Emerging Variability in HIV-1 Genetics among Recently Infected Individuals in Yunnan, China

Background Yunnan has the longest endured Human Immunodeficiency Virus-1 (HIV-1) epidemic in China, and the genetic diversity of HIV-1 constitutes an essential characteristic of molecular epidemiology in this region. To obtain a more comprehensive picture of the dynamic changes in Yunnan’s HIV-1 epidemic, a cross-sectional molecular epidemiological investigation was carried out among recently infected individuals. Methodology/Principal Findings We sequenced partial gag (HXB2∶781–1861) and env (HXB2∶7002–7541) genes from 308 plasma samples of recently infected patients. With phylogenetic analysis, 130 specimens generated interpretable genotyping data. We found that the circulating genotypes included: CRF08_BC (40.8%), unique recombinant forms (URFs, 27.7%), CRF01_AE (18.5%), CRF07_BC (9.2%), subtype B (2.3%) and C (1.5%). CRF08_BC was the most common genotype, and was predominant in both intravenous drug users (IDUs) and heterosexually transmitted populations. CRF08_BC and CRF07_BC still predominated in eastern Yunnan, but CRF08_BC showed increasing prevalence in western Yunnan. Strikingly, the URFs raised dramatically in most regions of Yunnan. Seven different types of URFs were detected from 12 prefectures, suggesting that complicated and frequent recombination is a salient feature of Yunnan’s HIV-1 epidemic. Among URFs, two BC clusters with distinctive recombination patterns might be potential new CRF_BCs. CRF01_AE was no longer confined to the prefectures bordering Myanmar, and had spread to the eastern part of Yunnan, especially the capital city of Kunming, with a large number of infections in the transient population. The ratios of the main genotypes showed no statistical differences between infected IDUs and heterosexually transmitted infections. Conclusions/Significance The changing patterns of the dominant HIV-1 genotypes in Yunnan indicate the complex evolving dynamic nature of the epidemic. Understanding new trends in molecular epidemiology of HIV-1 infection is critical for adjusting current prevention strategies and vaccine development in Yunnan.


