High-level resistance to bictegravir and cabotegravir in subtype A- and D-infected HIV-1 patients failing raltegravir with multiple resistance mutations

Abstract

Objectives

The second-generation integrase strand transfer inhibitor (INSTI) bictegravir is becoming accessible in low- and middle-income countries (LMICs), and another INSTI, cabotegravir, has recently been approved as a long-acting injectable. Data on bictegravir and cabotegravir susceptibility in raltegravir-experienced HIV-1 subtype A- and D-infected patients carrying drug resistance mutations (DRMs) remain very scarce in LMICs.

Patients and methods

HIV-1 integrase (IN)-recombinant viruses from eight patients failing raltegravir-based third-line therapy in Uganda were genotypically and phenotypically tested for susceptibility to bictegravir and cabotegravir. Ability of these viruses to integrate into human genomes was assessed in MT-4 cells.

Results

HIV-1 IN-recombinant viruses harbouring single primary mutations (N155H or Y143R/S) or in combination with secondary INSTI mutations (T97A, M50I, L74IM, E157Q, G163R or V151I) were susceptible to both bictegravir and cabotegravir. However, combinations of primary INSTI-resistance mutations such as E138A/G140A/G163R/Q148R or E138K/G140A/S147G/Q148K led to decreased susceptibility to both cabotegravir (fold change in EC₅₀ values from 429 to 1000×) and bictegravir (60 to 100×), exhibiting a high degree of cross-resistance. However, these same IN-recombinant viruses showed impaired integration capacity (14% to 48%) relative to the WT HIV-1 NL4-3 strain in the absence of drug.

Conclusions

Though not currently widely accessible in most LMICs, bictegravir and cabotegravir offer a valid alternative to HIV-infected individuals harbouring subtype A and D HIV-1 variants with reduced susceptibility to first-generation INSTIs but previous exposure to raltegravir may reduce efficacy, more so with cabotegravir.

Introduction

HIV-1 drug resistance remains a global threat, even in the era of second-generation HIV-1 integrase strand transfer inhibitors (INSTIs). In low- and middle-income countries (LMICs), the increasing prevalence of HIV-1 drug resistance in the treatment-naive population has required the reduced use of first-line combined ART (cART) based on an NNRTI, such as efavirenz or nevirapine, and a change to a tenofovir/3TC (lamivudine)/dolutegravir (DTG) regimen.¹ Nevertheless, prolonged virological failure is common in LMICs due to limited virological monitoring of HIV-infected individuals, leading to accumulation of even INSTI-resistance mutations and reduced susceptibility to dolutegravir, raltegravir (RAL) and elvitegravir (EVG).^2–4

Bictegravir (BIC, formerly GS-9883) and cabotegravir (CAB, formerly S/GSK 1265744 or GSK 744), both structural analogues of dolutegravir, are the latest second-generation INSTIs. Two clinical trials in ART-naive individuals and two trials in virologically suppressed patients^5–7 led to approval of bictegravir in 2018 by the FDA as a fixed-dose combination of BIC/emtricitabine (FTC)/tenofovir alafenamide (TAF) in cART-naive and suppressed patients (<50 copies/mL) with no history of drug resistance. Bictegravir is a potent unboosted once-daily INSTI with a higher in vitro barrier to resistance than raltegravir and elvitegravir, and with limited drug–drug interactions. Its structure has a distinct oxazepane ring attached to a metal-chelating scaffold (Figure 1), which increases flexibility, allowing for more adaptability in the presence of drug resistance mutations (DRMs).⁸ Cabotegravir is another analogue of dolutegravir, from a class of carbamoyl pyridones, which has recently been approved by the FDA and the EU as long-acting injectable cabotegravir/rilpivirine for use in HIV patients with undetectable viral loads, stable on current ART and with no drug resistance.^9–11 The unique physicochemical and pharmacokinetic properties of the cabotegravir formulation allow for its use as a single daily tablet or a long-acting nanosuspension for monthly or quarterly administration subcutaneously or intramuscularly. Cabotegravir was recently approved as a fixed-dose long-acting injectable combination of cabotegravir plus the NNRTI rilpivirine after successful clinical trials in HIV-infected individuals.^12,13

Figure 1.

