Identification of key genes and pathways associated with different immune statuses of hepatitis B virus infection

Abstract We aimed to identify key genes and pathways associated with different immune statuses of hepatitis B virus (HBV) infection. The gene expression and DNA methylation profiles were analysed in different immune statuses of HBV infection. Differentially expressed genes (DEGs) and differentially methylated genes (DMGs) were identified, followed by their functional and integrative analyses. The differential expression of IgG Fc receptors (FcγRs) in chronic HBV‐infected patients and immune cells during different stages of HBV infection was investigated. Toll‐like receptor (TLR) signalling pathway (including TLR6) and leucocyte transendothelial migration pathway (including integrin subunit beta 1) were enriched during acute infection. Key DEGs, such as FcγR Ib and FcγR Ia, and interferon‐alpha inducible protein 27 showed correlation with alanine aminotransferase levels, and they were differentially expressed between acute and immune‐tolerant phases and between immune‐tolerant and immune‐clearance phases. The integrative analysis of DNA methylation profile showed that lowly methylated and highly expressed genes, including cytotoxic T lymphocyte‐associated protein 4 and mitogen‐activated protein kinase 3 were enriched in T cell receptor signalling pathway during acute infection. Highly methylated and lowly expressed genes, such as Ras association domain family member 1 and cyclin‐dependent kinase inhibitor 2A were identified in chronic infection. Furthermore, differentially expressed FcγR Ia, FcγR IIa and FcγR IIb, CD3−CD56+CD16+ natural killer cells and CD14highCD16+ monocytes were identified between immune‐tolerant and immune‐clearance phases by experimental validation. The above genes and pathways may be used to distinguish different immune statuses of HBV infection.


| INTRODUC TI ON
Hepatitis B virus (HBV), an common human pathogen, spreads through the mucosal and percutaneous exposure to infected blood and other body fluids. 1 Acute and chronic infection, cirrhosis and hepatocellular carcinoma (HCC) are common sequelae of HBV infection that represents a major health problem worldwide. [2][3][4] An approximate estimate of 600 000 HBV-related deaths are annually reported, 5 and 73% deaths related to liver cancer are attributed to hepatitis virus infection. 6 Moreover, HCC is known as one of the most common causes of cancer-related death worldwide. 7,8 Therefore, a better understanding of molecular mechanisms underlying the progression of HBV infections may facilitate researchers to design specific biomarkers and effective therapeutic strategies for HBV-related liver diseases.
In general, the co-ordinate action of both innate and adaptive immune responses is believed to be involved in the sustained control of HBV infection. 9,10 Innate immunity is the early defensive line for viral containment and may efficiently induce virus-specific adaptive responses through the production of pro-inflammatory cytokines and chemokines. 11 Several innate effectors are found to exhibit adaptive-like features or exert defensive effects against HBV via immunoregulation of T cells. 12 Moreover, the frequency of natural killer (NK) cells is found to increase during the early stages of HBV infection and decreases with the decrease in viraemia, 13 indicative of the key roles of innate immune responses following HBV exposure. Adaptive immunity necessitates time to activate the functional maturation and expansion of distinct B and T cell clones, which can specifically recognize the infectious agent and generate a memory response to enhance the host ability to control the infection caused by the same pathogen. 14 HBV infection outcome is ultimately determined by the presence of functional HBV-specific T cells and antibody-producing B cells. 15 Despite great efforts, the immunopathogenesis of HBV infection is largely unknown.
Accumulating studies have identified key molecules that play a crucial role in regulating the immune response against HBV infections. HBV X protein has been shown to disrupt innate immunity through the down-regulation of mitochondrial antiviral signalling protein. 16 Toll-like receptors (TLRs) may activate intracellular antiviral pathways and promote the production of antiviral effectors such as interferons (IFNs) to further activate immune responses for controlling HBV infection. 17 In addition, aberrant cytosine-guanine dinucleotide (CpG) DNA methylation is shown to be correlated with the progression of liver diseases caused by chronic HBV infection. 18 A study has confirmed that HBV infection progression is associated with the methylation rate of exportin 4 (XPO4) promoter and that XPO4 methylation status may be used as a promising biomarker to predict HBV infection progression. 19 However, the molecular mechanisms that activate the immune responses to prevent HBV infections are questionable.
To identify key genes and pathways associated with the different immune statuses of HBV infection, integrative analysis of gene expression and DNA methylation profiles of HBV infection were performed in this study. Moreover, the differential expression of IgG Fc receptors (FcγRs) in chronic HBV-infected patients and immune cells at different stages of chronic HBV infection was investigated.
The findings of this study will aid in the development of a promising biomarker to predict HBV infection progression.

