Prevalence of HIV-1 integrase mutations related to resistance to dolutegravir in raltegravir naïve and pretreated patients


Corresponding author: F. Saladini, Department of Biotechnology, University of Siena, Policlinico S., Maria alle Scotte, Viale Bracci 16, 53100 Siena, Italy


Clin Microbiol Infect 2012; 18: E428–E430


The prevalence of HIV-1 integrase mutations related to resistance to the next-generation integrase inhibitor (INI), dolutegravir (DTG), was assessed in 440 INI-naïve subjects and in 120 patients failing a raltegravir (RTG)-containing regimen. Of the mutations selected by DTG in vitro, S153FY was not detected in any isolate while L101I and T124A were highly prevalent in both groups and significantly associated with non-B subtype. RTG-selected double and triple mutants, mostly the G140S/Q148H variant, were detected in only 32 (26.7%) RTG-treated patients. As L101I and T124A do not appear to exert any major effect in vivo and double and triple mutants resistant to DTG are infrequently selected by RTG, DTG can be effectively used in INI-naïve patients and may retain activity in many patients failing RTG.

Dolutegravir (DTG, formerly S/GSK1349572) is a next-generation HIV-1 and HIV-2 integrase inhibitor (INI) with an apparently increased genetic barrier to resistance with respect to first-generation INIs such as raltegravir (RTG) [1,2]. Although no data about selection of resistance by DTG in vivo have been reported so far, in vitro drug resistance selection experiments have shown that DTG selects for a defined set of integrase (IN) mutations (L101I, T124A, S153FY), yet these appear to confer only limited (within 2.5-fold) resistance to DTG in standard phenotypic assays [3]. Higher fold resistance levels to DTG have been documented with the G118R mutant selected in vitro by the prototype second-generation INI MK-2048 [1,4] and with some RTG-selected HIV variants containing at least two IN mutations, particularly those occurring at codon pairs 138/148, 140/148 and 148/155 [3,5]. In light of the possibility of using DTG as a first- or second-line INI, it is important to know how often IN mutations potentially decreasing DTG activity occur in the RTG-naïve and RTG-treated patient population.

The roughly nationwide Italian HIV drug resistance database ARCA (Antiretroviral Resistance Cohort Analysis, was searched to retrieve HIV-1 IN sequences obtained from patients with available treatment information. The sequences were grouped depending on whether they were obtained from RTG-naïve patients or patients failing RTG-based therapy. For each patient, the first IN sequence obtained before RTG treatment and the first IN sequence obtained following at least 12 weeks of RTG treatment were retained for further analysis, whichever available. Mutations putatively related to DTG resistance were defined from in vitro DTG resistance selection studies [3] (group DTG-S) and from in vitro DTG activity data on MK-2048-selected [1] (MK-2048-S) and RTG-selected [3,5] (group RTG-S) mutants. The prevalence of DTG resistance mutations in the two treatment groups was compared by Fisher’s exact test. Subtyping of IN sequences was performed through phylogenetic analysis by using the Kimura two-parameter distance matrix and the neighbour joining method as implemented in the PHYLIP software package version 3.69 ( The updated Los Alamos Laboratory HIV-1 subtype sequence file was used as a reference ( The prevalence of individual DTG resistance mutations in subtype B and non-B isolates was also compared by Fisher’s exact test.

A total of 592 IN sequences obtained from 550 distinct patients between 30 November 2006 and 14 November 2011 were available. Following the inclusion criteria, 440 and 120 sequences obtained from RTG-naïve and RTG-pretreated patients, respectively, were analysed. Paired sequences obtained before and after RTG treatment were available for only 10 patients, therefore only a cross-sectional comparison was performed. Of the 550 distinct patients, 445 and 105 harboured a subtype B and non-B virus, respectively. RTG treatment cases were significantly associated with subtype B virus (111/120 in RTG-treated vs. 344/440 in RTG- naïve patients, p < 0.001). In the RTG-treated group, the median (IQR) duration of RTG treatment at the time of IN genotyping was 70 (31–100) weeks.

Overall, 263 (59.8%) IN sequences from RTG-naïve patients and 68 (56.7%) IN sequences from RTG-treated patients harboured at least one of the DTG-S mutations (Table 1). Actually, only L101I and T124A were detected and their prevalence in the two groups was not significantly different. The only double mutant detected (L101I/T124A) was also evenly distributed with respect to RTG treatment and was not among those reported in the original DTG resistance in vitro selection experiments, namely T124A/S153F and T124A/S153Y [3]. Interestingly, both L101A and T124A were more common in non-B than in B subtype sequences (86.5% vs. 41.0% and 54.2% vs. 15.7%, respectively, when considering RTG-naïve patients; both p < 0.0001). The only MK-2048-S mutation, G118R, was not detected in any of the IN sequences and none of the mutations included in the RTG-S double or triple mutants were detected in RTG-naïve patients.

Table 1.  Prevalence of integrase mutations related to resistance to dolutegravir in raltegravir-naïve and raltegravir-pretreated patients. Double and triple mutants involving codons 138, 140, 148 and 155 are shown only when detected at least in one sequence
Mutations groupMutation(s)Raltegravir-naïve patients (n = 440)Raltegravir treated patients (n = 120)p (Fisher exact test)
No. of cases%No. of cases%
  1. DTG-S, selected by dolutegravir in vitro; MK-2048-S, selected by MK-2048 in vitro; RTG-S, selected by raltegravir in vitro and/or in vivo.

