on an analysis of the correlation between CGs and BGs in terms of their shared involvement in molecular pathways, the following associations with maximum statistical significance (p < 0.05) were identified:

CD81: CLDN12 (0,69), MMP2 (0,69), CLDN3 (0,47);

SCARB1: APOB (0,73), APOA1 (0,46), APOE (0,47), LRPAP1 (0,30), MYLIP (0,30), LDLRAP1 (0,29), LIPC (0,29), PCSK9 (0,29);

LDLR: APOA1 (0,78), APOA2 (0,64), LIPC (0,67), APOB (0,61), APOE (0,51), LRPAP1 (0,37), PCSK9 (0,37);

CLDN1: CLDN3 (0,70), CLDN5 (0,60), CLDN12 (0,48), MMP2 (0,48), ITGB1 (0,47), ADAM10 (0,33), CLDN2 (0,33), CLDN6 (0,33), CLDN9 (0,33), CLDN11 (0,33), CLDN17 (0,33);

OCLN: CLDN11 (0,55), CLDN2 (0,55), CLDN17 (0,55), CLDN6 (0,55), CLDN9 (0,55), F11R (0,55), TJP1 (0,63), CLDN12 (0,35), MMP2 (0,35), TJP2 (0,35), CLDN5 (0,25).

The total points of involvement of CGs in molecular pathways with BGs: CLDN12 (1,52), MMP2 (1,52), CLDN3 (1,18), CLDN11 (0,89), CLDN17 (0,89), CLDN2 (0,89), CLDN6 (0,89), CLDN9 (0,89), CLDN5 (0,85), APOA1 (0,78), APOB (0,73), LIPC (0,67), APOA2 (0,64), TJP1 (0,63), F11R (0,55), APOE (0,51), ITGB1 (0,47), LDLRAP1 (0,37), PCSK9 (0,37), TJP2 (0,35), ADAM10 (0,33), LRPAP1 (0,30), MYLIP (0,30).

Assessment of the involvement of candidate genes in biological processes

Functional mapping of the analyzed genes revealed their involvement in 384 biological processes, ranked by p-values ranging from 1.01e-2 to 9.92e-4. During the analysis, processes were selected in which the BGs (CD81, CLDN1, LDLR, OCLN, SCARB1) and/or their products are involved, as well as processes associated with viral activity (Table 2).

 

Table 2. The role of BGs and CGs in biological processes. Ranked by level of evidence