The chemical structures of different INSTIs. (a) Elvitegravir; (b) raltegravir; (c) dolutegravir; (d) cabotegravir; and (e) bictegravir. The coplanar oxygen atoms are highlighted in red; green circles highlight extended linkers, which are a common feature of all second-generation INSTIs. This figure appears in colour in the online version of JAC and in black and white in the printed version of JAC.

Bictegravir has shown broad activity in vitro as an INSTI against recombinant viruses and clinical isolates carrying primary mutations associated with resistance to raltegravir and elvitegravir.^14,15 However, mutations at position Q148 of the HIV-1 integrase (IN) confer a wide range of resistance to bictegravir in vitro when combined with other amino acid substitutions.¹⁶ Selection of the Q148R mutation was observed in two patients on either oral or long-injectable cabotegravir in the LATTE clinical trials;^13,17 however, cabotegravir has shown activity against viruses harbouring T66I, Y143R, N155H, E92Q, Q148H/K/R and G140S/Q148H/K/R [fold change (FC) < 7].¹⁸

By mid-2020, transition to generic dolutegravir-based first-line therapy had been implemented in 100 LMICs; however, the distribution of dolutegravir at large scale in LMICs has coincided with the emergence of SARS-CoV-2, responsible for the COVID-19 pandemic.^19,20 Unfortunately, this may lead to increasing emergence of INSTI-resistance mutations associated with poor therapy adherence due to reduced access to pharmacies and healthcare providers, or disruptions in the distribution of antiretroviral drugs.²¹ Currently, patients failing third-line raltegravir-based regimens can only be switched to optimized dosage of dolutegravir. Twice-daily dolutegravir dosage of 50 mg or 100 mg may rescue viral suppression in patients carrying resistance mutations at residue G140 and Q148 of IN.^22,23 However, some cART-experienced patients infected with subtype B or non-B HIV-1 strains with multiple DRMs are already resistant to dolutegravir.^2–4,16 Second-generation INSTIs (dolutegravir, bictegravir and cabotegravir) have shown potency against subtype B and non-B HIV-1 viruses in vitro; however, with limited access to more potent antiretroviral options in LMICs, it remains crucial to assess susceptibility of viruses harbouring diverse INSTI-associated mutations affecting bictegravir and cabotegravir.^24,25

In this study, non-B IN-recombinant viruses derived from eight Ugandan patients failing raltegravir-containing regimens and carrying single or multiple primary INSTI mutations, with or without secondary INSTI mutations, were phenotypically tested for susceptibility to cabotegravir and bictegravir. Although patient-derived IN-recombinant viruses carrying single primary INSTI mutations (or in combination with secondary INSTI mutations) were susceptible to bictegravir and cabotegravir, viruses harbouring multiple primary INSTI-resistance mutations (i.e. E138A/G140A/G163R/Q148R and E138K/G140A/S147G/Q148K) led to reduced susceptibility to both novel INSTIs.

Materials and methods

Clinical samples

The study patients failing raltegravir-based third-line therapy with INSTI-resistance mutations (n = 8) were part of a cohort (n = 51) of HIV-1-infected patients failing raltegravir-based third-line therapy at the Joint Clinical Research Center (JCRC) in Kampala, Uganda, as described previously.^4,26 The JCRC was the first centre to provide generic cART in Uganda, starting in early 2000.^27,28 Samples of patients with virological failure detected during routine check-ups are sent to the Case Western Reserve University Center for AIDS Research (CFAR) laboratories at the JCRC for Sanger HIV-1 genotyping tests.²⁹ The CFAR laboratory is accredited by WHO, the College of American Pathologists (CAP) and the NIH Virology Quality Assurance Program (NIH-VQA). HIV-1 patients who provided written consent and were experiencing virological or immunological failure, described as plasma HIV-1 RNA load >1000 copies/mL and/or CD4+ T cell counts below 250 cells/mm³, were included in the study. Ethical clearance was obtained from the Institutional Review Boards at the JCRC and University Hospitals Cleveland Medical Center/Case Western Reserve University (EM-10-07 and 10-05-35). The samples were processed and sequenced and HIV-1 recombinant viruses constructed as described previously.^4,26

Cells and antiviral compounds

TZM-bl, U87.CD4.CXR4, MT4 and HEK293T cell lines were obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH. TZM-bl and HEK293T cells were maintained in DMEM supplemented with 10% FBS and 100 μg/mL penicillin/streptomycin. U87.CD4.CXR4 cells were maintained in DMEM supplemented with 10% FBS, 100 μg/mL penicillin/streptomycin, 300 μg/mL G418 and 1 μg/mL puromycin (Invitrogen, Carlsbad, CA, USA). MT4 cells were maintained in RPMI supplemented with 10% FBS. All cell lines were subcultured every 3 to 4 days at 37°C under 5% CO₂. Cabotegravir and bictegravir were purchased from Selleck Chemicals (Houston, TX, USA).

