Differentially regulated gene expression associated with hepatitis C virus clearance

Abstract

Human chronic hepatitis C virus (HCV) infections pose a significant public health threat, necessitating the development of novel treatments and vaccines. HCV infections range from spontaneous resolution to end-stage liver disease. Approximately 10–30 % of HCV infections undergo spontaneous resolution independent of treatment by yet-to-be-defined mechanisms. These individuals test positive for anti-HCV antibodies in the absence of detectable viral serum RNA. To identify genes associated with HCV clearance, this study compared gene expression profiles between current drug users chronically infected with HCV and drug users who cleared their HCV infection. This analysis identified 91 differentially regulated (up- or downregulated by twofold or more) genes potentially associated with HCV clearance. The majority of genes identified were associated with immune function, with the remaining genes categorized either as cancer related or ‘other’. Identification of factors and pathways that may influence virus clearance will be essential to the development of novel treatment strategies.

Introduction

Hepatitis C virus (HCV) is a major cause of liver disease in the USA, with approximately 4 million people chronically infected (Alter, 1997). Diseases resulting from HCV infections can range from a subclinical anicteric disease to fulminant hepatitis, although the latter is more rare (Hoofnagle, 1997). Cirrhosis of the liver presents in 20–30 % of persons with chronic HCV, leading to hepatocellular carcinoma or end-stage liver disease that can be treated only by liver transplantation (Hoofnagle, 1997). In the USA, 85–90 % of HCV infections are associated with injection drug use and related behaviours, including having sex with an injection drug user, sharing contaminated needles and other high-risk behaviours such as multiple sex partners or sex with partners with a history of sexually transmitted diseases (Alter, 1997).

Spontaneous HCV clearance, defined as individuals having antibodies specific to HCV antigens but no detectable viral RNA, occurs in ~10–30 % of HCV-infected, untreated individuals (Shah et al., 2012; Thomas et al., 2000; Villano et al., 1999). Persons who clear their HCV infections are more likely to be white, have jaundice, have lower peak viral titres and not be infected with human immunodeficiency virus (HIV) (Shah et al., 2012; Thomas et al., 2000; Villano et al., 1999). The HCV viral genotype has not been shown to have an impact on clearance (Alric et al., 2000; Villano et al., 1999), but investigations into immunological components associated with clearance have been shown to differ between patients who spontaneously clear HCV and patients who do not (Alric et al., 2000; Balagopal et al., 2010; Mosbruger et al., 2010).

Various genes have been implicated in mediating host susceptibility/resistance to HCV infections and in affecting disease presentation. However, identification of these genes has been primarily in the context of various states of HCV infection and not based on factors associated with clearance (Balagopal et al., 2010; Honda et al., 2001; Iizuka et al., 2003; Mosbruger et al., 2010; Shackel et al., 2002). Data from these studies indicated that impaired innate/acquired immunity associated with cytokine responses influence an individual’s ability to clear HCV infections.

As no vaccine options exist for the prevention of HCV infections, and patient responses to treatment are highly variable, extremely expensive and have embedded adverse effects (Yu & Chiang, 2010), additional studies designed to identify novel genetic targets associated with HCV clearance in humans are needed, as the identification of genes differentially expressed or associated with protection/clearance of HCV infections have primarily been carried out in vitro and not in human populations naïve to HCV therapies.

This study identified differentially expressed genes between participants presenting with anti-HCV antibodies negative for HCV RNA (defined as HCV viral clearance) and individuals positive for HCV RNA (defined as HCV viral persistence). Identification of these potential target genes or gene products will lead to the discovery of new pathways that can be targeted for the development of potential novel treatment regimens and to a better understanding of the natural progression of HCV infections.

Results

Study population

Participants in this study were 63 % male and 81 % African–American, with a mean age of 45 years. Of the 16 participants, six (38 %) had a history of injection drug use, three (19 %) were men who had sex with men and eight (50 %) had a history of trading sex for money or drugs (Table 1).