No.	Involvement	Genes	p
1	Organization of intercellular junctions	F11R, CLDN11, CLDN1, OCLN, CLDN3, CLDN12, TJP2, ESAM, CD9, TJP1, ADAM10, CLDN9, CLDN6, CLDN17, CLDN5, GJB1, CLDN2	3.22e-25
2	Organization of tight junctions	F11R, CLDN11, CLDN1, OCLN, CLDN3, CLDN12, ESAM, TJP1, CLDN9, CLDN6, CLDN17, CLDN5, CLDN2	1.80e-23
3	Assembly of apical junctions	F11R, CLDN11, CLDN1, OCLN, CLDN3, ESAM, TJP1, CLDN9, CLDN6, CLDN17, CLDN5, CLDN2	1.30e-21
4	Regulation of plasma lipoprotein levels	LDLRAP1, PCSK9, APOA2, APOB, LRPAP1, MYLIP, APOA1, SCARB1, LIPC, LDLR, APOE	9.08e-19
5	Cell adhesion	DAB1, F11R, CLDN11, CLDN1, CLDN3, CLDN12, TJP2, ITGB1, CD81, APOA1, ESAM, CD9, TJP1, ADAM10, CLDN9, CLDN6, MMP2, CLDN17, CLDN5, CLDN2	1.51e-14
6	Lipid homeostasis	LDLRAP1, PCSK9, APOA2, APOB, MYLIP, APOA1, SCARB1, LIPC, LDLR, APOE	9.94e-14
7	Entry into the host body	F11R, CLDN1, ITGB1, IFITM1, CD81, SCARB1, CLDN9, CLDN6, LDLR	3.38e-12
8	Biological process involved in host-pathogen interaction	F11R, CLDN1, ITGB1, IFITM1, CD81, SCARB1, CLDN9, CLDN6, LDLR	2.90e-11
9	Biological process involved in a symbiotic interaction	F11R, CLDN1, ITGB1, IFITM1, CD81, SCARB1, CLDN9, CLDN6, LDLR, APOE	4.18e-11
10	The life cycle of a virus	F11R, CLDN1, ITGB1, IFITM1, CD81, SCARB1, CLDN9, CLDN6, LDLR, APOE	5.78e-11
11	Receptor-mediated endocytosis	LDLRAP1, PCSK9, LRPAP1, ITGB1, CD81, CD9, C3, LDLR, APOE	2.40e-10
12	Control of the assembly of bicellular tight contacts	F11R, CLDN1, CLDN3, TJP1, CLDN5	2.40e-10
13	Cholesterol transport	APOA2, APOA1, SCARB1, LDLR	7.88e-10
14	Viral process	F11R, CLDN1, ITGB1, IFITM1, CD81, SCARB1, CLDN9, CLDN6, LDLR, APOE	9.25e-10
15	Vesicular transport	LDLRAP1, PCSK9, APOA2, LRPAP1, ITGB1, CD81, APOA1, CD9, SCARB1, NEDD4L, C3, LDLR, APOE, CSNK1E	4.95e-8
16	Endocytosis	LDLRAP1, PCSK9, LRPAP1, ITGB1, CD81, CD9, SCARB1, NEDD4L, C3, LDLR, APOE, CSNK1E	6.72e-11
17	Receptor-mediated endocytosis involved in cholesterol transport	LDLRAP1, PCSK9, LDLR	1.44e-7
18	Inflammatory response	F11R, APOA2, ITGB1, CD81, APOA1, C3, LDLR, APOE	4.62e-5
19	Defensive response	F11R, APOA2, CLDN1, ITGB1, IFITM1, CD81, APOA1, C3, LDLR, APOE	3.09e-4
20	Adhesion of the symbiont to the host	CD81, SCARB1	3.12e-4
21	Regulation of immune processes	APOA2, CD81, APOA1, CD9, ADAM10, CD19, C3, LDLR, APOE	3.69e-4
22	Cytokine response	LDLRAP1, APOB, CLDN1, IFITM1, APOA1, ADAM10, MMP2	6.01e-4

 

Final ranking

Table 3 presents the results of the final ranking of the identified CGs with a total score of 4 or higher.

 

Table 3. Interim scores for each ranking stage and the total score at a threshold of 4

General list	Expression	Localization	Pathways	Processes	Bonus points	Total
CLDN12	2.95	1.19	1.52	0.74	4	10.41
TJP1	2.72	1.62	0.63	0.63	4	9.60
F11R	3.21	1.20	0.55	0.61	4	9.57
CLDN5	2.62	1.35	0.85	0.51	4	9.32
TJP2	3.17	1.03	0.35	0.62	4	9.17
ITGB1	3.31	0.35	0.47	0.53	4	8.66
CLDN3	1.00	1.35	1.18	1.11	4	8.63
IFITM1	3.31	0.84	0.00	0.62	3	7.77
ESAM	2.76	1.00	0.00	0.74	3	7.50
CD9	3.28	0.49	0.00	0.38	3	7.15
LDLRAP1	3.30	0.47	0.37	0.00	3	7.14
APOE	3.24	0.26	0.51	0.00	3	7.01
ADAM10	2.67	0.29	0.33	0.00	3	6.29
MMP2	2.66	0.00	1.52	0.00	2	6.18
PTGFRN	2.39	0.32	0.00	0.33	3	6.04
CLDN6	0.00	1.63	0.89	0.47	3	5.99
LRPAP1	3.31	0.00	0.30	0.00	2	5.61
CLDN9	0.00	1.22	0.89	0.47	3	5.58
CLDN11	0.00	1.33	0.89	0.32	3	5.54
CLDN17	0.00	1.22	0.89	0.37	3	5.48
C3	3.20	0.26	0.00	0.00	2	5.46
CLDN2	0.00	1.22	0.89	0.35	3	5.46
MYLIP	1.76	0.00	0.30	0.25	3	5.31
PCSK9	0.50	0.47	0.37	0.00	3	4.34
APOB	0.00	0.26	0.73	0.34	3	4.32