Virus titration

The titres of the IN-recombinant viruses were measured using short-term infectivity assays in TZM-bl cells using the Galacto-Star chemiluminescent reporter gene assay (Thermo Fisher Scientific), which measures expression of the β-galactosidase enzyme under HIV-1 Tat expression. Briefly, 20 000 TZM-bl cells were seeded in the presence of polybrene (1 mg/mL) to each well of a 96-well plate. The cells were infected with two-fold serially diluted virus and cultured in a 37°C and 5% CO₂ incubator for 48 h. The cells were washed with 200 μl of 1× PBS and lysed with 10 μL of lysis buffer for 10 min. To the cell lysate, 100 μL of reaction buffer and Galacton-star substrate was added and incubated for 1 h at room temperature. The expression of β-galactosidase was measured on a Cytation 5 plate reader (BioTek, USA).

INSTI susceptibility assay

The susceptibility of patient-derived IN-recombinant viruses to bictegravir and cabotegravir was determined using short-term resistance assays based on TZM-bl cells. Briefly, 20 000 cells were seeded into each well of a 96-well plate and infected with the IN-recombinant viruses, or three control HIV-1 strains (i.e. subtype B NL4-3, subtype A UG14 and subtype D UG98) in the presence of 10-fold dilutions of bictegravir and cabotegravir (100 to 10⁻⁸μM) and polybrene (1 mg/mL), at an moi of 0.05 IU/mL in quadruplicate. After 48 h incubation at 37°C and 5% CO₂, virus replication was quantified using the Galacto-Star chemiluminescent reporter gene assay as described above. Drug dose–response curves were generated using non-linear regression curve-fitting features of GraphPad Prism 8.0 software (GraphPad Software, Inc., San Diego, CA, USA). Drug resistance was expressed as FC in 50% effective concentration (EC₅₀) between HIV-1 controls and IN-recombinant viruses.

HIV-1 integration assay

The relative HIV-1 integration into cellular DNA was determined by carrying out Alu-gag quantitative PCR (qPCR), as previously described^30,31 with some modifications. Briefly, 30 000 MT4 cells were infected with normalized viruses (moi of 0.05 IU/mL) in the presence of 1 mg/mL polybrene. The HIV-1 PI darunavir (1 μM) was added to allow only a single round of viral infection and cells were cultured in a 37°C and 5% CO₂ incubator for 72 h. The extracted genomic DNA was normalized using the β-globin gene and amplified using Alu-gag PCR (Table S1 and Figure S1, available as Supplementary data at JAC Online). The relative HIV-1 integration was quantified using a standard curve generated by a dilution series of cellular DNA from MT4 cells infected with NL4-3 WT (dilution using cellular DNA from uninfected cells).

Global occurrence of Q148H/K/R mutations

To assess global occurrence of Q148H/K/R mutations and associated INSTI primary resistance mutations in patients failing raltegravir, we analysed all the available HIV-1 INSTI-associated mutations from the Drug Resistance HIV Stanford Database (https://hivdb.stanford.edu/). All sequence data with original reference, patient identifier, isolate name, accession number and treatment history were retrieved for the analysis. The search was conducted on 28 August 2020 to obtain all available sequence information on patients who were infected by any of the HIV-1 subtypes and failing on an RAL-based regimen.

Statistical analyses

Statistical analyses were performed using non-linear regression in GraphPad Prism 8.1.2. The level of cross-resistance between INSTIs was analysed using Spearman’s rank order test. The means of EC₅₀ for bictegravir, cabotegravir, dolutegravir, raltegravir and elvitegravir were compared using one-way ANOVA and P values of <0.05 were considered statistically significant.