| Workflow of this study
This study mainly includes three parts ( Figure 1A  negative) (CH2) phases. In accordance with the diagnostic criteria as previously described, 22 the characteristic AH, CA1, CH1, CH2 and CA2 were listed in Table 1a. Besides, the inclusion criteria for enrolling patients and healthy individuals were also described in detail in Table 1a.

Isolation of peripheral blood mononuclear cells (PBMCs)
Whole blood (15 mL) was collected from each patient through peripheral venepuncture and then maintained in the sterile anticoagulant tubes containing sodium citrate (Becton, Dickinson and Company, Franklin Lakes, NJ). After centrifugation at 1200 g/min for 5 minutes, PBMCs were isolated by Ficoll-Hypaque density gradient separation, counted and saved in RNALater (Qiagen Inc., Valencia, CA) at −80°C.
Gene expression microarray preparation, pre-processing and functional analysis Using TRIzol (Invitrogen, Corporation, Carlsbad, CA) following the protocols provided by manufacturers, Total RNA was extracted from PBMCs. RNA integrity was then detected performed with Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara CA).
Equal amount of total RNA (5 µg) was then prepared as microarray targets performed with Ambion RNA amplification kit (Ambion, Austin, TX, USA) for its reverse transcription into cDNA and a single round of in vitro transcription into cRNA. Human WG-6 v3.0 whole-genome expression bead array was performed and scanned with Illumina's BeadLab system Array Scanner-iScan based on Illumina Human whole-genome gene expression array v3 BeadChip.
All the raw expression data were pre-processed, including data signal conversion, quality evaluation, probe filtration and normalization.  Table S1. We performed t test 23   DEGs corresponding to a special GO function or KEGG pathway, respectively. Significant P-value was adjusted by BH method, and the adjusted P-value < .05 was considered as the cut-off value.

Analysis of the correlation between clinical indicators and gene expression levels
To identify key genes associated with HBV infection, Spearman's test was performed for the determination of the correlation between gene expression levels and clinical indicators, including ALT levels and HBV DNA. Significant P-value was adjusted by BH method. The adjusted P-value < .05 and |spearman coefficient| > .45 were defined as the cut-off value.

