Emergence of Lamivudine-Resistant HBV during Antiretroviral Therapy Including Lamivudine for Patients Coinfected with HIV and HBV in China

In China, HIV-1-infected patients typically receive antiretroviral therapy (ART) that includes lamivudine (3TC) as a reverse-transcriptase inhibitor (RTI) (ART-3TC). Previous studies from certain developed countries have shown that, in ART-3TC, 3TC-resistant HBV progressively emerges at an annual rate of 15–20% in patients coinfected with HIV-1 and HBV. This scenario in China warrants investigation because >10% of all HIV-infected patients in China are HBV carriers. We measured the occurrence of 3TC-resistant HBV during ART-3TC for HIV-HBV coinfection and also tested the effect of tenofovir disoproxil fumarate (TDF) used as an additional RTI (ART-3TC/TDF) in a cohort study in China. We obtained 200 plasma samples collected from 50 Chinese patients coinfected with HIV-1 and HBV (positive for hepatitis B surface antigen) and examined them for the prevalence of 3TC-resistant HBV by directly sequencing PCR products that covered the HBV reverse-transcriptase gene. We divided the patients into ART-3TC and ART-3TC/TDF groups and compared the efficacy of treatment and incidence of drug-resistance mutation between the groups. HIV RNA and HBV DNA loads drastically decreased in both ART-3TC and ART-3TC/TDF groups. In the ART-3TC group, HBV breakthrough or insufficient suppression of HBV DNA loads was observed in 20% (10/50) of the patients after 96-week treatment, and 8 of these patients harbored 3TC-resistant mutants. By contrast, neither HBV breakthrough nor treatment failure was recorded in the ART-3TC/TDF group. All of the 3TC-resistant HBV mutants emerged from the cases in which HBV DNA loads were high at baseline. Our results clearly demonstrated that ART-3TC is associated with the emergence of 3TC-resistant HBV in patients coinfected with HIV-1 and HBV and that ART-3TC/TDF reduces HBV DNA loads to an undetectable level. These findings support the use of TDF-based treatment regimens for patients coinfected with HIV-1 and HBV.


Introduction
In 2011, the World Health Organization estimated that about 34 million people worldwide were infected with HIV-1, of whom approximately 12% (4 million) had chronic hepatitis B virus (HBV) infection (defined by positivity of hepatitis B surface antigen (HBsAg) in blood for >6 months) [1][2][3][4]. The guidelines for the use of antiretroviral drugs in HIV-1-infected patients recommend that all patients must be screened for HBsAg before antiretroviral therapy (ART) is initiated, and that if a positive result is obtained, ART with tenofovir disoproxil fumarate (TDF) plus lamivudine (3TC) or emtricitabine must be initiated [5]. However, HBsAg testing is either not available or prohibitively expensive in several developing countries, and thus the HBV infection status is frequently unknown in the case of HIV-1-infected patients who have received a standard ART regimen including 3TC [6].
The drug 3TC is a cytidine analog that inhibits the reverse transcriptase (RT) of both HIV and HBV [7]. The major disadvantage associated with the use of this drug is that resistance mutations progressively emerge at an annual rate of 15-20% in HIV-1-HBV-coinfected patients in developed countries [8][9][10][11]. Consequently, new-generation RT inhibitors (RTIs) that feature comparatively higher resistance barriers have now been produced, and in the specific case of TDF, no resistance mutation has been detected in chronically HBV-infected patients after 3 years of therapy [12,13]. However, in developing countries, most of these new drugs are not available or extremely expensive [14].
In China, government-supported ART including 3TC (ART-3TC) was started in 2003 [15], and numerous HIV-HBV-coinfected patients continue to remain on the ART-3TC regimen [16]. However, the effect of ART-3TC on HBV DNA loads and the prevalence of 3TC-resistant HBV mutants in HIV-HBV-coinfected patients have not been examined in China. Moreover, whether ART-3TC/TDF can reduce HBV DNA loads and the emergence of 3TC-resistant HBV mutants is unclear.
Here, we report the emergence of 3TC-resistance mutations in the HBV RT gene among HIV-HBV-coinfected Chinese patients during 96-week ART-3TC without TDF, and we describe the relationship between HBV DNA loads at baseline and the emergence of 3TC-resistant HBV.

