Original Article Safety and Efficacy of The

Antiviral Therapy 2017; 22:215–223 (doi: 10.3851/IMP3112)

Original article Safety and efficacy of the HIV-1 attachment inhibitor prodrug fostemsavir in antiretroviral- experienced subjects: week 48 analysis of AI438011, a Phase IIb, randomized controlled trial Melanie Thompson1*, Jacob P Lalezari2, Richard Kaplan3, Yvett Pinedo4, Otto A Sussmann Pena5, Pedro Cahn6, David A Stock7, Samit R Joshi7, George J Hanna8, Max Lataillade7, the AI438011 study team

1AIDS Research Consortium of Atlanta, Atlanta, GA, USA 2Quest Clinical Research, San Francisco, CA, USA 3Desmond Tutu HIV Foundation, Cape Town, South Africa 4Asociacion Civil Via Libre, Lima, Peru 5Asistencia Cientifica de Alta Complejidad SAS, Bogotá, Columbia 6Fundacion Huesped, Buenos Aires, Argentina 7Bristol-Myers Squibb, Wallingford, CT, USA 8Bristol-Myers Squibb, Princeton, NJ, USA

*Corresponding author e-mail: [email protected]

Background: Fostemsavir is a prodrug of temsavir, an intent-to-treat [mITT] and observed analysis, respectively); attachment inhibitor that binds directly to HIV-1 gp120, 71% and 88% for ATV/r subjects (mITT and observed). blocking initial viral attachment and entry into host CD4+ Observed virological response rates were 74–100% versus T-cells. Efficacy, safety and dose-response data of fostem- 96% (fostemsavir versus ATV/r) in subjects with baseline savir in treatment-experienced, HIV-1-infected subjects, viral load <100,000 copies/ml and 60–91% versus 71% through week 48, are reported. when baseline viral load was ≥100,000 copies/ml. Across Methods: AI438011 is an ongoing Phase IIb, randomized, fostemsavir arms, median CD4+ T-cell count increases active-controlled trial (NCT01384734). Subjects were from baseline were 145–186 cells/µl and 142 cells/µl for randomized 1:1:1:1:1 into five arms: fostemsavir (400 mg the ATV/r arm. Fostemsavir doses were generally well tol- twice daily, 800 mg twice daily, 600 mg once daily or erated and no fostemsavir-related adverse events led to 1,200 mg once daily) and a reference arm (ritonavir- discontinuation. boosted atazanavir [ATV/r] 300/100 mg once daily), each Conclusions: Through week 48, fostemsavir continued to with a backbone of raltegravir 400 mg twice daily plus be well tolerated and showed similar efficacy to ATV/r. tenofovir disoproxil fumarate 300 mg once daily. These results support the ongoing Phase III trial in heav- Results: In total, 251 subjects were treated. Through week ily treatment-experienced adults with limited therapeutic 48, the proportion of fostemsavir subjects with HIV-1 options (≤2 classes of active antiretrovirals remaining). RNA <50 copies/ml was 61–82% and 77–95% (modified ClinicalTrials.gov identifier: NCT01384734.

Introduction

Lifelong treatment with combined antiretroviral ther- order to minimize drug resistance, decrease progression apy has dramatically improved life expectancy in the to AIDS, improve life expectancy and reduce the risk HIV-1 population [1,2]. Current treatment guidelines of viral transmission to non-infected individuals [5–7]. recommend that antiretroviral-experienced patients However, there are limited therapeutic options for failing their current regimen be treated with at least treatment-experienced HIV-1-infected patients who two and preferably three fully active agents [3,4]. Such may have multi-class drug resistance, comorbidities and regimens have the highest likelihood of providing maxi- multiple concurrent medications with potential drug– mal and durable suppression of plasma HIV-1 RNA in drug interactions that must be managed [3,4,8–10].

AVT-16-OA-3925_Thompson.indd 215 02/06/2017 14:22:53 M Thompson et al.

Therefore, new classes of antiretroviral agents with previously [18]. The study was conducted at 53 hos- favourable tolerability and safety profiles are needed pitals and outpatient clinics across 10 countries in to attain potent and durable antiviral suppression in South America, North America, Europe and Africa. treatment-experienced populations. Eligible subjects were randomized 1:1:1:1:1 into five Fostemsavir (formerly BMS-663068) is a prodrug treatment arms: four dosing arms of fostemsavir that is metabolized to the active moiety temsavir (for- (400 mg twice daily, 800 mg twice daily, 600 mg once merly BMS-626529), a first-in-class HIV-1 attachment daily or 1,200 mg once daily) and a reference arm inhibitor that binds directly to the viral envelope pro- (ATV/r 300/100 mg once daily), each with a back- tein gp120, blocking initial attachment of the virus bone of RAL 400 mg twice daily plus TDF 300 mg to the host CD4+ T-cells (the proposed mechanism of once daily. As treatment guidelines do not specify a action of temsavir is shown in Additional file 1) [11,12]. specific standard-of-care regimen for treatment-expe- Unlike CCR5 antagonists, temsavir binds directly to rienced individuals, a combination of RAL plus TDF the virus rather than to host cells [12,13], and there- was selected at the time of study design as the opti- fore acts prior to co-receptor binding and fusion [12]. mized background therapy. It was anticipated that It is active against CCR5-, CXCR4- and dual-tropic the majority of screened subjects would have a virus (R5X4) strains of HIV-1 [12,14–16]. Preliminary in susceptible to both agents and, with fostemsavir, sub- vitro data have shown that with the exception of subjects would receive three fully active agents providing type AE and Group O, HIV-1 viruses are susceptible to a potent treatment regimen. The fixed backbone also temsavir regardless of CCR5 tropism and subtype [14]. allowed for fostemsavir dose differentiation. A subset Temsavir also has a unique resistance profile with no of subjects participated in a fostemsavir monotherapy in vitro cross-resistance to other classes of antiretrovi- study to confirm fostemsavir activity. Randomization rals [15]. A proof-of-concept study found that adminis- was performed with a computer-generated code and