DNA methylation microarray preparation, pre-processing and functional analysis
Genomic DNA was isolated from the peripheral blood by the overnight digestion of lysed peripheral lymphocytes with proteinase K and phenol/chloroform extraction. The concentration and quality of genomic DNA were determined with NanoDrop 2000 (Thermo Fisher Scientific) and agarose gel electrophoresis (1% wt/vol), respectively.
The sample preparation for DNA methylation microarray was carried out as previously described. 27 To produce 200 to 1000 bp fragments (keeping CpG islands intact), 2 to 6 μg of high-quality genomic DNA was digested with MseI (New England Biolabs, Ipswich, MA) and then purified using PCR purification kit (Qiagen). The sample was heat-denatured to form single strands; the methylated DNA fragments were immunoprecipitated (IP) with 1 μg of monoclonal mouse antibody against 5-methylcytidine (Eurogentec, Belgium) and subsequently captured with Protein A agarose beads. The DNA-antibody-bead mixture was digested with proteinase K and purified with phenol-chloroform.
The enrichment of the methylated IP DNA was completed performed with a whole-genome amplification kit (WGA2 kit, Sigma, USA). The labelling of IP and input DNA, microarray hybridization and scanning were conducted by NimbleGen Laboratories (Reykjavik, Iceland) as previous described. 28  All the methylation microarray data were generated based on Roche NimbleGen Human DNA Methylation 3x720K CpG Island Plus RefSeq Promoter Array and pre-processed by NimbleScan V2.6 software (Roche) following NimbleScan software user's guide V2.6, including data signal conversion, quality evaluation, probe filtration and normalization. Differentially methylated genes (DMGs) between two groups were identified with t test. 23 Significant P-value was adjusted as FDR by BH method (FDR = 1.11%). A value of P < .0001 and the difference in means between the two groups > 0.4 were set as cut-off values. GO and KEGG pathway enrichment analysis for DMGs were performed with the cumulative hypergeometric test as described above, and the adjusted P-value < .05 (by BH method) was considered as the threshold.  Table 1b in details. The characteristics of patients and healthy controls were represented in Table S2. All individuals provided informed consent for using their samples for research.

Isolation of PBMCs and qRT-PCR
Whole blood collection, PBMC isolation and total RNA extraction were performed as described in Section (a). Reverse

Analysis of the correlation between FcγRs expression and clinical indicators
To further investigate whether FcγRs were key regulators in different statuses of HBV infection, we analysed the correlation between the expression levels of FcγRs and clinical indicators (ALT and AST) by Spearman's test with the same cut-off value as described in Section (a).  Table   S3. All individuals provided informed consent for using their samples for research.

Cell cytokine detection
Cytometric bead array (CBA) was used to detect cytokine levels in the plasma of subjects. Briefly, the levels of interleukin (IL)-1β, IL-6, IL-10, IL-12p70, macrophage inflammatory protein-1 beta (MIP-1β) and tumour necrosis factor (TNF), in plasma of patients, were measured using CBA kit (BD Biosciences) as previously described. 29 During flow cytometry, data were acquired and analysed using Becton Dickinson CBA software.

Statistical analysis
The obtained data were expressed as mean ± SD, and the difference in FcγRs expression between any two groups was estimated by Mann-Whiney U test within SPSS 18.0 software (SPSS, USA). Statistically significant results were obtained when P < .0083.

| Identification and functional enrichment analysis of DEGs
As shown in Figure 1B The functional enrichment analysis was also performed for DEGs from CA1 and CH1 phases of chronic hepatitis B. We found that 309 DEGs significantly enriched between CA1 and CH1 exhibited important functions related to inflammatory process. Of these, expression levels of IFI27 and FcγRs (including FcγR Ib, FcγR Ia and FcγR IIIa) were markedly higher in CH1 patients as compared with CA1 patients.

| Analysis of the correlation between clinical indicators and gene expression levels
With the cut-off value of adjusted P-value < .05 and |spearman co-efficient| > 0.45, a significant correlation existed between ALT and gene expression levels ( Figure 3A), but no genes were markedly cor-    Figure S1).

| Integrative analysis of gene expression and DNA methylation profiles
In addition, 38 lowly methylated and highly expressed genes and 12 highly methylated and lowly expressed genes were identified in all C vs N group in combination with gene expression profile data.
These highly methylated and lowly expressed genes, such as Ras association domain family member 1 (RASSF1A) and cyclin-dependent

| The expression of FcγRs in different statuses of chronic HBV-infected patients
We performed qRT-PCR analysis to detect FcγRs expression in chronic HBV infection. As shown in Figure 4A, the expression levels of FcγR Ia, FcγR IIa, FcγR IIIa and FcγR IIIb in CH patients were increased as compared to CA patients and healthy control (all P < .05). Moreover, FcγR Ia and FcγR IIb expressions in CA group were markedly increased as compared with that of HBISC group, while FcγR Ib and FcγR IIIb expression was decreased in CA group (all P < .05). FcγR IIb expression was higher in CA group relative to that of CH group (P < .01).