Introduction
Yunnan is located in southwestern China and has 16 prefectures and 129 counties. Geographically, Yunnan borders Myanmar, Laos and Vietnam and is situated along the drug trafficking routes that channel heroin into China from southeast Asia's opiumproducing ''Golden Triangle'' region. In 1989, the first Human Immumodeficiency Virus (HIV) epidemic in China was identified among intravenous drug users (IDUs) in Ruili County of Dehong Prefecture [1]. Since then, Yunnan has been one of the areas most affected by HIV in China. Furthermore, Yunnan serves as a primary entry point for the introduction of different HIV-1 genotypes into China. Thus, Yunnan is considered as an epicenter of the HIV-1 epidemic in China.
In the late 1980s, China's first HIV-1 epidemic among IDUs was initiated by both subtype B (circulating in United States and Europe) and subtype B' (the Thailand variant of subtype B, also referred to as Thai-B) strains imported from Thailand to Yunnan by drug trafficking through Myanmar [2,3,4]. In the early 1990s, another epidemic was introduced into Yunnan by Indian IDUs with subtype C strains (India C) [5,6]. In 1994, subtype E strains (now referred to as CRF01_AE) were identified among commercial sex workers who had returned from Thailand to Yunnan [7]. Over time, the Thai-B subtype overtook the prototypical B subtype in frequency, increasing from 20% of all subtype B in 1990 to 90% in 1996 [2,3,4]. Thus, subtype B' and subtype C strains co-circulated in Yunnan's IDUs in the first half of 1990s. Meanwhile, intravenous drug injection was the predominant transmission route in the early HIV-1 epidemic of Yunnan [1,8]. These circumstances provided the opportunity for recombination between subtype B' and subtype C among the IDU population in Yunnan.
Though CRF07_BC and CRF08_BC were first detected from IDUs in Xinjiang Province and Guangxi Province in 1997, respectively [9,10], it is likely that these two CRFs were initially established in Yunnan Province [11,12] and spread through two different overland heroin trafficking routes: CRF07_BC northwestward to Xinjiang Province, and CRF08_BC eastward to Guangxi Province [13]. Sequence analysis suggests that CRF07_BC and CRF08_BC are closely related and may have evolved from a common parent [14]. However, understanding the origin of the two CRFs in Yunnan required additional retrospective molecular epidemiological investigations.
Subsequently,studies carried out from 2001 to 2006 showed that there were three major groups circulating in Yunnan: C/ CRF07_BC/CRF08_BC, CRF01_AE and B [15,16]; C/ CRF07_BC/CRF08_BC viruses included C, CRF08_BC, CRF07_BC and new BC recombinants. Although distributed widely, CRF08_BC and CRF07_BC largely dominated in eastern Yunnan. However, new BC recombinants were also found in western Yunnan [15,16,17]. CRF01_AE was mainly distributed in the western region bordering Myanmar, in prefectures such as Dehong, Baoshan, Xishuangbanna and Puer [15,16]. Statistically, C/CRF07_BC/CRF08_BC and CRF01_AE were dominant in IDUs and those with sexually transmitted infections, respectively, suggesting that HIV-1 subtype distribution was closely associated with risk factors.
HIV-1 molecular epidemiological surveillance in new infections is of prime importance for understanding the real-time dynamics of the HIV-1 epidemic in a region with complicated HIV-1 subtype prevalence, such as Yunnan. The ''gold standard'' method for measuring recent infections is a prospective cohort study. However, cohort studies are difficult and expensive to implement and prone to biases that could reduce the general applicability of the results. As an alternative method, the BED-capture enzyme immunoassay (CEIA) has been widely used to measure the proportion of HIV-1-specific IgGs among total IgGs in blood samples for the purpose of identifying infections that were acquired recently [18,19]. Because the last large-scale HIV-1 molecular epidemiological study in Yunnan was done between 2002 and 2004 [16], we decided to evaluate the new HIV-1 subtype propagation. In this work, we performed a cross-sectional molecular epidemiological investigation among the recently infected population identified by BED-CEIA to uncover the new prevalent trends of HIV-1 genetic strains in Yunnan. We found that CRF08_BC was the most common genotype in Yunnan, and the distributions of the main HIV-1 genotypes showed no statistical differences between IDUs and the heterosexually transmitted population. Moreover, diverse recombinations emerged rapidly with the multiple genotypes co-existing in Yunnan. HIV-1 infection status was determined by an enzyme immunoassay and confirmed by Western blot assay (HIV BLOT 2.2, MP Diagnostics, Singapore). A total of 308 plasma samples identified as recent infections among 3034 HIV-1-positives were used for genotype analysis. Plasma was separated from whole blood and used to obtain HIV-1 RNA for subsequent analysis. All HIV tests are informed and voluntary. Written consents for HIV testing were obtained, in which the subjects agreed that if they have HIV, their samples can be used in the researches for the purpose of controlling and preventing HIV. Because no human experimentation and no investigation of host genetics were conducted, the authors were exempted from approval by the local ethical review committee at the Yunnan Centers for Disease Control and Prevention.

Study Participants and Sample Collection
HIV-1 Seroconversion Identified with BED-CEIA BED-CEIA was performed according to the manufacturer's instructions (Calypte HIV-1 BED incidence EIA, Calypte Biomedical Corporation, Portland, OR) [20]. Test specimens were initially run singly. If the normalized OD (ODn) was .1.2, the specimen was classified as being from a long-term seroconverter. Specimens with ODn ,1.2 were tested again in triplicate to confirm the values. In confirmatory testing, specimens with ODn values ,0.8 were considered to have undergone recent seroconversion.

Sequence Analysis
The contig assembly of sequences was performed using DNA sequence analysis software Sequencher 4.9 (Gene Codes, Ann Arbor, MI). The ClustalW Multiple alignment and manual editing were performed using Bio-Edit 7.0 software. The reference sequences were obtained from the NIH/NIAID-funded HIV Databases (http://hiv-web.lanl.gov/content/index), covering the major HIV-1 subtypes/CRFs. Some reference sequences which were previously characterized in China and countries surrounding Yunnan were included. Phylogenetic tree analyses were performed using the neighbor-joining method based on Kimura 2-parameter model with 1000 bootstrap replicates, using MEGA version 4.0 [21]. To demonstrate possible intertype mosaicism, candidate sequences were analyzed using the Recombination Identification Program (RIP; version 3.5.1) which is available at the HIV sequence database (http://hiv-web.lanl.gov) using the appropriate parameters. Similarity plot analyses (version 3.5.1; S. Ray, Johns Hopkins University, Baltimore, MD) were further performed using reference strains of subtype A1, subtype A2, subtype B, subtype C, subtype D, subtype F1, subtype J, CRF01_AE and CRF08_BC. The conditions are further descried in the figure legends.

