Background

Guangxi is located in Southwest China, adjacent to Vietnam and the Yunnan and Guangdong provinces. It is one of the areas in China that is most heavily affected by the human immunodeficiency virus 1 (HIV-1) in China [4, 14, 16, 17], was used to generated phylogenetic tree based on the Kimura 2-parameter model with 1000 bootstrap replicates using the MEGA7.0 software (available at: http://www.megasoftware.net) [19].

Drug resistance analysis

We evaluated clinically relevant resistance to nucleoside reverse transcriptase inhibitors (NRTIs), non-nucleoside reverse transcriptase inhibitors (NNRTIs), or protease inhibitors (PIs) using the Stanford University HIV Drug Resistance Database Genotypic Resistance Interpretation Algorithm (version 8.8) and the International Antiviral Society Drug Resistance Mutation list [16]. The degree of drug resistance to each antiretroviral drug was categorised as susceptible, potential low-level resistance, low-level resistance, intermediate resistance, or high-level resistance.

Transmission cluster construction

The aligned sequences were analysed using the HyPhy software to calculate the genetic distance, and the Tamura-Nei 93 pairwise genetic distance was calculated for all pairs of sequences. A genetic distance of ≤1.5% between two sequences was considered to indicate potential transmission partners [14]. The data is converted into edge lists, and the network was constructed by identifying pairs of sequences (nodes) and their potential transmission relationships (edge) using the visualisation sofware Cytoscape 3.5.1. We described the characteristics of the network, including the number of sequences (nodes), links (edges), and clusters (groups of linked sequences) [20].

Statistical analysis

SPSS version 21.0 software (IBM, Chicago, IL, USA) was used for statistical analysis. The data in this study comprised categorical variables indicated with frequencies and percentages. Logistic regression analysis was performed to identify risk factors associated with TDR. A P-value < 0.05 was considered statistically significant. In case a P-value < 0.05 was obtained, the variable was included for further adjustment, otherwise, the variable was omitted from the adjustment.

Results

Distribution of HIV-1 subtypes and drug resistance

In our study subjects, CRF01_AE (42.14%) was the predominant subtype, followed by CRF07_BC (30.93%), CRF08_BC (15.90%), CRF55_01B (6.86%), and others (4.17%). The three main HIV-1 subtypes accounted for 88.97% of the cases among the newly diagnosed individuals. Other subtypes included the CRFs and unique recombinant forms CRF59_01B, CRF68_01B, CRF67_01B, and 85_BC, subtype A, B, and C. The overall prevalence of drug resistance was 7.21%, including 1.56% to PIs, 1.04% to NRTIs, and 4.87% to NNRTIs (Fig. 1).

Fig. 1
figure 1

Phylogenetic tree analysis based on the sequences of human immunodeficiency virus 1 pol gene. We constructed a phylogenetic tree based on the human immunodeficiency virus 1 pol gene sequences in the plasma samples of 1151 newly diagnosed individuals. CRF, circulating recombinant form; PI, protease inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor

Full size image

Characteristics of the subjects and the factors associated with drug resistance

We enrolled 1324 HIV-1-infected individuals in the study, and we successfully sequenced and analysed the samples collected from 1151 (86.93%) individuals. The mean age was 40.8 years (range: 2–86 years). Most of the participants were single (46.39%), 41.62% were married, and 10.95% were divorced or widowed. Their level of educational was primarily junior high school or below (47.87%) or college (28.58%). Most of the individuals belonged to the ethnic majority Han (52.22%) and the Zhuang minority (41.96%). The major route of infection was sexual transmission; heterosexual transmission (60.21%), followed by MSM (35.19%), intravenous drug user (IDU, 4.0%) and mother-to-child transmission (MTCT, 0.6%). There was significant difference between the prevalence of TDR and marital status, subtype (all p values < 0.05) (Table 1).

