Differential expression profile of long noncoding RNAs in chronic HBV infection: New insights into pathogenesis

1 INTRODUCTION

Chronic hepatitis B virus (HBV) infection remains a serious global public health problem with significant morbidity and mortality.¹ Globally, an estimated 257 million people have chronic HBV infection, accounting for nearly one million deaths annually.² The natural history of chronic HBV infection varies widely and is affected by a complex interplay between HBV replication and host immune response. With the progression of chronic HBV infection, it could evolve toward serious liver diseases, such as cirrhosis, liver failure, and hepatocellular carcinoma (HCC).³ Currently, antiviral treatment for HBV infection includes PEGylated interferon (IFN)-α and nucleoside/nucleotide analogs. However, largely due to the persistence of the covalently closed viral circular DNA and the failure of a sufficiently robust, functional, and sustained immune response, these treatments are less than satisfactory.⁴ Therefore, a better understanding of the pathogenesis of chronic HBV infection is urgently needed to improve the success of future therapies.

Long noncoding RNAs (lncRNAs) are RNA transcripts longer than 200 nucleotides, without protein translation capacity.⁵ Initially, they were considered as transcriptional noise without biological function. Recently, increasing studies revealed that lncRNAs, functioning as regulatory factors, might regulate gene expression by interacting with RNA, DNA, and protein at various levels, such as chromatin modification, transcription, and posttranscriptional processing, and played critical roles in various physiological and pathological processes, and associated with many diseases, including viral infectious diseases.^{6, 7} Recent studies suggested that lncRNAs might demonstrate diverse effects on the interactions between virus and host factors, and consequently determined the clinical outcome of infectious diseases.⁶ For instance, HULC could activate HBV replication by HBx/STAT3/miR-539/APOBEC3B signaling, leading to the growth of liver cancer.⁸ lncRNA NRON, highly expressed in resting CD4⁺ T lymphocytes, could be involved in HIV-1 latency by specifically inducing Tat protein degradation.⁹ Additionally, lncRNA#32 could regulate the level of IFN-stimulated genes under both unstimulated and type I IFN-stimulated conditions through interactions with hnRNPU and ATF2, and therefore plays an important role in host antiviral responses.¹⁰ However, the lncRNA profile and their biological functions in chronic HBV infection remain largely unknown.

Therefore, in this study, lncRNA microarray was conducted to comprehensively identify differentially expressed lncRNAs and messenger RNAs (mRNAs) in chronic HBV infection. Subsequently, comprehensive bioinformatics analyses were performed to investigate the potential functions of these genes. Finally, coexpression networks were further constructed to clarify the interaction between lncRNAs and mRNAs. These findings will provide a better understanding of the potential roles of lncRNAs in the pathogenesis and progression of chronic HBV infection.

2 MATERIALS AND METHODS

3 RESULTS

3.1 Overview of differentially expressed lncRNAs and mRNAs

To investigate the differentially expressed mRNA and lncRNA, three liver tissues from patients with chronic HBV infection and three age- and sex-matched normal liver tissues were used for microarray analyses. The scatter plots were created to assess lncRNA and mRNA expression variations between samples of two groups. As shown in Figure 1A,B, Pearson's correlation coefficient for the samples of two groups showed a high degree of correlation. Compared with controls, a total of 203 lncRNAs (170 were upregulated and 33 were downregulated) and 180 mRNAs (154 were upregulation and 26 were downregulation) were differentially expressed in chronic HBV infection (fold change [FC] ≥ 2; P < .05) (Figure 1C,D). The top 10 upregulated and 10 downregulated lncRNAs and mRNAs have been listed in Tables 1 and 2, respectively. The most significantly dysregulated lncRNAs were ENST00000553896 (DHRS2) (FC: up; 10.8856479) and ENST00000555383 (AL355102.1) (FC: down; 8.8937791), while the most markedly dysregulated mRNAs were ENST00000372930 (TMEM35A) (FC: up; 9.0263073) and ENST00000334287 (SLC51B) (FC: down; 11.6170243). Hierarchical clustering analysis was also performed to analyze the distinguishable lncRNA and mRNA expression pattern between these two groups and displayed as a heat map (Figure 1E,F).

