Phylogenetic Analysis of HIV-1 CRF65_CPX Reveals Yunnan Province Is Still a Source Contributing to the Spread of HIV-1 in China

Abstract

To the Editors: Yunnan, a south-western province of China bordering the “Golden Triangle” region of Southeast Asia, was considered as an epicenter of HIV/AIDS in China. The first human immunodeficiency virus type 1 (HIV-1) epidemic in China was detected among injecting drug users (IDUs) in Ruili City in the Dehong prefecture of Yunnan province in 1989.1 The HIV-1 epidemic in Yunnan was initiated with both subtype B and subtype B′ (the Thailand variant of subtype B), with subtype B′ later becoming the most prevalent isolate.2 In the early 1990s, subtype C strain closely related to Indian isolates was also identified and subsequently became main circulating strain among IDUs.3 The cocirculation of subtype B′ and C led to the generation of CRF07_BC and CRF08_BC, which subsequently spread outside Yunnan province. Drug trafficking activities in conjunction with local heroin use along the trafficking routes have been implicated in the spread of CRF07_BC and CRF08_BC across China.4 It was also suggested that subtype B′ epidemics among former plasma donors and heterosexuals in Mainland China were originated from a single founding B′ strain that had been circulating among IDUs in Yunnan province.5 Therefore, in the early stage, Yunnan was the main source contributing to the genesis and spread of HIV-1 strains in China. Besides CRF07_BC and CRF08_BC, 7 new circulating recombinant forms (CRFs) have been identified in China recently.6–126–126–126–126–126–126–12 Notably, 5 of 7 new CRFs were first identified in Yunnan, suggesting this region is a hotspot for HIV-1 recombination. CRF65_CPX, the first complex mosaic HIV-1 CRF in China, was identified by Feng.7 It showed a complex genome organization composed of parental subtypes B′, C, and CRF01_AE, with a total of 13 recombinant breakpoints in its genome. Three CRF65_CPX near full-length genome (NFLG) sequences (YNFL01, YNFL02, and YNFL05) were derived from 3 epidemiologically unlinked individuals infected through a heterosexual contact in Yunnan province. One NFLG sequence (ANHUIHF104), sampled in 2011 from a man who had sex with men (MSM) in Anhui province, was determined by extensive search of the Los Alamos HIV database (www.hiv.lanl.gov). These 4 NFLG sequences were representative sequences of CRF65_cpx. In addition to the 4 representative sequences of CRF65_cpx identified by Feng, we also determined 7 gag-pol CRF65_CPX sequences, 6 from our previous sequence data, and 1 from the Los Alamos HIV Sequence Database. This study was supported in part by the AIDS High Risk Cohort Program, in You'an Hospital, Beijing, China. This is a longitudinal prospective cohort to study the incidence and prevalence of HIV-1 among MSM in Beijing. All the participants in this cohort were assessed for HIV-1 antibodies and HIV-1 RNA every 2 months until HIV-1 antibodies or HIV-1 RNA showed positive results. Early HIV-1 infection was defined as (1) positive HIV-1 RNA with a negative or indeterminate HIV-1 antibody test, followed by HIV seroconversion within 6 months, or (2) a negative enzyme-linked immunosorbent assay and Western blot less than 180 days before a documented positive enzyme-linked immunosorbent assay or Western blot. If a patient was found to be infected at the beginning, he was assigned to the chronic HIV infection group.13 Three gag-pol CRF65_CPX sequences (sequence names: 2010CNBJ255, 2012CNBJ381, and 2012CNBJ392) were retrieved from 133 MSM in Beijing between 2010 and 2012. All the 133 patients met the above criterion and it could be defined as early HIV-1 infection. Despite its rarity, accounting for only 2.3% among MSM with early HIV-1 infection in Beijing, CRF65_CPX showed evidence of circulation in Beijing MSM. Two CRF65_CPX sequences (sequence names: 2010CNBJ022C, 2010CNBJ023C) were also obtained from 2 MSM with chronic HIV-1 infection in Beijing from the AIDS High Risk Cohort Program. One CRF65_CPX sequence (sequence name: 2011CN214C) was retrieved from a patient with chronic HIV-1 infection with the geographic region and transmission route unavailable. The 6 gag-pol CRF65_CPX sequences are available under GenBank accession numbers KP698503-KP698508. Apart from the above 6 CRF65_CPX sequences, we also identified 1 sequence (sequence name: 156, accession numbers: JQ898164 for gag, JQ898223 for pol) from the Los Alamos HIV Sequence Database. It was derived from a 34-year-old male IDU in Yunnan in 2009, who was confirmed as HIV-1 positive in 2004, indicating that CRF65_CPX might have existed in Yunnan for a long time. In short, CRF65_CPX strains have been detected in different regions and different high-risk groups of China. In other words, CRF65_CPX is another CRF, which was identified in China and did cause epidemics, like CRF07_BC, CRF08_BC, and CRF55_01B.14 Phylogenetic analysis was used to clarify the epidemiological relationship of CRF65_CPX strains in China. jModelTest was used to find the best-fit nucleotide substitution model,15 and the general time reversible model plus a gamma distribution among site rate heterogeneity was chosen in both phylogenetic maximum likelihood and Bayesian analyses on the basis of Akaike information criterion and Bayesian information criterion. PhyML 3.0 was used to estimate a maximum likelihood phylogenetic tree under GTR + G4 nucleotide substitution model. Bayesian analyses were performed with the selected nucleotide substitution model using a Bayesian uncorrelated lognormal relaxed molecular clock method as implemented