Trial design

HIV-uninfected, low-risk participants (Clinicaltrials.gov NCT02165267) were randomized in a 1:1 ratio to receive open-label IV VRC01 in treatment groups T1 or T2 (Fig 1A, and S1 CONSORT Checklist). Participants in treatment group T1 (n = 20) received VRC01 at 40 mg/kg IV at week 0 and then 20 mg/kg at weeks 4, 8, 12, 16, and 20; participants in treatment group T2 (n = 20) received VRC01 at 40 mg/kg IV at weeks 0, 8, and 16. Participants who consented to a more intensive visit schedule were assigned to a blinded third group in a 5:1 ratio to receive SC VRC01 (treatment group T3) or placebo (group 3, P3) injections. Participants in treatment/placebo group T3/P3 received SC injections of VRC01 (n = 20) every 2 weeks at 5 mg/kg or placebo (n = 4), following an IV loading dose of VRC01 at 40 mg/kg or IV placebo. Two additional dosing regimens were subsequently added via a protocol modification to evaluate potential candidate regimens for future efficacy trial testing. Participants in these latter 2 groups were randomized in a 1:1 ratio to open-label treatment groups T4 or T5 (n = 12 each) to receive IV infusions of VRC01 at 10 mg/kg and 30 mg/kg, respectively, at weeks 0, 8, and 16. Blood samples were collected before product administration and at 3 days, 2, 4 (except treatment/placebo group T3/P3), and 8 (treatment groups T2, T4, and T5 only) weeks after each infusion (or injection) as well as at 1 hour post last infusion, and at 10–16 weeks post last infusion/injection (Fig 1B). Specifically, trough measurements were taken at 4 weeks (for treatment group T1) or 8 weeks (for treatment groups T2, T4, and T5) following infusions and 2 weeks following SC injections (treatment/placebo group T3/P3), immediately prior to the next infusion or injection; peak measurements were taken at 1 hour post-infusion for IV groups (treatment groups T1, T2, T4, and T5) and 3 days after injection for the SC group (treatment/placebo group T3/P3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. HVTN 104 CONSORT flow diagram (A) and specimen collection schedule (B).

T1: 20 mg/kg IV q 4 weeks with 40 mg/kg IV loading; T2: 40 mg/kg IV q 8 weeks; T3 (P3): 5 mg/kg SC q 2 weeks with 40 mg/kg IV loading; T4: 10 mg/kg IV q 8 weeks; T5: 30 mg/kg IV q 8 weeks. ADCC, antibody-dependent cellular cytotoxicity; ADCP, antibody-dependent cellular phagocytosis; CONSORT, Consolidated Standards of Reporting Trials; HVTN, HIV Vaccine Trials Network; IV, intravenous; P3, placebo group 3; q, quodque; SC, subcutaneous; T1, treatment group 1; T2, treatment group 2; T3, treatment group 3; T4, treatment group 4; T5, treatment group 5.

https://doi.org/10.1371/journal.pmed.1002435.g001

Participants

A total of 88 participants were each enrolled into 1 of the 5 treatment groups in HVTN 104. Eligibility criteria included being low risk, HIV uninfected, aged 18 to 50 years; weighing between 53 kg and 115 kg; being deemed healthy based upon medical history, physical examination, and laboratory tests; having access to a participating HVTN clinical research site for the planned duration of the study; and demonstrating understanding of the study. Volunteers who were born female and had reproductive potential were screened for pregnancy and excluded if pregnant, and were required to agree to use effective contraception for the duration of the study. The Investigational New Drug (IND) application was sponsored by the National Institute of Allergy and Infectious Diseases (NIAID) Division of AIDS (DAIDS). The study was conducted in compliance with the principles of GCP (ICH E6) and according to DAIDS and HVTN policies and procedures. Clinical research site staff completed enrollment procedures and obtained written informed consent from all participants. The study was approved by the local institutional review boards at all participating sites: Brigham, Partners Human Research Committee; Fenway, Fenway Health IRB; Cleveland, University Hospitals, Cleveland Medical Center IRB; New York Blood Center, New York Blood Center IRB; College of Physicians & Surgeons, Columbia University IRB; Philadelphia, University of Pennsylvania IRB (Clinicaltrials.gov NCT02165267).

