Short Communication: Comparative Evaluation of Coformulated Injectable Combination Antiretroviral Therapy Regimens in Simian Immunodeficiency Virus-Infected Rhesus Macaques

The use of nonhuman primate (NHP) models of AIDS to evaluate virologic and immunologic processes in the setting of suppressive combination antiretroviral therapy (cART) has been limited due, at least in part, to challenges in identifying feasible cART regimens that can sustainably suppress viral replication in macaques infected with pathogenic simian immunodeficiency virus (SIV) to clinically relevant levels such as those seen in cART-suppressed HIV-infected individuals.¹ In addition to issues arising from differences in drug potency against SIV vs. HIV and differences in the bioavailability and pharmacokinetics of antiretrovirals in NHPs vs. humans, the feasibility of daily administration of multiple antiretroviral drugs to NHP research animals can pose practical and logistical challenges, especially for drugs given via the oral route more than once daily.

The development of safe and effective cART regimens providing potent antiviral activity against SIV in pathogenic infection models coupled with convenient administration methods that achieve consistent efficacious drug levels would be of great benefit to HIV treatment and cure research and would allow for the in vivo study of residual virus sources that persist in the face of suppressive cART as well as the in vivo evaluation of proposed viral eradication/functional cure strategies. Here, we compared in SIV-infected rhesus macaques the pharmacokinetic and antiviral activity of two cART regimens, one containing three drugs and the other containing four drugs, which were coformulated for administration as a single once daily subcutaneous injection.

We and others^2–5 have previously reported on the use of cART regimens containing the protease inhibitor darunavir (DRV), often with the addition of ritonavir (RTV) as a putative DRV boosting agent based on clinical practice in humans, in SIV-infected rhesus macaques. To specifically evaluate the anti-SIV activity of DRV in vivo in the absence of other antiretroviral drugs, we treated two Indian-origin rhesus macaques chronically infected with SIVmac239 with DRV/RTV “monotherapy” (Fig. 1) and monitored plasma viral RNA levels by quantitative RT-PCR as previously described.⁵ DRV and RTV (both purchased from the NIH pharmacy) were initially administered via the oral route at fixed doses of 400 mg (equivalent to 36 mg/kg for AZ32 and 54 mg/kg for AY37) and 100 mg (equivalent to 9 mg/kg for AZ32 and 13 mg/kg for AY37), respectively, twice daily. After 2 weeks of treatment, the DRV dose was increased to 600 mg (equivalent to 55 mg/kg for AZ32 and 81 mg/kg for AY37) twice daily while the RTV dose was maintained.

Prior to treatment, both animals had plasma viral loads >10⁵ viral RNA (vRNA) copies/ml. Within 1 week of treatment initiation, plasma viral loads for both animals declined by >1 log to 10⁴ vRNA copies/ml. Thereafter, viremia in animal AY37 continued to decline by another 1 log over the next 7 days of treatment before rebounding over the subsequent 26 days to pretreatment levels by day 40 despite continued treatment with 600 mg DRV twice daily. After an initial decline, plasma viral loads in animal AZ32 rebounded more quickly, returning to pretreatment levels by day 11 and remaining elevated until again transiently declining by ∼1 log between 2 and 3.5 weeks after switching to the elevated dose of DRV before returning to pretreatment levels by day 40.

To determine if virologic rebound in the setting of DRV treatment was associated with the acquisition of drug resistance mutations, suggestive of SIV selective pressure applied by DRV treatment in infected rhesus macaques, we performed single-genome amplification (SGA)-based viral genome sequencing⁶ using primers spanning the protease coding region. After 40 total days of DRV/RTV monotherapy, 10 of 10 and 11 of 14 sequences for AY37 and AZ32, respectively, contained the I50V mutation (Fig. 1), previously described as a DRV resistance mutation for HIV-1.⁷ This mutation was not identified in any of eight viral protease sequences derived from either animal just prior to DRV therapy, suggesting that DRV/RTV monotherapy conferred selection pressure in vivo and that virologic rebound was likely due to the emergence of drug resistance. There was also apparent selection for protease mutation H41Y, a change not previously described with DRV resistance, which was not present in any pretreatment sequences but was present in 10 of 10 posttreatment sequences in one animal, AZ32. The functional importance of the H41Y mutation is unclear, though it may represent a compensatory mutation to restore fitness in the presence of I50V.

We next measured the pharmacokinetic steady state plasma concentrations (C_ss) achieved in the two DRV/RTV-treated animals on day 12 (400 mg) and day 21 (600 mg) of therapy immediately prior to the next scheduled administration for each dose level (Table 1). Oral drug administration, such as that used here for a fixed dose of DRV, requires hiding drugs in food/treats provided ad libidum, which can lead to inconsistencies in dose timing and the amount of dose delivered, and also offers challenges for body weight-normalized drug dosing, which can result in animal-to-animal variability in achieved in vivo drug levels. Plasma DRV levels for both animals at both drug doses exceeded the in vitro protein adjusted (PA) EC₉₅ for DRV (measured in a multiround infection assay on CEMx174 target cells⁸); however, plasma DRV exposures were highly disparate between the two treated animals, with the measured plasma C_ss at the 600 mg dose for animal AZ32 more than 10-fold lower than that of animal AY37, consistent with the larger body weight and lower weight-normalized dose in animal AZ32 (Table 1).

