The safety, tolerability, and pharmacokinetic profile of GSK2838232, a novel 2nd generation HIV maturation inhibitor, as assessed in healthy subjects

Abstract This work aimed to assess the safety, tolerability, pharmacokinetics (PK), and relative bioavailability of GSK2838232, an investigational HIV maturation inhibitor. GSK2838232 was administered over four dose‐escalation studies in healthy subjects which assessed single oral doses (5‐250 mg) and repeat doses (up to 200 mg once or twice daily) ±100 mg ritonavir (RTV) once daily. GSK2838232 administration (up to 250 mg) to 124 subjects across four studies resulted in few mild adverse events (AEs) with similar frequencies to placebo. There were no clearly identified drug‐related AEs. GSK2838232 tested fasted was quickly absorbed with a t max of 2‐3 hours. With food, the absorption was delayed and more variable, with ~60% increase in AUC and C max. Overall, following single doses GSK2838232 AUC and C max generally exhibited proportional PK from 50 to 100 mg dose without RTV and from 50 to 250 mg with RTV and following repeated doses of 20‐200 mg with RTV. In relative bioavailability studies, a micronized formulation was found to be suitable for development. At steady state, RTV increased GSK2838232 AUC and C max by 10‐ and 3‐fold, respectively. Half‐life was prolonged from ~17 hours nonboosted to ~34 hours with RTV. This boosting effect was also seen in repeat‐dose GSK2838232 studies, which achieved the targeted plasma exposure with GSK2838232 as a once‐daily regimen of up to 200 mg with RTV. The results of these studies demonstrated a favorable safety and PK profile for GSK2838232 and support its investigation for the treatment of HIV infection.

Unfortunately, this drug was terminated in Phase IIb clinical development because of concerns over safety/GI tolerability and viral resistance. 5 GSK2838232 ( Figure 1) is derived from the triterpene betulin natural product that has demonstrated an excellent virological profile against a panel of viruses and retains activity to viruses with key resistance mutations. 6 The IC90 in cell-based assays has been calculated as 6.4 nmol/L (5 ng/mL), and there was no significant shift in this value in assays to which human serum albumin was added, even though plasma protein binding was assessed in vitro as >99.9% across all species. This value (termed PA-IC90) was considered a minimal threshold to achieve in Phase I studies, prior to evaluation in HIV-infected subjects.
In preclinical studies, GSK28382232 was shown to have low to moderate oral bioavailability (6%-40%, depending on species and formulation) and was excreted via bile (primarily as metabolites) or in feces (primarily unabsorbed drug). The major compound in plasma circulation was parent compound. GSK2838232 is subjected to metabolism via multiple pathways, including Phase I oxidation (primarily by cytochrome P450 [CYP] 3A4) and Phase II glucuronidation, with negligible urinary excretion of drug-related material (≤1% administered dose). Metabolite profiles observed preclinically were qualitatively similar to those observed in human plasma.
The current report presents the results from four Phase I dose escalation studies in healthy subjects to ascertain the safety, tolerability, and single-and repeat-dose pharmacokinetics (PK) of GSK2838232 with and without ritonavir (RTV) as well as the relative bioavailability of suspension and solid oral dosage formulations and the impact of food.

| MATERIALS AND METHODS
Each one of the four studies described were conducted and completed at a single center in the United States (Parexel International, Baltimore, MD) per the ethical principles of "good clinical practice" (GCP) and the Declaration of Helsinki after obtaining a written informed consent from each subject. The protocols and amendments were approved by Aspire Investigational Review Board (Santee, California, USA).