Interventions

VRC01 (VRC-HIVMAB060-00-AB) was formulated at a concentration of 100 (±10) mg/ml suspended in formulation buffer (25 mM sodium citrate, 50 mM sodium chloride, and 150 mM L-arginine hydrochloride at pH 5.8). The placebo for SC administration was VRC-PLAMAB068-00-AB, a sterile buffered aqueous solution of 25 mM sodium citrate, 50 mM sodium chloride, 150 mM L-arginine hydrochloride, 10% dextran 40 (w/w), and 0.005% polysorbate 80 (w/w) at pH 5.8. The purified product vials and the placebo vials were filled and labeled at the VRC Vaccine Pilot Plant operated by Leidos Biomedical Research, Inc. (Frederick, MD). The placebo was matched for color and viscosity to mimic the interventional product to avoid unintentional unblinding. The placebo for IV infusions was sodium chloride for injection 0.9%, USP, obtained locally by the sites. IV infusions were administered via an IV pump over 60 minutes for the first administration, then over at least 30 minutes for subsequent infusions. For each SC administration, up to 4 injections of 2 mL of the study product could be administered, with abdominal sites being used most frequently.

Outcomes

The prespecified primary clinical objectives of the study were to evaluate the safety and tolerability of VRC01 administered intravenously and subcutaneously at multiple time points. Primary endpoints were local and systemic reactogenicity signs and symptoms, laboratory measures of safety, adverse events (AEs) and serious AEs (SAEs), early discontinuation of infusions and reason(s) for discontinuation, and early study termination. The primary laboratory objective was to evaluate the serum levels of VRC01 at week 24, following IV and SC administration in 3 different regimens, with the primary end points being serum concentrations of VRC01 in each group at week 24. Secondary objectives included the determination of serum levels of VRC01 over time and whether anti-idiotypic antibodies could be detected. Exploratory objectives included the evaluation of in vitro functional humoral activities of VRC01 in serum and mucosal compartments. Specimens from participants were also tested for neutralizing activity against a panel of circulating and laboratory strains of HIV. Assessment of the binding of VRC01 to multiple Env proteins as well as the determination of VRC01 concentrations and functional activity in genital, rectal, and oral secretions are still underway and will be reported in a future manuscript.

Sample size

The sample size of the study was determined to provide reasonable precision in the estimation of safety event rates and the mean trough drug concentration after the administration series for each active treatment arm. Specifically, for an active arm of n = 12, there is a 90% chance of observing at least 1 safety event if the true rate of such an event is 17.5% or more, and there is a 90% chance of observing no events if the true rate is 0.88% or less. For each active arm of n = 20, there is a 90% chance of observing at least 1 event if the true rate of such an event is 10.9% or more, and there is a 90% chance of observing no events if the true rate is 0.52% or less. In addition, the chosen sample sizes provide reasonable precision for the estimation of serum concentrations, assuming a normal distribution for the log-transformed trough levels. Further discussion of the statistical considerations involved in the selection of sample sizes is provided in the full protocol (S1 Protocol).

Randomization

The randomization sequence was obtained by computer-generated random numbers and provided to each HVTN clinical research site through the Statistical Data Monitoring Center (SDMC) of the HVTN via a web-based randomization system. The randomization was conducted in blocks to ensure balance between T1 and T2, between T4 and T5, and within T3/P3 treatment groups. At each institution, the pharmacist with primary responsibility for dispensing study products was charged with maintaining security of the treatment assignments.

Blinding

Participants and site staff were unblinded as to participant treatment assignments among treatment groups T1, T2, T4, and T5, but blinded, except for site pharmacists, as to treatment assignment (active versus placebo) for treatment/placebo group T3/P3. Study product assignments were accessible to study pharmacists, contract monitors, and SDMC staff, who are required to know this information in order to ensure proper trial conduct.

