Serendipitous identification of natural Intergenotypic recombinants of hepatitis C in Ireland

Hepatitis C virus infects approximately 170 million individuals' world wide [1]. Chronicity develops in 50–80% of infections [2, 3]. Hepatitis C exists as a family of viruses, divided into 6 genotypes each with multiple subtypes [4]. The management and treatment of chronic hepatitis C virus [HCV] is in part guided by the genotype of the infecting virus [5, 6]. The requirement for liver biopsy can be guided by genotype. Current therapeutic options of pegylated interferon and ribavirin have a population efficacy of only 50% [7]. Genotype 2 and 3 respond with efficacy of 80% in clinical trials. Duration of therapy is determined by genotype. Individuals infected with either genotype 1 or 4 usually have to undergo treatment for 48 weeks, while genotypes 2 and 3 usually have treatment duration of 24 weeks. There are several methods which can be used to determine viral genotype among which, (a) DNA sequencing of amplicons from the 5'untranslated region [5'UTR] (b) restriction length fragment polymorphism analysis and (c) Reverse Line Probe Hybridization [RLPH] [8]. Determination of viral genotype using RLPH can readily identify rare mixed genotype infections [9]. On occasion non-specific reactivity between probes and amplicons can give an indication of an apparent mixed genotype infection. However, upon repeat with additional specimens these ambiguities usually resolve to a single genotype infection. RLPH may infrequently yield ambiguous patterns which required further investigation. It may be possible to discriminate between very similar isolates by extension of the analysis to additional regions of the genome [10]. Alternatively, there are genotype specific motifs present within NS5B which can then be used to delineate strain genotype.

Following routine RLPH of specimen HC9A98987 it was noticed that an unusual banding pattern indicative of a putative genotype 2 and 4 mixed infection was evident [Figure 1]. This interpretation is based on the fact that the banding at positions 9, 10, and 11 are indicative of a genotype 2a/c. However, a specimen of genotype 2a/c can also have a pattern of 5, 9, 10 and 11. In addition, a banding pattern of 6 and 16 is usually indicative of a genotype 4a, also, 5 and 16 is indicative of a specimen of genotype 4e. There is no banding pattern to support the existence of a genotype 1 within this isolate. The genotype determination was repeated on several occasions for this specimen and on an additional specimen from the same patient (results not shown) all of which yielded the same banding pattern. The DNA sequence analysis of the 5'UTR indicated that the specimen was of genotype 2, GenBank [DQ417427], with no definable subtype or indication of a mixed genotype infection with a strain of genotype 4. In light of this fact, clonal analysis of the NS5B region from HC9A98987 was undertaken. 14 clones were sequenced [DQ417439–DQ417452]. A search for similarities with previously published sequences was performed with the program Blast. [11] The results indicated that this region of the genome had homology to sequences of HCV 1b [e.g., D50485, AF31916] and not the expected genotype 2 or 4. The corresponding amino acid sequences analysis indicated that there were only 7 unique isolates, GenBank [DQ417439, DQ417443, DQ417446, DQ417447, DQ417449, DQ417451, DQ417452].

Subsequent clonal analysis of (i) the core sequences [DQ417428–DQ417433] and (ii) the HVR1 sequences [DQ417434–DQ417438] indicated that these regions of the genome were in fact homologous to viruses within the 2k subtype [e.g., AB031663]. Two clones from the core region had synonymous mutations [DQ417430, DQ417433]. The HCV 2k [e.g., AB031663] isolate has a relatively high prevalence in the countries of the former USSR. Kalinina et al reported in 2002 the identification of a natural intergenotypic recombinant form of HCV which had homology to 2k at the 5' end of the genome and to 1b at the 3' end of the genome [12]. The cross over point was identified as occurring within the NS2 region by Simplot by Kalinina [12].

