Emergence of Hepatitis C Virus Genotype 4: Phylogenetic Analysis Reveals Three Distinct Epidemiological Profiles

Study population.The participants included all patients who were diagnosed with HCV-4 infection at the hepatology ward of the Academic Medical Center (AMC) from 1999 through 2008, as well as persons who were found to be HCV-4 positive at the Amsterdam Health Service (AHS). All AHS participants were part of surveys performed to monitor the general population or specific risk groups for the presence of sexually transmitted and blood-borne infections. AHS establishes the HCV genotype using an in-house PCR targeting the HCV core region (33). AMC uses a commercial 5′ untranslated region-based HCV genotyping assay (Trugene 5′NC; Genelibrarian module 3.1.2; Bayer Healthcare, Berkeley, CA). For this study, we collected data on the age, country of birth, human immunodeficiency virus (HIV) infection status, and plausible route of HCV transmission for each HCV-4-infected patient.

RNA isolation, reverse transcription, and PCR.RNA was isolated from the first RNA-positive serum sample obtained from each patient by using 200 μl serum and a MagNa Pure total nucleic acid extraction kit (Roche Diagnostics, Almere, The Netherlands). HCV RNA was eluted in 100 μl of TE (Tris-EDTA) buffer and was subsequently transcribed to cDNA as described by Beld et al. (3). This cDNA was used as the input for two separate PCR assays targeting the HCV nonstructural 5B (NS5B) region, which resulted in the amplification of two overlapping NS5B fragments with a total length of 669 bp. Sera and RNA isolates were stored at −80°C.

NS5B PCR 1.NS5B fragment 1 (338 bp, nucleotides 7938 to 8275) was amplified by a single-round PCR assay, as described by Murphy et al. (13). In brief, 12.5 μl of cDNA was added to a 12.5-μl reaction mixture containing 1× PCR ΙΙ buffer; 200 μmol/liter each of dATP, dCTP, and dGTP plus 400 μmol/liter dUTP; 0.1 μg/μl bovine serum albumin; 0.9 μM sense primer (5′-TATGAYACCCGCTGYTTTGACTC-3′); 0.9 μM antisense primer (5′-TAYCTVGTCATAGCCTCCGTGAA-3′); 0.5 U uracil N-glycosylase; and 2.5 U Amplitaq Gold (Applied Biosystems, Nieuwerkerk a/d IJssel, The Netherlands). The final MgCl₂ concentration was 2.5 mM. The PCR cycling conditions were 5 min at 95°C, followed by 45 cycles, with each cycle consisting of 30 s at 95°C, 30 s at 55°C, and 60 s at 72°C, plus a final period of incubation of 10 min at 72°C.

NS5B PCR 2.NS5B fragment 2 (436 bp, nucleotides 8206 to 8642) was amplified by a nested PCR assay, as described by van de Laar et al. (33). The composition of the reaction mixture and the cycling conditions were identical to those described above, except some modifications were made so that 3 μl cDNA instead of 3 μl RNA could be used as the input for the first-round PCR. In the second-round PCR, we used only the reaction mixture containing HCV-4 genotype-specific primers 4HCV-IS (5′-CTGAGAGACTGCACSATGYTGGT-3′) and E1b (5′-AATGCGCTRAGRCCATGGAGTC-3′).

Sequencing and phylogenetic analysis.The sequencing reactions were performed as described earlier (33). The viral genotype was confirmed by phylogenetic analysis of the NS5B sequences obtained, in which the sequences were compared with established reference sequences in the GenBank database (27). Phylogenetic trees were constructed by the maximum-likelihood approach implemented in PHYML software (9). Evolutionary model selection was performed after comparison of the corrected Akaike information criterion scores of various models by using PHYML software and the MrAIC script (15). The Hasegawa-Kishino-Yano substitution model with a γ distribution of among-site-rate heterogeneity (HKY-Γ) provided the best fit and was subsequently used. The statistical robustness of the phylogenetic trees was tested by bootstrapping with 1,000 replicates (bootstrap values of 70% were used as a cutoff point for cluster analysis; only branching with a bootstrap value of ≥70% was defined as robust clustering). For each HCV cluster, we estimated the year of origin of the most recent common ancestor (TMRCA) using a molecular clock approach, which places the phylogenetic history of HCV onto a timescale of months and years. On the basis of previous estimates, the inferred rate of nucleotide evolution of our NS5B sequences was 5.0 × 10⁻⁴ (18). The year of origin for each cluster was subsequently calculated by using an HKY model with a variable rate for each codon position, as implemented in the Tipdate program (21).