tration of fostemsavir for 8 days resulted in maximum stratified by phenotypic sensitivity to temsavir (IC50

median decreases from baseline in plasma HIV-1 RNA <1.3 nmol/l or IC50 ≥1.3 nmol/l) as measured by the

of 1.21–1.73 log10 copies/ml [17]. Monogram Biosciences (San Francisco, CA, USA) Based on these findings, a Phase IIb study (AI438011) PhenoSense® Entry assay. Subjects and investigators was designed to investigate the efficacy, safety and dose were blinded to fostemsavir dose but not to alloca- response of fostemsavir versus ritonavir-boosted ataza- tion. Unblinding occurred after the last participant navir (ATV/r), each with raltegravir (RAL) and teno- reached week 24 for the sponsor and week 48 for the fovir disoproxil fumarate (TDF), in HIV-1-infected, participants and investigators. treatment-experienced subjects [18]. Thirty-two subjects The study was performed according to international participated in a 7-day fostemsavir monotherapy lead- laws and guidelines and was designed and conducted in substudy. Median changes in HIV-1 RNA from base- according to the protocol (and any amendments)

line at day 8 of -0.69, -1.40, -1.28, -1.44 log10 copies/ml jointly by the study investigators and sponsor. Each site were reported for the 400 mg twice-daily group, 800 mg conducted the study with oversight and approval from twice-daily group, 600 mg once-daily group and 1,200 an Institutional Review Board or Independent Ethics mg once-daily group, respectively [18]. We previously Committee. All participants were required to provide reported the week 24 analysis in which 69–80% of sub- written informed consent in agreement with Declara- jects receiving fostemsavir in the modified intent-to-treat tion of Helsinki principles. (mITT) population achieved HIV-1 RNA <50 copies/ml compared with 75% of subjects receiving ATV/r. In the Study population observed population, 78–87% of subjects receiving fos- Eligible subjects had plasma HIV-1 RNA ≥1,000 cop- temsavir achieved HIV-1 RNA <50 copies/ml compared ies/ml, CD4+ T-cell counts >50 cells/mm3 and were men with 86% of subjects receiving ATV/r [18]. There were and women aged ≥18 years who were antiretroviral no fostemsavir-related serious adverse events (SAEs) or treatment-experienced (defined as current or previous fostemsavir-related adverse events (AEs) leading to dis- exposure to at least one antiretroviral drug for at least continuation [17]. The efficacy and safety of fostemsavir 1 week). Subjects were required to be naive to treatment compared with ATV/r through week 48 are presented. with any integrase inhibitor (defined as <1 week prior exposure) and have a screening genotype and pheno- Methods type indicating susceptibility to ATV/r, RAL and TDF (PhenoSense® GT and Integrase assays, Monogram Study design Biosciences). Subjects were screened at entry using AI438011 (NCT01384734) is an ongoing Phase the PhenoSense® Entry assay (Monogram Biosciences)

IIb, randomized, active-controlled, blinded-to-fos- and eligible if they had a temsavir IC50 <0.1 mM. Sub- temsavir dosing trial. Methods have been described jects with a history of allergy to any study drugs were

AVT-16-OA-3925_Thompson.indd 216 02/06/2017 14:22:53 HIV-1 attachment inhibitor fostemsavir: week 48 analysis

excluded. Those with newly diagnosed HIV-1-related Results opportunistic infections were excluded as were those with hepatitis B surface antigen, HCV antibody or Subjects quantifiable HCV RNA. Between 26 July 2011 and 16 July 2012, 581 participants were screened, 254 were randomized into the End points and assessments study and 251 received treatment (Figure 1). Overall, The primary end points were proportion of subjects 303 (52%) of those screened did not meet study cri- with plasma HIV-1 RNA <50 copies/ml at week 24, teria and were excluded from study treatment. The and frequency of SAEs and AEs leading to discontinu- most common reason for exclusion was plasma HIV-1 ation through week 24. Protocol-defined virological RNA <1,000 copies/ml (29.1%). Other reasons for failure was confirmed as HIV-1 RNA ≥50 copies/ml not randomizing enrolled subjects included temsa-

