Lenacapavir: A Novel, Potent, and Selective First-in-Class Inhibitor of HIV-1 Capsid Function Exhibits Optimal Pharmacokinetic Properties for a Long-Acting Injectable Antiretroviral Agent

Abstract

Lenacapavir (LEN) is a picomolar first-in-class capsid inhibitor of human immunodeficiency virus type 1 (HIV-1) with a multistage mechanism of action and no known cross resistance to other existing antiretroviral (ARV) drug classes. LEN exhibits a low aqueous solubility and exceptionally low systemic clearance following intravenous (IV) administration in nonclinical species and humans. LEN formulated in an aqueous suspension or a PEG/water solution formulation showed sustained plasma exposure levels with no unintended rapid drug release following subcutaneous (SC) administration to rats and dogs. A high total fraction dose release was observed with both formulations. The long-acting pharmacokinetics (PK) were recapitulated in humans following SC administration of both formulations. The SC PK profiles displayed two-phase absorption kinetics in both animals and humans with an initial fast-release absorption phase, followed by a slow-release absorption phase. Noncompartmental and compartmental analyses informed the LEN systemic input rate from the SC depot and exit rate from the body. Modeling-enabled deconvolution of the input rates from two processes: absorption of the soluble fraction (minor) from a direct fast-release process leading to the early PK phase and absorption of the precipitated fraction (major) from an indirect slow-release process leading to the later PK phase. LEN SC PK showed flip-flop kinetics due to the input rate being substantially slower than the systemic exit rate. LEN input rates via the slow-release process in humans were slower than those in both rats and dogs. Overall, the combination of high potency, exceptional stability, and optimal release rate from the injection depot make LEN well suited for a parenteral long-acting formulation that can be administered once up to every 6 months in humans for the prevention and treatment of HIV-1.

Introduction

Oral ARV drugs have been a cornerstone of HIV treatment for more than three decades, and their continually improving safety and efficacy directly enabled reversing the initially tragic course of the HIV epidemic. The current oral drugs are easy to administer, highly effective, and safe for both the prevention and treatment of HIV-1 infection. However, life-long adherence to once-daily oral medications remains a significant challenge for many people with HIV (PWH) and individuals who could benefit from pre-exposure prophylaxis. In these cases, effective longer-acting treatment and prevention options that can be administered less frequently would provide alternatives and be advantageous to current oral ARV agents. The advantages include an improved quality of life, adherence, and overall success rate, including prevention options that may provide life-saving reduction of HIV transmission at the individual and population level and lead to critical advancement toward ending the HIV epidemic globally.¹

A critical aspect of innovation brought by biologic drugs has been their long duration of plasma exposure, which led to less frequent dose regimens via parenteral administration, and this has been readily accepted by providers and patients. Long-acting parenteral products have been difficult to achieve for small-molecule drugs without significant chemical modifications such as highly lipophilic prodrugs.^2,3 The first introduction of long-acting HIV products administered by intramuscular administration is one of the relatively recent advancements in ARV therapeutics.⁴ These initial products drew interest and support from both PWH groups and healthcare providers, encouraging additional innovation focused on the development of novel injectable drug products that could be administered with low frequency and without notable adverse effects.

To sustain efficacious plasma trough concentrations over an extended period of time, a long-acting drug must have high potency, a slow and continuous release rate from the parental depot that can be administered in an appropriate volume, and a slow systemic elimination rate (Scheme 1).² Maximizing the potency and minimizing the systemic elimination directly enable extension of the dosing intervals. LEN is a picomolar, multistage inhibitor of HIV capsid function and is highly effective against a broad range of HIV types and strains.^5,6 LEN has demonstrated high virologic efficacy in PWH, with up to a 2.3-log₁₀ decline in HIV-1 RNA copies/mL observed over 10 days after single SC doses of LEN monotherapy.⁵

Scheme 1. Lenacapavir (LEN) Structure and Considerations for a Long-Acting Injectable Drug Product.

Link et al.⁵

In this article, we present the nonclinical PK characterization of LEN that demonstrates its extended plasma half-life following SC administration of a suspension and solution formulations. PK modeling showed that LEN displays flip-flop kinetics due to a substantially slower release rate than the systemic elimination rate in rats and dogs, leading to an increased apparent terminal half-life (t_1/2). The nonclinical data predicted an extended t_1/2 in humans following SC administration and were in good agreement with the observed human PK from both formulations, supporting further development of LEN for both treatment and prophylaxis of HIV.

Materials and Methods

Materials

LEN (Scheme 1) and internal standards for mass spectrometry-based analyses were synthesized by Gilead Sciences (Foster City, CA). All other chemicals or reagents were of research grade or higher and were purchased from Sigma-Aldrich (St. Louis, MO) or BioIVT (Westbury, NY), unless specified otherwise.

LEN Formulations for Intravenous (IV) and SC Administration

The IV solution was formulated with LEN (free acid) at 0.5 mg/mL in 5% ethanol, 20% propylene glycol, 45% polyethylene glycol (PEG) 300, and 30% 10 mM phosphate buffer.

The SC aqueous suspension was formulated with amorphous LEN free acid at 100 or 200 mg/mL in normal saline with 2% poloxamer 188. The particle size distribution was bimodal with peaks at 0.4 and 2.0 μm, and a d₉₀ of 3.6 μm. The LEN solubility in the aqueous suspension was <1 μg/mL; therefore, amorphous LEN solids represented >99.99% of the total LEN in the formulation. The SC PEG/water solution was formulated with LEN (sodium salt) at 309 mg/mL in 68.2% PEG 300 and 31.8% water.