Table 1. Demographic characteristics and risk factors of HCV antibody-positive participants.

IDU, Injection drug use; MSM, men who have sex with men.

Characteristic	Total (N = 16)	HCV RNA positive (n = 8)	HCV RNA negative (n = 8)	P value
Demographics
Age
<40	1 (6 %)	1 (12 %)	0 (0 %)
40–49	11 (69 %)	5 (63 %)	6 (75 %)
≥50	4 (25 %)	2 (25 %)	2 (25 %)
Gender
Male	10 (63 %)	5 (63 %)	5 (63 %)
Female	6 (37 %)	3 (37 %)	3 (37 %)
Race
African–American	13 (81 %)	6 (75 %)	7 (88 %)
Caucasian	2 (13 %)	2 (25 %)	0 (0 %)
Hispanic	1 (6 %)	0 (0 %)	1 (12 %)
Risk factor for HCV
IDU				0.608
No	10 (62 %)	4 (50 %)	6 (75 %)
Yes	6 (38 %)	4 (50 %)	2 (25 %)
Trading sex in past 30 days				0.010
No	8 (50 %)	1 (12 %)	7 (88 %)
Yes	8 (50 %)	7 (88 %)	1 (12 %)
MSM in past 30 days				0.200
No	13 (81 %)	5 (63 %)	8 (100 %)
Yes	3 (19 %)	3 (37 %)	0 (0 %)

HCV viral load detection

The eight HCV antibody-positive/RNA-positive patients all had detectable viral loads at the time that the array was carried out. Specifically, the range of HCV RNA was 382–17 960 000 IU ml⁻¹. The eight HCV antibody-positive/RNA-negative patients were all negative for detectable HCV RNA at a cut-off of 43 IU ml⁻¹.

Differential gene expression

A total of 471 genes (Table S1, available in JGV Online) were differentially expressed between HCV RNA-positive and RNA-negative participants that were significantly up- or downregulated at a nominal value of P<0.05. Of the 471 genes, 91 candidate genes had significantly different expression profiles (up- or downregulated by twofold or more) between the two groups (P<0.05, Table 2). One-third of the genes were downregulated and two-thirds were upregulated.

Table 2. Differentially expressed genes (n = 91) associated with HCV clearance.

