HIV-1 drug-resistant mutations and related risk factors among HIV-1-positive individuals experiencing treatment failure in Hebei Province, China

Background To understand HIV-1 drug resistance in 11 prefectures of Hebei Province, China, we implemented a cross-sectional HIV-1 molecular epidemiological survey. Methods Blood samples were collected from 122 newly diagnosed drug-naïve HIV-1-positive individuals and 229 antiretroviral therapy (ART)-failure individuals from 11 prefectures in Hebei Province, China. Patient demographic data were obtained via face-to-face interviews using a standardized questionnaire when blood samples were collected. Genotyping of HIV-1 drug resistance (DR) was implemented using an in-house assay. Results In this study, the overall prevalence of HIV-1 DR was 35.5%. The prevalence of HIV-1 DR in participants experiencing treatment failure and ART-naïve participants was 51.9 and 5.9%, respectively. Mutations in protease inhibitors, nucleoside reverse transcriptase inhibitors (NRTIs), and non-NRTI (NNRTIs), as well as dual and multiple mutations were extensively seen in participants experiencing treatment failure. The proportions of NNRTI mutations (χ2 = 9.689, p = 0.002) and dual mutations in NRTIs and NNRTIs (χ2 = 39.958, p < 0.001) in participants experiencing treatment failure were significantly higher than those in ART-naïve participants. The distributions of M184V/I and M41L mutations differed significantly among three main HIV-1 genotypes identified. Viral load, symptoms in the past 3 months, CD4 counts, transmission route, and the duration of ART were found to be associated with HIV-1 DR. Conclusions Our results suggest that new prevention and control strategies should be formulated according to the epidemic characteristics of HIV-1-resistant strains in Hebei Province, where antiretroviral drugs are widely used. Electronic supplementary material The online version of this article (doi:10.1186/s12981-017-0133-3) contains supplementary material, which is available to authorized users.


Background
Human immunodeficiency virus (HIV) epidemics can be traced back to the 1920s in Kinshasa, the capital of the Democratic Republic of the Congo [1]. Among the first HIV-1 individuals in China were four hemophiliac patients in 1985 [2]. Some early cases of HIV infection were linked to imported blood products [3]. In 1989, an HIV outbreak occurred among injection drug users (IDUs) in Yunnan Province, China [4]. Since then, individuals with HIV or AIDS have been successively identified in provinces of mainland China [5][6][7], and an estimated 740,000 individuals in China are currently thought to be infected with HIV/AIDS [8]. Over the past 30 years, the most common route of transmission of HIV-1 infection in China has shifted from blood products to sexual contact [9], and the genetic diversity has rapidly increased because of HIV-1 gene hypermutability [10].
Hebei Province, China comprises 11 prefectures, surrounds the cities of Beijing and Tianjin, and neighbors Henan Province to the south. In 2014, it was inhabited by more than 73 million people [11]. The first case of HIV infection in Hebei Province was detected in Shijiazhuang in 1989 [12]. In the 1990s, local HIV outbreaks occurred in Xingtai and Langfang, and many individuals infected with HIV-1 through blood transmission were identified [13,14]. More recently, HIV infection has been detected in all 172 counties of Hebei Province, and sexual exposure, especially in men-who-have-sex-with-men population, has gradually replaced blood transmission as the most common transmission route [15]. By the end of 2014, a total of 5315 HIV/AIDS cases had been reported, including 3050 HIV-1-positive individuals and 2265 AIDS patients. The HIV/AIDS infection rate in Hebei was estimated to be 0.011%, which is significantly lower than the 0.059% reported for the whole of China and the 0.8% worldwide [16], representing a low HIV/AIDS epidemic.
Before 2002, it was not practical to use antiretroviral therapy (ART) in China due to a lack of drug access, and HIV-1 drug-resistant strains were rare [17]. Since 2003, the central government has provided free ART to HIV/ AIDS patients, and first-line regimens are commonly used in Hebei. By the end of October 2014, 167 of 172 counties in Hebei had carried out the "four free, one care" policy [18], and a total of 2893 HIV/AIDS patients received highly active ART. This represented a large increase in ART coverage, from 9.9% in 2003 to 96.6% in 2014, which coincided with a significant decrease in HIV/ AIDS patient mortality from 11.6 to 2.6% [16]. However, with the increase in antiretroviral drug use, the frequency of adaptive mutations in HIV-1 has also increased, generating drug-resistant strains [19]. This has created severe clinical and epidemiological problems [20].
The objective of the present study was to perform a detailed analysis of the prevalence and genetic mechanisms of HIV-1 drug resistance (DR) among participants experiencing treatment failure in Hebei, and to evaluate the underlying influencing factors associated with the development of HIV-1 drug-resistant strains.