Table 1. The top 10 upregulated and 10 downregulated lncRNAs in chronic HBV infection

Probe name	Seqname	Gene symbol	Chromosome	Start	End	P values	Fold change	Regulation	Relation
ASHG19LNC1A110657780V5	ENST00000553896	DHRS2	chr14	24099324	24111647	.022708533	10.886	Up	Exon sense
ASHG19LNC1A106830235V5	ENST00000560864	AC022898.1	chr15	61056939	61057824	.000237543	7.3905	Up	Intronic antisense
ASHGV40037340V5	T262735	G060477	chr4	27187414	27190488	.001309845	7.245	Up	Intergenic
ASHGV40004180V5	ENST00000512675	AC139491.1	chr5	175476681	175489024	.007679361	6.942	Up	Natural antisense
ASHGV40057253V5	ENST00000417805	AL160408.1	chr1	234782035	234796670	.005005722	6.1324	Up	Intergenic
ASHG19LNC1A100106365V5	ENST00000443030	AC092168.1	chr2	101338683	101339344	.002731133	6.0248	Up	Intergenic
ASHGV40034602V5	ENST00000484892	LINC00971	chr3	84687557	84918726	.004820202	5.9495	Up	Intergenic
ASHGV40032672V5	ENST00000425517	LINC01707	chr1	69521581	69650686	.00018729	5.4275	Up	Intergenic
ASHGV40034502V5	ENST00000485805	LINC00994	chr3	64064037	64073039	.013201746	5.3217	Up	Intergenic
ASHG19LNC1A110868358V5	ENST00000624327	AC004158.1	chr16	72463252	72567791	.000311768	4.8343	Up	Intergenic
ASHG19LNC1A100011776V5	ENST00000555383	AL355102.1	chr14	96677197	96681155	.006043097	8.8937791	Down	Intronic sense
ASHGV40025827V5	ENST00000598170	AC006262.1	chr19	46713499	46718094	.023666021	4.8221762	Down	Intergenic
ASHG19LNC1A100063611V5	ENST00000624272	C5orf66	chr5	134368970	134680375	.014622353	3.7770337	Down	Natural antisense
ASHGV40049489V5	ENST00000518143	ZFHX4-AS1	chr8	77523116	77595513	.014022802	3.4827895	Down	Natural antisense
ASHG19LNC1A107979790V5	ENST00000457901	AC018742.1	chr2	21444057	21486958	.011928929	3.2169485	Down	Intergenic
ASHG19LNC1A108440760V5	ENST00000642898	LINC01488	chr11	69300379	69309720	.005731354	3.1080119	Down	Intergenic
ASHG19LNC1ABL100000692V5	compmerge.2878.pooled.chrX	LINC01186	chrX	46322668	46327652	.037491139	3.0599907	Down	Natural antisense
ASHGV40044473V5	TCONS_00011631	XLOC_005119	chr6	230931	237179	.040471364	2.6646732	Down	Intergenic
ASHG19LNC1A107575874V5	ENST00000449405	LINC01939	chr2	897183	900983	.036997745	2.6634536	Down	Intergenic
ASHG19LNC1A103227854V5	ENST00000649049	AL390856.1	chr1	182096432	182097817	.038133246	2.6416402	Down	Intergenic

• Abbreviations: HBV, hepatitis B virus; lncRNA, long noncoding RNA.

Table 2. The top 10 upregulated and 10 downregulated mRNAs in chronic HBV infection