| Analysis of the correlation between clinical indicators and FcγR expression levels
The correlation between FcγR expression and clinical indicators, including ALT and AST, was analysed. As shown in Figure 4B

| Differential expression of FcγRs on immune cells in HBV-infected patients
The expression of FcγR III (CD16) on NK cells and their subsets in patients with different HBV infection was detected. As shown in Figure 5A and Table S4A, total NK and CD3 − CD56 + CD16 + NK cells in CH patients significantly decreased as compared with those in CA patients and healthy controls (all P ≤ .001). In addition, total NK and CD3 − CD56 + CD16 + NK cells in HBISC patients decreased in HBISC patients as compared with those in healthy controls (P = .002).
The expression of FcγRs on monocyte cells was also detected.
As indicated in Figure 1E and Table S4C, CD14 high CD16 + cells in CH patients were increased in comparison with those in CA patients and healthy controls (P < .001). Moreover, CD14 high CD16 + cells were markedly increased in HBISC patients as compared with those in healthy controls (P = .002). No significant differences were observed in CD14 + CD32 + and CD14 + CD64 + cells between any two groups.

| Analysis of cytokine levels in plasma of HBVinfected patients
Cytokine levels in plasma of HBV-infected patients were analysed.
We found that levels of IL-6, IL-1β, IL-10, TNF, MIP-1β and IL-12p70 were markedly increased in CH patients relative to those in CA patients and healthy controls (all P ≤ .001, Figure 7). Moreover, IL-10 and IL-12p70 levels in HBISC patients were obviously higher than those in healthy controls (P = .002 and .007, respectively) ( Figure 7).

