ARV pipeline: doravirine, GSK-744 and BMS-068

Simon Collins, HIV i-Base

CROI 2014 logo-190

HIV is still a fruitful market for new drug development and CROI 2014 included oral presentations from phase 2 data on three important pipeline compounds: doravirine, GSK-744 and BMS-068. 

Doravirine (MK-1439)

Doravirine (MK-1439) is a once-daily NNRTI in development at Merck which in previous studies reported -1.3 log reductions in viral load after 7 days monotherapy. Doravirine retains activity (less than 3-fold resistance) to major NNRTI-associated mutations, including K103N, Y181C, G190A and E138K, indicating likely lack of cross-resistance to rilpivirine and etravirine.

Results from a randomised, double-blind, phase 2 dose-finding study in treatment naïve patients looked at 20, 50, 100 and 200 mg doses compared to a control group using efavirenz 600 mg. [1] All patients also using tenofovir/FTC. There were approximately 40 patients in each arm and patients were stratified by baseline viral load above or below 100,000 copies/mL.

Baseline characteristics included median age 37 years (range 19-69); 90% men; 74% white, 20% black, 17% Hispanic. Approximately half the patients were from North America, 38% from Europe and 14% from Asia-Pacific. Median CD4 and viral load was approximately 380 cells/mm3 (range ~ 80–1140) and 4.6 log copes/mL (range 2.8–6.1), respectively, with 30% having viral load >100,000 copies/mL. Distribution was roughly similar between arms, although a higher percentage of people with CD4 counts <200 cells/mm3 were in the 25 mg doravrine group (17% vs ~10% in other arms).

Results for the primary endpoint of viral suppression <40 copies/mL at week 24, were 80%, 76%, 71% and 78% vs 64% for the increasing doravirine doses vs efavirenz arms respectively. When using a cut-off of <200 copies/mL the results were: 85%, 85%, 92% and 90% vs 81%, respectively. CD4 increases were +137 for the combined doravirine groups vs +121 for the efavirenz arm.

Virological non-responders using the <40 copy cut off across arms (n=1, 5, 6 and 5 vs 5) were undetectable using the <200 copy cut-off, all with reported >90% adherence. One person from the 25 mg doravirine arm had viral rebound: resistance results were not yet available. Response rates in patients with baseline viral load >100,000 copies/mL were 90%, 67%, 50% and 58% vs 54% using the <40 cut-off and 90%, 83%, 100% and 100% vs 92% using the <200 cut-off.

Neither viral efficacy nor tolerability results appeared to show a dose-related response. Few patients discontinued related to doravirine side effects and this was less than for efavirenz (approximately 2.5% vs 4.8%) and serious side effects were not considered drug-related (including suicidal ideation in a patient taking efavirenz). Most CNS-related side effects were more common with efavirenz.

Laboratory abnormalities were generally grade 1/2 with lipids and liver marker changes higher in the efavirenz group.

Participants in this study using doravirine roll over to the selected dose of 100 mg once-daily, for continued follow-up compared to efavirenz, out to 96 weeks.

GSK-744: dual maintenance therapy with rilpivirine

Although the headline news for GSK-744 (the follow-on integrase inhibitor to ViiV’s dolutegravir) were grabbed for data on the slow-release formulation for use as PrEP, [2, 3] results using daily oral dosing of 744 were presented from a phase 2 dose-finding study in treatment naïve patients. [4]

The half-life of GSK-744 is approximately 40 hours for the oral formulation and 40 days for the long-acting injection. In earlier studies, oral dosing of 5 mg and 30 mg GSK-744 produced a -2.0 log drop in viral load after 10 days monotherapy.

The current study compared 10 mg, 30 mg and 60 mg doses of oral GSK-744 compared to 600 mg efavirenz control (approximately 60 patients in each arm), all with investigator selected 2RTIs, with a roll over at week 24 to oral dual therapy using GSK-744 plus 25 mg rilpivirine.

This was generally a group in early infection. Baseline demographics included median age 32 (no range given); >95% men; race: 65% white, 30% African-American, 15% Hispanic [stet – totals >110%]. Median CD4 count was approximately 400 cells/mm3 (<5% <200 cells/mm3) with mean viral load or 4.2 – 4.4 log copies/mL (approximately 12% >100,000 copes/mL. No data on the range or variance for the baseline data were included in the presentation, limiting the ability to interpret the results. Choice of NRTIs was roughly 60% tenofovir/FTC vs 40% abacavir/3TC.

Results at week 24 were previously presented at the EACS conference in October 2013 with viral suppression <50 copies/mL in 88% vs 74% with efavirenz, and a selection of the 30 mg dose for weeks 24-48. [5]

At week 48, after 24-weeks dual maintenance treatment, viral suppression <50 copies/mL (ITT snapshot) was 93% for 744/rilpivirine vs 94% with efavirenz + 2NRTIs, with similar responses across arms and only a single discontinuation due to virologic failure (in the efavirenz arm). However, two patients had protocol defined virological failure in the dual therapy arm. One patient in the 10 mg arm was suppressed <40 copies/mL at week 40 but had resistance to both drugs with Q148R and E138Q at week 48. This person had suboptimal plasma levels of GSK-744 from week 2-36 and to rilpivirine from week 26-26. This was possibly related to a calorie-restricted diet (650 calories/day from week 40-48). No drug resistant mutations were found in the second person who had been randomised to the 30 mg arm.