Geographic Distribution Analysis of HIV-1 Genotypes
HIV-1 genetic geographic distribution was analyzed with the public health geographic information system (PHGIS, China CDC). A Dot Density Map was used to display the distribution density of each HIV-1 genotype within Yunnan Province. For each genotype, the number of patients with the genotype in each prefecture was divided by 130 (the total number of patients in this study) to obtain the percentage of each genotype in each prefecture. When using PHGIS to map the data, one dot was defined as 0.025% of the population.

Sequence Data
All the sequences obtained in this study were submitted to GenBank under accession numbers JX263434 to JX263661.

Statistical Analysis
Statistical analyses for this report were conducted using the SPSS 17.0 statistical analysis software package (SPSS Inc. Chicago, IL). Categorical variables were compared using x 2 . All tests were two-tailed and a p value ,0.05 was considered statistically significant.

Demographic Characteristics of Study Subjects
A total of 3034 newly diagnosed HIV-positive samples were collected from VCT, medical case reports and sentinel surveillance in Yunnan Province during the first quarter of 2009. Of these, 308 samples were identified as recent infections, all of which were used for the HIV-1 genetic analysis. For each sample, gag and env genes were amplified and sequenced. In total, we obtained 120 gag sequences and 108 env sequences. Combining the phylogenetic tree analysis of gag and env, we successfully genotyped 130 samples with a rate of 42.2% (130/308). The failure of amplification was most commonly due to poor storage and transportation conditions of samples. The constituent ratios of the 130 subjects genotyped and the total of 308 recently HIV-infected individuals showed no statistical differences by geographical area, gender, age, ethnicity, marital status and infection routes ( Table 1).
As shown in Table 1, among the 130 subjects, the ratio of males to females was 1:0.69. The mean age was 33.
Interestingly, the proportion of the URFs exceeded that of CRF01_AE and ranked second, which suggested that new recombination occurs frequently among the HIV-1 recentlyinfected population. Among the URFs, two types of recombination patterns existed. In pattern one, as mentioned above, mosaic construction was detected in gag or env region ( Fig. 3 and Fig. 4). In pattern two, URFs' gag and env region belonged to different subtypes or CRFs, respectively. As shown in Table 2

Distribution Characteristic of HIV-1 Genotypes by Infection Routes
To better characterize the distribution of HIV-1 genotype, we performed a detailed demographic study. The distribution of genotypes by gender, age, or marital status of patients showed no significant differences (   (Table 2). Finally, the distribution of different types of URFs in different infection routes showed no statistical difference (Table 4).

Geographic Distribution Characteristic of HIV-1 Genotypes
Finally we analyzed the geographic distribution of each HIV-1 genotypes. CRF08_BC was distributed almost throughout the whole province, except Baoshan, Lijiang and Diqing (Fig. 5A). In line with the previous report, this viral genotype predominated in east Yunnan, including 3 highly affected prefectures: Wenshan, Honghe and Kunming. Additionally, the prevalence of CRF08_BC increased in west Yunnan, especially in Dali and Lincang. Conversely, CRF07_BC was mainly distributed in the eastern region including Honghe, Kunming and Qujing (Fig. 5B). Previously, CRF01_AE was limited to the western prefectures (Dehong, Baoshan and Lincang). In this study, we found that CRF01_AE had spread to the eastern prefectures (Kunming, Honghe and Wenshan), and predominated in Kunming, the capital of Yunnan (Fig. 5C). Traditionally, subtypes B and C were confined in Dehong where these two subtypes were first introduced into Yunnan in the late 1980s and early 1990s. Here, we found that subtype B and C still predominated in Dehong, and separate subtype B strains were found in Honghe (Fig. 5D and 5E). The 36 URFs distributed widely, and were in 12 prefectures (Fig. 5F). There were five prefectures with more than one type of URF detected (Dehong, Baoshan, Chuxiong, Honghe and Qujing). In line with the distribution of subtype B and C, BC recombinants predominated in Dehong, while CRF01_AE/C and CRF08_BC/CRF01_AE were not limited to the local area, but were scattered widely.
Except for the three northwest prefectures, Nujiang, Lijiang and Diqing (HIV-1 genetic analysis in Diqing was missing because no recent infector was detected in Diqing), the other 13 prefectures have more than one viral genotype present, and Dehong, Honghe and Kunming showed more complex genetic diversity (Fig. 6).

