Monotherapy with either dolutegravir or raltegravir fails to durably suppress HIV viraemia in humanized mice

Abstract

Objectives

To compare the effectiveness of HIV integrase inhibitor monotherapy between raltegravir and dolutegravir as an approach to simplify therapy.

Methods

We evaluated and compared the efficacy of 20 week monotherapy with dolutegravir or raltegravir in humanized mice (HSC-NSG) infected with HIV_BaL. Plasma HIV RNA was measured by quantitative RT–PCR (limit of detection of 150 copies/45 μL of plasma) and drug levels by LC-MS/MS. Escape viruses were genotyped and analysed for replication capacity and drug susceptibility in tissue culture.

Results

Drug-untreated control mice maintained constant viraemia throughout the study. Virus isolates from these mice were susceptible to both raltegravir (EC₅₀ of <8 nM) and dolutegravir (EC₅₀ of <1 nM). Mice treated with raltegravir or dolutegravir had plasma drug levels comparable to those in humans. Monotherapy with raltegravir initially suppressed HIV viraemia, but failed to maintain suppression in 4/4 mice. Viruses from raltegravir failing mice developed mutations G140G/S and Q148H/K, and were resistant to both raltegravir (EC₅₀ values of >100 nM) and dolutegravir (EC₅₀ values ranging from 8.8 to 13.3 nM). Monotherapy with dolutegravir suppressed viraemia in 5/5 of mice, but viraemia rebounded in one animal. The virus from this mouse had mutations E138K, G140S, Q148H, N155H and S230R, was highly resistant to both raltegravir (EC₅₀ of >1000 nM) and dolutegravir (EC₅₀ of 550 nM), and replicated to levels similar to those of control viruses in PBMCs.

Conclusions

Monotherapy with either raltegravir or dolutegravir does not consistently maintain HIV suppression, suggesting that dual therapy may be required in simplification strategies.

Introduction

Current HIV ART regimens frequently consist of one HIV integrase strand transfer inhibitor (INSTI) plus two NRTIs.¹ Although data are conflicting, several studies suggest NRTIs are associated with kidney disease, reduced bone mineral density and a risk of myocardial infarction.¹ Unlike other ARTs, INSTIs have decreased toxicity and reduced potential for interaction with comorbid medications.² The first-generation INSTIs (raltegravir and elvitegravir) have a modest genetic barrier to resistance (i.e. mutations necessary for resistance occur during a relatively short period of drug pressure). In contrast, the second-generation INSTI dolutegravir has a higher barrier to resistance, due in part to slower dissociation rates from integrase–DNA complexes.^3,4 In tissue culture, emergence of resistance to dolutegravir is slow (∼26–40 weeks of drug pressure) and resistance is low (2- to 6-fold only).⁵ Resistance to dolutegravir during first-line therapy is rare, with only few cases reported to date in patients whose adherence to treatment may have been suboptimal.⁶ Together, these observations suggest that monotherapy with dolutegravir, and dual therapy in combinations with lamivudine, emtricitabine or tenofovir disoproxil fumarate, could be effective as maintenance regimens. Several small, single-centre, retrospective or observational studies suggest the use of dolutegravir monotherapy as a maintenance regimen.^7–13 Nevertheless, it is currently unclear whether dolutegravir-based mono- or dual-drug therapy will sustain HIV suppression, without emergence of resistance, in the long term.¹⁴ The paucity of information on paths of resistance to dolutegravir, mainly obtained in tissue culture, prompted us to conduct studies in vivo using humanized mice. We used NSG (NOD.Cg-Prkdcscid IL2rgtm1Wjl/SzJ) mice transplanted with human CD34+ cells, an animal model that upon infection with HIV resembles chronic infection in humans and its control by ARTs.¹⁵ We demonstrate that dolutegravir monotherapy is not sufficient to maintain HIV suppression and that resistance mutations differ from those reported in tissue culture experiments. Importantly, HIV escaped dolutegravir treatment in humanized mice by following the N155H, Q148H and S230R resistance pathways, which have been recently reported in patients failing dolutegravir monotherapy.^16,17

Methods

Generation and infection of humanized mice

Animal protocols were approved by the Institutional Animal Care and Use Committee, University of Maryland School of Medicine. We humanized 2–3-day-old irradiated (10 cGy) NSG pups by hepatic injection of 120 × 10³ human cord blood-derived CD34+ cells (Lonza, Walkersville, MD, USA). At week 12, mice were screened for human CD45 cells in peripheral blood. Mice successfully transplanted (i.e. >5% human CD45+) were infected, via intraperitoneal injection, with 15 000 TCID₅₀ units of the CCR5-tropic HIV reference strain BaL.

Drug treatments

Raltegravir and dolutegravir were given in food pellets, assuming that a 25 g mouse eats 3–4 g of food daily and taking into account the faster metabolism of mice compared with humans.¹⁵ Raltegravir was given at 768 mg/kg/day and dolutegravir at 48 mg/kg/day. Concentrations of INSTIs in plasma were determined by LC-MS/MS.

