First evidence of a HIV-1 M/O recombinant form circulating outside Cameroon

HIV-1 are now classified into four groups: M (major), O (outlier), N (non-M – non-O), and a new group P, the prototype of which is the divergent strain recently characterized in a Cameroonian patient [1]. Recombination between HIV strains is a very frequent phenomenon, greatly increasing genetic diversity, particularly for group M. HIV-M and HIV-O strains co-circulate in west-central Africa and dual infections have been reported in some individuals [2–4], leading to the emergence of M/O recombinants despite the great genetic divergence between these groups. Only three recombinant strains have been characterized to date in unrelated patients with dual infections, all in Cameroon [3,5,6]. The transmissibility of these rare recombinants and their ability to circulate remain unknown. We report in the present article the first M/O mosaic strain circulating outside Cameroon and potentially corresponding to direct transmission of a well established recombinant form. In May 2008, a 25-year-old Cameroonian woman (RBF208) living in France attended a consultation at Reims University Hospital (north-east France). She had been diagnosed with HIV infection in 2003 in the Paris suburbs, but the initiation of highly active antiretroviral therapy (HAART) was delayed because this patient was lost to follow-up a few months after diagnosis. On arrival at Reims Hospital, this patient was thus antiretroviral drug-naive and presented with severe immunodepression (CD4 cell count = 6 cells/μl) and AIDS, but no detectable plasma viral load (pVL) in the HIV-M-specific Cobas TaqMan test (Roche Diagnostics, Meylan, France). This immunovirological discordance was investigated with the plasma samples available, dating from May 2008, with two in-house Env serotyping assays, to identify the group to which the strain belonged [7,8], RNA quantification tools (RealTime HIV-1 assay, Abbott Molecular, Abbott Park, Illinois, USA; in-house HIV-O-specific technique [9]; and the Generic HIV viral load assay [Biocentric, Bandol, France] specific for HIV-M [10]), and group-specific PCRs targeting the Pol and Env regions [4,11,12]. The near full-length genome of strain RBF208 was amplified and sequenced, from RNA and proviral DNA, with procedures adapted from [1]. Env serotyping assays revealed a typical HIV-M infection, with no reactivity for group O-specific antigens. Conversely, pVL was 4.8 log copies/ml and 5.1 log copies/ml with the RealTime HIV-1 and in-house HIV-O-specific assays, respectively, whereas no virus could be detected in the plasma with the Generic HIV and Cobas Taqman tests, consistent with a group O strain. Using group-specific PCRs, only the HIV-O Pol region [protease, reverse transcriptase (RT) and integrase] was detected in plasma and peripheral blood mononuclear cells. The Env/Gp41 region was first not amplified by any classical group M or O primers, but with alternative primers, revealing a HIV-M to HIV-O switch in this region. The characterization of the near full-length genome showed a 5′-O-M-O-3′ recombinant form, with a first O/M breakpoint in the first codon of Vpr (3' end of Vif) and a second M/O breakpoint towards the 3′ end of Gp41 (Fig. 1a, b). Phylogenetic analyses confirmed that the Pol (integrase) and Env (Vpr-to-Gp120) regions, belonged, respectively, to group O (Fig. 1c) and to group M (subtype D, Fig. 1d), and that strain RBF208 was unrelated to the other M/O recombinants previously described. The full-length sequences obtained from plasma RNA and intracellular DNA displayed an identical mosaic pattern.Fig. 1: Phylogenetic and recombination analyses of the RBF208 strain (accession no. GQ351296). We used MEGA and SimPlot softwares for phylogenetic and bootscan analyses. Sequences representative of the closest strains of group M and O identified by HIV BLAST (http://www.hiv.lanl.gov/content/sequence/BASIC_BLAST/basic_blast.html), and of the three recombinant forms previously described were used [3,5,6]. (a) Mapping of the Vpr and Env breakpoints on the near full-length genome using HIV recombinant drawing tool (http://www.hiv.lanl.gov/content/sequence/DRAW_CRF/recom_mapper.html). (b) Bootscan analysis of the near full-length genome (8561 bp) compared with the sequence of HIV-1 group M HXB2 and with the closest group M and O sequences identified by HIV BLAST. (c) Phylogenetic trees of the integrase (411 bp) and (d) the Vpr-to-Gp120 regions (1941 bp), belonging respectively to group O and to group M subtype D. One thousand bootstrap replicates were performed to assess the reliability of the branching order; bootstrap values are shown only when significant (>70).Fig. 1: (Continued)This first characterization of an M/O intergroup recombinant circulating outside Cameroon was based on discordances between pVL quantification techniques, leading to the suspicion of an HIV-O infection and paradoxical Env serotyping reactivities indicating HIV-M infection. The lack of RNA and DNA detection with group-specific primers binding to the Gag-to-Vif region for HIV-M and the Vpr-to-3'Env region for HIV-O suggested the absence of putative parental strains. These results, supported by the lack of reactivities against group O Env antigens, suggest an infection with a transmitted recombinant. HIV recombination results from the RT switching between two different copackaged RNA templates, at non random breakpoints in the HIV-1 genome dependent on local sequence identity [13]. Interestingly, our results indicate the involvement of the Vpr region in the recombination pattern, as in two of the three M/O forms described to date, suggesting the presence of a recombination hotspot in this region. These recombinants may combine the inherent biological properties of both groups, resulting in higher level of fitness than for either of the individual parental strains, as previously suggested [5]. In particular, the natural resistance to nonnucleoside reverse transcriptase inhibitors (NNRTIs) described for HIV-O may have important consequences, leading to a risk of treatment failure. At the population level, the widespread use of NNRTIs in central Africa might result in the selection of resistant strains containing HIV-O sequences in the Pol region. In our case, resistance genotyping before treatment revealed the presence of a natural 181C mutation in the RT region. The treatment choosen, begun in May 2008, contained a protease inhibitor [Truvada (Gilead Sciences, Inc., Foster City, California, USA) + Kaletra (Abbott Laboratories)], and resulted in a rapid and sustained virological response (pVL undetectable), together with immune restoration (data not shown). In conclusion, discordances between clinical status, serological and molecular results should always be considered carefully and genotyping should be carried out before the initiation of HAART, to adapt monitoring and therapeutic management appropriately. This report indicates that surveillance with suitable tools can be used to detect these rare forms and for evaluation of the epidemiological impact of dual M/O infections and emerging recombinant forms. Acknowledgements We thank the Institut de Veille Sanitaire (InVS) and the Agence Nationale de Recherche sur le SIDA et les Hépatites virales (ANRS) for financial support, the technical staff at Rouen University Hospital and Dr Elisabeth Estrangin (CHI Créteil, France), for providing us with retrospective biological data for the patient.