Patients
This study was approved by the ethics committee at Peking Union Medical College Hospital (ClinicalTrials.gov Identifiers: NCT00618176 and NCT00872417). Patients were enrolled in Chinese cohort studies (National Key Technologies R&D Program for the 10th and 11th Five-year Plans) [16,17] for examination of the response to the Chinese standard treatment regimen and the effectiveness of the treatment for HIV-1-infected patient. In these studies, patient selection and treatment methods were completely randomized. Informed consent was obtained from every participant. In China, patient plasma is collected at 12 centers (collaborative hospitals and local CDCs) that cover 23 provinces and cities (Beijing, Henan, Shaanxi, Shanghai, Fujian, Guangdong, Yunnan, Tianjin, Hebei, Neimenggu, Shandong, Liaoning, Jilin, Heilongjiang, Jiangsu, Zhejiang, Jiangxi, Shanxi, Anhui, Hunan, Hubei, Guangxi, and Sichuan). The patient plasma samples were collected between 2005 and 2014, and patients were classified according to the HBV-status criteria of the US Centers for Disease Control and Prevention [18]; these criteria have been used previously [16]. The flowchart in Fig 1 presents the patientselection process, and Table 1 lists the clinical characteristics of the selected patients. We had access to stored patient blood samples collected before ART initiation (baseline) and after ART (12, 48, 96 weeks), and we selected 200 plasma samples from 50 patients who met previously described criteria [16]. Patients coinfected with hepatitis C virus were excluded from the analysis. The ART-3TC group comprised 38 patients who received treatment with 3TC/stavudine (d4T)/nevirapine (NVP) or 3TC/zidovudine (AZT)/NVP or 3TC/AZT/efavirenz (EFV) as first-line therapy, and the ART-3TC/TDF group included 12 patients who received 3TC/TDF/ ritonavir-boosted lopinavir (LPV/r) or 3TC/TDF/EFV as first-line therapy (Table 1).

Baseline characteristics
Of the 913 HIV infected patients who had enrolled in the national cohorts, 134 (14.7%) had tested HBsAg-positive, and 50 of these patients could be traced until 96 weeks of treatment. All 50 patients were Chinese, 88% of them were male, and the median age of this group was 36.0 years (range: 18-56 years) ( Table 1). Most of the patients (86%) were infected through sexual contact, and 51.2% of these patients were men who have sex with men ("MSM"); 2% of the patients were infected through blood transfusion; and the remaining 12% were infected through unknown routes. This study included no intravenous drug users. The CD4+ T-cell count of the participants was <350/μL, and the median CD4+ T-cell count was 162/μL (range: 2-343 cells/μL). The median HIV RNA load was 26,993 copies/mL (range: 917-4.8 × 106 copies/mL), and the median HBV DNA load was 7.8 × 105 IU/mL (range: <20-2.5 × 108 IU/mL). The CD4+ T-cell counts, HIV-1 RNA loads, and HBV DNA loads did not differ between the ART-3TC and ART-3TC/TDF groups ( Table 1). The HBV RT coding region could be amplified from 46 out of 50 patient samples collected before ART initiation. A phylogenetic analysis performed using the PCR-product sequences indicated that 43.5% (20/46) of the patients were infected with HBV B genotype, 54.3% (25/46) with C genotype, and 2.2% (1/46) with D genotype (data not shown). Currently, the viral pathogenesis of HBV is considered to depend on the genotypes and not on geographic distributions [25][26][27]. HBV genotypes B and C are the major epidemic genotypes in China [28], and our study confirmed this. Genotype B is reported to be more responsive to 3TC than genotype C [29,30], but we detected no large differences in the responsiveness of genotypes B and C to 3TC (S1 Fig). The crucial factor underlying treatment responsiveness was HBeAg in both genotypes: in our study, HBeAg-negative patients responded more strongly to the treatment than did HBeAg-positive patients (S1 Fig). The HBeAg positivity rate is also recognized to be high in HBV genotype C [31], but our results did not confirm this (S1 Table).
Because this study included 50 patients, we examined whether the sample size was appropriate for the statistics: we used Stata 11.0 software (StataCorp, College Station, TX, USA) and the "power and sample size analysis" method. In this analysis, we assumed that resistance rates were 50% in the ART-3TC group and 5% in the ART-3TC/TDF group after 2 years of treatment, based on previous reports [11,20]. The calculation result showed that the statistical power was 0.90 (α = 0.05), because we enrolled 38 patients in the 3TC-treatment group and 12 patients in the TDF-treatment group. In this method, a statistical power of >0.80 is indicated to be statistically significant. Thus, we conclude that the number of patients analyzed can be considered to be statistically significant.

Effect of ART including 3TC on HIV-1 and HBV replication
The median HIV-1 RNA loads of both the ART-3TC group and the ART-3TC/TDF group significantly decreased after treatment (P < 0.0001; Table 1), but HIV-1 viral-suppression efficacy did not differ between the treatments (P > 0.05; Table 1). The median CD4+ T-cell counts increased after both treatments (P < 0.0001; Table 1).
After treatment, the median HBV DNA loads of both groups were also significantly decreased (Table 1 and S2 Fig). At 48 weeks, the median HBV DNA load was below detectable levels, and at 96 weeks, the median HBV DNA load was still within the range considered "undetectable," although the level now slightly increased in the ART-3TC group, indicating a possible HBV DNA rebound in certain patients. No such increase in the HBV DNA load was observed in the ART-3TC/TDF group (Table 1 and S2 Fig). In terms of individual cases, the HBV DNA loads of 16 and 15 patients were still detectable at 48 and 96 weeks, respectively. We suspect that the HIV ART was closely adhered to, because the HIV RNA load reduced in all patients. Measurement of the plasma concentrations of 3TC in patients in whom HBV breakthrough or insufficient suppression of HBV DNA loads was observed revealed that these concentrations were within the standard range in both groups of patients (data not shown).