In addition, although the measured DRV plasma C_ss for animal AY37 was ∼1.7-fold higher at the 600 mg dose than at the 400 mg dose, in animal AZ32 the determined C_ss at 600 mg was more than 3-fold lower than at 400 mg, again highlighting the potential inconsistencies of oral drug administration both within animals and between animals. Consistent with the lower DRV plasma concentrations in animal AZ32, virologic rebound, suggestive of the development of drug resistance, occurred more rapidly following initial plasma viral load decline (Fig. 1).

Recent improvements in the feasibility and effectiveness of cART regimens in SIV-infected macaques have included the development of multidrug cocktails containing two nucleo(t/s)ide RT inhibitors (NRTIs), emtricitabine (FTC) and tenofovir (TFV, PMPA), and one integrase strand transfer inhibitor (INSTI), such as raltegravir (RAL) or dolutegravir (DTG), given with or without oral DRV^2,9,10 (Del Prete et al., unpublished results). Coformulating multiple drugs with demonstrated potency against SIV replication in vivo into a single injectable regimen offers several practical advantages, including improved ease and consistency of administration compared with regimens that require multiple daily injections and/or oral drug administration.

Although DRV has apparent anti-SIV activity in vivo (Fig. 1) it is unclear whether it might enhance the antiviral activity of the TFV/FTC/DTG cART regimen, particularly if better DRV dose consistency and in vivo drug levels can be achieved. Additionally, although TFV has become a core component of most cART regimens used in SIV-infected rhesus macaques, occasional renal toxicity has been noted in association with acute high dose plasma exposure of TFV coupled with chronic administration; chronic administration is often required for studies of persistent residual virus in the setting of suppressive cART and for evaluations of viral eradication/functional cure strategies.^5,11,12 We therefore sought to develop and evaluate two new coformulated injectable drug regimens in SIVmac239-infected Indian-origin rhesus macaques.

The first regimen, hereafter referred to as the “triple regimen,” contained FTC, DTG, and TDF (TFV prodrug; tenofovir disoproxil fumarate) coformulated as a once daily subcutaneous injectable; TDF can achieve in vivo antiviral potency similar to TFV but with lower plasma exposure.^13,14 Although TFV is typically administered to SIV-infected macaques at doses ranging from 15 to 30 mg/kg,¹ in the triple regimen TDF was administered at 5.1 mg/kg, with FTC and DTG administered at 50 mg/kg and 2.5 mg/kg, respectively. The second regimen, hereafter termed the “quad regimen,” contained the same compounds and dosages as the triple regimen with the addition of DRV at 40 mg/kg, with all four compounds again coformulated to enable administration via a single once daily subcutaneous injection.

For compounds used in the coformulated triple and quad regimens, TDF and FTC were manufactured by Gilead Sciences, Inc., DTG sodium salt and free base were purchased from Shanghai Medicilon, Inc., and DRV ethanolate was purchased from Toronto Research Chemical Inc. To make 1 liter of the triple regimen, first ∼1 liter of a 15% (w/w) Kleptose solution was prepared in sterile water using Kleptose HPB parenteral grade (Roquette). Next 2.5 g of DTG free base equivalent was placed in a 1-liter volumetric flask and filled to 80% with the 15% Kleptose solution and then stirred and sonicated until all material was dissolved, in about 30 min. Next 5.1 g of TDF was added and again the solution was stirred and sonicated until all material was dissolved, in about 5 min. Then, 50 g of FTC was added and the solution was again stirred and sonicated until all material was dissolved, in about 3 min. Finally, the total volume was brought up to 1 liter with the 15% Kleptose solution, pH confirmed/adjusted (final pH ∼4.2), filtered through a 0.22-μm filter, aliquoted (100-ml bottles), and stored frozen at −20°C.

To make 100 ml of the quad regimen, first a DRV/DTG solution was prepared in Polyethylene glycol 400 (PEG400; Spectrum Chemical) by adding 4.34 g of DRV-ethanolate and 0.25 g of DTG free base to a 50-ml volumetric flask and bringing up to 90% full with PEG400 followed by stirring for 1 h and then sonication for 20 min. The total volume was then brought up to 50 ml with PEG400, followed by stirring for an additional hour and sonicating for 10 min. Next, a TDF/FTC solution was prepared in 30% Kleptose; 0.51 g of TDF, 5.0 g of FTC, and 15 g of Kleptose were added to a 50-ml volumetric flask, then sterile water was added to bring to ∼90% full and the solution was stirred for 1 h. The total volume was then brought up to 50 ml with water and the solution stirred for an additional 3 h. Finally the 50 ml of DRV/DTG/PEG400 solution and 50 ml of TDF/FTC/Kleptose solution were combined in a 100-ml bottle, stirred, sterile filtered, and stored frozen at −20°C.

All cART formulations were made fresh on a monthly basis and stored frozen at −20°C until the day of administration, upon which individual bottles were thawed to room temperature prior to subcutaneous administration at 1 ml/kg animal body weight.