Following these observations, the core, the HVR1 and the NS5B amino acid sequences from HC9A98987 were aligned against reference strains of different genotypes [AY587845 (2k/1b), AB031663 (2k), AF313916 (1b) and D50485 (1b)] and a phylogenetic analysis was carried out. Phylogenetic trees were constructed with the NEIGHBOR program in the PHYLIP package and genetic distances were calculated with PRODIST program in the PHYLIP package based on Kimura's distance. [13] The results indicated clearly that sequences corresponding to specimen HC9A98987 are related to genotype 2k in the core and HVR1 region (data not shown) and closely related to genotype 1b in the NS5B region [Figure 2]. All of these three regions are highly homologous with AY587845 (2k/1b) the recombinant form initially identified in Russia [12].

To investigate if specimen HC9A98987 was a possible recombinant form of HCV, primers were designed with the Primer3 program [14] to span the previously identified recombination point. A product of the expected 346 bp [3168–3413, reference strain [AY587845] (2k/1b)] was recovered by nested RT-PCR [Table 1]. The amplicon was cloned and the DNA sequence from 5 clones examined. Four of the 5 clones were found to be identical at the level of amino acid sequence [DQ839386–DQ839389] with a minor quasispecies [DQ839390]. Sequence analysis indicated that this region of specimen HC9A98987 is homologous with the reference strain [AY587845] [Figure 3]. The recombinant form identified here likely shares the same cross over point as the RF_2k/1b reported by Kalinina [12]. This is the first independent identification of this natural intergenotypic recombinant of HCV between 2k/1b.

Following the identification of HC9A98987 as a recombinant form of HCV similar to AY587845 (2k/1b), the laboratory investigated whether an additional specimen (hence, HC9A99966) from another individual with the same RLPH pattern was also a candidate natural intergenotypic recombinant form. Clonal analysis of the NS2 region was performed on this specimen from the second patient (HC9A99966). The sequences of six clones were determined. All six clones were of identical nucleotide sequence [DQ904446] and were found by sequence analysis to be homologous to the natural intergenotypic recombinant related to HC9A98987, i.e., AY587845 (2k/1b) [Figure 3].

While, the prevalence of the 2k/1b recombinant form of HCV is globally low, its existence poses an interesting challenge as to what anti-viral regimen should be followed. Mechanistically the identification of a recombinant strain of HCV, a single stranded RNA virus, poses some interesting questions. The initial recombination event had to happen in a hepatocyte infected with a 2k and a 1b isolate. Recombination had to maintain the reading frame of the polyprotein. The majority of recombination events would be expected to yield fatal nucleotide lesions generating life cycle incompetent virions. It is interesting to hypothesize that perhaps the generation of this recombinant form of HCV is the result of "strand switching" during replication rather than classical recombination-ligation of viral RNA as previously proposed [15].

An additional natural intergenotypic recombinant between genotypes 2i and 6p has very recently been described [16]. The recombination break point was again narrowed down to a region at the 3' end of the NS2 gene, possibly extending into the NS3 gene. It is currently evident that the viable recombination occurs infrequently in HCV biology. The frequency of this phenomenon may be under estimated. In an attempt to estimate the prevalence of HCV recombinants Cristina and Colina recently reported the identification of an intragenotypic recombinant between 1a and 1c by examining only 89 full length sequences, representative of all genotypes present in the LANL database [17, 18]. These results may give an insight into the actual prevalence of HCV recombinant forms but would require prospective examination to verify the findings. Molecular characterisation of these natural chimeras may yet yield some unique insights into life cycle of hepatitis C. Recombinant forms for HIV do not have the same sensitivity to anti-retroviral therapy as wild type HIV-1 clade B isolates [19]. The sensitivity of recombinant forms of HCV to pegylated interferon based therapy is unknown. Intergenotypic recombinant isolates that have genotype 1 as part of the genomic backbone will likely require a conservative approach, with patients being treated for 48 rather than 24 weeks.

The diagnostic standard for determining HCV genotype, irrespective of methodology, is to interrogate the sequence of the 5'UTR. The serendipitous identification of two recombinant forms of HCV, related to the Russian 2k/1b recombinant form, in Ireland was the result of a detailed molecular investigation across the genome. The RLPH pattern observed in this study may be a signature indicative of this particular type of intergenotype recombinant of hepatitis C meriting clonal analysis of NS2. Perhaps diagnostic laboratories need to consider confirmation of genotype by interrogation of the NS5B to ensure that management of chronic HCV is optimal.