Functional category	Gene or gene product symbol	Description	Fold change	P value
Cancer	DYRK2	Dual-specificity tyrosine-(Y)-phosphorylation regulated kinase 2, transcript variant 1	2.04	0.0056
	CDC14A	CDC14 cell division cycle 14 homologue A (Saccharomyces cerevisiae), transcript variant 3	2.05	0.0002
	HBD	Haemoglobin δ	2.12	0.0025
	GSTM1	Glutathione S-transferase M1, transcript variant 1	2.16	0.0355
	MGST1	Microsomal glutathione S-transferase 1, transcript variant 1b	2.19	0.0190
	BHLHB9	Basic helix–loop–helix domain containing, class B, 9	2.20	0.0500
	CTAGE5	CTAGE family, member 5, transcript variant 4	2.25	0.0036
	PRPH	Peripherin	2.27	0.0317
	SF1	Splicing factor 1, transcript variant 4	2.30	0.0449
	SMPD3	Sphingomyelin phosphodiesterase 3, neutral membrane (neutral sphingomyelinase II)	2.38	0.0210
	GSTM3	Glutathione S-transferase M3 (brain)	2.44	0.0008
	FOLR3	Folate receptor 3 (γ)	2.45	0.0461
	HTRA1	Htra serine peptidase 1	2.55	0.0441
	TFF3	Trefoil factor 3 (intestinal)	2.68	0.0348
	HRK	Harakiri, BCL2 interacting protein (contains only BH3 domain)	2.82	0.0375
	ALOX15	Homo sapiens arachidonate 15-lipoxygenase	2.96	0.0171
	OLIG2	Oligodendrocyte lineage transcription factor 2	2.96	0.0010
	AMACR	α-Methylacyl-coa racemase, transcript variant 1	0.50	0.0263
	PARP10	Poly (ADP-ribose) polymerase family, member 10	0.49	0.0012
	CXCR7	Chemokine (C-X-C motif) receptor 7, transcript variant 1	0.48	0.0307
	CXCR7	Chemokine (C-X-C motif) receptor 7	0.50	0.0004
	HOXB2	Homeobox B2	0.47	0.0279
	PARP9	Poly(ADP-ribose) polymerase family, member 9	0.46	0.0019
	OSM	Oncostatin M	0.44	0.0126
	RASSF5	Ras association (ralgds/AF-6) domain family member 5, transcript variant 3	0.32	0.0321
	PPP2R3B	Protein phosphatase 2 (formerly 2A), regulatory subunit B′′, β, transcript variant 2	0.25	0.0338
	GATA2	GATA binding protein 2	0.23	0.0085
Immunity	CLC	Charcot-Leyden crystal protein	2.12	0.0065
	IL5RA	Interleukin 5 receptor, α, transcript variant 6	2.13	0.0126
	EMR4	Egf-like module containing, mucin-like, hormone receptor-like 4	4.14	0.0002
	CCL23	Chemokine (C-C motif) ligand 23, transcript variant ckβ8-1	6.26	0.0440
	PLSCR1	Phospholipid scramblase 1	0.50	0.0456
	CD8B	CD8b molecule, transcript variant 5	0.50	0.0006
	TRIM5	Tripartite motif-containing 5, transcript variant α	0.49	0.0060
	TRIM22	Tripartite motif-containing 22	0.48	0.0062
	CD38	CD38 molecule	0.48	0.0437
	FCGR1B	Fc fragment of igg, high affinity Ib, receptor (CD64), transcript variant 2	0.48	0.0057
	IFI35	Interferon-induced protein 35	0.47	0.0059
	IFIT2	Interferon-induced protein with tetratricopeptide repeats 2	0.47	0.0276
	IRF7	Interferon regulatory factor 7, transcript variant b	0.43	0.0022
	IRF7	Interferon regulatory factor 7, transcript variant b	0.49	0.0004
	DLL1	δ-like 1 (Drosophila)	0.43	0.0491
	EVI2A	Ecotropic viral integration site 2A, transcript variant 1	0.40	0.0252
	MX1	Myxovirus (influenza virus) resistance 1, interferon-inducible protein p78 (mouse)	0.40	0.0086
	C1QB	Complement component 1, q subcomponent, B chain	0.35	0.0453
	OAS1	2′,5′-Oligoadenylate synthetase 1, 40/46 kDa, transcript variant 2	0.21	0.0010
	OAS1	2′,5′-Oligoadenylate synthetase 1, 40/46 kDa, transcript variant 3	0.45	0.0371
	TNFRSF18	Tumour necrosis factor receptor superfamily, member 18, transcript variant 3	0.31	0.0028
	IFIT3	Interferon-induced protein with tetratricopeptide repeats 3	0.24	0.0051
	IFIT3	Interferon-induced protein with tetratricopeptide repeats 3	0.34	0.0003
	IFIT3	Interferon-induced protein with tetratricopeptide repeats 3	0.36	0.0002
	IFI6	Interferon, α-inducible protein 6, transcript variant 2	0.29	0.0034
	MS4A2	Membrane-spanning 4-domains, subfamily A, member 2 (Fc fragment of IgE, high affinity I, receptor for; β polypeptide)	0.27	0.0046
	ISG15	ISG15 ubiquitin-like modifier	0.23	0.0280
	IFIT1	Interferon-induced protein with tetratricopeptide repeats 1, transcript variant 2	0.23	0.0388
	SERPING1	Serpin peptidase inhibitor, clade G (C1 inhibitor), member 1, transcript variant 2	0.19	0.0233
	TMEM176B	Transmembrane protein 176B	0.16	0.0012
	TMEM176A	Transmembrane protein 176A	0.10	0.0005
Other	SERINC5	Serine incorporator 5	2.04	0.0214
	PRSS33	Protease, serine, 33	2.04	0.0059
	ZNF493	Zinc-finger protein 493, transcript variant 3	2.04	0.0088
	GYPE	Glycophorin E, transcript variant 1	2.44	0.0002
	ZNF121	Zinc-finger protein 121	2.69	0.0159
	NEBL	Nebulette, transcript variant 1	2.73	0.0147
	GYPB	Glycophorin B (MNS blood group)	2.78	0.0001
	KCNMB4	Potassium large conductance calcium-activated channel, subfamily M, β member 4	2.97	0.0300
	NAV1	Neuron navigator 1	5.92	0.0402
	TMEM100	Transmembrane protein 100, transcript variant 2	7.76	0.0300
	GBP4	Guanylate binding protein 4	0.49	0.0079
	LFNG	LFNG O-fucosylpeptide 3-β-N-acetylglucosaminyltransferase, transcript variant 2	0.49	0.0322
	RGPD2	RANBP2-like and GRIP domain containing 2 (RGPD2) XM_001134112 XM_001134114 XM_001134116	0.45	0.0279
	RPS6KA2	Ribosomal protein S6 kinase, 90 kDa, polypeptide 2, transcript variant 2	0.45	0.0216
	ARHGAP12	Rho GTPase activating protein 12	0.44	0.0208
	DHRS9	Dehydrogenase/reductase (SDR family) member 9, transcript variant 2	0.44	0.0431
	GBP1	Guanylate binding protein 1, interferon-inducible, 67 kDa	0.42	0.0008
	HERC5	Hect domain and RLD 5	0.40	0.0107
	GCH1	GTP cyclohydrolase 1, transcript variant 3	0.39	0.0106
	LMTK2	Lemur tyrosine kinase 2	0.39	0.0340
	SLC34A1	Solute carrier family 34 (sodium phosphate), member 1	0.37	0.0260
	OASL	2′-5′-Oligoadenylate synthetase-like, transcript variant 1	0.30	0.0026
	OASL	2′-5′-Oligoadenylate synthetase-like, transcript variant 2	0.36	0.0102
	TBRG4	Transforming growth factor β regulator 4, transcript variant 1	0.33	0.0280
	HDC	Histidine decarboxylase	0.31	0.0459
	GPRC5C	G protein-coupled receptor, family C, group 5, member C, transcript variant 2	0.22	0.0369
Undefined	PHF20L1	PHD finger protein 20-like 1, transcript variant 2	2.01	0.0104
	RUNDC2A	RUN domain containing 2A	2.02	0.0148
	GTSF1	Gametocyte specific factor 1	2.07	0.0003
	MGC26718	Similar to ankyrin repeat domain 20A	2.32	0.0342
	KIAA0367	Kiaa0367	2.78	0.0074
	C17ORF97	Chromosome 17 open reading frame 97	3.42	0.0117
	MOBK1B	MOB1, Mps One Binder kinase activator-like 1B (yeast)	0.47	0.0327
	FAM26F	Family with sequence similarity 26, member F	0.43	0.0025
	DDX60	DEAD (Asp-Glu-Ala-Asp) box polypeptide 60	0.43	0.0295
	CPA3	Carboxypeptidase A3 (mast cell)	0.38	0.0478
	BATF2	Basic leucine zipper transcription factor, ATF-like 2	0.35	0.0218
	SPRYD5	SPRY domain containing 5	0.29	0.0062
	SAMD8	Sterile α motif domain containing 8	0.22	0.0462