Results

Individuals failing ART with INSTI-associated mutations

We previously described a cohort of 60 HIV-1-infected patients failing a raltegravir-based third-line regimen in Uganda.²⁶ From this group of individuals, we identified eight patients carrying HIV-1 strains with a variety of mutations associated with resistance to INSTIs (the most common being N155H), as well as two patients infected with HIV-1 strains lacking INSTI-associated mutations (Table 1). The patients’ median age was 30.5 years (IQR 26.25–36.25) and the median plasma HIV-1 RNA load was 85 091 copies/mL (IQR 3687.5–570 480). As expected, most patients were infected with subtype A (sub-subtype A1) HIV-1 strains (62.5%), followed by subtype D (25%) and recombinant A/D (12.5%). Five of the eight patients were treated with RAL/lopinavir (LPV)/ritonavir (RTV), while the other three individuals were experienced with RAL/darunavir (DRV)/RTV, RAL/tenofovir disoproxil fumarate (TDF)/3TC/LPV/RTV and RAL/TDF/FTC/DRV/RTV (Table 1).

Table 1.

Clinical and demographic characteristics

Patient ID	HIV-1 subtypea	Age (years)	HIV-1 RNAb (copies/mL)	Antiretroviral regimen	INSTI-associated major mutationsc	INSTI-associated minor mutationsd
UG23	A	51	850	LPVr/RAL	E138A	T97A, V151A
UG35	A	31	2 293 840	DRVr/RAL	Y143R	T97A, M50I, L74IM
UG42	D	23	155 982	LPVr/RAL	N155H	E157Q, G163R, M50L, L74I, V151I
UG206	D	ND	1 515 000	LPVr/RAL	E138K, G140A, Q148K, S147G	None
UG1059	A	ND	14 200	TDF/3TC/LPVr/RAL	E138A, G140A, Q148R	G163R
UG481	A	38	3914	TDF/FTC/RAL/DRVr	Y143R	TA97AT, G163R
UG537	A/D	ND	255 641	LPVr/RAL	N155H	None
UG1179	A	30	ND	LPVr/RAL	N155H	None
UG14	A	ND	3008	ART naive	None	None
UG98	D	ND	ND	ART naive	None	None

ND, not determined; DRVr, DRV/ritonavir; LPVr, LPV/ritonavir.

^aHIV-1 subtype was predicted using the SCUEAL subtype classification algorithm.

^bViral loads were assayed using Abbott m2000sp/rt or Roche COBAS Amplicor Monitor ultrasensitive tests, v1.5.

^cMajor INSTI-resistance mutations confer a larger reduction in susceptibility.

^dMinor INSTI-associated mutations occur later during infection, after emergence of major mutations, and increase resistance and/or restore viral fitness.

Bictegravir and cabotegravir retain activity in the presence of single N155H or Y143R and added secondary mutations

Our previous report showed that subtype A and D IN derived from treatment-naive patients, or those failing treatment, cloned into the NL4-3 backbone did not result in a complementation defect. The INSTI-resistant subtype A and D IN within the NL4-3 backbone showed the same FC in INSTI susceptibility, compared with just the NL4-3 virus, as the WT IN of subtype A and D cloned into NL4-3.⁴ In this study, IN-recombinant viruses carrying diverse INSTI-associated resistance mutations from patient-derived IN (n = 8) were tested for their susceptibility to bictegravir and cabotegravir (Table 1). Subtype A and D patient-derived INs were compatible with subtype B NL4-3 vector backbone, as described previously.⁴ The raltegravir- and elvitegravir-resistant mutant N155H remained susceptible to both bictegravir (FC 1–2.3) and cabotegravir (FC 1.3–6.3), and emergence of secondary mutations (E157Q, G163R, M50L, L74I and V151I) in the context of N155H did not affect susceptibility to bictegravir (FC 0.8) and cabotegravir (FC 1.3) (Table 2; Figures 2 and 3). Viruses carrying the primary mutation to raltegravir, Y143R, in combination with secondary mutations T97AT, G163R, M50I and L74IM remained susceptible to bictegravir (FC 1.2–1.3) and cabotegravir (FC 1.7–2.6). The mutation E138A selected by raltegravir, elvitegravir and dolutegravir did not reduce susceptibility to either bictegravir (FC 0.8) or cabotegravir (FC 3.3) in the presence of secondary mutations T97A and V151A (Table 2; Figures 2 and 3).