Statistical analysis.Differences in epidemiological characteristics between HCV-4-positive patients allocated to phylogenetically distinct HCV-4 clusters were tested by the χ² test in the case of sex, country of birth, most plausible route of HCV transmission, and HIV infection status and by analysis of variance in the case of age. The STATA statistical software package (StataCorp, College Station, TX) was used.

Nucleotide accession numbers and reference sequences.The sequences reported in this study have been submitted to the GenBank database and can be retrieved under accession numbers FJ807047 through FJ807168.

The GenBank accession numbers of the full-genome reference sequences used to determine the HCV genotype were as follows: Y11604 (subtype 4a), FJ462435 (subtype 4b), FJ462436 (subtype 4c), DQ516083 (subtype 4d), EU392175 (subtype 4f), FJ462432 (subtype 4g), EU392171 (subtype 4k), FJ839870 (subtype 4l), FJ462433 (subtype 4m), FJ462441 (subtype 4n), FJ462440 (subtype 4o), FJ462431 (subtype 4p), FJ462434 (subtype 4q), FJ462439 (subtype 4r), and FJ839869 (subtype 4t). The GenBank accession numbers of the reference sequences of the HCV subtypes omitted from the list above were available only for NS5B fragment 1 and are as follows: L29590 (subtype 4e), L29611 (subtype 4h), L36437 (subtype 4i), and L36438 (subtype 4j).

Study population.A combined search of the AMC and AHS databases identified 135 patients infected with HCV-4: 65 who had attended AMC, 48 who had attended AHS, and 22 who had attended both institutions. Two patients who had attended AMC were excluded when amplification and sequencing of the NS5B region revealed that they were infected with subtypes 1b and 1g. As shown in Table 1, the remaining 133 HCV-4-infected patients had a median age of 43 years (interquartile range [IQR], 36 to 48 years), 85% were male, 50% were born in Europe, and 28% were coinfected with HIV. The predominant underlying risk factors for infection were birth in a country where HCV is endemic (mostly Egypt and countries in Africa) (37%), injecting drug use (IDU; 32%), and high-risk sexual behavior (21%).

Reverse transcription-PCR, sequencing, and genotyping.Sera were available from 123/133 (93%) patients. Amplification and sequencing of NS5B fragment 1 succeeded in 119/123 (94%); the remaining 4 patients had HCV loads below 1,000 IU/ml. NS5B fragment 2 was obtained from 120/123 (98%) patients. Failure in one sample was due to a low viral load (<1,000 IU/ml), and in two samples, failure was most likely due to subtype-specific (HCV-4o and an unrecognized HCV-4 subtype) mutations at the primer locus. For four patients for whom serum samples were lacking, HCV NS5B fragment 2 sequencing data were available from an earlier study (31).

In total, the combined HCV NS5B fragment was obtained from 117/133 patients. Moreover, sequence data for at least one NS5B fragment were available for 126/133 patients. HCV genotyping of these 126 partial HCV NS5B sequences revealed eight distinct HCV-4 subtypes. The vast majority of patients were infected with HCV subtypes 4d (n = 71; 56%) and 4a (n = 46; 37%). The isolates from the remaining nine patients (6%) belonged to subtypes 4 h (n = 1), 4k (n = 2), 4n (n = 1), 4o (n = 1), 4r (n = 3), and a subtype not previously recognized (n = 1).

Phylogenetic analysis.Phylogenetic trees were constructed for the HCV NS5B fragments combined and separately for fragment 1 and fragment 2. Figure 1 shows the HCV-4 phylogeny of the NS5B fragments combined. It represents the results for 117 HCV-4-infected patients and their assumed mode of HCV transmission, plus the results for 15 available HCV reference sequences. Phylogenetic analysis revealed three monophyletic clusters (bootstrap value, >70) containing 116/126 (92%) HCV-4-infected patients for whom a partial NS5B sequence was available. Table 2 summarizes the epidemiological characteristics of these patients according to the phylogenetic cluster to which they were allocated. The clusters were significantly different by sex, country of birth, HIV infection status, and plausible route of transmission (P < 0.001) but not by age (Table 2). Clearly, each cluster was linked to a specific epidemiological profile: immigrants from Egypt (cluster 1 [C1]), individuals reporting IDU (C2), and HIV-positive men who have sex with men (MSM) (C3).