Participants
Between October 2012 and April 2013, 351 whole blood samples were collected from 122 newly diagnosed drugnaïve HIV-1-positive individuals confirmed in 2012 and 229 participants experiencing treatment failure in 11 prefectures of Hebei (Fig. 1). We selected participants experiencing treatment failure according to the following criteria: (1) viral load (VL) ≥1000 copies/ml, (2) duration of therapy >6 months, (3) CD4 count lower than the level before ART, and (4) genotyping had not been previously performed. The local centers for disease control and prevention were responsible for the delivery of antiretroviral drugs and sample collection. Controls were 122 newly diagnosed HIV-1-positive individuals who had not received treatment. The study design was cross-sectional.
Demographic data were collected via face-to-face interviews when blood samples were collected, using a standardized questionnaire. A total of 50 µl of whole blood was used to measure the CD4 count using a FACSCount reagent kit (Becton-Dickinson, Franklin Lakes, NJ, USA). Plasma samples were obtained by centrifuging whole blood, and used to detect VL with the COBAS TaqMan 48 analyzer (Roche, Basel, Switzerland).

HIV-1 genotyping and drug resistance
HIV-1 RNA was extracted from 500 µl of blood plasma using the High Pure Viral RNA kit (Qiagen, Valencia, CA, USA All original pol sequence fragments were assembled, edited, and aligned as previously described [21], and used to construct an HIV-1 pol phylogenetic tree using the neighbor-joining method with 1000 bootstrap replicates, based on the Kimura 2-parameter Model (MEGA5.0). The online jpHMM Program (http://jphmm.gobics.de/ submission_hiv.html) and RIP 3.0 (http://www.hiv.lanl. gov/content/sequence/RIP/RIP.html) were used to further analyze the possible intertype mosaicism of unique recombinant forms (URFs). Finally, HIV-1 pol sequences were submitted to the HIV DR database (http://hivdb. stanford.edu/) to analyze HIV-1 DR mutations.

Statistical analysis
Statistical analyses were implemented using SPSS software version 21.0 (SPSS Inc., Chicago, IL, USA). Means or frequencies of demographic data (such as age, CD4 counts, and VL) were calculated. Categorical variables were analyzed using the Chi square test. When more than 20% of cells had an expected count of <5, Fisher's exact test was used. Multivariable logistic regression analysis was used to identify risk factors associated with DR. A stepwise approach was used for variable selection in the multivariate regression model. All tests were twosided, and a statistical result was considered significant when p < 0.05. Table 1 shows the demographic characteristics of participants. The sex ratio of males to females was 1:0.27. The median values of age, CD4 counts, and VL were 37.0 (range 6-71) years, 220 (range 2-1149) cells/μl, and 4.2 (range 3-6.8) log RNA copies/ml, respectively. Sexual Among all therapy regimens in 214 participants experiencing treatment failure (Fig. 2), the 3TC + AZT + NVP regimen was the most frequent, accounting for 59.3%. The percentage of participants treated with 3TC + D4T + NVP, 3TC + TDF + LPV/r, 3TC + AZT + EFV, 3TC + TDF + EFV, 3TC + D4T + EFV, and 3TC + TDF + NVP was 11.2, 10.3, 9.3, 4.2, 2.8 and 2.8%, respectively.

HIV-1 drug-resistant mutations in participants experiencing treatment failure
Compared with the low prevalence of HIV-1 DR in ARTnaïve controls, 51.9% (111/214) of participants experiencing treatment failure showed resistance to at least one antiviral drug. Mutations in PIs, NRTIs, and NNR-TIs, and dual and multiple mutations were common in participants experiencing treatment failure. As shown in Table 2, the mutation classes showed significant differences in frequency between ART-naïve participants and participants experiencing treatment failure (p = 0.014). The proportions of NNRTIs mutations (χ 2 = 9.689, p = 0.002) and dual mutations in NRTIs and NNRTIs (χ 2 = 39.958, p < 0.001) in participants experiencing treatment failure were significantly higher than those in ART-naïve participants. Furthermore, dual mutations in NRTIs and NNRTIs were the most common mutation class in participants experiencing treatment failure, accounting for 29.4% (63/214), followed by NNRTI mutations (10.7%, 23/214).   Table 4), accounting for 9.9% (11/111) of participants identified as DR. As shown in Table 4, the mean therapeutic duration of the 11 participants with TAMs was 42.8 (range 10-113) months, the mean VL was 4.3 (range 3.2-5.3) log copies/ml, and the mean CD4 count was 108.9 (range 7-187) cells/μl. Sexual transmission accounted for 90.9% (10/11) of cases, with heterosexual transmission accounting for 81.8% (9/11). TAMs were distributed in five CRF01_AE strains and six subtype B strains.