Nonclinical Pharmacokinetic (PK) Studies

All nonclinical studies were conducted at contract research laboratories in accordance with the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the U.S. National Institutes of Health and were approved by the Institution’s Animal Care and Use Committee or a local equivalent. Male Wistar Han (WH) rats and beagle dogs were administered the target IV dose by infusion over 30 min via a peripheral vein. The rate of infusion was adjusted according to the body weight of each animal to deliver the 1 mg/kg target dose. Male WH rats and beagle dogs were administered the target SC doses in the intrascapular region via a syringe injection. Individual doses were calculated based on body weights recorded on the day of dose administration. Serial blood samples were collected for plasma pharmacokinetic (PK) analysis from 90 to 364 days in rats and from 40 to 150 days in dogs. The blood samples were centrifuged at approximately 5 °C. The plasma was separated, frozen, and maintained at approximately 70 °C until analysis.

Human PK Studies

Clinical PK studies in healthy volunteers following SC administration of the aqueous suspension formulation (study GS-US-200-4070) and the PEG/water solution (study GS-US-200-4358) were previously described.^5,7−9 LEN SC PK was comprehensively characterized following a single administration of either aqueous suspension (at 30, 100, 300, 450 mg)^5,9 or PEG/water solution formulations (927 mg).^7,8 LEN PK in healthy volunteers following a single IV infusion administration of a modified solution formulation was also characterized in an ADME study (study GS-US-200-4329).¹⁰

Quantification of LEN in Plasma

LEN plasma concentrations were determined using sensitive high-performance liquid chromatography-tandem mass spectrometry (LC-MS/MS) method as exemplified previously.⁵ The methods were evaluated for selectivity, sensitivity, and linearity (1–10,000 nM), as well as intra-assay accuracy and precision. Briefly, a protein precipitation procedure was followed to prepare plasma samples for bioanalysis. A 50 μL aliquot of each plasma sample was added to a clean 96-well plate followed by the addition of 200 μL of cold acetonitrile solution containing an internal standard. The contents in the 96-well plate were vortexed and centrifuged at 2164g. An aliquot of 100 μL of the supernatant from each well was transferred to a clean 96-well plate and diluted with 100 μL of water. An aliquot of 20 μL of the above solution was subjected to LC-MS/MS analysis.

PK Modeling of SC Absorption Kinetics

PK modeling was performed to inform LEN systemic input (depot release) and exit rate(s) from the parenteral depot. SC PK across all species and formulations showed two input rates from the parenteral depot: an initial direct release process attributed to the soluble fraction of LEN in the administered dose and an indirect release process of the suspended or precipitated solid depot that was formed thereafter.

The PK parameters were determined by noncompartmental analysis (NCA) and compartmental analysis (CA) using Phoenix software (version 6.3 or higher). NCA of the IV PK profile was used to derive the systemic PK parameters—clearance (CL; L/h/kg), volume of distribution at steady state (V_ss; L/kg), and the terminal half-life (t_1/2, h) were used to compute the systemic exit rate (k_elimination; 1/h). NCA of the apparent terminal phase in the SC PK profile was used to estimate the slow depot release rate. The fraction of dose released from the parenteral depot over time was computed from the ratio of dose-normalized SC PK AUC_0–t (t = sampled time point) to IV PK AUC_0–inf.

Compartmental analysis (Scheme 2), in addition to the NCA-derived parameters, also informed k_direct and the fraction of dose released via direct and indirect release processes. Naive pooled engine and the first-order conditional estimation-extended least-squares (FOCE-ELS) method (Phoenix software) were used for parameter estimation. The PK parameters were assumed to be log normally distributed, and between-subject variability was modeled using an exponential error model. The residual error model was described by a proportional error model. The adequacy of the base model was assessed using Akaike Information Criteria (AIC), the precision of the parameter estimates, diagnostic plots, and individual predictions. IV and SC PK data sets were simultaneously fit for each species.

Scheme 2. Compartmental Analysis Model-Informed Input and Exit Rates.

A two-compartment model with parallel first-order slow and fast-release absorption and first-order elimination was sufficient to describe LEN SC suspension and solution PK. The slow-release component was sufficiently described with a lag time (T_lag) for the aqueous suspension PK and with transit compartments for the PEG/water SC PK. Legend: Dose (mg/kg); Frac_direct, fraction of dose released via a direct process; Frac_indirect, 1-Frac_direct; fraction of dose released via an indirect process; F_SC, total fraction of dose released following subcutaneous administration; IV, intravenous; k_direct, direct release absorption rate constant (1/h); k_indirect, indirect release absorption rate constant (1/h); k_tr, transit rate constant between compartments (1/h); MTT, mean transit time (h); T_lag, lag time (h); Q, intercompartmental CL (L/h/kg); SC, subcutaneous; V_c, volume of central compartment (L/kg); V_p, volume of peripheral compartment (L/kg).

SC PK data from the aqueous suspension formulation were fitted to a systemic two-compartment PK model with parallel first-order absorption from a fast direct release rate (k_direct; 1/h) and a slow indirect release rate (k_indirect; 1/h) with a lag time (T_lag; Scheme 2). Modeling was performed on rat and dog PK data sets at dose levels (10 and 6 mg/kg, respectively) that represented the human body surface area equivalent doses and the human PK data sets at all doses (30–450 mg).

SC PK data from the PEG/water solution formulation were initially fit with the CA model used for SC PK from an aqueous suspension; however, the fit was poor. The CA model with transit compartments (Scheme 2) resulted in a better visual fit of the prolonged indirect release absorption from the depot and with a lower AIC score. Modeling was performed on rat and dog PK data sets at dose levels (50 and 12 mg/kg, respectively) that were close to a human body surface area equivalent dose.