Probe name	Seqname	Gene symbol	Description	Chromosome	Start	End	P values	Fold change	Regulation
ASHG19AP1B123494791V5	ENST00000372930	TMEM35A	Transmembrane protein 35A	chrX	100333709	100351353	0.024003	9.026307	Up
ASHG19AP1B103900526V5	ENST00000344777	DHRS2	Dehydrogenase/reductase 2	chr14	24105582	24114848	0.03054	8.694347	Up
ASHG19AP1B105316397V5	ENST00000645002	SNRPN	Small nuclear ribonucleoprotein polypeptide N	chr15	25068872	25223724	0.036523	6.608694	Up
ASHG19AP1B100172403V5	ENST00000064571	CBLN4	Cerebellin 4 precursor	chr20	54572496	54580528	0.043373	6.272659	Up
ASHG19AP1B100083243V5	ENST00000303921	GPR37	G protein-coupled receptor 37	chr7	124386051	124405681	0.016513	6.267627	Up
ASHG19AP1B119068821V5	ENST00000360948	NTRK3	Neurotrophic receptor tyrosine kinase 3	chr15	88419988	88799962	0.000747	5.499896	Up
ASHG19AP1B124285250V5	ENST00000368845	OAT	Ornithine aminotransferase	chr10	126085876	126107505	0.021101	4.249839	Up
ASHG19AP1B113708091V5	ENST00000306318	GBX2	Gastrulation brain homeobox 2	chr2	237073879	237077012	0.032053	4.240156	Up
ASHG19AP1B103262351V5	ENST00000287139	NODAL	Nodal growth differentiation factor	chr10	72192071	72201423	0.004643	4.230555	Up
ASHG19AP1B100001215V5	ENST00000508487	CXCL2	C-X-C motif chemokine ligand 2	chr4	74962752	74965010	0.045589	4.085023	Up
ASHG19AP1B100141581V5	ENST00000334287	SLC51B	Solute carrier family 51 beta subunit	chr15	65337708	65345734	0.0208708	11.6170243	Down
ASHG19AP1B100062875V5	ENST00000294312	FGF19	Fibroblast growth factor 19	chr11	69513000	69519410	0.0137204	9.7112621	Down
ASHG19AP1B100199917V5	ENST00000326577	TNFRSF12A	TNF receptor superfamily member 12A	chr16	3070313	3072384	0.0347484	6.6562845	Down
ASHG19AP1B122125283V5	ENST00000302913	ACHE	Acetylcholinesterase (Cartwright blood group)	chr7	100487616	100493541	0.0152625	4.8140662	Down
ASHG19AP1B124038577V5	ENST00000402182	APOBEC3B	Apolipoprotein B mRNA editing enzyme catalytic subunit 3B	chr22	39378404	39388809	0.0069478	4.794671	Down
ASHG19AP1B100088306V5	ENST00000239461	PRRX1	Paired related homeobox 1	chr1	170633047	170708560	0.0109187	4.0384107	Down
ASHG19AP1B100060467V5	ENST00000252809	GDF15	Growth differentiation factor 15	chr19	18496968	18499986	0.022066	3.3410947	Down
ASHG19AP1B100085273V5	ENST00000560104	BLID	BH3-like motif containing, cell death inducer	chr11	121986062	121986923	0.036857	2.8834914	Down
ASHG19AP1B101559793V5	ENST00000261597	NDC80	NDC80, kinetochore complex component	chr18	2571510	2616634	0.0333044	2.5954002	Down
ASHG19AP1B100127172V5	ENST00000262713	AJUBA	Ajuba LIM protein	chr14	23440383	23451851	0.031267	2.5511673	Down

• Abbreviations: HBV, hepatitis B virus; mRNA, messenger RNA.

3.2 Characteristics of differentially expressed lncRNAs

According to the location of lncRNA in coding genes, 203 differentially expressed lncRNAs were further classified. These lncRNAs were mainly from intergenic regions or natural antisense to protein-coding loci (Figure 2B). Furthermore, most of lncRNAs were distributed between 200 and 2000 bp in length (Figure 2A). Chromosome distribution revealed that these lncRNAs were located on different chromosomes (Figure 2C). Additionally, 13 of these lncRNAs might possess enhancer-like function, namely, T262735, ENST00000417805, ENST00000443030, ENST00000484892, ENST00000425517, ENST00000485805, ENST00000624327, ENST00000598170, ENST00000457901, ENST00000642898, TCONS_00011631, ENST00000449405, and ENST00000649049.