at week 24 or later (second measurement within 2–4 vir IC50 >0.1 mM (16.8%), genotypic or phenotypic weeks of the original sample), or virological rebound resistance to study drugs (15.3%) and CD4+ T-cell (confirmed HIV-1 RNA ≥50 copies/ml at any time count <50 cells/mm3 (9.2%). The PhenoSense® Entry after confirmed suppression to <50 copies/ml, or con- assay failed to produce a result in 26% of screen-

firmed increase in HIV-1 RNA of ≥1 log10 copies/ml ing samples in this study. Of the subjects successfully above the nadir level, for nadir ≥50 copies/ml). Safety tested in the PhenoSense® Entry assay, approximately

assessments, including vital signs, physical examina- 6% had a temsavir IC50 >100 nM at screening. Of tion, AEs, laboratory measurements and electrocar- those receiving treatment, 48/200 (24%) across the diographs were recorded throughout all phases of the fostemsavir arms, and 14/51 (28%) in the ATV/r arm study. AEs were coded according to MedDRA version failed to complete 48 weeks’ treatment (Figure 1). 15.1. Investigators assessed the severity and relation- Baseline demographic and disease characteristics ship to study drug. Major secondary end points at were broadly similar across all treatment arms [18]. week 48 included proportion of subjects with plasma Median age was 39 years, 151/251 (60%) subjects HIV-1 RNA <50 copies/ml, change from baseline in were male and 76/251 (30%) were Black/African- CD4+ T-cell count and frequency of SAEs, and AEs American. The median baseline HIV-1 RNA was + leading to discontinuation through week 48. Adher- 4.85 log10 copies/ml and the median baseline CD4 ence to treatment was evaluated by the investigative T-cell count was 230 cells/ml. Approximately 50% of staff at every treatment visit through interviews with subjects had ≥1 major protease inhibitor (PI), nucleo- the subjects and through examination of returned side reverse transcriptase inhibitor (NRTI) or non- medication. Subjects were instructed to bring all nucleoside reverse transcriptase inhibitor (NNRTI) unused medication back in the original container to resistance-associated mutation at baseline (M184V/I, each visit. 31%; K103N, 29%; thymidine analogue mutations, A pre-specified subgroup analysis was conducted 13%; major PI mutations, 2%). These data were to assess the number of subjects achieving plasma reported previously [18]. HIV-1 RNA <50 copies/ml according to the pre-

defined temsavir IC50 strata (< or ≥1.3 nM), and a Virological and immunological efficacy

series of additional temsavir IC50 breakpoints (< or The proportion of subjects receiving fostemsavir who ≥0.1, 1.0 and 10 nM) covering the range of baseline achieved HIV-1 RNA <50 copies/ml at week 48 was

IC50 values included in the study. Post-hoc analyses between 61% and 82% and was 71% for those receiv- were conducted to assess the number of subjects with ing ATV/r (mITT population). Response rates within plasma HIV-1 RNA <50 copies/ml per participation the observed population were between 77% and 95% in the monotherapy substudy or baseline viral load for fostemsavir-treated subjects compared with 88% (< or ≥100,000 copies/ml). of subjects in the ATV/r arm (Table 1 and Figure 2A). When stratified by baseline HIV-1 RNA level, response Statistical analysis rates (HIV-1 RNA <50 copies/ml) were higher for sub- This was an estimation study and was not pow- jects with <100,000 copies/ml than those with ≥100,000 ered to show statistical differences between treat- copies/ml. Observed response rates were 73.9–100% ment arms. The mITT population consisted of sub- versus 96.3% (fostemsavir versus ATV/r arms) in subjects who received ≥1 dose of fostemsavir or ATV/r. jects with baseline viral load <100,000 copies/ml and The observed population consisted of subjects who 60.0–90.5% versus 71.4% when baseline viral load was received ≥1 dose of fostemsavir or ATV/r and who ≥100,000 copies/ml. The proportion of subjects achiev- had plasma HIV-1 RNA measurements at week 48. ing HIV-1 RNA <50 copies/ml appeared to be similar A target sample size of 50 subjects per treatment arm within strata across all four fostemsavir arms and the was selected [18]. ATV/r comparator arm (Figure 2B). Response rates were

Antiviral Therapy 22.3 217

AVT-16-OA-3925_Thompson.indd 217 02/06/2017 14:22:53 M Thompson et al.