Results

LEN IV PK

The plasma PK profiles following a single IV infusion administration of LEN in rat and dog are shown in Figure 1, and the systemic parameters are summarized in Table 1. LEN showed a biphasic IV PK in all species studied (Table 1). The average plasma clearance (CL) of LEN was low in rat (0.045 L/h/kg) and dog (0.057 L/h/kg). The average volume of distribution (V_ss) in rat (2.63 L/kg) and dog (1.12 L/kg) was larger than that of the total body water. The elimination half-life (t_1/2) ranged from 19 h in dog to 43 h in rat. In humans, the CL was low (0.057 L/h/kg) and the V_ss was larger than total body water (23.0 L/kg) (Table S1) and the elimination half-life was approximately 270 h (11.3 days).

Figure 1.

Plasma concentration–time profiles of LEN following intravenous infusion administration of LEN to rats at 3 mg/kg and dogs at 1 mg/kg (n = 3; mean ± SD).

Table 1. Plasma NCA PK Parameters for LEN Following Intravenous Infusion Administration to Nonclinical Species^ab.

species	dose (mg/kg)b	AUClast (μM·h)	AUCinf (μM·h)	CL (L/h/kg)	Vss (L/kg)	t1/2 (h)	MRT (h)
WH rat	3	53.2 ± 8.08	60.1 ± 9.79	0.045 ± 0.004	2.63 ± 0.37	42.6 ± 6.14	50.2 ± 4.38
beagle dog	1	18.9 ± 5.9	19.1 ± 6.1	0.057 ± 0.017	1.12 ± 0.12	19.3 ± 2.5	20.6 ± 6.4

^aValues are the mean ± SD from 3 animals. AUC_last = area under the concentration–time curve from 0 to the last quantified time point; AUC_inf = area under the concentration–time curve extrapolated to time infinity; CL = systemic clearance; MRT = mean residence time; NCA = noncompartmental analysis; t_1/2 = terminal elimination half-life; SD = standard deviation; V_ss = volume of distribution at steady state; WH = Wistar Han.

^bDoses were administered via a 30 min IV infusion.

LEN SC PK Following Aqueous Suspension Formulation Administration

LEN Rat PK

The plasma PK profiles are shown in Figure 2, and the NCA parameters following SC administration in an aqueous suspension are summarized in Table 2. LEN PK following SC administration to rats showed a sustained drug release at all administered doses with measurable concentrations for at least 90 days after dose administration. The fraction of dose released (F) across all dose groups ranged from 63 to 118%. The mean apparent terminal half-life (t_1/2) following SC administration in the rat ranged from 9.1 to 16.8 days, which was substantially longer than the mean t_1/2 following IV administration (Table 1), indicating flip-flop kinetics following SC administration. The average observed time of the maximum plasma concentration (T_max) ranged from 1.0 to 9.3 days.

Figure 2.

Plasma concentration–time profiles of LEN following subcutaneous administration of an aqueous suspension formulation at 100 mg/mL to male rats (n = 3; mean ± SD).

Table 2. PK Parameters of LEN Following Subcutaneous Administration of LEN in Aqueous Suspension to Male WH Rats^abcde.

group ID	dose (mg/kg)b	dose volume (mL/kg)	AUC0–72 h (μM·h)	AUC0–672 h (μM·h)	AUClast (μM·h)	AUCinf (μM·h)	Cmax (μM)	Tmax (h)	t1/2 (h)	% of AUCinf on day 3	% of AUCinf on day 28	F%e
1	10	0.10	32.5 + 20.8	194 + 14	236 + 36	236 + 36	0.671 + 0.198	25 + 23	219 + 81	14	82	118 ± 18
2	30	0.30	20.6 ± 3.9	343 ± 63	663 ± 83	664 ± 83	0.724 ± 0.222	224 ± 97	263 ± 31	3	52	110 ± 14
3	100	1.0	68.0 ± 27.2	900 ± 89	1545c	1554c	1.91 ± 0.33	224 ± 97	397c	4	58c	78c
4	200	1.0 × 2d	180 ± 93	1498 ± 223	2499 ± 188	2517 ± 194	3.81 ± 1.08	120 ± 83	403 ± 50	7	60	63 ± 5

^aValues are the mean ± SD from 3 animals. AUC_0–72 h = 3-day AUC; AUC_0–672 h = 28-day AUC; SD = standard deviation.

^bFormulation concentration at 100 mg/mL LEN.

^cOnly mean value was reported because one animal died at 2112 h postdose. This animal had appeared normal at the daily regular check time of 2088 h.

^dTwo injections, each at a distinct site.

^eF% = total fraction dose released = 100 × dose-normalized SC AUC_inf-to-IV AUC_inf ratio.

Maximum plasma concentration values (C_max) of LEN were similar between 10 and 30 mg/kg but increased dose proportionally from 30 to 100 mg/kg. Plasma C_max nearly doubled following two doses of 100 mg/kg compared with one dose of 100 mg/kg. Total LEN plasma exposure (AUC_inf) increased with dose in the range of 10, 30, and 100 mg/kg; an increase of 3.0- to 3.33-fold in dose showed 2.3- to 2.8-fold increase in AUC and was, thus, slightly less than dose proportional. LEN AUC_inf observed following a total of 200 mg/kg (2 doses of 100 mg/kg each) was greater than the exposure achieved with one 100 mg/kg dose; however, the increase was 1.6-fold and less than dose proportional.