Table 1 Demographic characteristics and factors associated with drug resistance (n = 1151)
Full size table

Analysis of HIV Drug Resistance Mutation Sites (DRMs) with subtype

We identified 52 kinds of DRMs among the drug-resistant strains; 25 were associated with resistance to NNRTIs, 15 to NRTIs, and 12 to PIs. Individuals infected with the subtype CRF01_AE were the most likely to develop a PI-associated mutation, followed by those infected with CRF07_BC; the mutation sites primarily comprised M46V (0.35%) and Q58E (0.43%) mutations. Patients infected with the CRF01_AE subtype were the most likely to develop NRTI-associated mutations; the most common mutations were M41L (0.17%) and D67G (0.17%). Individuals infected with the subtypes CRF01_AE and CRF08_BC were the most likely to develop NNTRI-associated mutations; the most common mutations were V179E (7.21%), V179D (3.21%), and V106I (1.91%). The genotype CRF55_01B was the most likely to harbour V179E mutations (Fig. 2).

Fig. 2
figure 2

Analysis of drug resistance mutation sites with subtype

Full size image

Level of resistance to different antiretroviral drugs

The PI-associated mutations were predicted to confer resistance to atazanavir (0.09%), darunavir (0.17%), fosamprenavir (0.26%), indinavir (0.09%), Kaletra (also known as lopinavir/ritonavir; 0.17%), nelfinavir (1.13%), saquinavir (0.14%), and tipranavir (0.61%). The NRTI-associated mutations were predicted to confer resistance to abacavir (0.52%), zidovudine (also known as azidothymidine; 0.52%), stavudine (0.78%), didanosine (0.43%), emtricitabine (0.35%), lamivudine (0.35%), and tenofovir (0.26%). The NNRTI-associated mutations were predicted to confer resistance to doravirine (2.87%), efavirenz (1.04%), etravirine (0.87%), nevirapine (1.22%), and rilpivirine (2.17%). The drug resistance level was mostly categorised as low (Table 2).

Table 2 Analysis of resistance level against antiretroviral drugs
Full size table

Drug resistance-associated transmission cluster analysis

We constructed an HIV-1 transmission network (Fig. 3). Of the 1151 subjects evaluated, 585 (50.8%) were segregated into 119 clusters, which included 490 men and 95 women. In the network, the cluster sizes ranged between 2 and 205; there were 96 (80.7%) clusters with size < 5, 14 (11.8%) clusters with 5 ≤ size < 10, and 9 (7.5%) clusters with size ≥10. The biggest cluster had 205 individuals (35.0%). (Fig. 3). We analysed the infection routes and found that 56.9% of the transmission cases occurred in heterosexual individuals, 40.7% in MSM, and 2.4% were of the IDU and MTCT. We also observed that 43.4% of TDR cases were included in the transmission network, and most (21, 58.3%) of which were primarily concentrated in four clusters. These TDR included 11 cases with resistance to PIs, 4 to NRTIs and 21 to NNRTIs, with the predominant gender being male (77.8%) and the major infection route being heterosexual transmission (83.3%). There were three clusters in which the TDR cases were connected with each other. We observed that there were four TDR females connected with 12 TDR males. To explore the factors associated with TDR, we performed bivariate and multivariate analyses. There was significant difference observed between the risk of TDR and unmarried status (aOR:4.24; 95%CI:1.51–11.9; P:0.006), level of junior high school or below (aOR:5.89; 95%CI:1.37–24.97; P:0.017), or infection with the CRF08_BC subtype (aOR:3.69; 95%CI:1.5–9.04; P:0.004) (Table 3).

Fig. 3
figure 3

Drug resistance within human immunodeficiency virus 1 transmission clusters. PI, protease inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor; sensitive, sensitive to antiretroviral drugs

Full size image
Table 3 Factors associated with drug resistance transmission within clusters
Full size table

Discussion

We aimed to reveal the genetic characteristics and prevalence of transmitted HIV TDR among the newly diagnosed individuals in Guangxi, China. Our study revealed that the major epidemic HIV-1 subtypes detected in Guagnxi were CRF01_AE (42.14%), CRF07_BC (30.93%), CRF08_BC (15.9%) and CRF55_01B (6.86%). The distribution of HIV-1 subtypes has changed in the last two decades. Subtypes B and E were the major prevalent strains before 2000 [2]. In 2013, Liu et al. found that CRF01_AE (77.6%), CRF08_BC (10.7%), and CRF07_BC (7.4%) were the major strains in Guangxi [23]. To our knowledge, in this study, we observed the first incidence of CRF68_01B, CRF85_BC, and unconfirmed unique recombinant strains in Guangxi. Their routes of transmission warrant further study. These findings indicated the existence of high genetic heterogeneity and subtype/CRF diversity in HIV-1 in Guangxi, and showed that the new CRFs had spread to provinces with population movement. Furthermore, the continuation of HIV recombination led to the production of new CRFs and URF, which made the HIV subtypes more diverse and complex. Additionally, the introduction of the new CRFs had a profound impact on the local HIV epidemic, and made the spread across a floating population easier. Therefore, the surveillance of HIV subtype should be strengthened further.