Figure 1. Subject disposition through week 48

A total of 581 subjects were screened

Fostemsavir Fostemsavir Fostemsavir Fostemsavir ATV/r 400 mg 800 mg 600 mg 1,200 mg 300/100 mg twice daily twice daily once daily once daily once daily + RAL + TDF + RAL + TDF + RAL + TDF + RAL + TDF + RAL + TDF

Randomized 52 50 51 50 51

Treated 50 (96%) 49 (98%) 51 (100%) 50 (100%) 51 (100%)

Discontinued 6 (12%) 17 (35%) 9 (18%) 16 (32%) 14 (28%)

1 AE 2 AEs 1 pregnancy 2 AEs 2 AEs 1 consent 2 consent 2 lost to follow-up 2 consent 3 consent withdrawal withdrawals 2 no longer met withdrawal withdrawals 2 lost to follow-up 2 pregnancies study criteria 5 lost to follow-up 2 pregnancies 1 poor/ 1 lost to follow-up 1 lack of ef cacy 1 subject request 2 lost to follow-up non-compliance 1 no longer met 2 poor/ to discontinue 1 no longer met 1 other study criteria non-compliance 1 no longer met study criteria 5 lack of ef cacy 1 other study criteria 1 admin related 4 poor/ 5 lack of ef cacy 2 lack of ef cacy non-compliance 1 poor/ non-compliance

An additional 16 subjects discontinued across fostemsavir treatment arms along with an additional five subjects in the ritonavir-boosted atazanavir (ATV/r) arm, at week 48. One additional subject in the fostemsavir 1,200 mg arm discontinued due to an adverse event (AE) and seven additional subjects receiving fostemsavir discontinued due to lack of efficacy between week 24 and 48. No additional subjects in the ATV/r arm discontinued treatment due to AEs between week 24 and 48 and two subjects in this treatment arm discontinued due to lack of efficacy. RAL, raltegravir; TDF, tenofovir disoproxil fumarate.

Table 1. Proportion of subjects achieving HIV-1 RNA <50 (week 48 Snapshot): mITT

Fostemsavir + TDF + RAL ATV/r + TDF + RAL 400 mg twice 800 mg twice 600 mg once 1,200 mg once 300 mg/100 mg Parameter daily (n=50) daily (n=49) daily (n=51) daily (n=50) once daily (n=51)

HIV-1 RNA <50 copies/ml, n (%) 41 (82.0) 30 (61.2) 35 (68.6) 34 (68.0) 36 (70.6) HIV-1 RNA ≥50 copies/ml, n (%)a 2 (4.0) 9 (18.4) 10 (19.6) 8 (16.0) 5 (9.8) Discontinued (lack of efficacy), n (%) 0 2 (4.1) 0 1 (2.0) 0 Discontinued (other reasons), n (%)b 3 (6.0) 1 (2.0) 3 (5.9) 6 (12.0) 3 (5.9) No virological data at week 48 Discontinued due to AE or death, n (%)c 1 (2.0) 2 (4.1) 0 1 (2.0) 2 (3.9) Discontinued for other reasons, n (%) 2 (4.0) 5 (10.2) 3 (5.9) 0 5 (9.8) Missing data during window, but on study, n (%) 1 (2.0) 0 0 0 0

aIncludes participants who: had their last, on-treatment, HIV-1 RNA viral load ≥50 copies/ml within the week 48 window or discontinued before the week 48 window with their last, on-treatment, HIV-1 RNA viral load ≥50 copies per ml; the week 48 window is ±6 weeks (42–54 weeks). bIncludes withdrawal of consent, loss to follow-up, pregnancy or change of residence. cIncludes participants who discontinued because of an adverse event (AE) or death at any time point from day 1 through the week 48 window, if this resulted in no virological data on treatment during the week 48 window. ATV/r, ritonavir-boosted atazanavir; mITT, modified intent-to- treat; RAL, raltegravir; TDF, tenofovir disoproxil fumarate.

AVT-16-OA-3925_Thompson.indd 218 02/06/2017 14:22:54 HIV-1 attachment inhibitor fostemsavir: week 48 analysis

Figure 2. Virological response and association with baseline HIV-1 RNA

A Proportion of subjects with <50 copies/ml, % 100 Fostemsavir 90 400 mg 91 twice daily 80 Fostemsavir 70 800 mg 73 60 twice daily 50 Fostemsavir 40 600 mg 69 once daily 30 20 Fostemsavir 1,200 mg 79 10 <50 copies/ml (% with 95% CI) once daily 0 Proportion of subjects with HIV-1 RNA of subjects with HIV-1 Proportion ATV/r 300/ 0 4 8 12 16 20 24 32 40 48 100 mg 88 Week once daily B

100 Baseline HIV-1 RNA, 90 copies/ml 80 <100,000 70 ≥100,000 60 50 40 30 20

<50 copies/ml (% with 95% CI) 10

Proportion of subjects with HIV-1 RNA of subjects with HIV-1 Proportion 0 400 mg 800 mg 600 mg 1,200 mg ATV/r twice daily twice daily once daily once daily 300/100 mg once daily Fostemsavir dose

(A) Proportion of subjects achieving HIV-1 RNA <50 copies/ml through week 48 (observed). (B) Proportion of subjects achieving HIV-1 RNA <50 copies/ml through week 48 by baseline HIV-1 RNA (observed). ATV/r, ritonavir-boosted atazanavir.