Neither a prominent initial burst release nor unintended rapid drug release was observed. By day 3 (72 h postdose) when an initial release can be estimated, approximately 14% of the bioavailable dose (calculated as percent of AUC_0–72 h to AUC_inf) was released at the 10 mg/kg dose and a lower percentage (3–7%) was released at the higher dose levels. By Day 28 (672 h postdose), approximately 82% of the bioavailable dose (calculated as percent of AUC_0–672 h to AUC_inf) was released at the 10 mg/kg dose and a lower percentage (52–60%) was released at the higher dose levels.

LEN Dog PK

The plasma PK profiles following SC administration in aqueous suspension are shown in Figure 3, and the NCA parameters are summarized in Table 3. LEN PK showed sustained slow drug release at all administered doses with measurable LEN concentrations for at least 30 days. F% across all dose groups ranged from 49 to 114%. The mean apparent t_1/2 following SC administration in dogs ranged from 2.8 to 21.9 days, which was substantially longer than the mean t_1/2 following IV administration (Table 1), indicating flip-flop PK following SC administration. No unintended rapid drug release was observed; i.e., less than 30% of the bioavailable dose was released by day 3 (72 h postdose). The average observed T_max ranged from 1.3 to 4.0 days.

Figure 3.

Plasma concentration–time profiles of LEN following subcutaneous administration of an aqueous suspension formulation to male dogs (n = 3; mean ± SD). (A) 3–100 mg/kg of a 100 mg/mL formulation; (B) 6 mg/kg of a 100 or 200 mg/mL formulation; (C) 6 mg/kg of a 100 mg/mL formulation as one or two equal-volume injections.

Table 3. PK Parameters of LEN Following Subcutaneous Administration of LEN in an Aqueous Suspension Formulation to Male Beagle Dogs^abcd.

group ID	dose (mg/kg)b	dose volume (mL/kg)	AUC0–72 h (μM·h)	AUC0–672 h (μM·h)	AUClast (μM·h)	AUCinf (μM·h)	Cmax (μM)	Tmax (h)	t1/2 (h)	% of AUCinf on day 3	% of AUCinf on day 28	mean F%d
11	3	0.03	3.0 ± 2.1	18.6 ± 9.0	25.4 ± 7.5	28.0 ± 6.4	0.056 ± 0.037	72 ± 83	525 ± 147	11	66	49 ± 11
12	6	0.06	20.8 ± 7.4	117 ± 37	128 ± 41	129 ± 42	0.410 ± 0.141	40 ± 28	322 ± 51	16	91	114 ± 37
13	6	0.03 × 2c	18.8 ± 2.4	85.0 ± 7.0	88.4 ± 6.2	88.9 ± 6.2	0.417 ± 0.014	64 ± 28	174 ± 38	21	96	78 ± 5
14	6	0.03	6.6 ± 3.4	63.9 ± 27.3	75.9 ± 34.6	79.1 ± 32.2	0.175 ± 0.068	96 ± 63	354 ± 329	8	96	70 ± 28
15	30	0.30	145 ± 9	496 ± 49	497 ± 51	497 ± 51	2.96 ± 0.30	40 ± 14	66 ± 5	29	100	88 ± 9
16	100	1.0	340 ± 76	1336 ± 504	1356 ± 500	1357 ± 500	7.24 ± 1.28	32 ± 14	94 ± 59	25	100	72 ± 26

^aValues are the mean ± SD from 3 animals. AUC_0–72 h = 3-day AUC; AUC_0–672 h = 28-day AUC; SD = standard deviation.

^bFormulation concentration at 100 mg/mL LEN (groups 11–13; 15, 16) and at 200 mg/mL LEN (group 14).

^cTwo injections, each at a distinct site.

^dF% = Total fraction dose released = 100 × dose-normalized SC AUC_inf-to-IV AUC_inf ratio.

Following a single SC injection of the 100 mg/mL formulation, the average LEN AUC_inf increased with dose in the range of 3–100 mg/kg (Figure 3A). The increase in average AUC_inf was more than dose proportional from 3 to 6 mg/kg but less than dose proportional from 6 to 100 mg/kg. The increase in average C_max was more than dose proportional from 3 to 30 mg/kg and less than dose proportional from 30 to 100 mg/kg. By day 28, >90% of the total drug exposure (AUC_inf) was achieved in all three groups except the 3 mg/kg SC dose group. The average LEN AUC_inf following a single SC injection at 6 mg/kg was higher, but more variable compared to that from two equal-volume injections of 3 mg/kg (Table 3). The average C_max was similar, and the percent of total exposure by day 28 was higher in the two equal-volume injection groups. LEN systemic concentrations were detectable for at least 30 days. The mean apparent t_1/2 at dose levels ≤6 mg/kg was higher than the t_1/2 values observed at doses ≥30 mg/kg (Figure 3A).

Following a single SC injection at 6 mg/kg at two different concentrations (100 mg/mL, 0.06 mL/kg vs 200 mg/mL, 0.03 mL/kg), the average LEN plasma exposure (AUC_inf and C_max) was higher and the average T_max was achieved earlier in animals dosed at the lower formulation concentration (100 mg/mL). The percent of AUC_inf on day 28 was similar across both groups. LEN systemic concentrations were detectable for at least 56 days in both groups (Figure 3C).

LEN Human PK

The plasma PK profiles following a single ascending dose SC administration of the aqueous suspension formulation in healthy volunteers were previously published, and an adapted plot of the observed PK (30–450 mg) is shown in Figure 7C.^5,9 LEN PK profiles showed slow, sustained drug release, with a median apparent terminal t_1/2 of 32–45 days.⁶ LEN exposure increased approximately proportional to the increase in dose.⁵

Figure 7.