Discussion
In the present work, we conducted an HIV-1 molecular epidemiological study to disclose new trends in recently-infected individuals of Yunnan Province, the province most severely affected by HIV/AIDS in China. By analyzing the HIV-1 genes gag and env, we described the distribution characteristic of HIV-1 subtypes, CRFs and URFs in this area. We found that CRF08_BC was the most common CRF, which predominated in both east and west Yunnan among IDUs and the sexually transmitted population. Exceeding the frequency of CRF01_AE, URFs became the second most common HIV-1 strain which suggests continuous risk behavior still exists among certain groups. Importantly, the distribution of these genotypes was not closely associated with transmission route.
By the end of 2011, the cumulative number of reported HIV/ AIDS in Yunnan was 95296, accounting for 21.0% of the total national figure. Yunnan has shown more diverse HIV-1 genetics than any other HIV-1 high-prevalent province in China, including Guangxi, Sichuan, Xijiang, Guangdong and Hennan [22,23]. The early reports on characterization of HIV-1 genotypes in Yunnan were confined to the IDU population, which revealed multiple HIV-1 genotypes, including B, B', C, CRF01_AE, CRF07_BC, CRF08_BC and URF [2,3,5,9,11,16,17,24]. During the transition of main transmission route from intravenous injection to sexual contact, diverse HIV-1 strains were also identified among individuals who acquired HIV-1 through sexual contact [15,16,24]. Additionally, phylogenetic analysis showed that sequences from IDUs intermingled with those from individuals infected through sexual contact within each subtype of HIV-1 [24], which suggested HIV-1 strains circulating in the sexually transmitted population might come from those in IDUs through commercial or non commercial sex contacts.
In the early 2000s, CRF08_BC and CRF01_AE were dominant in IDUs and those with sexually transmitted infections, respectively [1,16]. At that time, it was presumed that the prevalence of CRF01_AE would increase with sexual transmission rising [16]. However, in recently infected people, we found that CRF08_BC predominated not only in IDUs, but also in the heterosexually transmitted population. The influx of CRF08_BC into the heterosexually transmitted population may occur through male IDUs who visit female sex workers (FSWs) as well as FSWs who inject drugs. These groups have played a crucial role in the transition of the epidemic from being primarily IDU-driven to sexually driven [25]. Another possible reason is the transmission from HIV-1 positive IDUs to their spouses and regular sexual partners through unprotected sex. Furthermore, the distributions of the main HIV-1 genetic strains showed no statistical differences between IDUs and the heterosexually transmitted population. This suggests that these strains tended to randomly distribute into these two populations, and the phenomenon observed in previous studies [15,16] of different subtypes/CRF predominating in different risk groups appears to have diminished. The changing characteristics of HIV-1 molecular epidemiology suggest that the bridging population is still the primary force behind the development of the HIV-1 epidemic in Yunnan. Thus, measures for effective epidemic control should be further developed and performed among the high-risk groups, particularly IDUs, commercial sex workers and the spouses of HIV-1 positive individuals. Key strategies should include behavior intervention, scaling up identification of HIV-1 infected persons among the high risk groups, and providing antiretroviral therapy (ART) to them to effectively reduce HIV-1 incidence by decreasing the viral load level. Recombination between HIV-1 genotypes is an important mechanism that contributes substantially to the genetic complexity of HIV-1 and may result in establishing epidemiologically important founder strains. The coexistence of multiple genotypes in the same area always causes the formation of CRFs. Typically, CRF07_BC and CRF08_BC have been thought to originate among IDUs in Yunnan [11,12], but how this occurred was not well-understood. This past lack of knowledge may have been the result of the use of the env region only in earlier studies [2,6]. CRF07_BC and CRF08_BC are composed mostly of subtype C and contain a few small segments of subtype B'. Thus, distinguishing between CRF07_BC/CRF08_BC and subtype C based on the env region alone would have been difficult. It has been reported that the nearly full-length genome analyses of the envbased subtype C samples revealed that they were actually BC recombinants [11]. In this study, we found two discrete BC clusters. The sequences in these two clades also clustered together with those of the BC recombinants detected in western Yunnan in the late 1990s and early 2000s [11,17]. Furthermore, each cluster has its own recombination pattern in gag region, which is distinct from those of CRF07_BC and CRF08_BC. All of this suggests that new CRFs_BC might have been established in this area. To prove this hypothesis, the complete sequencing of full-length HIV-1 genomes is required in future research.