Six Indian-origin rhesus macaques were intravenously inoculated with 300 IU (based on infectious titer determined on TZM-bl reporter cells¹⁵) of SIVmac239X¹⁶ (Fig. 2). On day 13 postinfection (dpi), three animals began treatment with the triple regimen while the remaining three animals began treatment with the quad regimen. Pretreatment viral loads on day 13 ranged from 8.6 × 10⁶ to 7.0 × 10⁷ vRNA copies/ml plasma, with mean pretreatment plasma viral loads of 4.2 × 10⁷ vRNA copies/ml and 3.1 × 10⁷ vRNA copies/ml for the triple and quad regimen treatment groups, respectively. The addition of DRV in the quad regimen conferred no apparent increase in the kinetics of first phase viral load decline over the first 12 days of treatment when compared with the triple regimen (p = 0.26), and plasma viral loads declined by a similar amount, 4.5–5 logs, after 4 weeks on cART for both treatment groups (Fig. 2). After 4 weeks of treatment the quad treatment group was switched to the triple regimen and all animals continued to receive the triple regimen for an additional 36 weeks of treatment with plasma viral load monitoring. Durable suppression of plasma viral loads to below the limit of quantification (<30 vRNA copies/ml) was achieved in all six animals after 12–20 weeks of treatment and was maintained through 40 weeks of follow-up on cART. The animal that achieved suppression to <30 vRNA copies/ml plasma most rapidly (animal ZI69 in the triple regimen group) notably had the lowest pretreatment viral load of 8.6 × 10⁶ vRNA copies/ml plasma. Complete blood counts (CBC) and serum chemistries were monitored throughout treatment. No systemic reactions or blood abnormalities were noted during cART treatment, and both regimens were well tolerated.

Plasma drug concentrations at 2 h postdose (C_2hr) on day 1 (start of cART) and at 24 h post-dose on day 30 (i.e., steady state trough levels, C_ss) were determined for all regimen drugs in the six treated animals (Table 1). Consistent plasma drug concentrations were achieved for all of the coformulated compounds, with mean C_ss drug levels that exceeded the target levels for DRV and DTG. Importantly, peak (C_2hr) plasma tenofovir levels in both treatment groups (mean = 1,704 nM both groups combined) were substantially lower than peak levels in animals treated with TFV at 20 mg/kg (mean = 36,000 nM; unpublished results) and were similar to maximum tenofovir levels measured in humans following a single oral dose of TDF.¹⁷ Because TDF can achieve higher active intracellular drug levels at lower plasma exposures than TFV,¹⁸ this lower peak plasma concentration may mitigate the potential renal toxicity associated with high dose plasma exposures of TFV while achieving therapeutic intracellular concentrations of active drug metabolites.

Although, TDF and FTC concentrations were comparable between the triple and quad regimen-treated animals, mean plasma DTG C_2hr and C_ss levels were 3.0-fold (66%) and 1.7-fold (40%) lower, respectively, in the quad regimen-treated animals, likely due to the presence of DRV in the quad regimen as has been noted in human studies when DTG has been coadministered with DRV.^19,20 Notably, trough (C_ss) DRV levels were higher and more consistent between animals in the quad regimen-treated macaques than in the monotherapy-treated macaques (Table 1), despite twice daily administration of DRV at comparable or higher weight-normalized doses in the monotherapy-treated animals. The higher levels and improved consistency achieved with once daily subcutaneous dosing were likely due to improved drug bioavailability and greater control over dose timing and weight-dependent dosing than is afforded by the oral route.

Although we show that DRV is active against SIV replication in infected rhesus macaques, and despite the achievement of higher and more consistent plasma DRV levels with subcutaneous administration, the addition of DRV to the coformulated TDF/FTC/DTG triple cART regimen did not lead to a measurable increase in the rate of plasma viral load decay in treated animals. Although DTG C_ss levels were lower in quad regimen-treated animals compared with triple regimen-treated animals, raising the possibility that the lower DTG levels may have offset an added antiviral benefit from the addition of DRV, we note that the DTG levels in quad regimen-treated animals were still ∼4-fold greater than the in vitro PAEC₉₅ for DTG against SIV. It thus seems unlikely that the 1.7-fold reduction in DTG plasma levels with the DRV-containing quad regimen resulted in a marked reduction in the overall antiviral activity of the regimen. In either case, for these specific coformulated injectable drug regimens, the addition of DRV to the regimen does not appear to confer any additional virologic benefit.

Overall, our findings show that a two class (NRTI/INSTI), three drug cART regimen formulated for coadministration of all three drugs as a once daily subcutaneous injection can effectively and durably suppress SIVmac239 replication in Indian-origin rhesus macaques to clinically relevant levels (<30 vRNA copies/ml plasma). The feasibility of this dosing regimen, coupled with its pharmacokinetic consistency and potential for improved safety through the use of the TDF prodrug form of TFV, makes it a highly attractive regimen for studies of AIDS virus persistence and latency and for the evaluation of viral eradication strategies in a relevant in vivo model system.