72–94% versus 90.3% (fostemsavir versus ATV/r arms) group, 149 cells/ml (IQR 62–255) for the 800 mg twice-

in subjects with temsavir IC50 ≤1.3 nmol/l and were daily group, 145 cells/ml (IQR 98–182) for the 600 mg 70.6–100% versus 80.0% (fostemsavir versus ATV/r) once-daily and 182 cells/µl (IQR 106–231) for the

in subjects with temsavir IC50 ≥1.3 nmol/l (Additional 1,200 mg once-daily group. These were similar to the file 2). Response rates through week 48 were similar mean CD4+ T-cell increase of 142 cells/ml (IQR 95–223), across all study arms, and across the additional base- observed in the ATV/r arm. CD4+ T-cell count increases

line temsavir IC50 categories assessed (Additional file 2). over time appeared to be similar across the fostemsavir These data indicate that response rates at week 48 were and ATV/r arms (Figure 3).

similar regardless of baseline fostemsavir IC50 stra- tum for the fostemsavir dosing arms; median increases Safety in CD4+ T-cell count from baseline at week 48 were Through week 48, fostemsavir was generally well tolerated 186 cells/ml (IQR 122–274) for the 400 mg twice-daily with no apparent dose-related safety signals (Table 2).

Antiviral Therapy 22.3 219

AVT-16-OA-3925_Thompson.indd 219 02/06/2017 14:22:54 M Thompson et al.

Figure 3. Median change in CD4+ T-cell count from baseline through week 48 (observed)

Median increase through week 48, cells/µl 200 180 Fostemsavir 400 mg 186 160 twice daily 140 T-cell count T-cell

+ Fostemsavir 120 800 mg 149 100 twice daily 80 Fostemsavir 60 600 mg 145 once daily from baseline (cells/ µ l) from 40 20 Fostemsavir 1,200 mg 182 Median change in CD4 0 once daily 8 16 24 32 40 48 ATV/r 300/ Week 100 mg 142 once daily

ATV/r, ritonavir-boosted atazanavir.

Table 2. Safety summarya and Grades 3–4 laboratory abnormalities through week 48 (≥2 subjects)b

Fostemsavir + TDF + RAL ATV/r + TDF + RAL 400 mg twice 800 mg twice 600 mg once 1,200 mg once 300 mg/100 mg Parameter daily (n=50) daily (n=49) daily (n=51) daily (n=50) once daily (n=51)

SAEs, n (%)c 3 (6.0) 5 (10.2) 4 (7.8) 3 (6.0) 5 (9.8) AEs leading to discontinuation, n (%)d 1 (2.0) 2 (4.1) 0 2 (4.0) 2 (3.9) Grade 2–4-related clinical AEs Total subjects with an event, n (%) 4 (8.0) 4 (8.2) 3 (5.9) 6 (12.0) 15 (29.4) Present in ≥2 subjects Abdominal pain, n (%) 0 0 0 1 (2.0) 2 (3.9) Nausea, n (%) 0 0 0 0 2 (3.9) Hyperbilirubinaemia, n (%) 0 0 0 0 4 (7.8) Jaundice, n (%) 0 0 0 0 2 (3.9) Blood bilirubin increased, n (%) 0 0 0 0 3 (5.9) Headache, n (%) 0 1 (2.0) 0 0 2 (3.9) Grade 3–4 laboratory abnormalities (≥2 subjects) Neutrophils (absolute), n (%) 1 (2.1) 1 (2.0) 2 (3.9) 1 (2.1) 1 (2.0) Alkaline phosphatase, n (%) 0 0 0 1 (2.1) 0 Alanine aminotransferase, n (%) 0 0 0 2 (4.2) 1 (2.0) Aspartate aminotransferase, n (%) 1 (2.1) 1 (2.0) 0 2 (4.2) 2 (4.0) Total bilirubin, n (%) 0 0 0 0 29 (58.0) Creatine kinase, n (%) 0 1 (2.0) 1 (2.0) 2 (4.2) 2 (4.0) Glucose fasting serum, n (%) 0 1 (2.0) 2 (3.9) 0 0 Uric acid, n (%) 0 0 0 2 (4.2) 0

aAs reported by investigators. A single subject may have had more than one event. bData not available for all subjects. cAnal abscess, herpes encephalitis, overdose (n=3), extrapulmonary tuberculosis (n=2), herpes zoster, abdominal pain, myalgia, spontaneous abortion, acute renal failure, cellulitis (n=2), lymphangitis, chronic cholecystitis, back pain, influenza, pneumonia, pyelonephritis, diarrhoea, cholelithiasis, migraine. dTwo subjects in the ritonavir-boosted atazanavir (ATV/r) arm discontinued treatment due to adverse events (AEs) of abdominal distention, flatulence, nausea and jaundice. Five subjects in the fostemsavir arms discontinued treatment due to AEs of illegal substance use (n=1), renal failure linked to tenofovir disoproxil fumarate (TDF)-treatment (n=1), bone tuberculosis (n=1), disseminated tuberculosis (n=1) and lymph node tuberculosis (n=1). None of the discontinuations were considered related to fostemsavir by the investigator. RAL, raltegravir; SAE, serious adverse event.

