Hepatitis C Virus Genotypes in the United States: Epidemiology, Pathogenicity, and Response to Interferon Therapy

Infection with hepatitis C virus (HCV) has been identified as the major cause of post-transfusion non-A, non-B hepatitis [1]. Chronic liver disease occurs in at least 50% of patients with acute HCV infection, and cirrhosis develops in 20% of these patients [2]. The virus has a single-stranded RNA genome that is approximately 10 Kbp long. A comparison of HCV genomic sequences from around the world has shown substantial heterogeneity of nucleotide sequences within several regions of the viral genome [3]. Hepatitis C virus has been classified into multiple strains or genotypes on the basis of the identification of these genomic differences. It has been suggested that the heterogeneity in sequence seen among HCV genotypes may be associated with variant antigenic and biological properties [4]. In addition, outcome of liver disease and rates of response to interferon therapy may vary according to HCV genotype [5, 6]. Therefore, understanding the distribution and properties of HCV genotypes may have important implications for prognosis and therapy. We evaluated the distribution of HCV genotypes in distinct geographic regions of the United States and determined the clinical characteristics of and response to interferon therapy in patients with one of several HCV genotypes. We used the classification system developed by Simmonds and colleagues [7] because it was recently adopted by consensus at the Second International Conference of HCV and Related Viruses (August 1994, San Diego, California). In this system, HCV genotypes are classified into six major genotypes (1 to 6, ordered according to when they were discovered) and 11 subtypes (1a, 1b, 1c, 2a, 2b, 2c, 3a, 3b, 4a, 5a, and 6a). Methods Serum Samples We analyzed serum samples of 208 patients who were positive for antibody to HCV and had chronic liver disease. The samples were retrospectively obtained from four tertiary referral centers in the United States (59 consecutive samples from the Mayo Clinic, Rochester, Minnesota; 48 consecutive samples from the University of Vermont, Burlington, Vermont; 49 consecutive samples from the University of Miami, Miami, Florida; and 52 consecutive samples from the University of Washington Virology Laboratory, Seattle, Washington [this last center provided samples from Washington State, Idaho, Utah, Oregon, and California]). Twenty-nine patients were excluded from the study: Nineteen had no detectable products for DNA sequencing, and 10 had ambiguous sequencing results. The remaining 179 samples were the focus of this study. No clinical information was available on the patients whose samples were obtained from the University of Washington Virology Laboratory; thus, these samples were used only to study the geographic distribution of HCV genotypes. Data on interferon treatment were available for 78 patients from the Mayo Clinic and the University of Vermont. Samples from these two institutions were obtained from patients who had agreed to participate in trials of interferon treatment. Reverse Transcriptase and Polymerase Chain Reaction We selected the direct sequencing technique because it remains the gold standard and the only way to definitively identify all HCV genotypes and subtypes. Hepatitis C virus RNA was extracted from 100-L aliquots of serum after the addition of 1 mL of RNAzol B solution (Biotecx Laboratories, Houston, Texas) (2 mol of guanidinium thiocyanate per L, 12.5 mol of sodium citrate per L, 0.25% N-laroylsarcosine, 0.05 mol of 2-mercaptoethanol per L, 100 mmol of sodium acetate per L, and 50% water-saturated phenol). After the addition of 100 L of chloroform, samples were spun for 15 minutes at 14 000 g and the aqueous phase was extracted. Total RNA was precipitated by the addition of isopropanol and 2 L of glycogen and incubation at 4 C for 45 minutes. An RNA pellet was recovered by centrifugation at 14 0006 g, washed in 1 mL of 70% ethanol solution, dried, and resuspended in 10 L of RNAase-free water (Promega, Madison, Wisconsin). Ribonucleic acid was reverse-transcribed into complementary DNA by using reverse transcriptase and an antisense oligonucleotide primer (5-CGCGGAATTCCTGGTCATAGCCTCCGTGAA-3) in the presence of reverse-transcriptase buffer (100 mmol of tris-HCl per L, 500 mmol of KCl per L, 1% Triton X-100, and a pH of 8.6 at 25 C) (Promega) and 3.0 mmol of magnesium per L. Hepatitis C virus complementary DNA was amplified by polymerase chain reaction (PCR) in the presence of the sense oligonucleotide primer (5-TGGGGATCCCGTATGATACCCGCTGCTTTGA-3), PCR buffer (500 mmol of KCl per L, 100 mmol of tris-HCl per L, and a pH of 8.3) (Perkins-Elmer-Cetus, Norwalk, New Jersey), 2.0 mmol of magnesium per L, and Amplitaq DNA polymerase (Perkins-Elmer-Cetus). The PCR assay was done in a DNA thermal cycler for 50 cycles (94 C for 1 minute, 58 C for 1 minute, and 72 C for 5 minutes). Products of the PCR assay were analyzed by gel electrophoresis in 3% agarose