Clinical assessments

Participants were monitored for type 1 hypersensitivity reactions, cytokine release syndrome (CRS), and serum sickness during and post infusion. Reactogenicity assessments were performed at baseline (prior to each study product administration) and following each IV infusion/SC injection at the early assessment time point (within 25–120 minutes) and daily for the subsequent 3 days. Participants were instructed to record symptoms using a postproduct symptom log and to contact the site daily during the assessment period. Local injection site reactions of pain, tenderness, erythema, and induration and systemic symptoms of malaise and/or fatigue, myalgia, headache, nausea, vomiting, chills, arthralgia, and body temperature were assessed. Following the first study infusion, safety monitoring included the performance of abbreviated physical exams and the assessment of AEs and concomitant medications at each study visit. Clinical laboratory testing included the collection of complete blood count and chemistry panel at 2 weeks (in all groups), then monthly until month 6 (in treatment and placebo groups T1 and T3/P3) and at month 2, 3, 4, and 6 (treatment groups T2, T4, and T5) and urine dipstick/urinalyses at 2 weeks and 3 and 6 months in all groups. Pregnancy testing was performed in female participants and results confirmed as negative prior to each study product administration. AEs were reported for the duration of the study. Safety reports were monitored daily by the HVTN Core for events meeting safety pause criteria and prompt Protocol Safety Review Team (PSRT) AE review rules. Weekly safety data reviews by the PSRT led to teleconferences when indicated, and the independent HVTN Safety Monitoring Board (SMB) reviewed unblinded safety data thrice yearly. Research sites reported clinical safety data to the Statistical Center for HIV/AIDS Research & Prevention (SCHARP) at the Fred Hutchinson Cancer Research Center. The number and percentage of participants experiencing each type of reactogenicity sign or symptom was tabulated by maximum severity and treatment arm and the percentages were displayed graphically by arm. Severity of reactogenicity and AEs was graded using the DAIDS US NIH Table for Grading the Severity of Adult and Pediatric Adverse Events, Version 1.0, December, 2004; clarification August, 2009 [36].

Pharmacokinetic analysis

VRC01 concentration in participants’ sera was quantified using the murine anti-VRC01 mAb 5C9 as previously described [33]. Briefly, 5C9 was coated onto Immulon-4HXB microtiter plates overnight at 4°C. Plates were washed and then blocked with 10% fetal bovine serum in phosphate-buffered saline. Duplicate serial 3-fold dilutions covering the range of 1:100–1:24,300 of the test sample were added and incubated for 2 hours at 37°C, followed by the addition of horseradish peroxidase-labeled goat antihuman antibody and TMB substrate. Color development was stopped by the addition of sulfuric acid, and plates were read within 30 minutes at 450 nm via the Molecular Devices Paradigm plate reader. Four-parameter logistic curve regression of a standard curve of VRC01 covering the range from 0.98 to 1,000 ng/ml was utilized to quantitate sample concentrations based upon the average of sample dilutions within the range of the assay. Concentration values below the limit of quantification (1.10 μg/ml) were replaced by 0.55 μg/ml in all calculations.

Neutralization assay

Neutralizing activity was measured against HIV Env-pseudotyped viruses as a function of reductions in Tat-regulated luciferase (Luc) reporter gene expression in TZM-bl cells as described [37,38]. Assays for measurements of serum VRC01 concentration were performed with 3 Env-pseudotyped viruses that exhibit either a highly sensitive tier 1A phenotype (clade B: MN.3, clade C: MW965.26) or a moderately sensitive tier 2 phenotype that is typical of most circulating strains (clade B: PVO.4). As these 3 strains are highly sensitive to neutralization by VRC01, they were selected for optimal sensitivity for the assays quantifying levels of the antibody. The 50% inhibitory concentration (IC50) and 80% inhibitory concentration (IC80) of the clinical lot of VRC01 were determined against each virus. The stock concentration of VRC01 used in this analysis (0.34 mg/ml) was determined by the same ELISA used for pharmacokinetic (PK) measurements. The concentration of VRC01 in serum samples was calculated by multiplying the 50% infectious dose (ID50) (or 80% infectious dose [ID80]) neutralization titer (as a dilution factor) of the serum sample against each isolate by the IC50 (or IC80) of the clinical lot VRC01 (in μg/ml) against the corresponding isolate, in which the IC50 and IC80 of the clinical lot of VRC01 was the geometric mean value obtained in repeated assays. Additional assays were performed with a global panel of 11 tier 2 Env-pseudotyped reference strains [39] to assess the magnitude and breadth of serum neutralizing activity in greater detail. These latter results were expressed as the serum dilution that resulted in either a 50% or 80% reduction in infectivity (ID50 and ID80 neutralization titers). Assays with this global panel of tier 2 viruses were performed for 6 participants from treatment group T5 (selected based upon sample availability for all time points) at 4 weeks following each infusion and at 1 hour and 8 weeks following the last infusion.