Emergence of 3TC-resistant HBV
We investigated the prevalence of 3TC-resistant mutants of the HBV RT gene by performing PCR-amplification and directly sequencing the PCR products; the results are presented in Table 2. The RT coding region was amplified from the samples of the 38 patients in the ART-3TC (without TDF) group: 34 plasma samples (89.5%) at baseline, 11 samples (28.9%)   Table 2). The main mutation sites identified here agree with previously published data [9-11, 20, 21]. However, the mutation rate recorded after 96 weeks of treatment, 23.5% (8/34 patients), was lower than that in Western countries, 50% [20]. This difference could depend on the distinct environments related to the HBV infection routes: In China, infection occurs mainly through vertical infection, which differs from the lateral infection prevalent in Western countries.
ART-3TC/TDF was administered to fewer patients (12 cases) than was ART-3TC (38 cases); however, the addition of TDF appeared to have effectively suppressed HBV DNA loads in 96 weeks ( Table 2). Because the HBV RT coding region could not be PCR-amplified after ART-3TC/TDF, we were unable to investigate whether mutations related to TDF or 3TC resistance would appear in the ART-3TC/TDF cases after treatment for >96 weeks. However, no resistance mutation in the HBV RT coding region has been reported in patients chronically infected with HBV even after 3 years of TDF therapy [12,13]. Thus, the HBV antigen test is recommended for the identification of HBV-coinfected patients and the use of TDF would be favorable for HIV-HBV-coinfected patients. The new treatment guideline for HIV-1 patients was issued by the Chinese Ministry of Health [32]. Although TDF is now available in China in accordance with this new guideline, no long-term follow-up study has previously been conducted in China to evaluate HBV drug resistance in the setting of HIV coinfection in Chinese patients. Thus, this is the first report of a long-term follow-up study of TDF treatment used for HIV-1-HBV-coinfected patients in China.
Intriguingly, HBV breakthrough or insufficient suppression of HBV DNA load was observed in 10 patients of the ART-3TC group after 96-week treatment, and in 8 of these patients, mutations associated with 3TC resistance were detected (Table 2, Fig 2 and S3 Fig). In the patients harboring these mutants, DNA loads at baseline were significantly higher than those in the remaining patients (P < 0.05; Fig 3). However, with regard to the emergence rate of 3TC-resistance mutations, we detected no significant differences between genotypes B and C (S1 Fig). In this study, 16 patients (32%) were HBeAg-positive, and the HBV DNA load of these patients was significantly higher than that of HBeAg-negative patients (P < 0.0001; S4 Fig). In both cases, the HBV DNA load was markedly reduced after ART with or without the inclusion of TDF (S4 Fig). However, in the patient group in which the 3TC mutants emerged, both ART-3TC and ART-3TC/TDF did not effectively suppress HBV DNA loads (S4 Fig). Furthermore, in the HBeAg-positive patients, ART-3TC suppressed HBV DNA loads significantly more weakly than did ART-3TC/TDF (P < 0.0001; S4 Fig). Recent clinical studies on chronic hepatitis B have revealed that high HBV DNA loads are correlated with HBeAg positivity, which plays a crucial role in the progression of hepatitis [20,[33][34][35]. Thus, our data also strongly suggest that HBV DNA loads must be measured and, if possible, HBeAg tests must be conducted in order to check for and avoid the emergence of 3TC-resistant HBV. We could not analyze a large number of patients for whom a TDF-based regimen has been used, because treatments that include TDF have only recently become available for HIV patients in China.  However, we used Stata 11.0 software and calculated the statistical power based on previous data obtained in the case of HIV-HBV coinfection, and our results showed that the statistical power here reached 0.90. This indicates that the number of patients included in our analysis yielded statistically significant data.

Conclusions
Our findings have clearly demonstrated that ART-3TC is associated with the emergence of 3TC-resistant HBV in patients coinfected with HIV-1 and HBV. The emergence of 3TC-resistance mutations was closely associated with high HBV DNA loads. To avoid the emergence of 3TC-resistant HBV, the monitoring of HBV DNA loads and the use of ART including TDF are recommended for patients coinfected with HIV and HBV. In the genotype B group (F), 14 patients harbored wild-type HBV during the treatment period (wild-type group) and 4 patients harbored 3TC-resistant HBV mutants after 96 weeks of treatment (mutant group). In the genotype C group (G), 11 patients harbored wild-type HBV (wild-type group) and 4 patients harbored 3TC-resistant HBV mutants after 96 weeks (mutant group). In the case of both genotypes, the median HBV DNA loads at baseline did not differ significantly between the wild-type and mutant groups (P > 0.05), but the DNA load decreased markedly after the 96-week treatment in the wild-type group (P < 0.05). However, we did not detect any suppressive effect of the treatments in the mutant group (P > 0.05). Error bars indicate the interquartile range. The statistical significance of differences was calculated using the Mann-Whitney test. (TIF)  Table. HBV genotype correlation with HBeAg positivity and mutant rate. (DOCX)