in BEAST v1.7.5. The analyses were performed using 4 parametric models (constant size, exponential growth, expansion growth, and logistic growth) and 1 nonparametric model (Bayesian skyline model). As is shown in the maximum likelihood tree (Fig. 1A), all the 11 CRF65_CPX sequences formed a monophyletic cluster with high bootstrap support. Four CRF65_CPX sequences from Yunnan were located near the root of this cluster whereas 6 CRF65_CPX sequences from Beijing and Anhui MSM formed a small monophyletic cluster with 97% bootstrap support and positioned inside the CRF65_CPX cluster. Moreover, the mean genetic distance of Yunnan viruses (2.4%) was larger than that of Beijing and Anhui MSM viruses (1.4%), suggesting that CRF65_CPX viruses probably hit Yunnan earlier than other regions and continued to evolve over time. The maximum clade credibility tree (Fig. 1B) corroborated topology structure of CRF65_CPX as described in the maximum likelihood tree. The robustness of monophyletic nature of CRF65_CPX viruses from Beijing and Anhui MSM was supported by high posterior probability (P = 1.0). The time to the most recent common ancestor of CRF65_CPX was inferred to be 2000.4 (95% highest posterior density: 1997.1–2003.7), indicating the origin time of CRF65_CPX. The time to the most recent common ancestor of monophyletic CRF65_CPX viruses from Beijing and Anhui MSM was dated to 2007.3 (95% highest posterior density: 2005.8–2008.4). This supports that CRF65_CPX initially originated in 2000.4 in Yunnan province and spread to the MSM population in other regions of China about 7 years later.FIGURE 1: Phylogenetic analysis of CRF65_CPX sequences. A, Phylogenetic maximum likelihood tree of CRF65_CPX. Maximum likelihood phylogeny was estimated from HIV-1 nucleotide sequences for the gag-pol region (HXB2: 790–4259) by PhyML 3.0. Reference sequences were retrieved from the Los Alamos HIV Sequence Database. The outgroup (O.CM.91.MVP5180) branch-length is 5 times reduced to facilitate visualization. Sequences are indicated by symbols corresponding to their sample region: Yunnan (square), Beijing, and Anhui MSM (circle). The scale bar represents 2% genetic distance. The bootstrap values of the 100 replicates are labeled on the branches. B, Maximum clade credibility tree of CRF65_CPX. The maximum clade credibility tree was constructed by Bayesian Markov chain Monte Carlo analysis based on HIV-1 nucleotide sequences (HXB2: 1600–3483, subtype C origin) implemented in BEAST v1.7.5. Subtype C sequences of India, China, and Myanmar were obtained from the Los Alamos HIV Sequence Database. The analysis was performed using an uncorrelated lognormal relaxed clock model in GTR + G4 nucleotide substitution model under a constant coalescent model. The Markov chain Monte Carlo analysis was run for 100 million generations, with sampling every 8000 generations. The CRF65_CPX tree branches were colored according to their respective geographical locations.CRF65_CPX is a new CRF circulating among MSM in China after CRF01_AE, CRF07_BC, CRF55_01B, and CRF59_01B.12,14,1612,14,1612,14,16 Although only a few CRF65_CPX strains retrieved from MSM have been identified in only 2 cities so far, it did cause an essential epidemic among MSM with early infection in Beijing reaching a prevalence of 2.3%. Unprotected sex, having multiple sexual partners, and limited knowledge of HIV make MSM significantly more vulnerable to infection. Furthermore, the high mobility presented by the MSM population may facilitate the rapid dissemination of CRF65_CPX across China. Consequently, we can forecast that more and more CRF65_CPX strains will be detected among MSM in China in the near future. Unlike CRF55_01B, which was generated among MSM in Shenzhen and subsequently spread to other high-risk groups and regions of China,14 CRF65_CPX was not generated among MSM but was spread to MSM after its generation in Yunnan. It reveals that Yunnan province is still serving as a source contributing to the spread of HIV-1 in China nowadays, like its role in the spread of CRF07_BC, CRF08_BC, and B′ in the early phase. As an epicenter of HIV/AIDS in China, Yunnan was characterized by circulation of multiple subtypes of HIV-1, which resulted in a significant prevalence of intersubtype recombinants in this region. A molecular survey, conducted in the first quarter of 2009, has indicated that unique recombinant forms (URFs) accounted for 27.7% of recently infected individuals and became the second most common HIV-1 strains in Yunnan.17 Under some circumstances, both the newly identified CRFs and various URFs in Yunnan may spread from Yunnan to other regions of China and contribute to the increasing complexity of HIV-1 in China like CRF65_CPX. The complex genetic structure of HIV-1 epidemics can be seen in China, where various HIV-1 strains are cocirculating within and among different risk populations.18 Genetic subtyping combined with phylogenetic analysis has enabled researchers to estimate the prevalence and time trends of subtypes, to confirm any risk factors associated with the increasing prevalence of certain subtype, to phylogenetically investigate the origin and spread of HIV-1 strains, and to identify new CRFs and URFs in China. Our results highlight the importance of monitoring the molecular epidemiological HIV-1 diversity and the need for implementation of effective preventive measures to reduce HIV transmission in China, especially in Yunnan province. ACKNOWLEDGMENT The authors acknowledge support from You'an Hospital, Beijing, for allowing the authors to gain access to the sequences in this letter.