ADCC assay

ADCC activity against CEM.NKRCCR5 cells infected with the subtype C HIV-1-infected CH505 transmitted/founder infectious molecular clone (IMC) (provided by Dr. Ochsenbauer, University of Alabama–Birmingham) was measured in 96-well plates as a function of reductions in Renilla Luc reporter gene expression as previously described [40,41]. One positive control in duplicate and 1 standardized negative control in duplicate were used per plate. The readout was reduction in Relative Luminescence Units (RLU), referred to as percentage specific killing. For each sample, percent specific killing was measured in 2 wells at a dilution of 1:25. The CD4bs bnAbs 3BNC117 and CH31 (both generated by Duke Human Vaccine Institute, Protein Production Facility, Dr. Haynes) and the clinical product VRC01-13-123 were also tested starting at the concentration of 50 μg/ml. The anti-RSV Synagis mAb (MedImmune, LLC, Gaithersburg, MD) was utilized as a negative control. The analyses of plasma ADCC responses examined peak percent specific killing (%SK), defined as the maximum %SK across the 6 dilution levels (“peak %SK”) and averaged over the 2 replicates. A positive response is defined as an increase in peak percentage over baseline greater than or equal to 10%.

Antibody-dependent cellular phagocytosis assay

Antibody-dependent cellular phagocytosis (ADCP) was measured as previously described [42,43], with the following modifications. Briefly, biotinylated ConS gp140CF antigen (Duke Human Vaccine Institute, Protein Production Facility, Dr. Haynes) [44] was conjugated to neutravidin fluorescent microspheres and 9×10⁵ microspheres equivalent were mixed with 10 μl of 25 μg/ml mAbs or 25 μg/ml purified plasma immunoglobulin G (IgG); VRC01 represented approximately 1% of the total purified plasma IgG when quantified by Binding Antibody Multiplex Assay (BAMA). After incubation, 50,000 (THP-1) or 500,000 (monocyte) cells were added to each well and spinoculated. Cells were fixed and phagocytosis was analyzed by flow cytometry from at least 2,000 cell events per sample. A phagocytic score was determined by the ratio of samples (percent bead positive multiplied by the MFI bead positive values for samples) to the no-bead antibody controls. Samples were considered positive if the average phagocytic score was(1) greater than the 95th percentile of the score for all baseline visits tested (visit 2) and (2) the follow-up visit response value was ≥3 times the baseline visit response value. Data presented are representative of 2 independent experiments. Positive controls included the VRC01 clinical lot, another VRC01-class CD4bs bnAb, VRC-CH31 [27], and the Env glycan-specific bnAb 2G12 [45] and negative controls included palivizumab (RSV fusion protein-specific mAb) [46] and CH65 (H1N1-specific mAb) [47].

Infectious virion capture assay

Infectious virion capture assay (IVCA) was measured as previously described [48,49]. Briefly, 25 μg of total purified IgG was mixed with BaL.LucR stock at a final concentration of 20 ug/mL to form antibody–virion immune complexes (ICs), which were passed through a protein G column. The infectivity of the flow-through was measured by a TZM-bl infectivity assay. The percentage of captured infectious virions (iVirion) was calculated as follows: iVirion = [(100—flow-through infectivity) / (virus no-Ab infectivity)] × 100%. VRC01 clinical lot, VRC-CH31 [27], and 2G12 [45] were used as positive controls. Negative controls included CH65 [47] and palivizumab [46] as well as a no-Ab antibody control. Samples were considered positive if the average IVCA was(1) greater than the 95th percentile of the score for all baseline visits tested (visit 2) and (2) follow-up visit response value was ≥3 times the baseline visit response value.