3.3 Bioinformatics analysis of differentially expressed lncRNAs

The potential functions of differentially expressed lncRNAs were predicted by GSEA base on Pearson's correlation coefficient between the selected lncRNA and all mRNAs. To better understand the enrichment results, the gene sets of each lncRNA were clustered and displayed using a heat map. As indicated in Figure 3, the top 10 upregulated and 10 downregulated lncRNAs might be mainly involved in biological processes including fatty acid oxidation, mitochondrial translational elongation and termination, mitochondrial electron transport (NADH to ubiquinone), mitochondrial respiratory chain complexes I and IV assembly, ATP biosynthesis process, branched-chain amino acid catabolism, regulation of cellular amino acid metabolism, gluconeogenesis, ribosomal RNA processing, protein folding, protein targeting to peroxisome and mitochondrion, proteasomal protein catabolic process, endoplasmic reticulum to Golgi vesicle transport and lymphocyte and monocyte chemotaxis. Moreover, these lncRNAs were primarily related to biotransformation pathways: fatty acid metabolism, nonalcoholic fatty liver disease, amino acid metabolism (such as valine, leucine, isoleucine, glycine, serine, threonine, and tryptophan), carbon metabolism, drug metabolism, protein processing in endoplasmic reticulum, protein export, oxidative phosphorylation, ribosome, and proteasome.

3.4 GO, KEGG, and GSEA analysis of differentially expressed mRNAs

According to the results of GO analysis, these differentially expressed mRNAs were primarily enriched in a biological process linked to the organic/carboxylic/amino acid metabolic process, and in cellular components, principally including mitochondrion, as well as in molecular function, such as transmembrane transporter activity and transferase activity (Figure 4A). The KEGG pathway analysis revealed that these mRNAs were significantly enriched in cytokine-cytokine receptor interaction, fatty acid degradation, fatty acid elongation, biosynthesis of unsaturated fatty acids, valine, leucine, and isoleucine degradation, carbon metabolism, and peroxisome proliferator-activated receptor signaling pathway (Figure 4B).

Additionally, GSEA was also performed (Tables S3 and S4). Consistently, these mRNAs were significantly correlated with valine, leucine, and isoleucine degradation (normalized enrichment score [NES] = 1.60; P < .001) and fatty acid degradation (NES = 1.52; P < .001) (Figure 4C).

3.5 Construction of lncRNA-mRNA coexpression network

Based on the correlations of all differentially expressed lncRNAs and mRNAs, a coexpression network was constructed to identify hub regulatory factors associated with chronic HBV infection (Figure S1). According to the degree analysis, 8 of the top 10 upregulated and 10 downregulated lncRNAs with higher core degrees might be critical to the pathogenesis of chronic HBV infection. Therefore, a subgroup coexpression network of these eight lncRNAs and interacted mRNAs was further constructed and shown in Figure 5A. In addition, subgroup networks based on KEGG terms were also constructed. As presented, 36 lncRNAs interacted with 7 mRNAs in the KEGG term of cytokine-cytokine receptor interaction (Figure 5B) and 21 lncRNAs interacted with 3 mRNAs in the KEGG term of fatty acid degradation (Figure 5C).

3.6 Validation of differentially expressed lncRNAs

Five lncRNAs with more significant difference in expression and higher core degree were selected for quantitative real-time polymerase chain reaction (qRT-PCR) analysis to validate the microarray data. Consequently, the expression of ENST00000485805, T262735, and ENST00000560864 were upregulated (FC = 4.12, P = .02; FC = 3.01, P = .003; and FC = 1.50, P = .037, respectively) in chronic HBV infection compared with controls. ENST00000417805 and ENST00000443030 were also upregulated in chronic HBV infection; however, they did not reach statistical significance (FC = 1.41, P = .102 and FC = 1.61, P = .795, respectively) (Figure 6).

4 DISCUSSION

Recent studies have identified many lncRNAs associated with the regulation of immune responses during host-pathogen interactions,^15-17 and an understanding of their functional roles in the pathogenesis of infectious, inflammatory, and autoimmune diseases is just emerging.⁶ However, to date, the expression profile and functions of lncRNAs in chronic HBV infection have not been well demonstrated yet. To the best of our knowledge, this is the first study to systematically compare the expression profiles of lncRNAs and mRNAs in liver tissues between patients with chronic HBV infection and normal controls with microarray analysis. Consequently, a total of 203 lncRNAs, including 170 upregulated and 33 downregulated lncRNAs, were identified to be differentially expressed in chronic HBV infection. Simultaneously, 180 differentially expressed mRNAs were also identified, including 154 upregulated and 26 downregulated mRNAs. Then, five candidate lncRNAs were further chosen for qRT-PCR confirmation. Consequently, ENST00000560864, T262735, and ENST00000485805 showed the same expression pattern as those analyzed by microarray. These results, to a certain extent, indicated a low FDR and the good reliability of the microarray data.