| Study population and design
For all four studies, the inclusion/exclusion criteria allowed for the enrollment of healthy men and women of nonchild bearing potential aged between 18 and 55 years of age with a body weight ≥ 50 kg for men and ≥ 45 kg for women, body mass index (BMI) within the range 18.5 to 31.0 kg/m 2 (inclusive), and creatinine clearance >80 mL/min at time of screening. Other screening assessments included 24-hour Holter monitoring and baseline echocardiograms. This study was a first-in-human, double-blind, randomized, placebocontrolled, single dose escalation study of GSK2838232 in healthy subjects conducted in two sequential cohorts with four dosing visits in each cohort. At each visit in each cohort, 6 out of the 8 subjects were randomized to receive the active dose, and 2 out of the 8 F I G U R E 1 Structure of GSK2838232 (MW 809.6) subjects were randomized to placebo. Cohort 1 tested escalating doses of 5, 10, and 20 mg GSK2838232 as a Spray Dried Dispersion (SDD) delivered as a powder-in-bottle (PiB) suspension; Cohort 2 tested the same formulation at doses of 50 and 100 mg, the effect of a high-fat breakfast on a 50 mg dose, and the impact of steadystate dosing of 100 mg once-daily (QD) RTV on the PK of a single 10 mg dose of GSK2838232 (administered on Day 10 of the 12 days of RTV dosing). All doses were administered in a fasted state, except in the food effect portion of the study. Part B was conducted in a separate group of subjects and was not dependent on the results of Part A. This was a randomized, open-label, 3-period, crossover design to assess the relative bioavailability of the nonmicronized API PiB versus the original SDD formulation at a 100 mg dose level (Periods 1 and 2). Based on an interim analysis of these data confirming that crystalline API was viable, a single dose of 20 mg API was assessed after RTV predosing for 10 days. Subjects were randomized 1:1 to each dosing sequence.

| Multiple-ascending-dose escalation study
(20-50 mg) AE rtv -spray dried dispersion formulation (study 200207, NCT02289495) This study was conducted in parallel with Study 200912 and was planned as a two-part, double-blind, placebo-controlled, repeat dose escalation study of GSK2838232 with and without RTV for 8-11 days. Subjects were randomized 3:1 to receive GSK2838232 or placebo. Part A was designed to evaluate GSK2838232 given alone for 8 days (20 mg and 50 mg GSK2838232 once-daily (QD) for 8 days, respectively) and Part B was designed to evaluate GSK2838232 when co-administered with 100 mg RTV QD for 11 days. Because of the occurrence of an SAE that was eventually determined to be unlikely to be related to GSK2838232 (see Safety discussion), the study was terminated early after one cohort in Part B (10 mg GSK2838232 + 100 mg RTV QD) had completed only 5 days of the planned 11-day regimen.

| Single-, multiple-dose escalation, and
relative bioavailability (50-250 mg) AE RTV-powder-in bottle API versus powder blend capsule formulation (Study 204953, NCT02795754) This study was planned as an extension of the previously conducted single-and multiple-dose escalation studies described above. It was a double-blind, placebo-controlled, single and repeat dose escalation study of micronized API GSK2838232 with RTV conducted in three parts (Part 1A, Part 1B, and Part 2).
In Part 1A, a cohort of subjects was given single doses of 50, 100, and 250 mg of GSK2838232 with RTV, following 2 days of RTV predosing, in a sequential dose-escalation design. Part 1B studied the relative bioavailability of GSK2838232 given as 100 mg (2 9 50 mg) capsules versus the reference API (PiB) in a two-period crossover design with RTV, following 2 days of RTV predosing. In Period 3, subjects took the capsule formulation and RTV with a moderate fat meal (30% of calories from fat).
In Part 2, GSK2838232 was given to separate cohorts as oncedaily 20, 50, 100, and 200 mg doses with RTV for 11 days, with GSK2838232 and RTV dosing starting at the same time (i.e., no RTV predosing). GSK2838232 was administered as PiB for the 20 and 50 mg cohorts and as the capsule formulation for the 100 and 200 mg cohorts. Another separate cohort of subjects received GSK2838232 alone (i.e., unboosted) at a dose of 200 mg twicedaily (BID; given every 12 hours) for 11 days using the capsule formulation to assess trough concentrations without RTV co-administration.  GSK2838232 (the isotopically labeled internal standard), respectively.
Urine was not collected in any of the clinical studies, as preclinical evaluation indicated negligible renal excretion (≤1% of administered dose) of parent or drug-related material.

| Pharmacokinetic data analysis
For all four studies, the primary objectives were to describe the safety, tolerability, and pharmacokinetics of single and repeated doses of GSK2838232 with or without co-administration of RTV. No formal statistical hypotheses were tested. All data were descriptively summarized.
The PK parameters were calculated by standard noncompartmental analysis based on actual sampling times using WinNonlin Version 6.3 or higher. Estimation of AUC used the linear up/log down trapezoidal rule.
For the assessment of relative bioavailability of the capsule formulation compared to the PiB suspension and of the food effect using the capsule formulation (all with RTV), PK parameters were separately analyzed for each treatment using a mixed effects model with fixed effect terms for period and treatment following log e transformation. Subject was treated as a random effect in the model. Point estimates and their associated 90% CIs were constructed for the differences in PK parameter values between the test and reference treatments. The point estimates and their associated 90% CIs were then back-transformed to provide point estimates and 90% CIs for the ratios of PK parameters from test and reference treatments on the original scale.