• Open in new tab
• Download powerpoint

FIG. 1.

HCV phylogenetic tree (combined NS5B fragments) and clustering of patients diagnosed with HCV-4 infection in the Amsterdam region. The risk factors allocated for each patient are listed opposite each sequence: born in Egypt (⋄), born in other areas where HCV is endemic area (A), IDU (•), high-risk sexual behavior of MSM (▿), blood transfusion in Europe before 1991 (×), and other or unknown (−).

The largest cluster, C1, contained 46 sequences of HCV subtype 4a. The vast majority of the patients in this cluster were male (96%), of Egyptian descent (70%), and presumed to be HIV negative (91%). Within this so-called Egyptian cluster, phylogenetic analysis revealed one subcluster (C1a) of isolates from nine patients of non-Egyptian descent: six patients reporting IDU (four of them Dutch), two Dutch hemophiliacs, and one patient from Pakistan for whom information on the mode of transmission was missing. In addition, we identified a pair (C1b) of highly similar HCV-4a strains. These two strains were obtained from a female sex worker born in Suriname and a heterosexual HIV-positive Egyptian male who reported commercial sex contacts in The Netherlands.

Both C2 and C3 contained sequences of HCV subtype 4d. C2 comprised isolates from 44 patients; the majority were male (73%), reported IDU as the primary risk factor for HCV transmission (80%), and were of European descent (77%). In C2, 18% of the patients were coinfected with HIV. C3, which was actually a robust subcluster of C2, consisted of 26 sequences obtained from MSM who were diagnosed with acute HCV infection in The Netherlands after the year 2000. HIV-HCV coinfection was present in 24/26 (92%) of the MSM in C3; 1 MSM without HIV infection reported needle sharing during ketamine use and was therefore allocated to the IDU risk category.

In addition to these three clusters, phylogenetic analysis identified 10 unique unrelated isolates, all of which were from immigrants from areas in Africa or the Middle East where HCV is highly endemic. The median age of these 10 patients was 48 years (IQR, 38 to 60 years); 7 of them were male and 1 was coinfected with HIV. The isolates included three strains of HCV-4r (Somalia, Ethiopia); two strains of HCV-4k (Kenya, Rwanda); and one strain each of HCV-4d (Eritrea), HCV-4h (Democratic Republic of Congo), HCV-4n (Egypt), HCV-4o (Egypt), and a subtype not previously recognized (Egypt).

To gain additional insight into the spread of HCV-4 over time, we used a molecular clock approach to estimate the year of origin of TMRCA for each of the HCV-4 clusters identified (Table 3). For C1, TMRCA was traced back to 1881 (95% confidence interval [CI], 1833 to 1912), with 42/43 (98%) strains sharing one ancestor after 1909. Viral spread was reflected by the temporal distribution of lineage splits in each cluster. Within C1, 28/42 (67%) lineage splits occurred over the period from 1920 to 1955. Virtually no spread occurred after 1980, suggesting that cluster 1 reflects a random sample of HCV-4a strains imported by immigrants who acquired their infections in Egypt.

In contrast, the majority of patients infected with HCV-4d (C2 and C3) were of Dutch origin. Evolutionary analysis suggests the introduction of HCV-4d in Europe in 1954 (95% CI, 1936 to 1965). However, the exclusion of patients 7, 28, 55, 69, and 81, all of whom were born and/or infected in Italy or Spain and who represent old single branches within C2, suggests that HCV-4d was introduced in southern Europe before it reached The Netherlands somewhat later. On the basis of the temporal distribution of the lineage splits, 70% of the spread of HCV-4d strains among Dutch individuals reporting IDU occurred before 1980, and 90% occurred before 1990. The two percentages indicate a withdrawing epidemic, for which rapid transmission occurred in the past but for which the current incidence is low. MSM C3 originated from the same HCV-4d strain circulating among individuals reporting IDU, probably in about 1985 (95% CI, 1977 to 1989). However, C3 could be further subdivided into three distinct subclusters (C3a, C3b, C3c) that varied in origin from 1992 to 1997. The temporal distribution of the lineage splits of C3 suggests the rapid and ongoing transmission of HCV-4d among MSM, as 90% of the lineage splits occurred after 1990 and even 57% occurred after 2000.