Journal of Viral Hepatitis, 2012, 19, e1–e10
doi:10.1111/j.1365-2893.2011.01471.x
Preactivation of the interferon signalling in liver is correlated with nonresponse to interferon alpha therapy in patients chronically infected with hepatitis B virus C. Xiao,1 B. Qin,1 L. Chen,2,3 H. Liu,4 Y. Zhu1 and X. Lu1
1
Department of Infectious Diseases, The First Affiliated
2
Hospital, Chongqing Medical University, Chongqing, China; Banting and Best Department of Medical Research, University of Toronto, Toronto, ON, Canada; 3Institute of Blood Transfusion, Chinese Academy of Medical Sciences and Peking Union Medical College, Chengdu, Sichuan, China; and 4
Shanghai Biochip Co., LTD, Shanghai, China
Received January 2011; accepted for publication March 2011
SUMMARY. Interferon alpha (IFN-a) therapy is widely used to treat patients with chronic hepatitis B (CHB) but the sustained response rate is low, and the molecular mechanisms for the ineffectiveness of IFN-a treatments are not known. We screened differentially expressed genes between responders (Rs) and nonresponders (NRs) in patients with CHB treated with IFN-a to explore the molecular basis for treatment failure. Expression profiling was performed on percutaneous needle liver biopsy specimens taken before therapy. Gene expression levels were compared between seven patients who did not respond to therapy (NR) and six who did respond (R). Gene ontology category and KEGG pathway were analysed for differentially expressed genes, and the selected differentially expressed genes were confirmed using real-time polymerase chain reaction. We identified 3592 genes whose expression levels differed significantly between all Rs and NRs (P < 0.05); many of
INTRODUCTION Hepatitis B virus (HBV) infection is a global public health problem with up to one million HBV-infected patients dying from HBV-associated liver disease annually [1]. It is well Abbreviations: ALT, alanine aminotransferase; CHB, chronic hepatitis B; CTL, cytotoxic T lymphocyte; GO, gene ontology; HBV, hepatitis B virus; IFN, interferon; ISG, IFN-stimulated genes; NR, nonresponders; RVM, random variance model. Correspondence: Dr. Bo Qin, Department of Infectious Diseases, The First Affiliated Hospital, Chongqing Medical University, No.1 Youyi Road, Yuzhong District, Chongqing 400016, China. E-mail:
[email protected] Dr. Limin Chen, Banting and Best Department of Medical Research, University of Toronto, 112 College Street, Toronto, ON M5G 1L6, Canada. E-mail:
[email protected]
2011 Blackwell Publishing Ltd
these genes are IFN-stimulated genes (ISGs) and immunerelated genes. The ISGs were more highly expressed, while immune-related genes were inhibited in NRs before IFN-a treatment. Two ISGs (CEB1 and USP18) that are linked in an IFN inhibitory pathway are highly expressed in NRs, and a potential antiviral gene ISG20 was inhibited in NRs, suggesting a possible rationale for treatment nonresponse. Patients who do or do not respond to IFN have different liver gene expression profiles before IFN-a treatment. Preactivation of the IFN signalling pathway leading to the increased expression of inhibitory ISGs and inhibition of immune response in the pretreatment livers was associated with treatment failure. Keywords: alpha interferon, chronic hepatitis B, differential gene expression, microarray, no response to treatment.
known that active HBV replication is the key driver of liver injury and disease progression; therefore, the aim of treatment for chronic hepatitis B (CHB) is to achieve sustained suppression of HBV replication and remission of liver disease. Two types of therapy are currently available for the treatment of CHB: (i) nucleoside analogues and (ii) interferon alpha (IFN-a), which has a dual mode of action, with both antiviral and immunomodulatory effects. Both types of therapy have less than optimal efficacy, but the advantages of IFN-a therapy include a limited treatment course, a lack of resistance development and even clinical cure with hepatitis B surface antigen (HBsAg) seroconversion in a few patients [2]. However, IFN-a therapy can result in only 30–40% hepatitis B e antigen (HBeAg) seroconversion, and over 60% of patients still continue to suffer from chronic active hepatitis B even with IFN-a therapy [3]. There are many factors associated with IFN-a treatment failure, such as therapeutic