Observed and model-predicted plasma concentration–time profiles of LEN following subcutaneous administration of a LEN aqueous suspension formulation to rat (10 mg/kg; A), dog (6 mg/kg; B), and human (30–450 mg; C). Observed concentration values are mean ± SD; n = 3 for rat and dog; n = 8 for human.⁵ LEN: 1 nM = 0.968 ng/mL.

LEN SC PK Following PEG/Water Solution Formulation Administration

LEN Rat PK

The plasma PK profiles for LEN following SC administration are shown in Figure 4, and the NCA parameters are summarized in Table 4. LEN PK showed sustained drug release at all administered doses with measurable LEN concentrations for at least 250 days. The average observed T_max ranged from 29.9 to 53.7 days. The F% across all dose groups ranged from 48 to 79%. The mean apparent terminal half-life (t_1/2) following SC administration in the rat ranged from 27.0 to 59.2 days, which was substantially longer than the mean terminal t_1/2 following IV administration (Table 1), indicating flip-flop kinetics following SC administration.

Figure 4.

Plasma concentration–time profiles of LEN following subcutaneous administration of a PEG/water solution formulation at 309 mg/mL to male rats (n = 3; mean ± SD).

Table 4. PK Parameters of LEN Following Subcutaneous Administration of LEN in a PEG/Water Solution Formulation to Male WH Rats^abc.

group ID	dose (mg/kg)	dose volume (mL/kg)	AUC0–72 h (μM·h)	AUC0–672 h (μM·h)	AUClast (μM·h)	AUCinf (μM·h)	Cmax (μM)	Tmax (h)	t1/2 (h)	% of AUCinf on day 3	% of AUCinf on day 28	mean F%c
21	50	0.12	7.86 ± 1.47	152 ± 49	774 ± 74	789 ± 61	0.469 ± 0.120	718 ± 257	1056 ± 504	0.68	13	79 ± 6
22	100	0.33	16.2 ± 6.2	216 ± 113	1330 ± 238	1340 ± 228	0.500 ± 0.138	1064 ± 257	1240 ± 281	0.71	9.4	67 ± 11
23	150	0.5	21.7 ± 3.6	282 ± 14	2160 ± 211	2190 ± 205	0.765 ± 0.092	1120 ± 194	1420 ± 403	0.63	8.2	73 ± 7
24	450	1.5	75.7 ± 12.1	926 ± 164	5360 ± 787	5400 ± 837	1.95 ± 0.306	840 ± 444	1330 ± 263	0.73	9.0	60 ± 9
25	200	0.33 × 2b	32.3 ± 4.7	507 gf± 147	2410 ± 60	2410 ± 60	1.09 ± 0.17	728 ± 256	647 ± 28	0.71	11	60 ± 1
26	300	0.5 × 2b	41.1 ± 1.3	550 ± 54	3630 ± 189	3660 ± 184	1.35 ± 0.09	896 ± 775	1110 ± 299	0.60	8.0	61 ± 3
27	900	1.5 × 2b	99.4 ± 22.2	1140 ± 239	8400 ± 1160	8620 ± 1250	2.22 ± 0.35	1288 ± 423	1360 ± 156	0.48	5.5	48 ± 7

^aValues are the mean ± SD from 3 animals. AUC_0–72 h = 3-day AUC; AUC_0–672 h = 28-day AUC; SD = standard deviation.

^bTwo injections, each at a distinct site

^cF% = Total fraction dose released = 100 × dose-normalized SC AUC_inf-to-IV AUC_inf ratio.

LEN C_max was similar across 50–150 mg/kg dose levels but increased approximately dose proportionally from 150 to 450 mg/kg. Plasma C_max nearly doubled following 2 doses of 100 and 150 mg/kg. The plasma C_max following 450 mg/kg was similar following 1 vs 2 doses. Total LEN plasma exposure (AUC_inf) increased with dose; the mean increases were 1.6- to 2.5-fold and close to dose proportional. LEN AUC_inf observed following 2 doses of 100 mg/kg each (for a total of 200 mg/kg), 150 mg/kg each (for a total of 300 mg/kg), or 450 mg/kg each (for a total of 900 mg/kg) was greater than the exposure achieved with respective single dose levels; however, the increase was 1.6- to 1.8-fold and thus slightly less than dose proportional.

Neither a prominent initial burst release nor unintended rapid drug release was observed. By day 3 (72 h postdose), approximately 1% of the bioavailable dose was released at all of the dose levels. By day 28, up to 21% of the bioavailable dose was released at all of the dose levels.

LEN Dog PK

The plasma PK profile following SC administration is shown in Figure 5, and the NCA parameters are summarized in Table 5. LEN PK showed a sustained slow drug release at all administered doses with measurable LEN concentrations for at least 90 days. The average observed T_max ranged from 12.3 to 30.3 days. F% across both dose groups ranged from 47 to 103%. The mean apparent terminal t_1/2 following SC administration in dogs ranged from 17.5 to 25.8 days, which was substantially longer than the mean terminal t_1/2 following IV administration (Table 1), indicating flip-flop PK following SC administration. No unintended rapid drug release was observed, with less than 3% of the bioavailable dose released by day 3 (72 h postdose). By day 28, up to 43% of the bioavailable dose was released at the two dose levels. LEN C_max and AUC_inf were increased more than dose-proportionally from 6 to 12 mg/kg.

Figure 5.

Plasma concentration–time profiles of LEN following subcutaneous administration of a PEG/water solution formulation at 309 mg/mL to male dogs (n = 3; mean ± SD).

