Low-level viraemia despite emergence of dolutegravir-resistant variants

Dolutegravir (DTG), an integrase strand transfer inhibitor (INSTI)-based HIV-1 therapy, is widely recommended in first-line and second-line regimens.1,2 Integrase strand transfer inhibitor resistance mutations associated with DTG-containing regimens have been well described, most often occurring after DTG monotherapy or in INSTI-experienced patients.3Although rare, emergence of these mutations has also been described in patients on DTG-containing triple-drug regimens4 and INSTInaïve patients.5,6 The R263K mutation is commonly associated with the emergence of DTG resistance but reduces viral fitness and DNA integration.7,8 Here we describe a case of very slow viral decline (~42 months) in a treatment-experienced, INSTI-naïve patient on a DTG-based triple therapy regimen.


Introduction
Dolutegravir (DTG), an integrase strand transfer inhibitor (INSTI)-based HIV-1 therapy, is widely recommended in first-line and second-line regimens. 1,2 Integrase strand transfer inhibitor resistance mutations associated with DTG-containing regimens have been well described, most often occurring after DTG monotherapy or in INSTI-experienced patients. 3 Although rare, emergence of these mutations has also been described in patients on DTG-containing triple-drug regimens 4 and INSTInaïve patients. 5,6 The R263K mutation is commonly associated with the emergence of DTG resistance but reduces viral fitness and DNA integration. 7,8 Here we describe a case of very slow viral decline (~42 months) in a treatment-experienced, INSTI-naïve patient on a DTG-based triple therapy regimen.

Case
A 43-year-old man was diagnosed with HIV-1 in 2012 and commenced on an antiretroviral treatment (ART) regimen consisting of tenofovir disoproxil fumarate (TDF), emtricitabine (FTC) and efavirenz (EFV). Initial poor adherence to treatment and virologic failure was reported for 2012-2015 (viral load range 14 269 copies/mL -173 455 copies/mL), and due to concerns of having possibly acquired resistance to this regimen, treatment was empirically switched according to national guidelines to lamivudine (3TC), zidovudine (AZT) and lopinavir/ritonavir (LPV/r). He remained on this regimen with his family practitioner from 2015 until early 2018, when he was referred back to our hospital for persistently high HIV viral loads. He admitted to poor adherence secondary to LPV/r-associated diarrhoea, and was subsequently switched to 3TC, AZT and DTG. Patient clinical data and treatment history are shown in Table 1. After initial viral load decline to 6500 copies/mL the viral load returned to 11 700 copies/mL and nonadherence was suspected. Although admitting poor adherence during 2012-2018 the patient claimed perfect adherence since starting his DTG regimen. Adherence was confirmed initially by means of directly observed therapy, and then by unannounced random (same-day unscheduled clinic visit requests) 3TC and DTG drug concentration testing on two and four occasions (3TC 1.24 µg/mL -4.45 µg/mL, DTG 0.583 µg/mL -1.354 µg/mL in plasma). Drug resistance testing on multiple occasions ( Figure 1) identified M184V which, results in resistance to 3TC and increases susceptibility to AZT; K65R was also detected and also increases AZT susceptibility. 9 Integrase sequencing indicated susceptibility to DTG from January 2018 until September 2019 (21 months), after which resistance to DTG (N155H and R263K) was detected (January 2020) in a single sample. Follow-up INSTI drug resistance testing could not confirm DTG resistance. Even though DTG resistance was detected at one time point, the viral load continued to decline (Figure 1) despite the patient remaining on a DTG-based regimen. Plasma samples from three time points and cells (buffy coat/peripheral blood mononuclear cells [PBMCs]) from two time points were collected (S1, S2 & S3) for further investigation (Table 1 and Figure 1). All samples used in this study were collected after DTG initiation, with the relevant regimen over sampling period S1 and S2 being 3TC/AZT/DRV/r and S3 being 3TC/AZT/DTG.

Molecular workup
PBMCs were isolated from freshly collected whole blood ethylenediaminetetraacetic acid specimens according to the HIV/AIDS Network Coordination Cross-Network Peripheral Blood Mononuclear Cell Processing Standard Operating Procedure (www.hanc.info/labs/ labresources/procedures/Pages/pbmcSop.aspx). Cell-associated DNA and RNA were isolated from PBMCs and viral RNA was isolated from plasma as reported by Hong et al. 10 Single genome amplification products covering the p6-protease-reverse transcriptase (p6PrRT) region were http://www.sajhivmed.org.za Open Access generated and Sanger sequencing was performed as described previously. 11 Additionally, single genome amplicons covering the entire polymerase region were analysed for drug resistance mutations. Population and single genome nucleotide sequences were analysed for drug resistance mutations.