Table 2.

Susceptibility of IN-recombinant viruses to bictegravir and cabotegravir

	Bictegravir			Cabotegravir
Patient ID	EC50 (nM)	95% CI for EC50 (nM)	FC in EC50	EC50 (nM)	95% CI for EC50 (nM)	FC in EC50
NL4-3	2.78	1.12–1.92	1	1.36	2.34–3.33	1
UG14	3.8	3.05–4.76	1.4	1.38	1.14–1.68	1
UG98	5.0	3.93–6.38	1.8	1.87	1.58–2.23	1.3
UG481	3.4	1.79–3.26	1.2	2.4	2.53–4.62	1.76
UG537	2.7	1.39–2.41	1	1.8	2.28–3.29	1.3
UG23	2.4	1.67–3.6	0.8	4.7	2.99–7.62	3.3
UG42	2.5	2.15–3.12	0.8	1.8	1.47–2.29	1.3
UG35	3.7	2.75–5.08	1.3	3.6	2.52–5.37	2.6
UG1179	3.9	2.82–5.37	1.4	2.8	1.91–4.35	2.0
UG1059	166.1	495.3–686.9	59.7	584.1	126.3–217.5	429.1
UG206	369.1	2015–3493	132.7	2650	319.5–426.2	1948.5

The EC₅₀ and 95% CI for EC₅₀ were determined using non-linear regression analysis in GraphPad Prism. The NL4-3 was used as WT in the assays. The FC values are relative to NL4-3 WT.

Figure 2.

The FC in EC₅₀ of IN-recombinant viruses. The susceptibility of IN-recombinant viruses UG537 (with mutation N155H), UG1179 (N155H), UG481 (Y143R, T97AT, G163R), UG23 (E138A, T97A, V151A), UG42 (N155H, E157Q, G163R, M50L, L74I, V151I), UG35 (Y143R, T97A, M50I, L74IM), UG1059 (E138A, G140A, Q148R, G163R) and UG206 (E138K, G140A, Q148K, S147G) to bictegravir and cabotegravir was determined using TZM-bl cells. The mean EC₅₀ (nM) values from independent experiments run in quadruplicate were used to determine FC in EC₅₀ (nM) of recombinant viruses harbouring INSTI-resistance mutations relative to subtype B, A and D references (a, b and c). The error bars represent ±SD of FC in EC₅₀ values between replicates of independent experiments. The horizontal line represents an FC of 1. This figure appears in colour in the online version of JAC and in black and white in the printed version of JAC.

Figure 3.

The susceptibility of recombinant viruses to cabotegravir and bictegravir. The susceptibility of recombinant viruses to the INSTIs bictegravir and cabotegravir was determined using a short-term infectivity assay in TZM-bl cells. Each experiment was done in quadruplicate. The change in EC₅₀ (nM) of recombinant viruses harbouring INSTI-resistance mutations was determined in reference to NL4-3 WT. (a) UG206; (b) UG1059; (c) UG537; (d) UG42; (e) UG35; and (f) UG481: drug susceptibility to bictegravir (left panel) and cabotegravir (right panel). This figure appears in colour in the online version of JAC and in black and white in the printed version of JAC.

Combination of multiple primary INSTI-associated mutations leads to high-level resistance to both bictegravir and cabotegravir

Severe reduced susceptibility to dolutegravir (FC >100) was previously shown with IN-recombinant virus UG1059 carrying the primary E138A, G140A and Q148R mutations as well as the secondary mutation G163R.⁴ In the context of bictegravir and cabotegravir, UG1059 showed decreased susceptibility to bictegravir (FC 60) and cabotegravir (FC >100). Interestingly, the subtype D virus UG206, also resistant to dolutegravir and carrying E138K, G140A, Q148K and S147G mutations, was even more resistant to bictegravir and cabotegravir (FC >100 and >1000, respectively) (Table 2; Figures 2, 3 and 4). The NL4-3 alone or subtype B HIV-1 carrying a WT subtype A or D IN-coding region showed no difference in susceptibility to bictegravir and dolutegravir with EC₅₀ values as previously reported for WT HIV-1.^18,32 Overall, bictegravir showed more potency against viruses carrying various INSTI-associated resistance mutations compared with cabotegravir (P = 0.03). IN-recombinant viruses carrying multiple primary INSTI-resistance mutations showed substantially reduced susceptibility to both bictegravir and cabotegravir but EC₅₀ values for bictegravir were significantly lower compared with those for cabotegravir for both UG206 (P = 0.003) and UG1059 (P = 0.0016; Figure 4a and b).