 

The number of genes with scores above the threshold was 25. It is worth noting separately all genes that scored more than 1 point: CSNK1E (3.31), IGSF8 (3.28), and CD19 (2.56). They were not selected during the intermediate stages of the analysis, but their potential biological significance may be due to their ability to interact with several BGs (or their products) within a single studied parameter. This is confirmed by the similarity of their expression profiles to those of a number of BGs, which allows them to be considered as CGs for further study.

Discussion

The interaction of HCV with the host cell via receptors and co-receptors triggers a cascade of molecular responses that determine the efficiency of viral infection, the progression to chronic infection, and the development of complications. Individual elements of this cascade can influence the efficiency of viral entry, intracellular replication, virion assembly, and their release. Assuming that potential CGs are capable of directly or indirectly influencing viral processes by binding to receptors, participating in their regulation, or modulating related molecular pathways, then their products must be in close proximity to the targets and present in sufficient concentration for such interaction, which is reflected in expression levels and subcellular localization. As a rule, such characteristics are possessed by participants in a single biological process or molecular pathway; however, the multifunctionality of individual genes expands the possibilities for interaction between the products of hypothetical CGs and BGs.

The high expression of several glycoproteins (APOA1, APOA2, APOB, APOE, C3, etc.) in the liver is consistent with their potential role in modulating the primary target of HCV—hepatocytes [17]. The products of most CGs (CD9, CLDN2/3/5/6/9/11/12/17, ESAM, F11R, TJP1/2, etc.) are concentrated in areas of intercellular junctions, including tight junctions and the apical junctional complex, which are particularly well-developed in hepatocytes. In these plasma membrane domains, transmembrane components (CLDN, CD, integrins) are mechanically linked to the cytoskeleton via TJP1/2 adaptors, forming a stable and dynamic cellular barrier [18, 19]. High expression of these molecules in the liver creates a specialized epithelial barrier that HCV uses as a functional target for entry: virions sequentially interact with receptors on the basolateral surface, followed by lateral translocation to the tight junction region, where the proteins CLDN and OCLN, which are part of the apical junction complex, are required to complete penetration [9, 10]. Consequently, CG products are localized in the same membrane microdomains where HCV receptor complexes form, suggesting their involvement in the reorganization of cell-virus interactions and the initiation of viral endocytosis [20]. This observation is correlational in nature and requires functional validation. The overlap in localization with known HCV entry receptors allows us only to suggest that some of these proteins may modulate the cell’s susceptibility to infection—either as part of the receptor complex or by influencing the integrity and dynamics of the barrier.

A statistically significant enrichment of CGs (APOB, APOE, LDLRAP1, LRPAP1, MYLIP, PCSK9) was identified in two key areas for HCV pathogenesis: the metabolism of LDL, HDL, and triglycerides (including associated diseases) and cholesterol metabolism. HCV is capable of causing persistent infection in more than 70% of infected individuals, leading to the development of chronic hepatitis C [21], which is closely associated with impaired lipid metabolism [22], manifested by abnormal lipid accumulation in hepatocytes and a decrease in serum β-lipoprotein levels [23]. Viral particles in the blood of infected individuals circulate as complexes with LDL, forming a heterogeneous population of lipoviral particles with very low or low buoyant density (< 1.03–1.25 g/mL), which varies depending on the stage of infection [24]. These particles are enriched with triglycerides and contain viral RNA, capsid protein, apolipoproteins ApoB (product of the APOB gene) and ApoE (product of the APOE gene)—key structural components of β-lipoproteins (very low-density lipoproteins (VLDL) and LDL) [25, 26]. The process of HCV assembly and secretion closely resembles lipoprotein biogenesis. It is important to note that primarily low-density particles undergo efficient secretion, whereas immature forms degrade in the post-endoplasmic reticulum, independent of the proteasomal system. Thus, HCV co-opted cellular mechanisms of lipoprotein assembly and secretion, which ensures hepatotropism and persistence of the infection [27]. HCV entry into the cell also critically depends on the lipid microenvironment. The key entry receptors CD81 and SR-BI are localized in cholesterol-enriched lipid rafts [28, 29], depletion of which inhibits viral entry [30]. The physical interaction of CD81 with cholesterol has been experimentally confirmed [31], and the fusion of the viral envelope with liposomes in vitro is enhanced in the presence of cholesterol in the target membrane [32]. These data indicate an important role for plasma membrane cholesterol in the process of HCV entry into hepatocytes.