AVT-16-OA-3925_Thompson.indd 220 02/06/2017 14:22:54 HIV-1 attachment inhibitor fostemsavir: week 48 analysis

Across the fostemsavir and ATV/r arms, 15/200 (7.5%) and week 48 across all fostemsavir arms and the ATV/r and 5/51 (9.8%) subjects had SAEs, respectively, most arm; median CD4+ T-cell count increases from baseline attributable to infections. SAEs related to study drug(s) through week 24 ranged between 100 and 138 cells/ml were TDF-induced renal failure (fostemsavir 800 mg across fostemsavir and was 103 cells/ml for subjects in twice-daily arm) and influenza-like symptoms (ATV/r the ATV/r arm, and through week 48 increases ranged arm). There were no fostemsavir-related AEs leading between 145 and 186 cells/ml (fostemsavir arms) and to discontinuation. AEs that led to discontinuation and 142 cells/ml (ATV/r arm). were considered to be related to study drug(s) included When stratified by baseline HIV-1 RNA level, the pro- abdominal distention, flatulence, nausea and jaundice portion of subjects achieving HIV-1 RNA <50 copies/ (two subjects in the ATV/r arm) and TDF-induced acute ml was higher for those who had baseline HIV-1 RNA renal failure (observed in one subject in the fostemsavir <100,000 copies/ml rather than ≥100,000 copies/ml. 800 mg twice-daily arm). Other AEs leading to discon- This was observed across all four fostemsavir arms tinuation in the fostemsavir arms were illegal substance and the ATV/r comparator arm. Similarly, virological use (n=1), bone tuberculosis (n=1), disseminated tubercu- response rates at week 48 were generally higher in sub- losis (n=1) and lymph node tuberculosis (n=1). These four jects with a baseline CD4+ T-cell count ≥200 cells/ml; a discontinuations were not considered related to fostem- trend observed across all fostemsavir and ATV/r arms. savir by the investigator. The most commonly reported Within baseline strata, the response rates were similar AE across the fostemsavir arms was headache (30/200 across all study arms, again indicating low intra-dose [15.0%]; mostly Grade 1). Diarrhoea and jaundice were variability. At both week 24 and week 48, additional the most frequent AEs in the ATV/r arm, both reported in subanalyses showed that efficacy was not affected by a 8/51 (16%) subjects. Across the fostemsavir and ATV/r subjects’ gender, age or race [19,20]. arms, 17/200 (8.5%) and 15/51 (29.4%) subjects had Our findings in this study continue to demonstrate

Grade 2–4 treatment-related AEs, respectively, and for that baseline temsavir IC50 was not predictive of fos- the fostemsavir arms, most of these were single instances temsavir antiviral activity because efficacy was similar with no relationship to dose (Table 2). For the ATV/r across all fostemsavir arms and the ATV/r arm when

arm most of the Grade 2–4 treatment-related AEs were pre-specified IC50 cutoffs were applied. In the ongo- related to gastrointestinal and hepatobiliary disorders. ing Phase III study of fostemsavir treatment in heavily Elevated total bilirubin was the most notable Grade 3–4 treatment-experienced subjects, there are no restric-

laboratory abnormality (29/51 [58%] subjects) observed tions to entry based on baseline temsavir IC50. Findings in the ATV/r arm. Across fostemsavir arms the frequency from the current Phase IIb study and a planned retro- of Grade 3–4 laboratory abnormalities was low; neutro- spective exploratory analysis of the Phase III data will

penia was reported in 5/200 (2.5%) subjects; increases determine any potential effect of baseline temsavir IC50 in aspartate aminotransferase and creatine kinase were on the antiviral efficacy of fostemsavir. each reported in 4/200 (2%) subjects. No notable trends Through week 48, fostemsavir was generally well in Grade 2–4-related or Grade 3–4 laboratory abnormali- tolerated with no fostemsavir-related SAEs or fostem- ties were observed across the fostemsavir arms. savir-related AEs that led to discontinuation. Across fostemsavir arms no new safety signals were identified Discussion through week 48; Grade 1 headache was the most common AE and Grade 2–4 treatment-related AEs were Consistent with the findings through week 24 [18], observed as single instances across arms. AEs including fostemsavir continued to show similar antiviral effi- hyperbilirubinaemia observed in the ATV/r arm were cacy and immunological responses relative to the consistent with the known safety profile of ATV [21]. comparator drug, ATV/r (both given in combination This Phase IIb study was not powered to show a sta- with RAL and TDF) through week 48. Across fostem- tistical difference between dosing arms. However, there savir arms, the proportion of subjects achieving HIV-1 was no observable fostemsavir dose response, as evi- RNA <50 copies/ml through week 24 was 69–80% denced by the overlapping confidence intervals shown and 78–87% for the mITT and observed popula- in Figure 2A. Based on data obtained from pharma- tions, respectively [18]. Importantly, these response cokinetic modelling fostemsavir 1,200 mg once-daily rates remained generally consistent through week 48, was selected as the open-label continuation dose with proportions of 61–82% and 77–95% reported beyond week 48. This dose mirrors the total daily expo- in the mITT and observed populations, respectively. sure of the selected Phase III dose, 600 mg twice daily. Given the small sample size of each treatment arm the Although not tested clinically, pharmacokinetic model- rates of efficacy were generally similar across fostem- ling and computational predictions of antiviral efficacy savir dosing arms. In addition, there were numerical identified fostemsavir 600 mg twice daily as the most increases in mean CD4+ T-cell counts between week 24 efficacious dose [22].