Table 5. PK Parameters of LEN Following Subcutaneous Administration of LEN in a PEG/Water Solution Formulation to Male Beagle Dogs^abc.

group ID	dose (mg/kg)	dose volume (mL/kg)	AUC0–72 h (μM·h)	AUC0–672 h (μM·h)	AUClast (μM·h)	AUCinf (μM·h)	Cmax (μM)	Tmax (h)	t1/2 (h)	% of AUCinf on day 3	% of AUCinf on day 28	mean F%b
31	6	0.02	0.738 ± 0.037	15.8 ± 5.3	50.2 ± 3.6	53.2 ± 2.9	0.050 ± 0.006	728 ± 194	421 ± 95	1.1	16	47 ± 3
32	12	0.04	6.02 ± 2.90	101 ± 69	166 ± 11	233c	0.267 ± 0.176	296 ± 231	620c	3.0	50	103c

^aValues are the mean ± SD from 3 animals. AUC_0–72 h = 3-day AUC; AUC_0–672 h = 28-day AUC; SD = standard deviation.

^bF% = Total fraction dose released = 100 × dose-normalized SC AUC_inf-to-IV AUC_inf ratio.

^cTerminal slope defined in two animals; only mean value reported.

LEN Human PK

The plasma PK profiles following SC administration of LEN at 927 mg (2 × 1.5 mL and 3 × 1 mL) in a PEG/water solution formulation were previously published.^7,8 An adapted plot at 927 mg (3 × 1 mL) is shown in Figure S2.⁸ LEN PK profiles showed slow, sustained drug release, with a median apparent terminal t_1/2 of 65 days.⁸

PK Modeling

The cumulative fraction of the dose released over time from the aqueous suspension and PEG/water solution formulations is presented in Figure 6. A summary of the compartmental PK modeling-derived input and exit rates is presented in Table 6, and the systemic PK parameters are shown in Table S1. The k_indirect values derived from NCA apparent terminal t_1/2 estimates and the k_indirect values obtained from CA were similar (data not shown).

Figure 6.

Cumulative extent of dose release following SC administration of aqueous suspension (A) and PEG/water solution formulations (B) to rat, dog, and human (mean ± SD; n = 3 for rat and dog, n = 8 for human).

Table 6. PK Model-Informed Release Rates and Total Fraction Dose Release^abcde.

			systemic elimination	SC depot release via direct process			SC depot release via indirect process
formulation	species	dose	kelimination (h–1)	kdirect (h–1)	t1/2 (days)b	Fracdirect	kindirect (h–1)	t1/2 (days)b	Tlag/MTT (days)	Fracindirect	F%c
aqueous suspension (100 mg/mL)	rat	10 mg/kg	0.016	0.018	1.6	0.09	0.00309	9.3	6.5d	0.91	136
	dog	6 mg/kg	0.023	0.0065	4.4	0.13	0.00326	8.8	0.6d	0.87	92
	human	30–450 mg	0.003	0.007	4.1	0.14	0.000623	46.3	0.3d	0.86	112
PEG/water solution (309 mg/mL)	rat	50 mg/kg	0.016	0.069	0.4	0.02	0.000818	35.3	15e	0.98	75
	dog	12 mg/kg	0.023	0.014	2.1	0.06	0.0015	19.2	7.3e	0.94	80

^aParameter definition shown in Scheme 2; k_elimination = 0.693/terminal t_1/2 from IV PK.

^bt_1/2 = 0.693/k_{direct or indirect}.

^cModel-estimated total fraction release.

^dT_lag.

^eMTT, given by MTT = (no. of transit compartments + 1)/k_tr.

Aqueous Suspension PK Modeling

The administered suspension doses were fully released in rat, dog, and human (mean F% of 84–110%; Figure 6A). An overlay of the observed and model-predicted PK plots is shown in Figure 7. The model-predicted SC PK were in good agreement with observed PK in all species. The model-predicted IV PK in rat, dog, and human are shown in Figure S1 and were in good agreement with observed PK in all species.

The suspension PK in all three species showed well-defined C_max peaks and fit well with a first-order parallel absorption compartmental model with two depot release rates – k_direct and k_indirect with a lag time (Figure 7). Compartmental modeling showed that the systemic elimination rate (k_elimination) was similar to the k_direct depot release rate and substantially higher (i.e., faster) than the k_indirect depot release rate across all species, indicating dissolution-limited flip-flop PK following SC administration of LEN in suspension. Modeling showed that approximately 87–91% of the dose was released via a slow indirect process in nonclinical species at similar depot release rates—k_indirect values of 0.00309 h^–1 (t_1/2 of 9.3 days) and 0.00326 h^–1 (t_1/2 8.8 days) was observed in rat and dog, respectively. The remaining small fraction (<13% of the dose) was estimated to be released more rapidly with a k_direct of 0.018 h^–1 (t_1/2 = 1.6 days) in the rat and 0.0065 h^–1 (t_1/2 = 4.4 days) in the dog. The estimated k_elimination, derived from the IV PK data, was 0.016 h^–1 (t_1/2 = 1.8 days) in the rat and 0.035 h^–1 (t_1/2 = 0.8 days) in the dog. In humans, 86% of the dose at all administered doses was released via a slow process with a k_indirect of 0.000623 h^–1 (t_1/2 = 46 days). The remainder of the dose (∼14%) was inferred to be released at a more rapid rate with a k_direct value of 0.007 h^–1 (t_1/2 = 4.1 days). As seen in animals, the predicted human PK data also displayed flip-flop kinetics due to a substantially faster k_elimination compared to the slow depot (k_indirect) release rate.

PEG/Water Solution PK Modeling

The administered solution dose was substantially released in rat, dog, and human (mean F% of 73–99%; Figure 6B). An overlay of the observed and model-simulated rat and dog PK plots is shown in Figure 8. The simulated PK were in good agreement with the observed PK in both species.