The findings regarding the involvement of CGs in biological processes logically complement data from previous studies, as the functional role of these receptors is not limited to passive virus binding but also includes participation in signaling cascades that regulate the cytoskeleton, lipid homeostasis, and contact integrity. This confirms the status of CGs as promising targets, whose link to HCV pathogenesis is justified not only by spatial proximity but also by functional coupling. A strong correlation is observed between the subcellular localization of CG products and their functional annotation. In particular, for genes whose protein products are localized in intercellular contact structures, bioinformatics analysis confirmed their involvement in key biological processes directly related to the formation and maintenance of these specific membrane domains. This functional-spatial consistency confirms the accuracy of both experimental localization data and in silico predictions, and highlights their potential role in modulating the barrier function of HCV target cells. The most illustrative examples are the following associations: the general process of organizing intercellular contacts; the specific process of organizing tight junctions; the process of assembling apical contact complexes. This overlap is not coincidental and suggests that the identified genes are part of a coordinated functional module that regulates the organization and remodeling of intercellular junctions. Since tight junction proteins (CLDN, OCLN) are essential receptors for HCV entry, the genes regulating their organization become critical points of influence on the cell’s susceptibility to infection. Alterations in the expression or function of any of these genes can lead to remodeling of the contact complex and, consequently, to changes in the accessibility of viral receptors, which may ultimately modulate the efficiency of HCV entry.

The analysis revealed a conceptual overlap between the results of molecular pathway enrichment and biological processes, highlighting the central role of lipid metabolism in the network of interactions under study. In particular, the identified set of CGs (APOB, APOE, LDLRAP1, LRPAP1, MYLIP, PCSK9), associated with the regulation of plasma lipoprotein levels and lipid homeostasis, also partially features as key participants in the LDL and HDL metabolic pathways. These genes are key regulators of lipoproteins: APOB and APOE serve as the primary structural and ligand apolipoproteins of LDL and VLDL particles, PCSK9 and MYLIP are negative regulators that control LDLR degradation, while LDLRAP1 and LRPAP1 are involved in the endocytosis and recycling of lipoprotein receptors. Thus, enrichment in this process indicates that the identified CGs are actively involved in the fine-tuning of lipoprotein concentrations in the bloodstream. However, only BGs (SCARB1, LDLR) are associated with the cholesterol transport process, whereas the cholesterol metabolism pathway involves numerous CGs. This selective enrichment suggests that the effect of the identified host genes on HCV entry is not mediated by direct cholesterol uptake but rather through indirect modulation—the regulation of the expression and activity of LDLR and SR-BI receptors, which directly mediate this process. The consistency of the identified patterns of functional enrichment allows us to propose a two-level model of host gene involvement in the HCV infection cycle. At the first, structural-functional level, the BGs (LDLR and SCARB1) directly interact with viral particles, facilitating their initial binding and internalization. At the second, regulatory level, CGs (including PCSK9, APOE, and LDLRAP1) influence the efficiency of this process indirectly by controlling the expression, stability, and functional activity of the receptors themselves, as well as the concentration of their endogenous ligands in the systemic circulation.

Special attention should be given to the identification of CGs involved in processes critical to viral infection: entry into the host organism; the biological process involved in host interaction; the viral life cycle; the viral process; receptor-mediated endocytosis; and endocytosis. The presence of the APOE gene in both processes directly related to the virus and in endocytosis pathways underscores the link between HCV entry mechanisms and cellular lipoprotein metabolism. The consistent appearance of the CLDN6/9, F11R, IFITM1, and ITGB1 gene cluster in categories representing the early stages of infection indicates their potential role as modulators of the infectious process. Concurrently, a separate cluster of genes (LDLRAP1, PCSK9, LRPAP1, CD9, C3) was identified, specifically associated with the receptor-mediated uptake mechanism. The last two genes in this group are also involved in the body’s immune response.