Previously, Yilmaz Susluer et al¹⁸ applied qRT-PCR to analyze the expression of 135 lncRNAs in plasma from 82 patients with chronic HBV infection (classified as chronic patients, inactive carriers, or resolved patients) and 81 healthy volunteers. Compared with healthy controls, they identified 15 upregulated and 7 downregulated lncRNAs from initial diagnosis to the 12th month of treatment/follow-up in patients with chronic HBV infection. Ruan et al¹⁹ showed that lncRNA HEIH and lncRNA HULC were upregulated in peripheral blood and liver tissues in hepatitis B patients, particularly those with hepatitis B-related HCC. However, the differential expression of these lncRNAs was not found in our study. The possible reasons for this discrepancy might be mainly due to differences in expression profile being analyzed, specimen type, detection methods, and characteristics of study populations. Despite, these results suggest that chronic HBV infection is also accompanied by changes in the expression of some lncRNAs, and these lncRNAs may be implicated in the pathogenesis and progression of chronic HBV infection.

To explore the functions of the identified genes and their relationships in the development of chronic HBV infection, a comprehensive bioinformatics analysis was performed. According to GO and KEGG analysis, differentially expressed mRNAs were mainly involved in cytokine-cytokine receptor interaction, organic/carboxylic/amino acid metabolic process and biotransformation pathways, particularly fatty acid metabolism. Moreover, similar results were also found by GSEA. Therefore, we inferred that these differentially expressed mRNAs might play vital roles in the underlying metabolic pathogenesis of chronic HBV infection.

In contrast to miRNAs, lncRNAs can play biological functions via a variety of complex mechanisms.⁷ Until now, there is no uniform approach to predict the potential functions of lncRNAs. However, many research studies indicated that lncRNAs might mediate the expression of neighboring and distantly located genes via cis/trans regulatory effects, respectively, and the genomic context could, therefore, provide important information about the potential functions of lncRNAs.^{7, 20} Therefore, most researchers always indirectly predict the functions of lncRNAs by performing GO and pathway analysis of their neighboring and distantly target genes.^{21, 22} However, to investigate the integrated functions of lncRNAs, we performed GSEA base on Pearson's correlation coefficient between the selected lncRNA and all mRNAs in this study. Largely consistent with the results of enrichment analysis of the differentially expressed mRNAs, the top 10 upregulated and 10 downregulated lncRNAs were also primarily involved in the biotransformation process, such as fatty acid metabolism, amino acid metabolism, carbon metabolism, and drug metabolism. Collectively, all the above results suggested that these lncRNAs might participate in the progression of chronic HBV infection by interacting with the differentially expressed mRNAs in this study. Then, to further study this complicated regulatory relationship, we conducted coexpression network analysis. In the network, numerous connections between lncRNAs and mRNAs indicated the complexity of molecular mechanisms of CHB. Subsequently, eight hub lncRNAs with higher core degrees and fold change were identified accordingly. Besides, we also selected two important pathways and linked them to some differentially expressed lncRNAs through a coexpression network, which might provide new directions for future experimental research.