Antiviral Therapy 22.3 221

AVT-16-OA-3925_Thompson.indd 221 02/06/2017 14:22:55 M Thompson et al.

This study had several strengths. Baseline character- Kowa Research Institute, Pfizer, Merck & Co, Tobira istics were broadly similar across all treatment arms. Therapeutics and ViiV Healthcare; and served on data A total of 100/251 (40%) of subjects were female, safety monitoring boards for ViiV Healthcare. JPL 107/251 (43%) had a baseline HIV-1 RNA ≥100,000 received research support from Bristol-Myers Squibb. copies/ml and 96/251 (38%) had baseline CD4+ T-cell YP has received personal fees from Via Libre. PC has count <200 cells/mm3, these groups were therefore con- received research support from AbbVie, Merck & sidered well represented. In addition, ≥50% of subjects Co and ViiV Healthcare; and has served on advisory had ≥1 major NRTI, NNRTI or PI resistance-associated boards for AbbVie, Gilead, Merck & Co, Teva and mutation at baseline, providing evidence of fostemsavir ViiV Healthcare. DAS is an employee and shareholder efficacy in a treatment-experienced population. of Bristol-Myers Squibb. GJH is a former employee of Throughout the study the fostemsavir arms required Bristol-Myers Squibb and current employee of Merck nine pills daily prior to unblinding, while the ATV/r & Co. ML and SRJ were former employees at Bristol- arm required five pills daily. This pill burden may have Myers Squibb and current employees of ViiV Health- affected adherence and discontinuation rates in the fos- care. RK and OASP declare no competing interests. temsavir arms; however, this was not formally assessed. Additionally, the PhenoSense® Entry assay failed in Additional files 26% of samples screened for this study. Given the high failure rate and the similarity of observed virological Additional file 1: A figure of temsavir proposed

response regardless of fostemsavir IC50, the decision mechanism of action can be found at https://www.

was made not to include IC50 in the ongoing Phase III intmedpress.com/uploads/documents/3925_Thompson_ study, either as an entry criterion or a stratification vari- Addfile_1.pdf able. The antiviral activity of fostemsavir observed in both the monotherapy substudy and in this Phase IIb Additional file 2: A table of proportion of subjects study, along with the Phase IIb safety profile, support achieving HIV-1 RNA <50 copies/ml through week

continued investigations with this compound. A Phase 48 by baseline temsavir IC50 category (observed) can III trial (NCT02362503) is underway to evaluate fos- be found at https://www.intmedpress.com/uploads/ temsavir in heavily treatment-experienced adults with documents/3925_Thompson_Addfile_2.pdf two or fewer active antiretroviral classes remaining due to resistance, intolerabilities and/or contraindications. Additional file 3: A list of the AI438011 investigators can be found at https://www.intmedpress.com/uploads/ Acknowledgements documents/3925_Thompson_Addfile_3.pdf We thank all of the AI438011 clinical trial participants References and their families, and the AI438011 investigators. We 1. May MT, Gompels M, Delpech V, et al. Impact on life also thank Anna Rightmire, Michelle DeGrosky, Lia expectancy of HIV-1 positive individuals of CD4+ cell count Donahue, John Coumbis, Neela Ray, Luna Zaru, Mark and viral load response to antiretroviral therapy. AIDS 2014; 28:1193–1202. Krystal, Carey Hwang, Todd Correll and Eric Y Wong at 2. Samji H, Cescon A, Hogg RS, et al. Closing the gap: Bristol-Myers Squibb, and Peter Lill and John Riefler at increases in life expectancy among treated HIV-positive ICON CRO for assistance with the study. Professional individuals in the United States and Canada. PLoS One 2013; 8:e81355. medical writing and editorial assistance was provided 3. Department of Health and Human Services. Panel on by Amanda Gallagher at MediTech Media, Manchester, Antiretroviral Guidelines for Adults and Adolescents. UK, with funding from ViiV Healthcare. This study was Guidelines for the use of antiretroviral agents in HIV-1- infected adults and adolescents. 2015. (Accessed November sponsored by Bristol-Myers Squibb. ViiV Healthcare has 2015.) Available from http://aidsinfo.nih.gov/ContentFiles/ acquired and will continue clinical development of fos- AdultandAdolescentGL.pdf temsavir. ClinicalTrials.Gov Number: NCT01384734. 4. European AIDS Clinical Society. European AIDS Clinical Society Guidelines Version 8, 2015. (Accessed November EUDRACT Number: 2011-000437-36. 2015.) Available from http://www.eacsociety.org/files/ A list of the AI438011 investigators can be found in guidelines_english_71_141204.pdf Additional file 3. 5. Attia S, Egger M, Muller M, et al. Sexual transmission of HIV according to viral load and antiretroviral therapy: This study was funded by Bristol-Myers Squibb. systematic review and meta-analysis. AIDS 2009; 23:1397–1404. 6. Cozzi-Lepri A, Phillips AN, Ruiz L, et al. Evolution of Disclosure statement drug resistance in HIV-infected patients remaining on a virologically failing combination antiretroviral therapy MT received research contracts paid to the AIDS regimen. AIDS 2007; 21:721–732. 7. Tarwater PM, Gallant JE, Mellors JW, et al. Prognostic Research Consortium of Atlanta from Bristol-Myers value of plasma HIV RNA among highly active Squibb, Cepheid Inc, Gilead Sciences, GeoVax Inc, antiretroviral therapy users. AIDS 2004; 18:2419–2423.