Figure 4.

FC in EC₅₀ of recombinant viruses carrying multiple primary INSTI-resistance mutations. FC in EC₅₀ (nM) of bictegravir and cabotegravir for recombinant virus UG1059 (a) and UG206 (b). FC in EC₅₀ (nM) of recombinant viruses harbouring INSTI-resistance mutations relative to WT NL4-3 was determined in a short infection assay in TZM-bl cells. The mean EC₅₀ (nM) values for cabotegravir and bictegravir were compared using non-parametric two-tailed t-test; P = 0.005 was considered statistically significant. This figure appears in colour in the online version of JAC and in black and white in the printed version of JAC.

Both bictegravir and cabotegravir are second-generation INSTIs targeting the IN gene for inhibition of HIV-1 replication. To determine the degree of cross-resistance between bictegravir and cabotegravir for different viral genotypes tested, log (FC in EC₅₀) values for bictegravir resistance were correlated with log (FC in EC₅₀) values for cabotegravir resistance. Significant cross-resistance was observed, as shown by the strong correlation coefficient (r = 0.995), P = 0.0001 and slope = 0.69 between bictegravir and cabotegravir (Figure S2).

IN-recombinant viruses carrying single or multiple primary INSTI-resistance mutations exhibit impaired integration capacity into cellular DNA

We quantified the relative amounts of viral integration using an Alu-gag qPCR (Figure 5) as described previously.³¹ Viruses UG35 and UG42 carrying the primary N155H or Y143R mutations with additional secondary mutations showed only 25%–29% of the capacity to integrate into the host genome compared with WT NL4-3 or NL4-3 carrying the WT subtype A or D IN. Presence of a single N155H mutation in the UG537 or UG1179 viruses led to a 50% reduction in integration. The UG1059 virus carrying mutations E138A, G140A, Q148R and secondary G163R had the lowest capacity of integration at only 14% compared with WT NL4-3 (Figure 5a). The relative integration of IN mutants significantly correlated with replication fitness of these viruses (r = 0.9; P = 0.0006) (Figure 5b); however, there was no correlation between relative integration and susceptibility to the INSTIs bictegravir (r = −0.31; P = 0.39), cabotegravir (r = −0.74; P = 0.01) and dolutegravir (r = −0.3; P = 0.37) (the dolutegravir analysis is based on our previous report).⁴

Figure 5.

The relative integration capacity of IN-recombinant viruses with diverse INSTI-resistance mutations. The relative integration capacity of mutant viruses compared with controls (UG14 and UG98) and WT (NL4-3) was determined in MT4 cells. The integrated HIV-1 long terminal repeat (LTR) was amplified and quantified using Alu-gag qPCR. (a) Relative integration of mutant viruses; (b) correlation between the relative integration and replicative fitness (the latter being based on our previous reports).⁴ Mean ± SD values are shown from two independent experiments carried out in triplicate for each sample. The correlation was determined by Spearman’s correlation coefficient. qPCR results were normalized relative to NL4-3 WT, arbitrarily set at 100%. This figure appears in colour in the online version of JAC and in black and white in the printed version of JAC.

Prevalence of Q148H/K/R mutations in raltegravir-failing patients infected with subtype B and non-B HIV-1 isolates