Figure 8.

Observed and model-predicted plasma concentration–time profiles following subcutaneous administration of a LEN PEG/water solution formulation to rat (50 mg/kg, A) and dog (12 mg/kg, B). Observed concentration values are mean ± SD; n = 3 for rat and dog. Insets show expanded (0–14 day) views.

The PEG/water solution PK in rat and dog show profiles with a flatter C_max (and a delayed T_max compared to the suspension PK) and fit well with a first-order parallel absorption compartmental model with two depot release rates—k_direct and k_indirect with a transit compartment model. Compartmental modeling showed that the systemic elimination rate was similar to the k_direct depot release rate and substantially faster (i.e., shorter t_1/2) than the depot (k_indirect) release rates across all species, reaffirming flip-flop PK following SC administration of LEN PEG/water solution formulation (Figure 8). In rat and dog, the depot direct and indirect release rates and the fraction of dose released via each process were well differentiated: the k_direct was markedly faster than the k_indirect with ≥94% of the dose released via the slower indirect process.

Discussion and Conclusions

Long-acting ARV regimens are needed to maintain viral suppression in PWH who have expressed a strong desire to relieve pill fatigue and/or avoid the potential stigma associated with daily oral regimens. The development of novel long-acting ARVs remains an unmet need in the fields of HIV treatment and prevention.¹¹ Small-molecule ARVs with the ability to address these needs would ideally not only allow for infrequent administration but also be able to quickly achieve clinically relevant concentrations. In this work, we present the nonclinical PK properties of LEN, a first-in-class HIV capsid inhibitor, and the modeling of the nonclinical and human PK that informed its long-acting properties and an additional course of clinical development.

During the discovery of LEN, medicinal chemists were faced with a considerable challenge to design a molecule that could interfere effectively with HIV capsid functions, be stable to hepatic drug-metabolizing enzymes, and also be easily administered in vivo. The optimized molecule with a molecular weight of 968 g/mol, 10 fluorine atoms, high lipophilicity, and low aqueous solubility was very atypical for what is considered “a drug-like agent”. The molecular design imparted the important attributes of low clearance and high potency. Collectively, these attributes of LEN are optimal properties for consideration as a long-acting injectable drug product. The dose and duration of the PK for LEN following its SC administration are governed by three key parameters (Scheme 1). The most important parameter is the high potency of the molecule because it informs the required sustained plasma concentrations to achieve persistent viral suppression, and any insufficiently potent agent would be less attractive due to improbably large dosing volume limitations.² LEN is a picomolar inhibitor of HIV capsid functions with a human serum protein binding adjusted 95% effective concentration (paEC₉₅) of 4.0 nM.⁵ The systemic PK governs the LEN exit rate and was appropriately characterized from the IV PK data across species. Low CL for LEN was observed in nonclinical species and in humans. The V_ss for LEN was moderate in nonclinical species and high in humans. The central compartment volume (V_c, Table S1) was similar across species, whereas the peripheral compartment volume (V_p, Table S1) was substantially higher in humans compared with the nonclinical species tested. The moderate to high V_ss and low clearance combine to yield a low exit rate, which is favorable to support a long duration of exposure. The LEN input rate from the SC depot is governed by its physicochemical properties, and these are also exemplary—LEN has low aqueous solubility (<1 μg/mL in water at pH 7.4) and is strongly lipophilic with a log D_7.4 value of 3.7.¹²

The LEN systemic input rate from the SC depot was found to be dissolution-limited with no unintended rapid drug release and occurs with simple dissolution-based formulations without the requirement for any significant coexcipients (such as a biodegradable polymer matrix)³ or prodrug strategies.¹³ During the nonclinical phase of early development, a significant number of SC PK studies in rats and dogs were conducted to understand the factors that would support parenteral administration with sufficient exposure and a long duration (dose volume, injection site release profile, tolerability). These studies suggested two strategies for the delivery of LEN: (1) as an aqueous suspension or (2) as a solution formulation, both form a LEN depot following SC administration. The studies presented here represent the extensive and lengthy PK characterization of an aqueous suspension formulation of amorphous LEN (free acid) in normal saline with 2% poloxamer 188 and a PEG/water solution formulation of LEN (sodium salt) in 68.2% PEG 300 and 31.8% water in rats, dogs, and human.

The initial expectation was that an aqueous suspension of LEN would release too slowly from the injection site to achieve efficacious exposure and that a solution formulation would be absorbed too quickly, leading to insufficient duration of exposure. The initial PK experiments in rat and dog with both formulations caused us to reconsider the mechanisms at play in the subcutaneous depot following the administration of the two formulations. The PK profiles for LEN following SC administration of an aqueous suspension and PEG/water solution formulations were clearly different and somewhat unexpected. The SC PK profile following the aqueous suspension administration displayed a two-phase absorption profile in all three species with an initial fast-release absorption phase and a short T_max (1–9 days in rat and dog; 14–35 days in human⁵) and a prominent C_max followed by a slow-release absorption phase. This observation would be consistent with the initial release of the soluble fraction in the administered dose volume followed by the dissolution-limited slower release of the suspension, which appears to be rapidly settled to form the drug depot. In contrast, the SC PK profile following the PEG/water solution administration was comparatively more complex across all species. In rat and dog, a short first absorption phase was observed (most apparent in dog PK; Figure 5) followed by a more prolonged second phase with a longer T_max (12–54 days in rat and dog) and a less prominent C_max. In human, the PK profile shows a shallow prolonged absorption phase (with no discernible C_max peak) with a drawn-out T_max (77–84 days⁸). One hypothesis for this observation would be that LEN administered in a solution formulation has a short initial absorption phase of LEN dissolved in the vehicle with the majority of the drug precipitating in situ to form the drug depot, but the nature of the depot is more fluid over a much longer duration as the PEG/water vehicle dissipates over time.