Besides CRFs, another striking result is that diverse URFs arise continually. URFs reported before mainly included new BC recombinants, CRF07_BC/CRF08_BC recombinants, and C/ CRF01_AE recombinants [11,16,17,26,27,28,29], most of which were found among IDUs. In this work, we found that the types and quantity of URFs increased among recent infections. The frequency of URFs was exceeded only by that of CRF08_BC. Of the seven types of URFs detected, nearly 70% were identified among heterosexually-infected individuals, and 55.5% were CRF01_AE relative URFs. In Yunnan, CRF01_AE was first found among FSWs [7], and was the preponderant strain in this population in the early stage of the HIV-1 epidemic [1,16]. The high proportion of CRF01_AE relative URFs suggests that active recombination took place through commercial sex behavior. The geographic distribution of URFs is growing and covered 12 out of the 16 prefectures. All of these contributed substantially to the genetic complexity of HIV-1 in Yunnan. The formation and linkage of these URFs will be our research focus in the future, which will elucidate the development of the HIV-1 epidemic of Yunnan, and provide references for HIV control in this area.
Furthermore, the HIV-1 strains showed different temporal and spatial dynamics in the process of spreading. We found that CRF07_BC and CRF08_BC were both highly prevalent in east Yunnan, but CRF08_BC was distributed more widely than CRF07_BC, covering almost the entire province. Presently, we do not know whether the transmission capacities of these two CRFs differ. CRF01_AE spread to east Yunnan, particularly in Kunming where the proportion of CRF01_AE remained at a low   Since the first acknowledged HIV-1 epidemic in China began in Yunnan, this region has become a critical area bridging HIV-1 epidemics in southeast Asia with the subsequent inland epidemic. Through durg trafficking, some genotypes originally found in Yunnan had spread not just to the neighboring provinces of Guangxi (CRF08_BC) [12,31,32,33,34] and Sichuan (CRF07_BC) [35], but also to the north-western province of Xijiang (CRF07_BC) [12,35] and the central province of Henan (subtype B') [36]. On the other hand, the countries bordering Yunnan display their own characteristics of HIV-1 molecular epidemiology. In Vietnam, CRF01_AE is the predominant strain [37]. However, in Myanmar, besides B', C and CRF01_AE, recombinant forms between B', C and CRF01_AE are increasing dramatically [38]. A similar trend is seen in Dehong, where the proportion of URFs is higher than any other subtype/CRF. These suggest that the China-Myanmar border area constitutes a hot spot of active recombination in Asia. The bidirectional transmission of HIV-1 between Yunnan and neighboring countries means that the epidemic of Yunnan could potentially influence neighboring countries. All of this suggests that the main strains prevailing in China and neighboring countries could be found in Yunnan. Thus, comprehensively understanding HIV-1 molecular epidemical characteristics in Yunnan plays an important role in efficient control of the HIV-1 epidemic, as well as vaccine design and evaluation for China and the neighboring countries.
One limitation of this work was the relatively low PCR amplification rate (gag: 39.0%, 120/308; env: 35.1%, 108/308), which also occurrd in our previous work [16]. The reasons may be RNA degradation owing to poor serum storage and transportation conditions. In fact, the plasma specimens had been used for HIV-1 screening, confirmation and BED-CEIA before genetic analysis was performed, during which time the samples underwent several freeze-thaw cycles. Other reasons may include low viral load in some cases and sequence variations at the primer binding sites. The low amplification rate meant that some low-frequent viral genotype might be missed, however it is unlikely to compromise the interpretation of the general genetic distribution. This is because the constituent ratios of the 130 subjects with definite genotyping and the total of 308 recently HIV-infected individuals showed no statistical differences by area of source, gender, age, ethnicity, marital status and infection routes.
In summary, we found that six HIV-1 genotypes circulated in the recently infected individuals of Yunnan, and the distribution of the main genotypes was not closely associated with transmission route. HIV-1 genetics became more diverse because of frequent intersubtypes/CRFs recombinations. These findings showed new trends of HIV-1 molecular epidemiology with dynamic changes in the HIV-1 epidemic and will contribute to a better understanding of the distribution and evolution of HIV-1 in Yunnan as well as help to establish and modify public health efforts to prevent new infections.