The lower response rates in the efavirenz control group were largely due to higher CNS-related side effects, responsible for 13% discontinuation (compared to 2%, 2% and 7% with GSK-744). However, headache was more commonly reported with GSK-744 (22% vs 11% with efavirenz), although these were predominantly grade 1/2, with no discontinuations.

Responses by choice of NRTI were slightly higher with tenofovir/FTC vs abavacir/3TC (84% vs 79% and 72% vs 70%, in the GSK-744 vs efavirenz groups, respectively).

Future clinical studies will use dual long-acting formulations of GSK-744 and rilpivirine as injection-based ART, although an induction period using oral dosing is still likely due to safety concerns in the event of a hypersensitivity reaction.

BMS-663068: attachment inhibitor

BMS-068 (a prodrug of BMS-626529) is an attachment inhibitor that binds to gp-120 causing conformational changes that block CD4 attachment, making it active irrespective of viral tropism (although notably not to sub-type AE or Group O). After 8 days monotherapy, BMS-068 reduced viral load by up to -1.6 logs in both treatment-naïve and -experienced patients.

The current study included randomised patients to one of four twice (BD) or once (QD) daily doses of BMS-068: 400 mg BD, 800 mg BD, 600 QD and 1200 QD) and an atazanavir/r QD control group. There were 50 people in each arm (including 10 patients in each arm using 7 days monotherapy) and background therapy was with raltegravir plus tenofovir for all patients. Sensitivity to BMS-529 is an entry criteria (requiring IC50 < 100 nM and ~ 5% screening failures were due to higher IC50s). This was an international study with sites in the USA, Mexico, Peru and Germany. [5]

Approximate baseline characteristics included median age 39 years, 60% men and 40% women, 65% sub-type B and 15% sub-type C, with race categorised as ~ 60% non-white. Mean CD4 and viral load was ~ 250 cells/mm3 (with ~ 40% < 200) and 4.7 log copies/mL (with ~ 40% >100,000). Approximately 30% patients had one of more resistant mutation to one or more classes (but sensitivity to all study drugs was an entry criteria).

Although discontinuations ranged from 11–22%, these were largely due to withdrawal of consent, loss to follow-up, pregnancy or poor adherence, with very few dues to either lack of efficacy or side effects.

Viral response rates from 8 days of monotherapy were dose-related, ranging from -0.69 logs (400 mg BD) to -1.47 (1200 mg QD). Unlike other ART classes, a transient small increase in viral load occurred during the first two days of treatment prior to the viral load drop.

At week 24, viral load suppression to <50 copies/mL (ITT analysis) was achieved by 80%, 69%, 76% and 72% across the BMS-068 arms compared to 74% in the atazanavir/r control. When stratified by baseline viral load, all groups except the 1200 mg QD group had at least 15-20% lower response rates at above compared to below 100,000 copies/mL. No pattern in response was seen by sensitivity to BMS-068 by IC50 levels (+/-0.1 vs +/-1.0 vs +/-10 nM). CD4 increases were similar across all arms.

In the BMS-068 arms, none of the 15 serious adverse events (in 123 people) were attributed to the study drugs (with four discontinuations due to 1 non-specific EKG changes, 2 TB cases and 1 tenofovir-associated acute renal failure).

All participants on BMS-068 have now been rolled over to the 1200 mg QD dose for continued follow up, although the presenter emphasised that this has not necessarily been selected as the development dose for future studies.


Unless indicated otherwise, all references are to the Programme and Abstracts of the 21st Conference on Retroviruses and Opportunistic Infections (CROI), 3-6 March 2014, Boston. CROI 2014 links are to webcasts.

  1. Morales-Ramirez JO et al. Safety and antiviral effect of MK-1439, a novel NNRTI (+FTC/TDF) in ART-naive HIV-infected patients. 21st CROI, 3-6 March 2014, Boston. Oral late-breaker abstract 92LB.
  2. Andrews C et al. Correlating GSK1265744 plasma levels to prevention of rectal SHIV transmission in macaques. 21st CROI, 3-6 March 2014, Boston. Oral abstract 39.
  3. Garcia Lerma JG et al. Monthly GSK744 long-acting injections protect macaques against repeated vaginal SHIV exposures. 21st CROI, 3-6 March 2014, Boston. Oral late breaker abstract 39.
  4. Margolis D et al. 744 and rilpivirine as two-drug oral maintenance therapy: LAI116482 (LATTE) week 48 results. 21st CROI, 3-6 March 2014, Boston. Oral late-breaker abstract 91LB.
  5. Margolis D et al. Once-daily oral GSK1265744 (GSK744) as part of combination therapy in antiretroviral naïve adults: 24-week safety and efficacy results from the LATTE study (LAI116482). 14th EACS, 2013, Brussels. Oral abstract PS7.1.>
  6. Lalezari J et al. Attachment inhibitor prodrug BMS-663068 in ARV-experienced subjects: week 24 analysis. 21st CROI, 3-6 March 2014, Boston. Oral abstract 86.>

Links to other websites are current at date of posting but not maintained.