The distribution of HIV-1 DR mutations among different genotypes
As shown in Table 5, there was no significant difference in the overall distribution of 15 main mutations in the RT coding region in CRF01_AE, subtype B, and CRF07_BC (p > 0.05). These 15 mutations largely resided in CRF01_AE and subtype B. Mutations T74S and M46L in the PI coding region were found in CRF01_AE. However, the distributions of M184 V/I (χ 2 = 7.289, p < 0.05)    and M41L (p < 0.05) were significantly different among CRF01_AE, subtype B, and CRF07_BC, respectively.

Factors associated with HIV-1 drug resistance
As listed in Table 6, 14 potential risk factors were considered in the analysis of univariate logistic regression. Of these factors, VL, symptoms in the last 3 months, CD4 count, transmission route, duration of ART, and genotype were clearly related to HIV-1 DR (p < 0.05). To identify risk factors associated with HIV-1 DR, multivariable logistic regression analysis was implemented using stepwise selection. Five factors were found to be significantly associated with the progress of HIV-1 DR in participants experiencing treatment failure: transmission route (compared with sexual contact, blood: odds ratio (OR) 0.

Discussion
Following the phylogenetic analysis of HIV-1 pol sequences in the present study, we successfully identified two HIV-1 subtypes, four CRFs, and two URFs in 11 prefectures of Hebei Province, China. The HIV-1 genotype distribution was shown to be closely related to the route of transmission. Moreover, the prevalence of HIV-1 genotypes in this study differs significantly from that in Sichuan, Yunnan, and Xinjiang provinces, where IDUs are the common high risk group [10], suggesting that the prevalence of HIV-1 genotypes in different provinces of China reflects the geographical difference of HIV-1 highrisk populations. Traditionally, HIV-1 subtype B was dominant in contaminated blood in the cities of Langfang and Xingtai [14,22], and our work provides new evidence to support this. CRF01_AE strains in China were identified in IDUs for the first time in Yunnan [4]. Since the first CRF07_BC epidemic in 2002 [23], the prevalence of CRF07_BC has increased significantly, from 4.5% in 2002 to 13.6% in this study, and it has been identified in all transmission routes. From 1989 to 2013, a shift in transmission routes became apparent [15,21], from which subtype B, CRF01_AE, and CRF07_BC spread out through sexual contact [21,24] with an increasing diversity of high-risk behaviors and the growing size of the floating population. Currently, sexual transmission is the most common route of transmission in Hebei, accounting for 98.1% of HIV-1-positive cases in 2013 [21]. Subtype B, CRF01_AE, and CRF07_BC are the three main genotypes, and mainly circulate through sexual contact. The co-circulation of these three genotypes has resulted in the occurrence and spread of novel recombinant strains, as evidenced by the detection of recombinant strains CRF01_AE/B and CRF01_AE/BC in this study. To our knowledge, this is the first report of HIV-1 subtype specialty and DR mutations in Hebei.
In our work, the mutation classes of HIV-1 DR showed significant differences between ART-naïve controls and participants experiencing treatment failure. The prevalence of single, dual, and multiple mutations in participants experiencing treatment failure was significantly higher than in ART-naïve participants, which is consistent with previous findings in Yunnan [25]. The dual NRTI and NNRTI DR prevalence (29.4%) was highest, followed by that of NNRTIs (10.7%), NRTIs (4.2%), and PIs (2.8%) in participants experiencing treatment failure. However, in ART-naïve participants, the PI DR prevalence (2.5%) was higher than that of NRTIs (1.7%), NNRTIs (0.8%), and NRTIs and NNRTIs (0.8%), in contrast to an earlier report [26]. Our observed DR rate of 51.9% in participants experiencing treatment failure was higher than that seen in Henan Province (47.1-64.7%) [27,28] and Switzerland (37-45%) [29], suggesting that the higher prevalence of HIV-1-resistant strains is closely related to the widespread use of antiviral drugs. This has occurred in China since 2003, after which time more HIV-1 drugresistant variants were identified and have spread.
The prevalence of NNRTI mutations was higher than that of other mutations in this study, which might reflect the replicative fitness of the virus. For example, Y181C can increase HIV-1 subtype B replicative capacity [30]. Moreover, our study also revealed significant differences in the distributions of M184V/I and M41L mutations among three main genotypes, with M46L/V and T74S only found in CRF01_AE. The differences of HIV-1 mutation distribution in three main genotypes provide some clues of replicative fitness of the virus and renewal of the therapeutic regime. By contrast, the distributions of the remaining mutations were not significantly different among three main genotypes, suggesting that they are randomly distributed in these genotypes.