Since individual variations in the immune response are a key factor determining the outcome of HCV infection, a separate analysis was conducted to examine the involvement of the identified CGs in biological processes directly related to immune function. The results showed a significant enrichment of several CGs in fundamental immune processes. A strong association was identified between the C3 and CD19 genes and the general category “Regulation of Immune Processes.” This indicates a potential role for these molecules not only in their classical functions but also in the broader modulation of the immune response during HCV infection. These findings are particularly relevant given HCV’s ability to infect not only hepatocytes but also immune system cells, specifically peripheral blood mononuclear cells, including B-lymphocytes. Tetraspanin CD19 is a marker of B-lymphocytes and participates in B-cell differentiation, which may be important for the chronicity of infection and the production of antibodies against HCV [33]. Several CGs (APOA1, APOA2, APOE, C3, F11R, ITGB1) are associated with processes related to the inflammatory response and the immune response. The F11R (JAM-A) gene encodes an adhesion protein involved in leukocyte transmigration through the endothelium, which may influence the intensity of inflammatory infiltration in the liver during fibrogenesis [34]. This function of F11R underscores its importance in modulating inflammatory processes and maintaining vascular integrity in chronic liver diseases, including those associated with HCV infection. The IFITM1 protein (interferon-induced transmembrane protein 1), known for its direct antiviral activity, was identified in processes associated with the cytokine response and immune defense, underscoring its role in interferon-mediated protection against HCV [35]. Thus, the analysis identified a group of genes among the selected CGs whose products may influence HCV pathogenesis not only through viral entry but also through the modulation of key components of both innate (complement, interferon response, inflammation) and adaptive (B-cell response) immunity. This broadens the potential functional context of the identified CGs and justifies the need to study them from an immunological perspective.

Of particular interest are the CGs that were assigned high ranks in the study but for which direct evidence of their involvement in HCV infection is limited or absent. The most indicative in this regard are the genes that ranked highest in the final composite score ranking (CLDN3/5/12, ITGB1, F11R, TJP1/2). Their high scores are due to strong associations with HCV across several criteria: co-expression, shared subcellular localization, and involvement in common molecular pathways and biological processes. It is known that CLDN6/9 can function as alternative HCV entry co-receptors [36], and many claudins (CLDN1/3/4/5/6/7/9/10/11/14/17) are abnormally expressed in hepatocellular carcinoma [37], which is a common outcome of chronic HCV infection. Recent studies have shown that F11R plays a key role in regulating leukocyte infiltration and fibrogenesis in the liver [28]. This suggests that dysregulation of these proteins may be associated not only with oncogenesis but also with viral persistence and chronic liver damage. The ITGB1 gene directly interacts with CD81 and is considered a cofactor for HCV entry [38]. Its high ranking in our analysis confirms this potentially key role and points to the need for further functional studies.

The results of the analysis clearly point to two central axes of functional enrichment of the CGs. The first axis involves lipid and lipoprotein metabolism, which is linked to a distinctive feature of the HCV life cycle — the formation of lipoparticulates. The APOE, APOB, PCSK9, LDLRAP1, and MYLIP genes are key regulators of this process. The second axis comprises genes involved in the dynamics of intercellular contacts and tissue remodeling. This axis reflects the impact of HCV on the integrity of the hepatocellular barrier. The virus utilizes tight junction proteins (CLDN, OCLN) for entry, which inevitably disrupts their function. The MMP2, ADAM10, ITGB1 genes, as well as numerous claudins and cytoskeletal proteins (TJP1/2), are active participants in the remodeling of intercellular junctions and the extracellular matrix, which underlies fibrogenesis in chronic hepatitis C [39, 40].