Quantification of HIV RNA levels in mouse plasma

Plasma HIV RNA levels were quantified by an in-house real-time RT–PCR using HIV Gag primers SK38/SK39 and SYBR green dyes. The assay has a sensitivity of 150 copies of HIV RNA/45 μL of plasma.

Genotyping

Viral RNA from 45 μL of plasma was isolated using the viral mini RNA kit (Qiagen). An aliquot of 10 μL was reverse transcribed into DNA using the SuperScript III one-step RT–PCR and amplified with a Platinum Taq kit (Invitrogen). The PCR primers have been described previously.¹⁸ PCR amplicons were purified with E.Z.N.A.^® Cycle Pure Kit (OMEGA bio-tek) and submitted to McLab (South San Francisco, CA, USA) for population-based sequencing in an ABI 3730xl DNA sequencer (Applied Biosystems, CA, USA). Mutations were identified by sequencing the integrase region of the pol gene. All sequences were aligned against HIV-1_BaL (accession number AY713409). Integrase mutations were identified using a sequence alignment editor/analysis program (BioEdit) and the HIV drug resistance database (https://hivdb.stanford.edu).

Virus isolation

Virus isolation was performed by co-culturing PBMCs isolated from mouse spleen with phytohaemagglutinin-activated CD8-depleted donor PBMCs at a ratio of 1:3. Co-cultures were maintained in the presence of IL-2. On day 7, supernatants were assayed for HIV content by p24 ELISA (PerkinElmer Life Sciences, Inc., Boston, MA, USA) and stored frozen.

Drug susceptibility and virus replication capacity assays

Drug susceptibility assays and virus replication capacity assays were performed in PBMCs, measuring p24 production in culture supernatants by ELISA (PerkinElmer Life Sciences, Inc.). For drug susceptibility assays, data were normalized to p24 levels in drug-untreated controls (100%) and EC₅₀ values were determined by fitting the data to sigmoidal dose–response (variable slope) curves using GraphPad Prism 5.0 software.

Results

HIV viraemia suppression by INSTI monotherapy

We evaluated the potential of dolutegravir monotherapy in first-line therapy in humanized mice infected with HIV BaL. For comparison, we evaluated monotherapy with raltegravir. Six weeks after infection, mice were randomly assigned to dolutegravir monotherapy (n = 5), raltegravir monotherapy (n = 5) or untreated control (n = 3) for 20 weeks. Treatment was well tolerated as suggested by changes in weight of treated versus untreated animals (not shown). However, one mouse in the raltegravir group and another in the control group died prematurely, which is not uncommon in this animal model.¹⁵ Mice were monitored for changes in plasma HIV RNA levels every 2–3 weeks (Figure 1). As expected, control mice had detectable viraemia throughout the study (Figure 1a). Mice in the raltegravir group had decreases in HIV viraemia, which became undetectable in 3/4 mice. However, HIV suppression by raltegravir lasted for 2–3 weeks only, rebounding in all animals (Figure 1b). In contrast, dolutegravir suppressed viraemia to undetectable levels in 5/5 mice and this effect was durable in all but one mouse (mouse #286) (Figure 1c). Drug levels in plasma ranged from 1042 to 2466 nM for raltegravir and from 9879 to 24 246 nM for dolutegravir (Table 1), consistent with patient observations.² Together, these data demonstrate that monotherapy with raltegravir and dolutegravir suppresses HIV viraemia, that suppression lasts longer with dolutegravir and that neither drug maintains durable suppression over 20 weeks in humanized mice.

Figure 1.

Differential long-term antiviral effects of raltegravir and dolutegravir in monotherapy of HIV_BaL-infected HSC-NSG mice. HSC-NSG mice were infected with HIV_BaL. Six weeks later, mice began receiving normal food pellets (a), pellets containing raltegravir (RAL) (b) or pellets containing dolutegravir (DTG) (c) for 20 weeks. Levels of HIV RNA were measured by quantitative RT–PCR in plasma samples separated from retroorbital bleeds at the indicated timepoints. The dotted line indicates the limit of detection of the RT–PCR (150 copies of HIV RNA/45 μL of plasma); undetectable samples were assigned an arbitrary value of 50 copies. Shaded boxes indicate duration of treatment.

Table 1.