| D ISCUSS I ON
The present study investigated key genes and pathways associated with different immune statuses of HBV infection. The results showed that TLR signalling pathway and leucocyte transendothelial migration pathway (including ITGB1) were enriched during acute HBV infection. Key DEGs such as FcγR Ia, FcγR Ib and IFI27 showed differential expression between AH and CA1 phases and between CA1 and CH1 phases. Furthermore, the differential expression of  30 Besides, there were increasing evidence showed that HBV inhibits natural immune response by regulating Toll-like receptors on the surface of immune cells. [31][32][33] In our study, the key different pathway between AH and all C groups was Toll-like receptor pathway also known as TLR signalling pathway. The innate immune response mediated by TLR signalling pathway is found to play a role in the control of HBV infection. 17 HBeAg induces the pathogenesis of HBV infection through targeting TLR-mediated signalling pathways to evade innate immune responses. 34 Moreover, higher level of HBsAg may attenuate TLR-mediated immune responses to evade innate and adaptive immune responses and maintain persistent HBV infection. 35 In our study, TLR signalling pathway was significantly enriched during acute infection of HBV. Thus, innate immune responses mediated by TLR signalling pathway may regulate the acute phase of HBV infection. In addition, leucocyte transendothelial migration is a key process to evoke the innate or adaptive immune response. 36 ITGB1 haplotype is shown to be correlated with the clearance of HBV infection. 37 Our results showed that leucocyte transendothelial migration pathway was enriched during AH phase of HBV infection, in which ITGB1 was included. Therefore, we speculate that leucocyte transendothelial migration pathway may be involved in HBV clearance during acute infection via ITGB1 regulation.
HBV-specific T cell dysfunction and decreased T cell numbers contribute to the chronic of HBV infection. 10  profile. It is reported that defective antigen presentation may induce T cell tolerance by combining CTLA4 and PD 1. 40 DNA immunization with the fusion of CTLA-4 to HBV core protein has the ability to promote Th2-type responses for HBV elimination. 41 In addition, MAPK3 is important for the induction of T cell energy and has the capacity to activate T cell responses via dendritic cells in autoimmunity. 42 Based on our results, we suggest that the epigenetic regulation of CTLA4 and MAPK3 by DNA methylation may be involved in the acute phase of HBV infection via regulation of T cell receptor signalling pathway.
In addition, highly methylated and lowly expressed genes (including RASSF1A and CDKN2A) were identified in chronic HBV infection. It has been confirmed that the reduced expression and hypermethylation of RASSF1A were frequently observed in HCC and played crucial roles in HCC tumorigenesis and metastases. 43 The combination of serum RASSF1A methylation and α-fetoprotein level is suggested as a promising biomarker to discriminate HCC patients with chronic HBV infection. 44  pathways. The T cell receptor (TCR) -CD3 complex formed by CD3E and TCR plays a key role in antigen recognition, linking downstream signalling pathways and T cell maturation. 52 Besides, the mutation or non-expression of CD3E is closely related to immunodeficiency. 53 LCK belongs to the Src family of tyrosine protein kinases, which are involved in T cell maturation and proliferation, and mediate the key downstream molecules of MAPK and NFKB through TCR-CD3 complex. 54 Clinically, elevated ALT is used as a marker of immune activation and as a basis for immune staging and treatment. 56,57 Our results suggested that genes associated with ALT levels were mainly enriched in immune responses, antigen processing and presentation, MHC class I receptor activity a and proteasome activities.
Functional enrichment and Spearman's correlation analysis showed that the expressions of HLA-F, IFI27 and PSME2 were associated with ALT. Among them, HLA-F acts as a surface marker for activated lymphocytes, whereas PSME2 (also known as PA28β) is associated with maturation of dendritic cells and MHC class I antigen presentation. 58,59 Immune cells recognize viruses and secrete IFN through pattern recognition receptors (PRRs), which induce the production of a series of Interferon-stimulated genes (ISGs). 60 The primary role of ISGs is to amplify the IFN signalling pathway, induce the production of cytokines capable of activating adaptive immune responses and directly inhibit the virus. 61 Our results showed that there were differential expressions of ISGs between AH and all C groups, as well as between different stages of chronic infection, such as IFI27, IFI30, IFI35, IFIT3 and IFITM3. Among of them, the FC of IFI27 was the highest with FC=10.44 and P = .0014225. In the early stages of viral infection, IFITM3 inhibits virus entry into cells, whereas IFI27 is produced in the late stages of viral infection, mediating apoptosis by disrupting mitochondrial membrane stability. 62,63 IFI30 is induced by type II IFN and mainly assists in the antigen presentation process restricted by MHC class II antigens. 64 We identified the DEGs and DGMs in HBV infection and predicted the possible roles by enrichment analysis. One limitation in this study was that we did not validate the roles of DEGs and DMGs on the HBV infection. However, in the next study, we will collect peripheral blood PBMC from patients with different HBV infection status, and the expression levels of main DEGs will be detected by qRT-PCR and Western blot. In addition, the mechanisms related to DEGs and DGMS in different HBV infection states will be further studied.
In conclusion, our study reveals key genes and pathways that may be used to distinguish between different immune statuses of HBV infection. The innate immune response mediated via TLR signalling pathway may regulate the acute phase of HBV infection.
Leucocyte transendothelial migration pathway may be involved in HBV clearance during acute infection via ITGB1 regulation. In addition, FcγR Ia, FcγR IIa and FcγR IIb may be used to distinguish different immune statuses of HBV infection and serve as potential targets in immunotherapy of HBV-related liver diseases. The epigenetic regulation of CTLA4 and MAPK3 by DNA methylation may be involved in the acute phase of HBV infection through regulating T cell receptor signalling pathway, while the methylation of RASSF1A and CDKN2A may exhibit an important role in chronic HBV infection.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.