e2
C. Xiao et al.
regimen, host and virus interaction, including age, sex, genotype, HBV-DNA [4–10]. However, the exact molecular mechanism of the ineffectiveness of IFN-a treatment in some CHB is not known. Considering the limitations of IFN-a treatment, such as less than optimal response rate, expensive and inevitable side effects, it is important to identify pretreatment parameters for predicting whether patients with CHB will respond to IFN-a treatment or not. Recently, some studies indicated that cell lines infected with HBV have different gene expression pattern after treatment with IFN-a, and this difference was closely associated with IFN-a treatment outcome [11,12]. It was also reported that 13 short tandem repeat markers (STR) displayed different allele and/or genotype frequency between IFN-a treatment responders and nonresponders (NRs) [13]. This pilot study has developed a new approach to identify genetic markers that allow us to predict the IFN-a response in CHB, but these data were generated from cell culture or peripheral blood, not fully reflecting the efficiency of IFN-a treatment on patients with CHB in the liver tissue. Another interesting study, using cDNA microarray technology, identified differential gene expression patterns in pretreatment liver tissues of patients chronically infected with hepatitis C virus. They identified 18 genes whose expression levels were statistically different between treatment responders and NRs [14]. Further study from this same group identified preactivation of the type I IFN signalling and ISG15/USP18 pathway were involved in treatment nonresponse. These studies provided novel molecular mechanisms for treatment nonresponse in patients chronically infected with HCV [15]. We hypothesized that there are differential hepatic gene expression patterns between treatment responders and NRs in the pretreatment liver tissues of CHB. In this study, we profiled the pretreatment hepatic gene expression levels in treatment responders and NRs in patients with CHB using Agilent Whole Human Genome Oligo Microarray (4 · 44 K; provided by Shanghai Biochip Co., LTD). Our purpose is to use a fully unbiased high-throughput genomic approach to reveal the key genes that are involved in IFN-a resistance and to predict treatment outcomes before the initiation of IFN-a therapy.
MATERIALS AND METHODS Patients and biopsies Thirteen treatment-naive patients with chronic HBV were treated at the First Affiliated Hospital of Chongqing Medical University from December 2008 to June 2010. The inclusion criteria are the following: (i) the diagnosis of CHB was based on the clinical diagnostic criteria in 2008 APASL CHB prevention and cure guideline [16]. (ii) All the patients had hepatitis B at least for 6 months, and HBsAg and/or HBVDNA is also positive at enrolment. (iii) All of these patients
meet the IFN-a antiviral criteria. All patients considered for treatment with IFN-a underwent percutaneous liver biopsy (via a 15-gauge needle) and had baseline viral loads determined. Treatment consisted of IFN-a1b (provided by Kexing Biotech Co., Shenzhen, China) 5 MU three times per week by subcutaneous injection for 48 weeks [16]. Quantitative HBV-DNA and liver function were determined at completion of therapy and 24 weeks after. Patients were designated as NRs or Rs according to the follow-up outcomes. Rs: alanine aminotransferase (ALT) returned to normal and HBeAg seroconversion at the end of treatment (all 13 patients were HBeAg positive before treatment) and having a sustained viral response if HBV-DNA (using polymerase chain reaction assay, the lowest detection limit is 103 copies/mL) was undetectable both at the end of therapy and at 6-month follow-up; NRs: do not achieve the above-mentioned criteria. Compliance was excellent (all patients completed therapy). The exclusion criteria: patients who had hepatitis induced by other virus, metabolism, medicine and autoimmunity.
RNA extraction and purification A portion of each liver biopsy specimen (0.5–1.0 cm) was immersed in liquid nitrogen. Total RNA was isolated from each liver biopsy specimen using the Qiagen RNeasy kit (Qiagen, Hilden, Germany). The concentration and quality of RNA were determined by spectrophotometry and by the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) according to the manufacturerÕs instructions. Only those samples that showed no degradation were used to generate labelled targets. The total RNA was further purified using an RNeasy Mini kit (Qiagen).
Complementary DNA microarrays Our experiment utilized the Agilent Whole Human Genome Oligo Microarray (4 · 44 K), which represents more than 41 000 human genes and transcripts. Single- and doublestranded cDNA were synthesized from total RNA samples (2 ug) as described in the Agilent GeneChip Expression Analysis Technical Manual. The cRNA was purified and fluorochrome labelled with Cy3, and then fragmented for hybridization. Eight hundred and seventy-five nanograms of labelled cRNA was used per GeneChip array. Hybridization was performed at 65 C with rotation for 17 h. The GeneChips were washed and then scanned by Agilent scanner (G265BA; Agilent).
Statistics and clustering analyses P value and false discovery rate (FDR) were calculated using the t-test modified from random variance model (RVM-t-test) [17], and the differentially expressed genes between Rs and NRs were identified based on P value and FDR both 1.3. The unsupervised 2011 Blackwell Publishing Ltd
Preactivation of interferon signalling hierarchical cluster analysis was performed using the Cluster 3.0 software [18].
Gene ontology (GO) and pathway analysis The categorization of biological process was analysed using Gene Ontology project (http://www.geneontology.org) based on these differentially expressed genes. The enrichment calculative formula was given by: Re = ¼(nf/n)/(Nf/N), where nf is the number of differentially expressed genes within the particular category, n is the total number of genes within the same category, Nf is the number of differentially expressed genes in the entire microarray, and N is the total number of genes in the microarray. GO analysis classified the differentially expressed genes into biologically significant functional groups that can be easily interpreted. Pathway analysis was performed using the KEGG database. Two-side FisherÕs exact test and chi-square test were used to classify the GO category and pathway analysis, and FDR was calculated to correct the P value. P value