INSTI monotherapy treatment, plasma INSTI levels and virus genotypic and phenotypic data

Mouse	Treatmenta	Plasma drug levels (nM)b	Integrase resistance mutationsc	RAL EC50 (nM)d	DTG EC50 (nM)d
309	none	ND	none	3.33	0.24
324	none	ND	none	7.9	0.84
161	RAL	1420	G140S, Q148H	>1000	10.8
172	RAL	1987	G140S, Q148H	>1000	8.8
208	RAL	2466	G140G/S, Q148H	ND	ND
244	RAL	1042	E138K, G140S, Q148K, D232N	374	13.3
261	DTG	24 246	ND	ND	ND
265	DTG	14 978	ND	ND	ND
277	DTG	15 474	ND	ND	ND
280	DTG	17 740	ND	ND	ND
286	DTG	9879	E138K, G140S, Q148H, N155H, S230R	>1000	550

RAL, raltegravir; DTG, dolutegravir; ND, not done.

^aINSTIs were provided in food pellets. Treatment was initiated on week 6 after HIV_BaL infection and continued for 20 weeks.

^bDrug levels (nM) measured by LC-MS/MS in plasma samples collected on week 20 of treatment.

^cDetermined by population sequencing.

^dSusceptibility of viruses at treatment week 20 to raltegravir/dolutegravir in PBMC infectivity assays.

Genotypic and drug susceptibility analyses

Following 20 weeks of INSTI monotherapy, we euthanized the mice and collected plasma for HIV genotypic analysis and spleen PBMCs for HIV isolation. Virus isolates were evaluated for drug susceptibility and replication capacity. As expected, viruses from control mice did not have any INSTI-resistance mutations and were susceptible to both raltegravir (EC₅₀ values ranging from 3 to 8 nM) and dolutegravir (EC₅₀ values ranging from 0.2 to 0.8 nM) (Table 1). Mice failing raltegravir monotherapy had the major primary resistance mutations Q148H/K and G140G/S, which in combination reduce raltegravir susceptibility.¹ Viruses from these mice were resistant to raltegravir (EC₅₀ values of >100 nM) and were ∼10-fold less susceptible to dolutegravir (EC₅₀ values ranging from 8.8 to 13.3 nM) than viruses from control mice. The only mouse with viraemia during dolutegravir monotherapy (#286) had the major primary resistance mutations Q148H, N155H and G140S, and the accessory mutations E138K and S230R. Inspection of integrase sequences from this mouse showed that the first mutation to appear was N155N/H, detected at week 20, while N155N/H was observed together with Q148H, E138E/K and G140G/S at week 22. At week 26, N155H, Q148H, E138K, G140S and a new accessory mutation S230R were observed and virus isolated from this mouse showed resistance to both raltegravir (EC₅₀ of  >1000 nM) and dolutegravir (EC₅₀ of 550 nM) in PBMCs (Table 1).

Replicative capacity

We next compared the replicative capacity of viruses by infecting donor PBMCs at an moi of 0.001 (Figure S1, available as Supplementary data at JAC Online). All viruses had peaks of p24 production at day 8 after infection. Virus from mouse #244 that experienced raltegravir failure had the lowest levels of p24, consistent with the observation that Q148K decreases replicative capacity.¹⁹ The virus isolated from mouse #286 that failed dolutegravir replicated at control levels, suggesting that the mutations selected did not impair replication capacity.

Discussion

In initial treatment of HIV infection in patients, both dolutegravir and raltegravir are effective in three-drug combinations with NRTIs.² In our monotherapy study in humanized mice, resistance to raltegravir emerged more easily than to dolutegravir, but both drugs were ultimately ineffective at maintaining virus suppression. These data demonstrate that WT HIV can escape dolutegravir monotherapy in vivo. Unlike in vitro studies,⁵ and data from some of the patients in the SAILING study in which highly treatment-experienced patients received dolutegravir as an initial INSTI together with optimized background,²⁰ HIV escaped dolutegravir monotherapy in humanized mice by developing mutations N155H, Q148H and S230R, rather than the R263K mutation. Thus, mutations N155H, Q148H and S230R can be selected by dolutegravir in mice that have not received prior ART, suggesting that these mutations can also be selected by dolutegravir in treatment-naive patients if insufficient antiviral drug pressure, as in monotherapy, is applied. These dolutegravir resistance pathways have been recently reported in a large cohort of patients who had no prior evidence of INSTI failure and were switched to dolutegravir monotherapy,^16,17 confirming the humanized mouse data and supporting the relevance of the animal model in the study of therapy simplification. However, our study has some limitations. First, we did not measure drug levels in treated mice at multiple timepoints and thus we cannot rule out that intake of drug-containing food pellets and/or metabolism may have been lower in individual mice at some timepoints, possibly favouring virus escape. However, this is unlikely because the same method of drug delivery in humanized mice has been successfully used in previous studies.¹⁵ Second, we did not evaluate and compare dolutegravir monotherapy with other ARTs that may have a higher genetic barrier to resistance, such as the INSTI bictegravir or the PI darunavir.

In summary, our data suggest that dolutegravir-based dual therapy, rather than monotherapy, may be minimally necessary to maintain sustained viral suppression in vivo. Ongoing experiments are evaluating dual therapy with combinations of dolutegravir and NRTIs in humanized mice as a potentially effective simplified regimen.