Prevalence of Integrase Strand Transfer Inhibitors (INSTI) Resistance Mutations in Taiwan

Antiretroviral therapy containing an integrase strand transfer inhibitor (INSTI) plus two NRTIs has become the recommended treatment for antiretroviral-naive HIV-1-infected patients in the updated guidelines. We aimed to determine the prevalence of INSTI-related mutations in Taiwan. Genotypic resistance assays were performed on plasma from ARV-naïve patients (N = 948), ARV-experienced but INSTI-naive patients (N = 359), and raltegravir-experienced patients (N = 63) from 2006 to 2015. Major INSTI mutations were defined according to the IAS-USA list and other substitutions with a Stanford HIVdb score ≧ 10 to at least one INSTI were defined as minor mutations. Of 1307 HIV-1 samples from patients never exposed to INSTIs, the overall prevalence of major resistance mutations to INSTIs was 0.9% (n = 12), with an increase to 1.2% in 2013. Of these 12 sequences, 11 harboured Q148H/K/R, one Y143R, and none N155H. Of 30 sequences (47.6%) with INSTI-resistant mutations from raltegravir-experienced patients, 17 harboured Q148H/K/R, 8 N155H, and 6 Y143C/R. Other than these major mutations, the prevalence of minor mutations were 5.3% and 38.1%, respectively, in ARV-naive and raltegravir-experienced patients. The overall prevalence of INSTI mutations remains low in Taiwan. Surveillance of INSTI resistance is warranted due to circulation of polymorphisms contributing to INSTI resistance and expected increasing use of INSTIs.

The 12 INSTI-resistant HIV-1 sequences were all of subtype B; 6 of them were from ART-naive and 6 from ART-experienced/INSTI-naive patients. The amino acid mutations/polymorphisms detected in these individuals are listed in Supplementary Table 1. Among these sequences, Q148H/K/R mutations were found in 91.7% (N = 11) and Y143R mutation in one; yet none of them harboured an N155H mutation. One sequence was identified to harbour Q148R/L74V/P145R mutation, which was predicted to have medium-level resistance to dolutegravir. The rest of the sequences only harboured a single major mutation without other relevant mutations and were predicted to have low-level resistance to dolutegravir.
Transmitted drug resistance to NRTIs, nNRTIs, and PIs. Among the 1307 HIV-1 specimens from INSTI-naive (948 ART-naive and 359 ART-experienced) and 63 raltegravir-experienced patients, the PR and RT genes were successfully amplified and sequenced in 1292 specimens (94.3%). The baseline demographic and clinical characteristics of the patients with PR and RT genes successfully sequenced were similar to those without  High-level of resistance to dolutegravir was defined as having Q148H/R/K with 2 or 3 of G140A/ C/S, L74I and E138A/K/T; medium-level of resistance to dolutegravir was defined as having Q148H/R/K with 1 of G140A/C/S, L74I and E138A/K/T; low-level of resistance to dolutegravir was defined as having Q148H/R/K without other relevant mutations or having R263K mutation; no significant resistance to dolutegravir was defined as having none Q148H/R/K mutations.
analyzable RT/PR sequences (data not shown). Among the 6 samples from ARV-naive patients with INSTI major mutations, only one harboured M184V mutation (Supplementary Table 1), known to confer high-level resistance to lamivudine or emtricitabine; no other mutations related to non-nucleoside reverse-transcriptase inhibitors (nNRTIs) and protease inhibitors (PIs) were identified, which was significantly different from the prevalence of resistance mutations to nNRTIs (11.8%) and PIs (2.5%) observed in individuals without major mutations to INSTIs (Table 2). For the raltegravir-experienced patients, the prevalence of resistance mutations of HIV-1 sequences to any class of ART (96.6% vs 51.6%, P < 0.001), nucleos(t)ide reverse-transcriptase inhibitor (NRTIs) (96.6% vs 38.7%, P < 0.001), and multi-drug resistance (MDR) (69.0% vs 29.0%, P = 0.002) among patients with HIV-1 sequences harbouring INSTI mutations were all significantly higher than those without INSTI major mutations (Table 2).