We performed comprehensive bioinformatics analyses to characterize the 91 differentially expressed genes identified. These analyses consisted of three complementary components: (i) identification of overrepresented biological functions/pathways; (ii) identification of transcription factors affecting gene expression profiles; and (iii) gene network analysis to identify interactions between the 91 genes identified (e.g. physical protein–protein interactions), as well as identification of additional genes with the potential of interacting with some of the gene products expressed by the 91 genes identified that may also be predictive of HCV clearance. We first used the Database for Annotation, Visualization and Integrated Discovery (DAVID) method (Huang et al., 2008) to identify biological functional categories, for example, gene ontology (GO) (Ashburner et al., 2000) terms that are enriched with differentially expressed genes. After adjustment for multiple testing, two GO terms, GO:0006955 (immune response) and GO:0009615 (response to virus), were found to be significantly overrepresented among the differentially expressed genes (both P<1×10⁻⁵). Secondly, the TELiS method (Cole et al., 2005) was used to identify transcription factor binding motifs (TFBMs) overrepresented in the promoter regions of the differentially expressed genes identified (600 bp upstream of each gene’s transcription start site). TFBMs for nuclear factor κB (P = 0.00005), interferon-stimulated response element (P = 0.01) and Ras-responsive element binding protein 1 (P = 0.04) were found to be overrepresented in the promoters of those genes whose changes in expression were probably driven by these transcription factors. Finally, we performed gene network analysis using the GeneMANIA method (Mostafavi et al., 2008). GeneMANIA identifies known gene–gene interactions, for example, physical protein–protein interactions or co-expression profiles between the query gene set (i.e. the 91 differentially expressed genes) and predicts additional genes that may be involved in HCV clearance if they are shown to interact with a large number of genes in the query set (Fig. 1). The most noticeable feature was the subnetwork formed by the IFI gene family (IFI6, IFIT1, IFIT2, IFI35 and others) that were included in the query set and the identification of additional (predicted) genes such as STAT1, IRF9, XAF1, IFI27 and IFITM3 that may also be involved in HCV clearance.