Based on bioinformatics analysis, two pathways might be probably associated with the progression of chronic HBV infection: cytokine-cytokine receptor interaction and fatty acid degradation. Cytokines have been shown to control HBV replication and mediate a noncytolytic clearance of the virus.²³ In this study, some differentially expressed mRNAs were involved in cytokine-cytokine receptor interaction, such as ACVR1C, CNTF, CXCL1, CXCL2, GDF15, IL-10, NODAL, and TNFRSF12A. Das et al²⁴ reported that in IL-10 was the primary factor in chronic HBV infection. As an anti-inflammatory Th2 cytokine, IL-10 could restrain the host's anti-HBV activity, leading to sustained replication and expression, and association with HBV serum titers and degree of liver inflammation, determining the development and progression of hepatitis B disease.^24-26 Additionally, inhibition of IL-10 has been proven to recover specific CD8⁺ T-cell response and accelerate virus eradication in in vivo and in vitro studies.^{24, 27} As shown in Figure 5B, IL-10 was positively and negatively correlated with five lncRNAs and two lncRNAs, respectively, indicating that these lncRNAs might participate in the molecular mechanism of CHB through mediating the functions of IL-10. Thus, future studies are warranted to elucidate the interactions between these lncRNAs and IL-10 in the development of chronic HBV infection.

Liver is a vital metabolic organ for the synthesis and circulation of lipids (including fatty acids), oxidation of fatty acids, and production of ketone bodies.²⁸ Recently, increasing research studies demonstrated that HBV, a “metabolovirus,” could interfere with hepatic metabolic processes, especially lipid metabolism.^29-31 For instance, HBx could induce the activation of fatty acid oxidation and subsequently maintain NADPH and ATP levels in the absence of glucose, which was crucial for HCC cells to overcome glucose starvation.³² Meanwhile, the life cycle of HBV, in turn, is affected by fatty acid metabolism. Okamura et al³³ revealed that fatty acid biosynthesis might be associated with the formation and egression of infectious HBV particles from cells. Huang et al³⁴ found that fatty acid translocase and cluster of differentiation 36 could promote HBV replication by upregulating the levels of hepatic cytosolic calcium. However, the precise mechanisms involved in altered lipid metabolism by chronic HBV infection remain largely unknown. In this study, 21 lncRNAs interacted with three mRNAs which were associated with fatty acid degradation, suggesting an underlying role of lncRNAs in lipid metabolism in the development of CHB. Therefore, a clearer understanding of this newly emerging role of lncRNAs in the metabolic pathogenesis of chronic HBV infection may provide new insights into the prevention and treatment for HBV-induced disease, which is worth further exploration.

In addition, lncRNAs might be also implicated in the metabolic pathogenesis of chronic HBV infection through amino acid metabolism, like valine, leucine, and isoleucine degradation. However, to date, few studies about an association between HBV infection and amino acid profile changes have been reported.³⁵ Therefore, this finding in the current study still needs further confirmation.

Finally, the following limitations should be noted. First, the pathogenesis of chronic HBV infection is complex and involves complicated interactions between virus and the host. Of which, intrahepatic innate and adaptive immune responses play crucial roles in the persistent virus replication and liver damage during chronic HBV infection.^{36, 37} Surprisingly, most differentially expressed genes in this study did not significantly associate with some important pathways of immune responses. One possible explanation for this might be the source of normal liver tissues. As liver tissues from healthy people were unavailable, normal liver tissues from HCC patients were selected as controls instead of identifying differentially expressed genes, which, to a certain extent, might bias our results. Second, due to the limited sources of liver tissue, the sample size was relatively small and the representativeness of these samples is still limited. Hence, a larger and more diverse cohort is needed to further validate our findings. Finally, the functions of lncRNAs and associated pathways in the development of chronic HBV infection were predicted mainly based on bioinformatics analysis, which still require further experimental validation in vivo or in vitro.

In summary, this study identified the differential expression patterns of lncRNAs and mRNAs in liver tissues from patients with chronic HBV infection. Bioinformatics analysis revealed that these lncRNAs might be involved in cytokine-cytokine receptor interaction and multiple biotransformation processes, especially fatty acid degradation. These findings will provide new insights into the pathogenesis and progression of chronic HBV infection, and the roles of these lncRNAs in chronic HBV infection should be further explored.

CONFLICT OF INTERESTS

The authors declare that there are no conflict of interests.

AUTHOR CONTRIBUTIONS

Conceptualization: QH, ZW, ZL, HQ; experiments and data curation: QH, ZW, QW, WX; formal analysis: QH, ZW, HC, HQ; software: HC; resources: ZL; writing and original draft: QH, ZW; writing, review, and editing: QW, WX, HC, ZL, HQ.