| Subject disposition and demographics
Across the four studies, a total of 124 healthy subjects were exposed to GSK2838232 or placebo, of whom 9 subjects were female. The subject demographics were similar across all 4 studies with respect to age, weight and height (Table 1). Of the 124 subjects, one was withdrawn in Study 200207 (NCT02289495) due to an SAE of a cardiovascular nature that was eventually determined to be unlikely to be related to GSK2838232 (detailed in safety results section), and two subjects were withdrawn due to adverse events (AEs) in Study 204953 (NCT02795754), neither of which was considered related to GSK2838232.  Table 2).

|
In Study 204953, observed exposures in the presence of RTV appeared to be generally dose proportional in GSK283832 up to 250 mg ( Figure 3B).

| Multiple-dose pharmacokinetics
The PK profile of unboosted GSK2838232 after multiple dosing for 8 days was as expected based on single dose data. AUC and C max appeared slightly less than dose-proportional between 20 mg and 50 mg GSK2838232 given QD (approximately a 2fold increase after a 2.5-fold dose). The absorption profile was also similar to prior studies, with C max occurring approximately 2-3 hours after dosing.
The accumulation ratios following repeated daily dosing with 50 mg GSK2838232 based on AUC, C max , and Cs were 1.2, 1.   was considered more attractive for manufacturing and developability reasons, including a lower overall tablet weight compared to SDD.

| Formulation studies and food effect
Therefore, the crystalline API, initially as PiB suspension and ultimately as capsules, was chosen for continuation of the clinical development program through Study 204953.
In Study 204953, total exposures of GSK2838232 following administration of a single 100 mg GSK2838232 dose (as two 50 mg capsules) in combination with RTV were approximately 1.4-to 1.5fold of those seen after higher than a corresponding 100 mg dose as API PiB suspension with RTV (Table 4). C max and C24 values with the capsule formulation were 1.6-fold and 1.4-fold of those seen with the PiB suspension, and the geometric LS mean ratio (90% confidence interval) for t½ showed no difference between the two products (data not shown).
When administered in the fasted state, GSK2838232 was rapidly absorbed, with time to C max (t max ) values of 2-3 hours. Absorption was delayed and more variable after food. The bioavailability of the initial SDD PiB formulation after a high-fat (60%) meal was approximately 1.6-fold relative to the fasted state (Study HMI116787,

| Safety
Nonclinical evaluation in short duration toxicology studies initially gave rise to potential cardiovascular signals which were minimal, appeared sporadically, and were not reproduced in longer term stud-   In aggregate, the studies showed that a micronized API formulation of GSK2838232 in capsules was suitable for future clinical studies. All products were quickly absorbed with t max of 2-3 hours when administered in the fasted state. In the fed state (moderate or high fat meal), the absorption was delayed and more variable. The presence of food causes the release of bile that is expected to increase solubility of GSK2838232 due to bile surfactant properties. The transition to a capsule formulation included the addition of excipients (e.g., lactose and cellulose) as part of the blend (~75% of the total load), and the presence of these excipients is expected to delay GSK2838232 precipitation, allowing more drug to be in solution for absorption. 7 This phenomenon is widely known as the "parachute effect". 8  relative contribution to increased GSK2838232 bioavailability between hepatic and GI CYP3A4 inhibition and/or P-gp inhibition by RTV co-administration is unknown, but CYP3A4 and P-gp have largely overlapping substrate specificities and are co-localized in the intestine, working in a coordinated transport-metabolism manner. 11 Co-administration of GSK2838232 and RTV together on Day 1 without ritonavir predosing showed only about a 25% reduction of Day 1 AUC and C max value relative to the GSK2838232 AUC seen after the first GSK2939232 dose following predosing ritonavir to steady-state. These data indicated that there is no need to predose with RTV in HIV patients, as GSK2838232 plasma concentrations attain potentially therapeutic levels within a few days.