AVT-16-OA-3925_Thompson.indd 222 02/06/2017 14:22:55 HIV-1 attachment inhibitor fostemsavir: week 48 analysis

8. Hasse B, Ledergerber B, Furrer H, et al. Morbidity and 16. Ray N, Hwang C, Healy MD, et al. Prediction of virological aging in HIV-infected persons: the Swiss HIV cohort study. response and assessment of resistance emergence to the Clin Infect Dis 2011; 53:1130–1139. HIV-1 attachment inhibitor BMS-626529 during 8-day 9. Marzolini C, Back D, Weber R, et al. Ageing with HIV: monotherapy with its prodrug BMS-663068. J Acquir medication use and risk for potential drug-drug interactions. Immune Defic Syndr 2013; 64:7–15. J Antimicrob Chemother 2011; 66:2107–2111. 17. Nettles RE, Schurmann D, Zhu L, et al. Pharmacodynamics, 10. Nachega JB, Hsu AJ, Uthman OA, et al. Antiretroviral safety, and pharmacokinetics of BMS-663068, an oral therapy adherence and drug-drug interactions in the aging HIV-1 attachment inhibitor in HIV-1-infected subjects. HIV population. AIDS 2012; 26 Suppl 1:S39–S53. J Infect Dis 2012; 206:1002–1011. 11. Brown J, Chien C, Timmins P, et al. Compartmental 18. Lalezari JP, Latiff GH, Brinson C, et al. Safety and efficacy absorption modeling and site of absorption studies to of the HIV-1 attachment inhibitor prodrug BMS-663068 determine feasibility of an extended-release formulation of in treatment-experienced individuals: 24 week results of an HIV-1 attachment inhibitor phosphate ester prodrug. AI438011, a Phase 2b, randomised controlled trial. Lancet J Pharm Sci 2013; 102:1742–1751. HIV 2015; 2:e427–e437. 12. Langley DR, Kimura SR, Sivaprakasam P, et al. Homology 19. Brinson C, Lalezari JP, Latiff GH, et al. HIV-1 attachment models of the HIV-1 attachment inhibitor BMS-626529 inhibitor prodrug BMS-663068 in antiretroviral- bound to gp120 suggest a unique mechanism of action. experienced subjects: week 24 subgroup analysis. IDWeek. Proteins 2015; 83:331–350. 8–12 October 2014, Philadelphia, PA, USA. Oral: 540. 20. Feinberg J, Lalezari JP, Martins M, et al. HIV-1 13. Henrich TJ, Kuritzkes DR. HIV-1 entry inhibitors: recent attachment inhibitor prodrug BMS-663068 in development and clinical use. Curr Opin Virol 2013; antiretroviral-experienced subjects: week 48 subgroup 3:51–57. analysis. IDWeek. 7–11 October 2015, San Diego, CA, 14. Nowicka-Sans B, Gong YF, McAuliffe B, et al. In vitro USA. Poster: 1075. antiviral characteristics of HIV-1 attachment inhibitor 21. Bristol-Myers Squibb. Reyataz® prescribing information. BMS-626529, the active component of the prodrug 2013. (Accessed September 2015.) Available from http:// BMS-663068. Antimicrob Agents Chemother 2012; packageinserts. bms.com/pi/pi_reyataz.pdf 56:3498–3507. 22. Landry I, Zhu L, Abu Tarif M, et al. Model-based Phase 15. Li Z, Zhou N, Sun Y, et al. Activity of the HIV-1 3 dose selection for HIV-1 attachment inhibitor prodrug attachment inhibitor BMS-626529, the active component BMS-663068 in HIV-1-infected patients: population of the prodrug BMS-663068, against CD4-independent pharmacokinetics/pharmacodynamics of the active moiety, viruses and HIV-1 envelopes resistant to other entry BMS-626529. Antimicrob Agents Chemother 2016; inhibitors. Antimicrob Agents Chemother 2013; 60:2782–2789. 57:4172–4180.

Accepted 31 October 2016; published online 6 December 2016

Antiviral Therapy 22.3 223

AVT-16-OA-3925_Thompson.indd 223 02/06/2017 14:22:55