Antidrug antibody analysis

Anti-VRC01 antibodies were measured by using the Meso Scale Discovery (MSD) platform via electro-chemiluminescence as described [33]. Specifically, serial dilutions of serum samples were incubated with optimized concentrations of biotinylated VRC01 and SULFO-TAG labeled VRC01 at 37°C for 2 hours. Samples were transferred to a previously blocked Streptavidin-coated MSD plate and incubated for 3 hours at room temperature on a plate shaker. Plates were washed on an automated plate washer, read-buffer was added, and samples were read with the MSD-2400 plate reader. Mean signals from replicate wells were evaluated and the endpoint dilution was defined as the greatest sample dilution with a response above the positivity threshold for the assay. The positivity threshold is the lower limit of the linear range for VRC01 in the assay. Serial dilutions of the VRC01-specific monoclonal antibody 5C9 were utilized to establish the linear range of 5C9 (0.01221 to 3.5 μg/mL). Samples were initially tested in quadruplicate wells with an initial dilution of 1:4. Any samples positive at 1:4 were further tested with 8 serial 4-fold dilutions (covering a dilution range of 4 to 65,536) and reported as one of the following: <4; 4; 16; 64; 256; 1,024; 4,096; 16,384; or >65,536.

Antidrug antibody (ADA) assays were performed on randomly selected baseline samples from 10 VRC01 participants and on select post-infusion/injection time points from all VRC01 participants. The post-IV infusion time points were week 24 (4 weeks after the final IV infusion in T1; 8 weeks after the final IV infusion in T2, T4, and T5; 2 weeks after the final injection in T3) and week 32 (12 weeks after the final IV infusion in T1; 16 weeks after the final IV infusion in T2, T4, and T5; and 10 weeks after the final SC injection in T3).

Statistical methods

The modified intent-to-treat (MITT) cohort included all enrolled HIV-uninfected participants receiving the first infusion and was analyzed according to participants’ assigned treatment. The per-protocol (PP) cohort included participants who received all assigned treatments within the expected visit windows. Safety evaluation was done in the MITT cohort. The number and percentage of participants experiencing each AE or reactogenicity symptom and sign were analyzed by severity and treatment group. VRC01 levels were evaluated in both the MITT and PP cohorts. Geometric means and associated 95% confidence intervals of VRC01 levels were based on a t distribution. Functional activities and their correlations with VRC01 levels were evaluated in the PP cohort.

Individual-level non-compartmental pharmacokinetics analysis was performed based on the ELISA-based serum concentration-time data in the PP cohort. Specifically, VRC01 accumulation was computed based on the ratio of trough drug level measurements and the ratio of the area under the curve (AUC) within the infusion interval for each participant, after the third and second infusions, as compared to the first in groups T4 and T5. The estimates were crude due to the relatively sparse time points. The trough levels were either observed or predicted as the estimated mean concentration at 8 weeks post infusion based on the fitted individual-level log-linear line through the observed concentrations at day 28 and day 56 after the first infusion, at day 84 and day 112 after the second infusion, and at day 140 and day 168 after the third and last infusion. AUC was calculated using the linear trapezoidal method. The expected visit dates were used in these calculations. Lastly, the terminal elimination rate and half-life were calculated based on the log-linear portion (28 to 84 days after the last infusion) of the time-concentration curve for each participant in groups T4 and T5. The terminal rate (slope) was determined by fitting a linear regression line to the log-transformed concentrations over time. The half-life was estimated as ln2 (= 0.693) divided by the terminal slope. Analyses were performed using SAS and R 3.1.1 [50].

Safety and adherence

Eighty-eight healthy, HIV-uninfected, low-risk participants were enrolled in 6 US clinical research sites affiliated with the HVTN between September 9, 2014, and July 15, 2015. Participants were followed for 32 weeks after the initial study infusion, and all participants completed the trial by February 2016. The median age of enrollees was 27 years (range, 18–50); 52% were White (non-Hispanic), 25% identified as Black (non-Hispanic), 11% were Hispanic, and 11% were non-Hispanic people of diverse origins (Table 1). Volunteers received the following: T1, single IV infusion of 40 mg/kg VRC01 followed by 5 IV infusions of 20 mg/kg VRC01 every 4 weeks; T2, 3 IV infusions of 40 mg/kg VRC01 every 8 weeks; T3 (P3), single IV infusion of 40 mg/kg VRC01 (or placebo) followed by 11 SC injections of 5 mg/kg VRC01 (or placebo) every 2 weeks; T4, 3 IV infusions of 10 mg/kg VRC01 every 8 weeks; and T5, 3 IV infusions of 30 mg/kg VRC01 every 8 weeks (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographics and study product administration frequencies by group in the MITT cohort.