To assess the likelihood of a functional link between CGs and BGs, an integrated approach was employed, based on a comparison of tissue expression profiles, subcellular localization of protein products, and involvement in common molecular pathways and biological processes. This approach allowed us to account for conditions critical to interaction, such as the necessary spatial proximity and sufficient concentration of gene products in the relevant cellular compartments, as well as the similarity of their functional roles in the context of infection. The developed scoring system allowed us to identify CGs with the highest hypothetical probability of functional interaction with the HCV receptor complex.

The main limitation of this study is the indirect nature of the established link between the identified CGs and the course of HCV infection, based on bioinformatics prediction and correlation analysis. This study is of a hypothetical-predictive nature and serves as a basis for subsequent experimental validation.

The following steps are the highest priority for further investigation:

• In-depth bioinformatics analysis aimed at assessing the functional impact of non-synonymous polymorphic variants in identified genes using molecular docking methods and protein tertiary structure prediction (AlphaFold);
• experimental verification of the identified genetic associations;
• investigation of the contribution of the identified genetic variants to immunopathogenesis mechanisms;
• in vitro functional validation to confirm their biological role.

Thus, the results obtained support the need for further functional studies to experimentally verify the role of the identified CGs in the pathogenesis of HCV infection. An in-depth study of their molecular mechanisms of action will help elucidate the complex interactions within the virus–host system and may open new avenues for the development of targeted therapeutic and preventive strategies.

Conclusion

As a result of a comprehensive bioinformatics analysis, an integrated ranking system was developed and applied, enabling the assessment of the functional proximity of the CGs to key HCV entry receptors (CD81, CLDN1, OCLN, SCARB1, LDLR). This resulted in a priority list of 25 genes with the highest scores, among which CLDN12, TJP1, ITGB1, and F11R are prioritized, having demonstrated the strongest associations across all criteria considered.

The data suggest that the efficiency of HCV entry into the host cell appears to be determined not only by the presence of specific receptors but also by the state of entire functional modules of the cell: the integrity and dynamics of intercellular contacts (tight junction proteins, including claudins) and the homeostasis of lipid metabolism (apolipoproteins, receptor regulators). Furthermore, a comprehensive analysis allowed us to compile a list of CGs associated with key immunological processes. This provides a basis for a targeted study of the immunomodulatory functions of these genes and their role in the pathogenesis of HCV infection.

Thus, this study shifts the focus from identifying individual susceptibility factors to a systemic understanding of the cellular context required for viral infection. The identified CGs represent a well-founded list of targets for in-depth bioinformatics analysis of natural genetic variations at these loci, with an assessment of their potential structural and functional effects on target proteins. The resulting predictions, in turn, will serve as the basis for targeted association searches in clinical cohorts and for planning in vitro experiments.

About the authors

Ekaterina V. Anufrieva

St. Petersburg Pasteur Institute

Email: kate.an21@yandex.ru
ORCID iD: 0009-0002-1882-529X

junior researcher, Laboratory of virology and immunology of HIV infection

Russian Federation, St. Petersburg

Yulia V. Ostankova

St. Petersburg Pasteur Institute

Author for correspondence.
Email: shenna1@yandex.ru
ORCID iD: 0000-0003-2270-8897

Cand. Sci. (Biol.), Head, Laboratory of virology and immunology of HIV infection, senior researcher, Laboratory of molecular immunology

Russian Federation, St. Petersburg

Vladimir S. Davydenko

St. Petersburg Pasteur Institute

Email: vladimir_david@mail.ru
ORCID iD: 0000-0003-0078-9681

junior researcher, Laboratory of virology and immunology of HIV infection

Russian Federation, St. Petersburg

Alexander N. Shchemelev

St. Petersburg Pasteur Institute

Email: tvildorm@gmail.com
ORCID iD: 0000-0002-3139-3674

Cand. Sci. (Biol.), junior researcher, Laboratory of immunology and virology of HIV

Russian Federation, St. Petersburg

Areg A. Totolian

St. Petersburg Pasteur Institute; First St. Petersburg State I. Pavlov Medical University

Email: totolian@pasteurorg.ru
ORCID iD: 0000-0003-4571-8799

Dr. Sci. (Med.), Professor, RAS Full Member, Head, Laboratory of molecular immunology, Director, Head, Department of immunology

Russian Federation, St. Petersburg; St. Petersburg

References

Supplementary files