INSTI-naive
Raltegravir-experienced Others Phylogenetic analysis. Phylogenetic analysis was conducted to determine whether there were clusterings observed among these integrase sequences amplified from INSTI-resistant sequences (Fig. 3). Three transmission clusters of subtype B were identified. One consisted of four sequences (cluster C), two from raltegravir-experienced patients and two from ARV-experienced/INSTI-naive patients. The other two transmission clusters were identified among the HIV-1 strains from raltegravir-experienced patients (cluster A), and raltegravir-experienced and ART-experienced/INSTI-naive patients (cluster B).

Discussion
In this study conducted in Taiwan, where raltegravir was only available for clinical use since 2009, we demonstrate that the prevalence of major mutations to INSTI remains relatively low in INSTI-naive patients, 0.6% (6/948) among ART-naive and 1.7% (6/359) among ART-experienced/INSTI-naive patients, which is somewhat different from the absence of major mutations in ART-naive patients reported in the US and Europe [10][11][12] . Other than these major mutations, the prevalence of mutations at resistance-associated positions with a Stanford HIVdb score ≧ 10 was 5.3% and 38.1%, respectively, in ARV-naive and raltegravir-experienced patients.
In patients experiencing virological failure to the first generation of INSTIs, raltegravir and elvitegravir, three genotypic mutation pathways have been defined: Q148H/R/K (± G140S or E138A/K), N155H (± E92Q), and The 42 integrase sequences (6 from ART-naive, 6 from ART-experienced/INSTI-naive, and 30 from raltegravir-experienced patients) with major INSTI-related mutations were aligned with reference integrase sequences from database. ARV-naive was labeled as grey circle and the ARV-experienced/INSTI-naive was labeled as black square. The horizontal branch was drawn in accordance with their relative genetic distances. Bootstrap values greater than 700 of 1,000 replicates were considered significant and indicated at the nodes of the corresponding branches. The brackets at the right indicate the major sequence genotypes.
Y143C/H/R (± T97A). Except E92Q, the amino acid substitutions in the parenthesis representing the compensatory mutations that can help restore fully or partly viral fitness [14][15][16] . In general, the Q148 and N155 pathways are more frequently observed than the Y143 pathway in the raltegravir-experienced patients based on a French study (15.4% v.s. 19.1% v.s. 6.7%, respectively; N = 502) 17 . In a US study prior to the clinical use of dolutegravir, the prevalence of INSTI-mutations in 3012 specimens submitted for determinations of INSTI genotypic resistance, the Q148, N155, and Y143 pathway accounted for 6.5%, 6.5%, and 2.8%, respectively 12 . Of 30 subjects who failed INSTI-containing regimens in this study, 17 had Q148H/K/R mutations, 8 had N155H mutations, and 6 had Y143C/H/R mutations. In the 12 sequences with the Q148H mutation, all had G140S mutation, the predominant combination of INSTI-related resistance mutations as observed in raltegravir-experienced patients by Fourati et al. 17 . Among the 6 strains with Y143R mutations, 4 had T97A, which has been shown to increase Y143R/C-mediated resistance to raltegravir 18 .
The susceptibility of these INSTI-resistant sequences to the second generation of INSTI, dolutegravir, was also evaluated based on the results from in vitro studies 19,20 and the clinical observation in the VIKING-3 clinical trial 21 . Significantly different from the study by Doyle et al. in the UK 11 , 1.59% and 19.05% of the sequences investigated in our study were defined as having high-or medium-level of resistance to dolutegravir, respectively. One potential reason for the difference is that our patients remained on failing regimens containing raltegravir for a longer period of time, which might lead to the accumulation of secondary mutations over time. In our study, a significantly higher percentage of these patients received zidovudine/lamivudine as the backbone (67.9% v.s. 37.5%, P = 0.02). Whether decreased tolerance to zidovudine/lamivudine may contribute to the risk of accumulating secondary mutations to raltegravir remains to be investigated. Besides the Q148 mutation combined with one or more of G140A/C/S, L74I and E138A/K/T was identified to reduce viral susceptibility to dolutegravir, two amino acid mutations, G118R and R263K, have been reported to confer low-level resistance to dolutegravir 22,23 . In our study, we did not identify any G118R substitution in our specimens, yet the prevalence of R263K was 0.31% and 1.64% in INSTI-naive and raltegravir-experienced patients, respectively.