Ethical considerations
Ethical approval was obtained from the University of the Witwatersrand, where the patient was enrolled. Written informed consent was obtained from the patient.
Mutations N155K and R263K were only detected once in S3 and were confirmed with single genome sequencing (Table 3). Drug resistance analyses of integrase population and single genome amplicons from plasma RNA indicate two distinct populations: INSTI resistant and INSTI non-resistant (Table 3). These two integrase populations are reflected in phylogenetic analysis ( Figure 2).     At the time INSTI resistance was definitively confirmed, the patient's HIV viral load had already declined to 50 copies/mL. After much discussion with the patient, the patient elected to continue on his current regimen (rather than switch to a non-DTG-based regimen) with regular HIV viral load monitoring.

Discussion
We report a patient with viral load suppression despite a mixed viral population that includes INSTI-resistant variants with N155H and R263K. The patient was treatment experienced but INSTI naïve when initiated on a DTG-based regimen. DTG-resistant mutations N155H and R263K were first detected 24 months after DTG initiation; however, despite this the HIV-1 viral load continued decline to 50 copies/mL after 42 months. It is unclear why the DTG-resistant variants did not outcompete DTG-susceptible variants, but the low viral fitness could have contributed. Mutations R263K, M184V and K65R are all associated with decreased viral fitness and here occurred on the same viral variant. The slow but continued viral load decline could be due to the remaining activity and hypersusceptibility to AZT, in the presence of the M184V and K65R mutations. 12,13,14,15 Moreover CD4 + cell recovery and associated improved cytotoxic T-cell responses could have contributed to immune reconstitution to HIV and resulting viral load decline.
Although previously thought to be rare, several cases of virological failure as a result of DTG resistance in INSTInaïve patients have been described. 4,16,17 The factors associated with DTG resistance are poorly understood. Whereas DTG resistance when used in first-line therapy is exceedingly rare, it is not unusual when DTG is administered to treatment-experienced patients, 1,6 even when previously INSTI naïve. It is interesting that six of the nine cases of treatment-experienced, INSTI-naïve cases in the NADIA study 1 with drug resistance received the same regimen as this patient, albeit that in our case this was given as a third-line regimen in contrast to NADIA where the patients received DTG in second-line treatment.
Reported clinical risk factors associated with emergence of DTG resistance include poor treatment adherence, drug interactions and HIV factors such as a high baseline viral load. 6 In this case we report good adherence but the high viral load (> 100 000 copies/mL) prior to DTG initiation could be a contributing risk factor, even though INSTI resistance was only detected after 24 months. The combination of N155H and R263K is associated with virologic failure on DTG-based regimens, 4,18,19 in spite of the fitness cost of R263K. 7 Even though N155H partially compensates for the R263K viral fitness cost, development of compensatory mutations to the N155H, R263K strain may be unlikely or slow despite drug pressure. 20 No additional risk factors were identified in this case.
The presence of DTG resistance only partially explains the viraemia in this patient but offers no clarity on the susceptible sub-population. The detectable viraemia could be the result of intermittent adherence or compartmentalised replication or clonal viraemia. 11,21,22 Random unannounced drug concentration testing on various occasions suggest that intermittent adherence is an unlikely cause of the viraemia. Both compartmentalised replication and clonal viraemia could explain the relatively high viral load for an extended period of time. We investigated clonal viraemia in this case as described by Halvas et al., 11 but we were not able to find a matching proviral clone to the monotypic plasma virus population (data not shown); further investigation could be performed to shed light on this.
This case report provides further evidence that DTG is not impervious to drug resistance, especially when used in treatment-experienced patients. It also suggests that patients with DTG resistance may not present with immediate viral rebound due to the fitness cost of mutations and could present with delayed viral load suppression or low-level viraemia. It is not known whether this patient's HIV viral load will continue to remain low or whether additional compensatory mutations might allow the drug-resistant variant to regain fitness resulting in rebound. Continued vigilance is required when using DTG in treatmentexperienced patients and sensitive drug resistance assays may be needed to detect drug resistance at low viral load levels.