Fig. 1.

GeneMANIA analysis of the 91 differentially expressed genes between HCV cleared and persistent infections. Lines represent co-expression, co-localization, physical or predicted interactions between genes as indicated in the legend. Differentially expressed genes identified by this study are represented by grey circles. Genes identified as having potential interactions with the identified genes (based on the GeneMANIA analysis) are represented by white circles.

Discussion

This study identified 471 genes that were expressed differentially among anti-HCV antibody-positive individuals that cleared their HCV infection and those that remained chronically infected. This approach was chosen as a means of identifying genes and/or gene pathways that may serve as new gene targets associated with HCV clearance. Ninety-one genes were significantly differentially up- or downregulated by at least twofold, and 15 genes were identified previously by others using in vitro infection models or by studies designed to identify genes associated with positive outcomes following treatment for HCV infections. Seventy-six genes had not been identified previously in the context of HCV clearance and represent novel gene targets potentially associated with HCV clearance or targets that can form the basis for the design of new therapeutic interventions or used to modify existing treatment modalities.

Because flow cytometric analyses designed to define the cell populations present in the buffy coats of the respective infection groups were not carried out at the time RNA was isolated, it is possible that some of the differentially expressed genes identified may have been a result of different cell populations present in the PBMCs of chronically infected individuals (compared with PBMC populations present in individuals who cleared their infections). Previous reports have described increased and decreased numbers of peripheral natural killer and CD8⁺ T-cells, respectively, in individuals chronically infected with HCV (Chan et al., 1999; Chang et al., 2001; Morishima et al., 2006; Pár et al., 2002). Others have identified increased frequencies of CD4⁺ and CD8⁺ regulatory T-cells (Tregs) among individuals chronically infected with HCV compared with uninfected controls (Hartling et al., 2012), and antibody repertoire differences between chronically infected individuals compared with individuals who cleared their HCV infections have also been reported (Racanelli et al., 2011). Data from the above-described studies suggested that the cells comprising the PBMC population in HCV chronically infected individuals and individuals who cleared their infections are probably different. However, it is unlikely that all of the 471 differentially expressed genes identified (and the 91 genes differentially expressed by more than twofold) were due strictly to differences between the populations present. More importantly, many of the genes identified as up- or downregulated by more than twofold may not have been associated with cell populations present in different proportions between chronically infected individuals and individuals who cleared their respective infections (e.g. macrophages). Conversely, even if the significantly up- or downregulated genes identified were due entirely to the presence of different cell populations in the PBMC repertoire, this does not take away from the information generated by the analysis described, namely that some genes not previously associated with HCV infections remain potential targets that can be used in the modification of current treatment modalities or in the development of novel treatments.