Although raltegravir and elvitegravir have a relatively lower genetic barrier than PIs and most NRTIs, the prevalence of primary INSTI resistance mutations is relatively rare based on a few published reports [10][11][12] . No major INSTI-related mutations were observed in INSTI-naive patients by the CORONET study group and the ICONA Foundation study group, or in an epidemiological surveillance study using specimens collected by the European SPREAD programme in 2006-2007, before INSTIs were introduced into clinical care in Europe 10,11,13 . However, some minor mutations were observed. In the study by the European SPREAD programme, the prevalence of the integrase substitutions with a Stanford HIVdb score ≧ 10 to at least one INSTI was 4% of 278 specimens, which is comparable to 5.2% of 1307 specimens observed in this study. Casadella et al. who analyzed 56 specimens from the 278 specimens by 454 sequencing to detect the presence of scarce INSTI-related mutations found that the prevalence of mutations with Stanford HIVdb score ≧ 10 increased to 14.3% 10 . The impact of these mutations on the susceptibility of circulating HIV-1 strains to INSTIs is still unclear. However, some of the mutations are significantly associated with raltegravir exposure, such as L74M, T97A, E138K, V151I and G163R 11 . In our study, we also found that the prevalence of T97A, G140CAS, E157Q, or V151I was all significantly higher in HIV-1 strains from raltegravir-experienced patients than those from INSTI-naïve patients (Supplementary Table 3).
In INSTI-naïve patients, we observed that 0.7% of the sequences (N = 9) harboured the N155S/T mutation, which was not observed in any of the raltegravir-experienced patients in our study. N155S was first identified by in vitro passage experiment in the presence of raltegravir, and it has never been reported in sequences from patients receiving raltegravir 24 . Whether the presence of N155S/T only in INSTI-naive patients implicates its better transmissibility than N155H mutant viruses requires future investigations. Different from previous studies, we did not find significant association of L74M and G163R with raltegravir exposure and the prevalence of P145R was significantly higher in the strains from INSTI-naïve patients than those from raltegravir-experienced patients. These observations suggest that there are some sequence variations between HIV-1 strains worldwide, even for subtype B alone, and their influences on the emergence of INSTI-related mutations after initiation of INSTI-containing regimen warrant close monitoring.
In this study, an increased prevalence of INSTI-related genetic mutations was noted in 2013. It could be related to transmission of resistant strains among patients with risky behaviors since increasing incidence of HDV, HCV, syphilis, and Entamoeba histolytica have been observed in our MSM population in Taiwan [25][26][27][28][29] . Indeed, we identified 3 transmission clusters of subtype B in the phylogenetic analysis of integrase sequences with major mutations (Fig. 3). One cluster, cluster C, consisted of 4 sequences, 2 from raltegravir-experienced patients and 2 from ART-experienced/ INSTI-naive patients. Therefore, although raltegravir had not been available for clinical use in treatment-naive patients until 2012, we still observed a 0.9% of INSTI-related major mutations in the strains from INSTI-naive patients. The detection of INSTI-resistant major mutations in one ARV-naive patient in 2006 could be possibly related to clinical trial (BENCHMARK) of raltegravir in 2006 when a few patients were enrolled. Our study provides the first evidence that Q148H/K/R mutation can be transmitted from patients who failed raltegravir-containing regimens to raltegravir-naive patients. Three of the 6 treatment-experienced, but INSTI-naive patients were in two clusters of the phylogenetic analysis (filled squares in Fig. 3), suggesting that they might acquire the INSTI-resistant viruses from INSTI-experienced patients. In addition, the two study subjects in the cluster B were confirmed to be sexual partners. One of them, patient 4562, was naive to raltegravir treatment and was likely to acquire Q148H/K/R mutation from his partner, patient 4563, who had failed to respond to raltegravir-containing regimen with HIV-1 harbouring Q148H/K/R mutation.
The findings of this study should be interpreted with the necessary caution. First, the study was based on patients included mainly in hospitals in northern Taiwan, where the majority of the patients were MSM. While MSM have become the leading risk behavior for HIV infections and TDR is more prevalent among MSM in Taiwan 30,31 , our study results may not be generalized to the IDU and heterosexual populations. Second, the number of raltegravir-experienced patients included in this study remains small as compared to those of other studies, Scientific RepoRts | 6:35779 | DOI: 10.1038/srep35779 and continued surveillance studies are warranted when increasing use of raltegravir is likely and newer INSTIs are to be introduced into clinical use in the near future.
In conclusion, the prevalence of primary INSTI resistance of the HIV-1 strains identified from ARV-naive patients in Taiwan remains relatively low, which is comparable to the previous reports in high-income countries. Nevertheless, given our previous observation that TDR to other antiretrovirals remains high among MSM in Taiwan, close and vigilant monitoring of INSTI resistance is important before its widespread use in Taiwan.