Any of L101I T124A S153FY26359.86856.70.600
Any combination 138/148 (plus others)00.075.8<0.0001
Any combination 140/148 (plus others)00.03125.8<0.0001
Any combination 148/155 (plus others)

The three RTG resistance pathways based on substitutions at codons 143, 148 and 155 [6], either alone or combined with another pathway, were detected in 7 (5.8%), 34 (28.3%) and 32 (26.7%) cases, respectively, in RTG treatment failure isolates. Notably, as many as 50 (41.7%) of the IN sequences obtained at failure of RTG therapy, did not contain any mutation at codons 143, 148 or 155. RTG-S double or triple mutants were detected in 32 (26.7%) RTG-failing patients and the difference in prevalence with respect to RTG-naïve patients reached statistical significance for some, particularly for the highly prevalent G140S/Q148H (Table 1).

Similar to two previous surveys carried out in Spain [7] and France [8], this large study helps define the first and second-line INI use scenarios for DTG. A relevant proportion of INI- naïve subjects were found to harbour virus with L101I and/or T124A, two of the four DTG-S mutations [3], particularly in association with non-B subtype, as also documented in the Spanish patient population [7]. We could not confirm the increased rate of detection of T124A in RTG-treated patients with respect to RTG-naïve subjects, reported in the French [8] but not in the Spanish [7] study. Differences in HIV-1 subtype distribution in the patient populations may account for these discrepancies. Although DTG-S mutations neither confer measurable resistance to DTG in vitro [3,5] nor appear to reduce response to DTG in vivo [Vavro C, Underwood M, Madsen H et al. Polymorphisms at position 101 and 124 in the HIV-1 integrase (IN) gene: lack of effects on susceptibility to S/GSK1349572. Fiftieth Interscience Conference on Antimicrobial Agents and Chemotherapy, Boston, MA, 2010. Abstract H-935], it cannot be excluded that they play a role in the acquisition or modulation of resistance to DTG. Analysis of data derived from DTG therapy is required to clarify this possibility.

RTG-S mutants were not detected as natural polymorphisms in RTG- naïve patients, thereby confirming previous reports [9,10]. By contrast, major IN mutations at codons 143, 148 or 155 were detected in 70 of the 120 RTG failures and RTG-S double and triple mutants were detected in less than half of these 70 patients. The relatively high rate of RTG failures in the absence of IN mutations may be explained by treatment adherence issues or some other unexplained mechanism(s); however, this observation has been commonly reported in previous RTG treatment studies [6]. Reversion from drug-resistant to wild-type virus cannot explain a frequent lack of RTG resistance mutations at failure in our case file because only 8 of the 120 (6.7%) post-RTG sequences were obtained following RTG discontinuation. The median time elapsed between RTG interruption and genotyping in these cases was only 4 weeks and three of the eight sequences indeed showed major RTG resistance mutations. As many as 21 of the 32 RTG-S variants were G140S/Q148H mutants, which were originally reported to result in lower changes in DTG IC50 with respect to the more resistant isolates containing Q148R (five cases in our file) [3]. However, a more recent work showed similar DTG IC50 values for the Q148H and Q148R mutants [5]. By contrast, the mutants E138K/Q148K, Q148R/N155H and E138A/G140S/Q148H, previously reported to have the greatest changes in DTG IC50 [3,5], were detected in only one, no and three cases, respectively. In the absence of a validated clinical cut-off for DTG fold resistance, it is currently difficult to estimate what proportion of patients failing RTG would benefit from a second-line use of DTG. The recent 24-week analysis of the pilot Viking study [Soriano V, Cox J, Eron JJ, et al. Dolutegravir (DTG, S/GSK1349572) treatment of subjects with raltegravir (RAL) resistance: viral suppression at week 24 in the Viking study. 13th European AIDS Conference (EACS 2011). Belgrade, October 12–15, 2011. Abstract PS1/2] suggests that DTG can suppress viraemia in most of the RTG-S cases provided the twice-daily schedule is used. While current data support successful use of DTG as both a first and second-line INI, more extensive and prolonged treatment data are required to define the full potential of this next-generation INI.


This work has been presented at: the Tenth National Congress of the Italian Society for Virology, Orvieto, Italy, 12–14 September 2011; the Tenth National Congress of the Italian Society for Tropical and Infectious Diseases, La Maddalena, Italy, 5–8 October 2011; the Thirteenth European AIDS Conference, Belgrade, Serbia, 12–15 October 2011.

Transparency Declaration

This work was supported by grants from the Italian Ministry of Health (PRIN Grant 200887SYZ5 and AIDS Program Grant 40h81) and from the European Community under the Seventh Framework Program (CHAIN project Grant 223131). ARCA has been supported by unrestricted educational grants from Abbott, Boehringer-Ingelheim, Bristol-Myers Squibb, GlaxoSmithKline, Gilead Sciences, Janssen-Cilag Tibotec division and ViiV Healthcare. A. De Luca has been a member of advisory boards for Abbott, ViiV Healthcare, Janssen-Cilag and Gilead Sciences and has received research grants from ViiV Healthcare. M. Zazzi has been a consultant to or has received research support or lecture fees from Abbott Pharmaceuticals, Abbott Molecular, Gilead Sciences, Janssen-Cilag, Merck Sharp and Dome and ViiV Healthcare.