The genes identified following our analyses were divided into three broad groups: immune function, cancer related and other (e.g. genes associated with cell growth, inflammation, signalling pathways, central nervous system pathways, lipid metabolism, and energy and metabolism pathways). Our comprehensive bioinformatics analyses strongly support a central role for immune response-related genes in HCV clearance.

Of the immune-related genes identified, only genes encoding IFIT2, IFIT3, IRF7, ISG15, MX1, OAS1, OASL, OSM, PLSCR1, TMEM176A, TMEM176B and TRIM22 have been shown previously to play roles in anti-HCV immune responses. In addition, IFIT2, IFIT3, IRF7, ISG15, MX1, OAS1, OASL, OSM and TRIM22 genes were shown to be involved in innate immune responses or to influence HCV antiviral activities.

IFIT2 and IFIT3 have been studied in the context of HCV treatment. Specifically, IFIT1 has been shown to suppress HCV activity in the presence of interferon, and therefore may be beneficial when used in combination with HCV treatment regimens designed to achieve lasting viral suppression (Lalle et al., 2008). The gene encoding TLR7 (which helps in the regulation of IRF7) is downregulated following HCV infections (by altering TLR7 mRNA stability and function) as an HCV-mediated immune evasion strategy (Chang et al., 2010). Decreased expression of TLR7 following HCV infections can therefore lead to viral persistence leading to chronic infections (Chang et al., 2010). HCV also can avoid immune responses by interfering with interferon-stimulating genes (ISGs), for example ISIG15 and OASL, which represent two ISGs that play important roles in mediating innate antiviral responses. Interference with ISG function is likely to result in persistent infections (Jouan et al., 2012). Finally, interactions between ISG15 and alpha interferon (IFN-α) mediate anti-HCV responses associated with increased expression of the IFN-α-inducible genes OAS1 and MX1. Therefore, even though ISG15 does not mediate direct antiviral effects against HCV, its role in the modulation of IFN-α genes may mediate antiviral responses (Chua et al., 2009).

OSM is a cytokine that can play a variety of roles including mediating anti-HCV immune responses in association with IFN-α, suggesting that this cytokine could possibly be used as a new anti-HCV treatment option (Ikeda et al., 2009). Upregulation of TRIM22 expression has also been described by other microarray analyses in chronically infected HCV patients and shown to possess anti-virus replication activities (Nisole et al., 2005), including activity against HCV (Folkers et al., 2011). The PLSCR1 gene has been shown to play a role in protection by interfering with HCV entry into hepatocytes (Gong et al., 2011); therefore, interruption of PLSCR1 gene function could inhibit HCV entry into hepatocytes, resulting in diminished viral loads, in addition to its use in the development of novel treatment strategies. TMEM176A and TMEM176B have been shown to be associated with HCV recurrence in a gene expression study of patients with liver transplants (Gehrau et al., 2011).

Although these genes have been found to possess antiviral activities, the above-mentioned studies were conducted in vitro using various cell-culture approaches or during the course of human studies designed to identify genes associated with anti-HCV drug therapies; therefore, their role during the initial infection process can only be inferred. However, identification of some genes previously associated with HCV clearance served to validate the microarray analysis described in this report that identified additional gene targets. Many previously unidentified genes associated with viral clearance were identified in this study and may also be important in clearing HCV infections. Combining these components into existing therapies may enhance the body’s ability to clear HCV infections.