Materials and Methods
Study setting. In Taiwan, cART has been provided free of charge since its introduction in April 1997, but routine drug-resistance testing has not been available to clinicians. By the end of December 2015, there were a total of 31,036 people diagnosed as having HIV infection in Taiwan, and about 70% of those who survived were receiving cART. The prevalence of transmitted drug resistance (TDR) in Taiwan has increased from 1999 to 2006, with an overall prevalence of 9.4% 30 . Decline in antiretroviral resistance mutations to NRTI and PI had been reported, followed by the increased use of more potent NRTIs and ritonavir-boosted PI combinations 31 . Nevertheless, due to financial constraints on the provision of free-of-charge access to cART, the Centers of Disease Control in Taiwan implemented regulations on the prescription of cART to antiretroviral-naive HIV-1-infected patients who received their first cART on 1 June 2012. Four categories of cART were defined: the first category consisted of nevirapine or efavirenz plus zidovudine/lamivudine (coformulated); the second category, nevirapine or efavirenz plus abacavir/lamivudine (coformulated); or TDF/emtricitabine (coformualted) or TDF and lamivudine; the third category, zidovudine/lamivudine plus PIs or raltegravir; and the fourth category, TDF/ emtricitabine, TDF and lamivudine, or abacavir/lamivudine plus PIs or raltegravir. The fourth category required approval before prescription. Raltegravir was first introduced into Taiwan in 2006 when a few patients were enrolled in a clinical trial (BENCHMARK); however, it was not available to clinical use for ART-experienced patients until 2009; and in 2012, it was available for ARV-naïve patients to be combined with 2 NRTIs.

Study subjects.
HIV-infected Taiwanese patients receive HIV care according to the national treatment guidelines at designated hospitals around Taiwan by the Centers of Disease Control in Taiwan. Residual plasma samples for routine determination of plasma viral load (PVL) were obtained from HIV-1-infected patients seeking HIV care at the NTUH. A standardized case record form was used to collect information on demographics, antiretroviral medications, baseline CD4 lymphocyte count, and PVL. PVL and CD4 count were quantified with the use of Cobas Amplicor HIV-1 Monitor TM Test, version 1.5, (Roche Diagnostics Corporation, Indianapolis, USA) and FACSFlow (Becton Dickinson), respectively. The study was approved by the Research Ethics Committee of the National Taiwan University Hospital, Taipei, Taiwan (REC registration number, 200905045R) and the experiment procedures were carried out in accordance with the approved guidelines. The patients who were linked to HIV care at the hospital gave written informed consent for determination of HIV-1 resistance mutations.

Determination of drug resistance mutations. For those with confirmed HIV infection and linked to
clinical care at NTUH, the genotypic resistance assays were performed retrospectively as described previously 30,31 . Antiretroviral resistance mutations were identified using the HIVdb program of the Stanford University HIV Drug Resistance Database (http://hivdb.stanford.edu; date last accessed, 10 November 2015), in accordance with the drug resistance mutation list of the International AIDS Society-USA Consensus Guidelines 32 . Multi-drug resistance (MDR) was defined as having genotypic resistance to more than one class of antiretroviral agents including INSTIs. Resistance-associated mutations (RAMs) were defined by the presence of at least one mutation included in the 2009 WHO surveillance drug resistance mutation list 33 .
Phylogenetic tree analysis. Reference sequences of various subtypes and recombinants were retrieved from the Los Alamos database (http://hiv-web.lanl.gov/seq-db.html). Sequences were aligned with the Clustal W listed in the MEGA (molecular evolutionary genetics analysis) analytical package (version 3.0) 36 with minor manual adjustments. The phylogenetic trees were constructed by the neighbor-joining method based on the Kimura 2-parameter distance matrix listed in the MEGA software. Bootstrap values greater than 750 of 1,000 replicates were considered significant. Statistical Analysis. Data were analyzed using SAS 9.2 (SAS Institute, NC, USA). Categorical data were analyzed using chi-square or Fisher's exact tests, as appropriate, and continuous variables were compared using the Wilcoxon test. All tests were two-tailed and a P value < 0.05 was considered significant.