gel that was stained with ethidium bromide. The appearance of a band 401-base pair was considered a positive result. To avoid and monitor for possible contamination with exogenous sequences during extraction or amplification, extraction of nucleic acid and genomic amplification steps were done in separate laboratories. Ribonucleic acid samples from at least one negative and one positive sample were extracted, subjected to reverse transcription, and amplified in each batch of samples tested by PCR. No false-positive results were obtained in any of the negative controls. Sequencing and Genotyping Each fragment of the PCR product, which was approximately 401 base pairs long, was desalted before undergoing sequencing with a direct column-purification method (Wizard PCR Preps DNA Purification System, Promega). Automated sequencing was done by using a standard Sanger procedure, which involved the incorporation of fluorescein-labeled dideoxynucleotides and detection on an acrylamide gel (ABI model 373 A, Applied Biosystems, Hercules, California). Nucleotide sequences were aligned and compiled with the previously reported sequences by using the Pileup program (Wisconsin Genetic Computer Group, Madison, Wisconsin)[8]. Cluster analysis was done by using the unweighted-pair group mean average, which was included in the program. These methods allowed comparison of a 222-base pair fragment of DNA that was homologous to nucleotide positions 7975 to 8196 in the NS5 region of the prototype virus. Collection of Epidemiologic Data We studied the geographic distribution of the HCV genotypes identified in the blood samples. Data from all samples were combined to define the prevalence of the HCV genotypes in patients with chronic hepatitis C in the United States. When available, age, sex, risk factors for HCV acquisition, and liver histologic findings at the time of presentation were recorded for each patient. Risk factors for acquiring HCV included history of blood transfusion, history of injection drug use, and employment at a health care facility. Liver histologic findings were classified into three groups: mildly active hepatitis (portal inflammation without substantial hepatocyte necrosis), moderately active hepatitis (inflammation with hepatocyte necrosis), and liver cirrhosis. Accurate history of alcohol consumption was not available for many of these patients and thus was not included in the analysis. The investigator who did the genotyping was blinded to the clinical data of patients at the time of analysis. Liver biopsy specimens were independently interpreted at each center. Pathogenicity of Hepatitis C Virus Genotypes To study the possible differences in the pathogenicity of HCV genotypes, we divided patients into two groups: patients with mild hepatitis and patients with severe hepatitis. Mild hepatitis was defined as 1) pretreatment alanine aminotransferase levels that were less than three times the normal level and 2) no cirrhosis seen during examination of the liver biopsy specimen obtained before treatment. Severe hepatitis was defined as pretreatment alanine aminotransferase levels greater than three times the normal level or the presence of liver cirrhosis on pretreatment biopsy. Response to Interferon Seventy-eight patients received an average dose of 3 million U of interferon (interferon- or consensus interferon) for 6 months. Response to interferon was defined as the normalization of alanine aminotransferase levels at the end of therapy. Partial response to interferon (defined as decreased but not completely normal alanine aminotransferase levels) was considered to be a treatment failure. Sustained biochemical response was defined as a normal alanine aminotransferase level 6 months after the discontinuation of interferon treatment. Statistical Analysis We used the rank-sum and Kruskal-Wallis tests to compare continuous variables (such as age) between groups, and we used the Fisher exact test to assess associations in tabular data. Because few patients had genotype 2a, 3, or 4, all tests of association between genotype and other factors are based on data that were collapsed into four groups: genotype 1a, genotype 1b, genotypes 2a and 2b, and genotypes 3 and 4. Logistic regression was used to evaluate the association between response to interferon and the combined predictors of cirrhosis and genotype. We used the SAS statistical analysis package (SAS Institute, Cary, North Carolina) for all calculations. Results Geographic Distribution of Hepatitis C Virus Genotypes Hepatitis C virus genotype 1a was present in 104 of 179 (58%) patients with chronic HCV infection; genotype 1b was the second most common genotype encountered (38 of 179 patients [21%]). Genotype 2b was present in 23 patients (13%), and genotype 3a was present in 8 patients (5%). Four patients (2%) had HCV genotype 2a, and 2 (1%) had genotype 4a. Geographic region and distribution of genotypes were not significantly associated (P = 0.18). However, samples obtained from the western United States conta