A gene recently associated with HCV clearance is the type III interferon gene, IL28B, which encodes IFN-λ3. Most studies have examined the role IL28 in HCV clearance in relation to the efficacy of pegylated IFN-α plus ribavarin-based therapies (Balagopal et al., 2010). Furthermore, single-nucleotide polymorphisms (SNPs) identified within the IL28 gene in various populations have been shown to affect the effectiveness of pegylated IFN-α plus ribavarin (Shaker & Sadik, 2012; Xie et al., 2012). It has recently been demonstrated that a SNP (rs12979860) was associated with HCV genotype 1 clearance among vertically infected children and that IFN-λ levels in blood and liver were elevated in individuals chronically infected with HCV, resulting in the upregulation of antiviral innate immune responses (by inducing the generation of dendritic cells with the potential of expanding Foxp3⁺ Tregs) (Dolganiuc et al., 2012). In addition, IFN-λ-mediated control of ISGs (in both humans and chimpanzees) is associated with HCV clearance (Thomas et al., 2012). The present study identified up- or downregulated genes in individuals who cleared their HCV infections relative to the expression profile observed in chronically infected individuals. IL28 was not among the genes identified; however, all ISGs identified in our study (IFIT2, IRF7, IFIT3, IFI6, IFIT1, IFI35, MX1, ISG15 and OASL; Table 2) were significantly downregulated in individuals who had cleared their HCV infections, suggesting that these genes are probably upregulated in chronically infected individuals. As temporal effects relating to IL28 expression in the context of HCV infections have not been established, we cannot rule out the possibility that IL28 levels were upregulated earlier during the infection process or during the earlier stages of chronic infection. It is possible therefore that temporal effects relating to IL28 gene expression in the context of HCV infection may explain why this gene was not identified in our study.

Of the cancer-related genes identified in the present study, only HRK has been studied previously in relation to HCV. HCV can elicit cell death via apoptosis by influencing the activities of different genes responsible for regulating cell death (Walters et al., 2009), including HRK, and has been linked to cytochrome c release from mitochondria (Walters et al., 2009).

Two genes, GBP-1 and HDB, did not fit into the three primary categories defined. Itsui et al. (2009) tested the effects of GBP-1 in cell-culture assays and determined that HCV bound GBP-1. Binding of the HCV protein NS5B to GBP-1 resulted in GTPase inhibition (typically mediated by GBP-1), facilitating HCV replication. This may lead to resistance to innate, IFN-mediated antiviral defences and to the clinical persistence of HCV infections (Itsui et al., 2009). HDC has been studied in the context of treatment but is believed to decrease the oxidative stress that accompanies HCV infections, especially in the liver (Lurie et al., 2002). HDC binds histamine receptors on phagocytes and subsequently blocks the production of reactive oxygen species produced by monocytes or macrophages (Lurie et al., 2002). This reduces oxidative damage that optimizes natural killer cell and T-cell function, thereby enhancing immune responses to viral antigens (Lurie et al., 2002).

As prevalence data were used in this analysis, the time when the respective participants acquired their infections or when the infections were cleared was unknown. However, based on previously published data, participants probably had ample time to clear the virus (Cox et al., 2005; Hofer et al., 2003). The HCV viral loads were verified at the time of the microarray analysis, which occurred 4–5 years after HCV detection by HCV enzyme immunoassay (EIA). The HCV infection status of all participants was verified, that is, patients defined as infected had positive viral loads and patients identified as having cleared their respective HCV infections had undetectable viral loads. We are currently replicating these studies using an HCV-positive population where HCV RNA status can be verified and the genetic polymorphisms in these candidate genes identified as a means of understanding better the roles of these genes in HCV persistence/clearance.

Methods

Study population.