Viruses UG1059 and UG206, with high-level resistance to both bictegravir and cabotegravir, carry Q148K/R with additional E138A, G140A and G163R, and E138K, G140A and S147G INSTI-associated mutations, respectively (Table 2; Figures 2, 3 and 4). However, this combination of INSTI-resistance mutations has not been previously reported, which prompted us to assess global prevalence of Q148H/K/R and other mutations in subtype B- and non-subtype B-infected individuals failing raltegravir (Table S2; Figure S3). We used the Stanford HIV database (https://hivdb.stanford.edu/) to analyse HIV-1 INSTI-resistance mutations from HIV-1-infected patients failing on raltegravir (n = 1653). HIV-1 subtype B comprised the majority of sequences carrying Q148H/K/R (87%; 1442/1653) while non-B HIV-1 subtypes represented 13% of sequences (211/1653). Q148H/K/R, singly or in combination with other primary mutations, was found in 29% (419/1442) of HIV-1 subtype B- but only 5.2% (11/211) of non-B subtype-infected patients failing raltegravir-based treatments. The Q148H/K/R, G140S/C/A combination was most dominant in HIV-1 subtype B (19.5%) and non-B subtypes (4.9%), with the Q148H/K/R, E138A/K/T, G140S/C/A combination occurring in 3.7% of subtype B- and 1% of non-B HIV-1 subtype-infected patients. The combination of Q148H/K/R and N155H is uncommon in all HIV-1-derived INs from patients failing raltegravir (1.8%), regardless of subtypes (Table S2; Figure S3). Our detection of these multiple INSTI-resistance mutations in two HIV-1-infected individuals²⁶ would appear to be a chance and rare discovery. However, it is important to point out that our IN sequences represent 51/71 subtype A and D sequences in the Stanford HIV database and 51/211 of all non-subtype B sequences from individuals failing any INSTI-based treatment.

Discussion

Since approval of the first-generation INSTIs (raltegravir in 2007), INSTIs remain the best choice as a backbone of ART. Current WHO guidelines recommend switch of patients on first-line therapy with no history of INSTI treatment to TDF/3TC (or FTC)/DTG.³³ Despite wide roll-out of dolutegravir in LMICs, there are increasing case reports of dolutegravir failure in patients started on dolutegravir-based regimens, and in INSTI-experienced patients carrying mutations including Q148H/K/R.^4,22

The recently approved second-generation INSTI bictegravir is available as part of a single fixed-dose formulation of TAF/FTC/BIC (Biktarvy). This TAF/FTC/BIC treatment is indicated in ART-naive and virologically suppressed patients (viral load <50 copies/mL) with no history of treatment failure and drug resistance.³⁴ As previously reported with dolutegravir,^4,35,36 bictegravir also shows potent inhibition of a broad range of drug-resistant viruses in vitro.^32,37 Before recent cabotegravir approval as a long-acting injectable, cabotegravir was tested in combination with rilpivirine in both ART-naive and virologically suppressed patients as an oral treatment or a long-acting injectable in the LATTE 1, LATTE 2, ATLAS and FLAIR clinical trials. In LATTE 1 and 2, INSTI-associated Q148R mutation emerged, while in the FLAIR and ATLAS studies, G140R, Q148R, E138A/T/K mutations and the N155H mutation, respectively, were observed in patients with virological failure.^13,17,38,39

In this study, we tested the HIV-1 inhibition profiles of cabotegravir and bictegravir using virus carrying patient-derived subtype A and D INs. The NL4-3 alone or carrying a WT subtype A or D IN coding region showed no difference in susceptibility to raltegravir, elvitegravir and dolutegravir or differences in replicative fitness to those previously reported.⁴ Recombinant viruses carrying either single primary INSTI-resistance mutations (N155H, Y143R, E138A) remained susceptible to both bictegravir and cabotegravir. We have previously shown that N155H is the predominant INSTI-resistance mutation (17.6%) in HIV-1 patients failing a third-line raltegravir-based regimen in Uganda.²⁶ Viruses carrying N155H in subtype A or D IN were susceptible to bictegravir (FC 1–2.3), comparable with susceptibility observed with HIV-1 subtype B virus (FC 1).³² In HIV-1-infected patients failing raltegravir, Q148H/K/R mutation emerges later in the course of infection, replacing N155H mutation of higher replication fitness but exhibiting lower resistance to raltegravir.⁴⁰ In our study, subtype A recombinant virus carrying E138A, G140A, Q148R and G163R mutations and subtype D virus carrying E138K, G140A, Q148K and S147G mutations showed high-level resistance to both cabotegravir and bictegravir, similar to the resistance previously reported to dolutegravir (FC >100).⁴ In the studies presented herein, the triple and quadruple INSTI-resistance mutations emerged in two different patients infected with subtype A and D HIV-1, respectively, and failing a raltegravir-based treatment regimen. The high-level cross-resistance to bictegravir and cabotegravir by this patient-derived HIV-1 has been confirmed with the same set of mutations in a study screening for bictegravir and cabotegravir using a subtype B HIV-1 with these mutations introduced in vitro.¹⁴ Despite high-level resistance to both drugs caused by the raltegravir-selected primary mutations Q148K/R, G140A, E138K/A and S147G, the FC cross-resistance to cabotegravir was more pronounced. Bictegravir, compared with cabotegravir, appears to be better accommodated in the binding pocket of IN even with these mutations, possibly due to the oxazepane ring attached to the metal-chelating scaffold of bictegravir.⁸ The superiority of bictegravir compared with cabotegravir to inhibit viral replication in viruses carrying substitutions at G140 and Q148 has also been reported in studies of mainly subtype B viruses.^7,14,41