According to the categorisation method established by WHO [24], the overall prevalence of TDR in Guangxi was at a moderate level (5–15%). It was higher than the prevalence determined in the region in a previous study [25] as well as in other regions of China [17, 26, 27]. As the increase in TDR will affect the antiviral treatment and the spread of drug resistance, the surveillance of TDR should be strengthened, and measures to curb the increase of TDR should be adopted. A significant difference was observed between the prevalence of TDR and marital status and subtype, which contradicted from the results of a previous study [17, 28]. When marital status was used as a categorisation parameter, unmarried individuals were most likely to develop drug resistance. When categorised based on subtype, patients with the CRF08_BC subtype were the most likely to develop drug resistance. Therefore, these two high-risk factors should be considered in clinical settings. We found a significantly higher prevalence of mutations related to NNRTI resistance than of those associated with resistance to PIs or NRTIs. Mutations related to NNRTI resistance were common, especially V179E (7.21%) and V179D (3.21%). Most V179E mutations were detected in CRF55_01B, and most V179D mutations were detected in CRF01_AE, CRF07_BC, and CRF08_BC subtypes. In case of NNRTIs, the mutation V106I, which can cause low-level resistance to doravirine, was a major cause of drug resistance. High-level resistance to efavirenz and nevirapine primarily resulted from the mutations K103N, L100I, and P225H. In case of NRTIs, 0.86% of the individuals were predicted to be resistant to zidovudine, stavudine, and didanosine. In case of PIs, 1.74% of the individuals were predicted to be resistant to nelfinavir and tipranavir. Combinatorial therapy with lamivudine, efavirenz, and nevirapine has been prescribed as the first-line ART regimen in China, and the total drug resistance rate to these three drugs found to be 2.61% in our study. Therefore, the emergence of these mutations might be related to the use of this first-line regimen. The higher prevalence of TDR may be caused by prolonged ART, as noted in other studies [29, 30].

To further understand the transmission of drug resistance, we constructed transmission clusters based on HIV-1 sequences. Our cluster analysis revealed that 585 individuals could be segregated into 119 clusters, including MSM, heterosexual men, heterosexual women, and intravenous-drug users; this indicated that the transmission characteristics were complicated. In the largest cluster, which comprised 146 individuals (1 woman, 145 men), MSM was the major route of transmission. Analysis of this super cluster suggested that MSM were strongly associated with the local epidemic. Additionally, 43.4% (36/83) of the individuals infected by drug-resistant strains were included in 19 clusters, in which drug resistance was mostly associated with heterosexuality (88.89%). Three clusters were identified that included drug-resistant individuals sharing a transmission relationship with each other, which counted for 44.4% (16/36). The result indicated that the spread of TDR may have occurred in the transmission network. The major infection route in the three clusters was heterosexual transmission (83.3%), and there were five women carrying TDR included in the TDR transmission network. This shows that drug-resistant individuals, especially women should be studied further, because they may act as a potential source of TDR. Logistic regression analysis revealed that being unmarried, having educational level of junior high school or lower, and subtype CRF08_BC may be related to TDR to transmit in the clusters. The reasons for the association of these parameters with TDR to transmit should be investigated further.

Conclusions

The present study revealed a diverse, complex distribution of HIV-1 subtypes in Guangxi, and a moderate prevalence of TDR. However, there are several limitations in the study including the number of intravenous-drug users. Additionally, 13% of the samples could not be sequenced successfully, the major causes may include sample quality, low viral load in patients, or low sensitivity of reagents. The lack of data on contact between study participants was also another limitation. Future studies conducted on a larger, more diverse group with contact information available is necessary to design effective strategies for intervention in high-risk populations. Regardless, the risk of TDR was pronounced and must be factored into the treatment and prevention policies. Our findings will be helpful in determining the optimal primary ART and implementing effective interventions that target the major populations at risk in the area.