The study was a nested case–control design. Anti-HCV antibody-positive individuals (n = 16) were recruited from a longitudinal cohort of an intervention study in their last year of follow-up. Briefly, the intervention study utilized hepatitis B vaccination as a model for a future HIV and HCV vaccine trial that evaluated the efficacy of a hepatitis B self-efficacy behavioural intervention and an accelerated vaccination schedule as methods to increase acceptance/adherence to hepatitis B vaccination among not-in-treatment drug users (Hwang et al., 2010). HIV- and hepatitis B virus-negative participants were enrolled for hepatitis B vaccination and follow-up. All enrolled subjects were tested for anti-HCV antibodies by an EIA at enrolment and at the 6-month visit (for confirmation). If the participant did not have a 6-month sample, the screening sample was tested or a subsequent follow-up visit sample was tested for anti-HCV antibodies. If anti-HCV positive (at least two samples positive at least 6 months apart), the sample was tested for HCV RNA by PCR. For the current study, we recruited eight participants who were anti-HCV antibody positive and HCV RNA positive, and eight individuals who were anti-HCV antibody positive and HCV RNA negative, who were matched by age, sex and race from the larger longitudinal cohort. Informed consent was obtained from each participant and human experimentation guidelines of the USA Department of Health and Human Services and of the University of Texas Health Science Center at Houston Committee for the Protection of Human Subjects were followed in the conduct of this clinical research.

HCV testing.

For the previously described longitudinal cohort study (Hwang et al., 2010), samples were collected in 10 ml tubes, centrifuged within 8 h to separate the serum and immediately stored at −20 °C until tested (mean duration of 1 month). Antibodies to HCV were detected by microparticle EIA (Abbott Laboratories) in serum. Samples positive for anti-HCV antibodies were tested qualitatively for the presence of HCV RNA by PCR (Cobas Amplicor HCV; Roche Diagnostics). Plasma samples from the corresponding participants in this study were analysed for the presence of HCV RNA at the Diagnostic Laboratories of the Memorial Herman Hospital (Houston, TX, USA) using a quantitative PCR test conducted with a lower limit of detection set at <43 IU ml⁻¹ (Cobas Ampliprep, Cobas TaqMan HCV Real-Time RT-PCR; Roche Diagnostics). All assays were performed according to manufacturers’ directions, including running appropriate controls and repeat testing when necessary (Hwang et al., 2010).

Microarray analysis.

Blood was collected in PaxGene (Qiagen) tubes and stored at −80 °C within 8 h of collection, and RNA was extracted using a PAXgene RNA extraction kit (Qiagen) as instructed by the manufacturer, ~1 month after collection. Extracted RNA samples were frozen at −20 °C until analysis at the University of Texas Health Science Center at Houston Microarray Core Laboratory using an Illumina Gene Expression chip to identify candidate genes based on RNA expression profiles. The data were analysed using GenomeStudio 1.03 software (Illumina).

Data quality.

Raw signals of all built-in controls were checked as quality controls for the performance of the arrays. Sample-independent controls were used to check hybridization and signal generation. Housekeeping genes were used as sample-dependent controls. Data for patient no. 9, an HCV EIA-positive, HCV RNA-negative patient, was excluded due to poor data quality.

Background signals were subtracted and arrays were normalized using quantile normalization (Bolstad et al., 2003). The reproducibility of biological or technical replicates was checked by comparisons between individual samples. Outliers were removed if necessary, and the remaining samples were grouped and the mean signal intensities between groups used for differential expression analysis.

Statistical analysis.

Pre-processing utilized quantile normalization and background subtraction. Differential expression was determined at P<0.05 with multiple testing correction using the Benjamini–Hochberg method to control for false discovery rates (Benjamini & Hochberg, 1995). Clustering and pathway analysis were performed using the Multiple Array Viewer version 4.3.02 (Saeed et al., 2003) and Ingenuity ipa version 8.6.