The LEN input rates from the depot that is formed following SC administration were informed by NCA and CA. The NCA informed the overall extent of release (F%) following SC administration, which was high in nonclinical species and in human for both formulations. A fit-for-purpose CA enabled deconvolution of the fast and slow depot release process contributions to the overall PK profiles. The compartment model (Scheme 2) was conceptually set up with a parallel absorption model with two first-order release rates—absorption of the soluble fraction (minor) from a direct process (with a k_direct rate) was to characterize the fast-release process leading to the early PK phase and absorption of the precipitated fraction (major) from the indirect process (with a k_indirect rate) was to characterize the slow-release process leading to the later phase (Table 6). The CA-derived k_direct was faster (i.e., higher) than the k_indirect in rat, dog, and human following dosing in an aqueous suspension formulation and in rat and dog following dosing in a PEG/water solution formulation. The human PK profile following SC administration of the PEG/water solution formulation model is more complex^7,8 and model-informed release rates will be subject of a future publication.

The main purpose of the nonclinical PK studies of the suspension and solution SC formulations was to predict and simulate human release rates to increase confidence that LEN formulations could achieve a target product profile of at least a once-monthly dose regimen in HIV patients. The human PK results fared better than predicted with slower LEN depot release rates observed in humans compared to both rat and dog. For the aqueous suspension formulation, the human slow process release rate (k_indirect; Table 6) was approximately 5-fold slower than the nonclinical slow-release rates. For the PEG/water solution formulation, the human slow process release rate (estimated from apparent terminal t_1/2(8)) was approximately 1.8- to 3.7-fold slower than the nonclinical slow-release rates. The administered SC dose was predominantly released via the slow-release process in nonclinical species and humans and is thus the main contributor to the overall long-acting PK duration.

The observed nonclinical PK properties of LEN strongly supported its selection as a candidate for long-acting administration. The clinical data confirmed and, in some aspects, exceeded the predicted human PK properties of LEN confirming the feasibility of LEN administration as a long-acting injectable drug at exceptionally long intervals. After a single 927 mg SC dose of LEN in the PEG/water solution formulation, target plasma concentrations corresponding to a mean 6-fold over paEC₉₅ were sustained for at least 6 months.⁷ LEN (Sunlenca; 927 mg in PEG/water formulation), administered subcutaneously twice-yearly, was recently approved for the treatment of HIV-1 infection in heavily treatment-experienced adults with multidrug-resistant HIV-1 infection failing their current antiretroviral regimen due to resistance, intolerance, or safety considerations.¹⁴ In addition, LEN administered SC once every 6 months is currently being tested in virally suppressed PWH in combination with other ARVs and as a single agent for HIV pre-exposure prophylaxis.

Glossary

Abbreviations

ADME

absorption, distribution, metabolism, and elimination

AIC

Akaike Information Criteria

ARV

antiretroviral

AUC

area under the plasma concentration–time curve

AUC_0–t

AUC from time 0 to a specific sampling time point

AUC_last

AUC from time 0 to the last sampling time point

AUC_inf

AUC from time 0 extrapolated to infinity

AUC, CA

compartmental analysis

CL

clearance

FOCE-ELS

first-order conditional estimation-extended least-squares

Frac_direct

fraction of dose released via a direct process

Frac_indirect

fraction of dose release via an indirect process (1-Frac_direct)

F_SC

total fraction of dose released following SC administration

C_max

observed peak plasma concentration

F

bioavailability

HIV-1

human immunodeficiency virus type 1

IV

intravenous

k

rate constant, defined in each context

k_direct

direct release absorption rate constant

k_indirect

indirect release absorption rate constant

k_tr

transit compartment rate constant via

LC-MS/MS

high-performance liquid chromatography-tandem mass spectrometry

LEN

lenacapavir

MRT

mean residence time

MTT

mean transit time

NCA

noncompartmental analysis

PWH

people with HIV

PEG

polyethylene glycol

PK

pharmacokinetics

Q

intercompartmental CL

SC

subcutaneous

SD

standard deviation

t_1/2

half-life as defined in each context

T_lag

lag time

T_max

time to reach C_max

V_c

volume of central compartment

V_p

volume of peripheral compartment

V_ss

volume of distribution at steady state

WH

Wistar Han

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.molpharmaceut.3c00626.

• Observed and model-predicted plasma concentration over time profiles of LEN following IV administration of LEN to rats, dogs, and humans; systemic PK parameters of LEN in rat, dog, and human derived from a 2-compartment model; and adapted plot of the human plasma concentration over time profile of LEN following administered SC in a PEG/water solution (PDF)

Special Issue

Published as part of the Molecular Pharmaceuticsvirtual special issue “Research Frontiers in Industrial Drug Delivery and Formulation Science”.

Author Present Address

^† Neuron23, Inc., 343 Oyster Point Boulevard, Suite 120, South San Francisco, California 94080, United States

Author Present Address

^‡ Ventyx Biosciences, Inc., 662 Encinatas Boulevard, Suite 250, Encinatas, California 92024, United States.

Author Present Address

^§ Vir Biotechnology, Inc., 499 Illinois Street, San Francisco, California 94158, United States.

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

All authors were employees of Gilead for the body of work reported in this manuscript. This work was supported by Gilead Sciences, Inc.

The authors declare no competing financial interest.