Multiple INSTI-resistance mutations (>3) emerging in HIV-1-infected patients is rare but its detection during raltegravir failure is not surprising in highly cART-experienced patients, especially if raltegravir failure is prolonged due to the limited virological monitoring common in LMICs.^2,3 Despite increasing reporting of patients failing INSTIs with multiple primary mutations in non-B HIV-1 subtypes, we found most occurrence of Q148H/K/R and associated primary mutations to be in subtype B viruses. This may be associated with early use of INSTIs in high-income countries where subtype B virus is predominant. The HIV-1 genotypes with Q148H/K/R in association with multiple INSTI primary mutations is currently <3% in Uganda and 5.2% in non-B HIV-1 subtypes, which encourages use of bictegravir and cabotegravir, but increased INSTI resistance surveillance will be required as INSTIs become accessible in LMICs.

IN-catalysed integration into cellular DNA is a two-step process of 3′-processing and strand transfer process and is required for viral replication and infectivity.⁴² IN-resistance mutations including N155H and E138K, and added mutations, reduce efficiencies of 3′-processing and strand transfer activities, which impairs HIV-1 integration capacity.^43,44 We found that all IN-recombinant viruses tested exhibited <50% capacity to integrate into cellular DNA, which may explain our previous observation of loss of replication fitness associated with these viruses.⁴

Conclusions

Emergence of multiple INSTI-resistance mutations after prolonged virological failure on raltegravir presents a huge threat to the efficacy of bictegravir, cabotegravir and dolutegravir regimens, the latter being previously reported.⁴ Bictegravir and cabotegravir retain potency in patients carrying most single primary INSTI-resistance mutations with or without secondary mutations. However, with common use of raltegravir-based regimens for third-line treatments, failure of prolonged raltegravir-based treatment, as well as spread of raltegravir-resistant viruses carrying multiple INSTI-resistance mutations, could impact the use of these second-generation INSTIs. Given the infrequent use of drug resistance genotyping in almost all LMICs, we have not determined the prevalence of the specific INSTI drug resistance genotypes in both individuals failing INSTI treatment or in the treatment-naive population.

Funding

This work was supported by research grants from the National Institutes of Health (AI-49170), Canadian Institutes of Health Research (CIHR) (377790) and sponsored research grants from Gilead Sciences to E.J.A., the CIHR (BC-370929) to S.D.B., the Government of Canada through Genome Canada and the Ontario Genomics Institute (OGI-131) to M.A. Salaries for this work were supported by Webster Family Chair in Viral Pathogenesis at the University of Otago for M.E.Q.-M., the Tier I Canada Research Chair in HIV Pathogenesis and Viral Control for E.J.A. and the Queen Elizabeth II Diamond Jubilee scholarship (DLI O19375892122) as well as the International Ontario Graduate Scholarship international for E.N.

Transparency declarations

None to declare.

Author contributions

E.N., Y.L. and P.S.R. performed drug resistance assays and experimentation in this article. E.N., M.A., A.S.O. and M.E.Q.-M. performed data analyses. E.N., I.N., F.K. and C.M.K. recruited all the patients for this study. E.J.A., M.E.Q.-M. and S.D.B. procured the funding and E.J.A. designed, supervised and guided the overall direction of the study as the principal investigator.

Supplementary data

Tables S1 and S2 and Figures S1–S3 are available as Supplementary data at JAC Online.