https://doi.org/10.1371/journal.pmed.1002435.t001

In the MITT cohort of 88 participants, 249 IV infusions and 208 SC injections were administered and were generally well tolerated, with 28% (95% CI: 23%, 34%) of infusions and 14% (95% CI: 10%, 19%) of injections associated with mild pain or tenderness. Fifty-five percent (95% CI: 44%, 65%) of VRC01 and 50% (95% CI: 15%, 85%) of placebo recipients experienced mild pain and/or tenderness at the infusion or injection site sometime during the trial, 1 person (in T3) experienced moderate tenderness following one of the SC injections, and no participants reported severe pain and/or tenderness. Very few of the erythema/induration reactions that occurred met the NIH’s DAIDS criteria for mild reactions (>25–81 cm²); these included 2 participants following receipt of IV VRC01 (1 in T3 and 1 in T5) and 1 participant following IV placebo administration (P3). Only 2 met the criteria for moderate reactions (>9 cm in diameter or >81 cm²); these included 1 participant following IV VRC01 (T2) and 1 participant following an SC injection (T3) (Fig 2). For 76% of the infusions and injections administered in the trial, no systemic symptoms were reported. Fifty-six percent (95% CI: 45%, 66%) of VRC01 and 75% (95% CI: 30%, 95%) of placebo recipients experienced systemic reactogenicity symptoms following at least 1 study product administration. The symptoms were all graded as mild in the placebo recipients, while, of those VRC01 recipients who experienced any systemic symptoms, the maximum severity was mild in 70%, moderate in 23%, and severe in 6% (3 participants). In the 3 participants with severe reactogenicity symptoms, 2 had concurrent viral infections, while 1 reported severe malaise lasting 1 day, with no sequelae. The most commonly reported systemic symptoms were malaise/fatigue, headaches, or myalgias.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Frequency of maximum severity of systemic (A) and local (B) reactogenicity symptoms by group in the MITT cohort.

T1: 20 mg/kg IV q 4 weeks with 40 mg/kg IV loading; T2: 40 mg/kg IV q 8 weeks; T3 (P3): 5 mg/kg SC q 2 weeks with 40 mg/kg IV loading; T4: 10 mg/kg IV q 8 weeks; T5: 30 mg/kg IV q 8 weeks. Grading per DAIDS Table for Grading the Severity of Adult and Pediatric Adverse Events, Version 1.0, December 2004; clarification August 2009 [36]. DAIDS, Division of AIDS; MITT, modified intent-to-treat; P3, placebo group 3; q, quodque; SC, subcutaneous; T1, treatment group 1; T2, treatment group 2; T3, treatment group 3; T4, treatment group 4; T5, treatment group 5.

https://doi.org/10.1371/journal.pmed.1002435.g002

There were 235 AEs occurring in 70 participants (79.5%), with similar rates of occurrence across all treatment groups. Seventy-four percent of AEs were graded as mild, 22.5% moderate, and 3.4% severe. There were 9 AEs (3.8% of all AEs, 95% CI: 2%, 7%) occurring in 8 participants that were deemed product related by the investigators; all were mild and transient, and those occurring following VRC01 administration included elevations of hepatic transaminases (aspartate transaminase [AST] and alanine transaminase [ALT] elevation in 1 participant), elevated creatinine, neutropenia, localized injection site pruritus, diarrhea, generalized rash, and varicella zoster virus reactivation. An AE of chest tightness deemed related to study product occurred in a placebo recipient. No VRC01-related hypersensitivity reactions or CRS symptoms were observed during the study.

Study product was discontinued in 8 participants, including 2 participants who relocated from study sites, 1 participant who became pregnant, 1 who did not adhere to study visits, and 1 who was unable to be contacted. One participant developed a mild, generalized, pruritic, maculopapular rash 3 days after receiving the first SC injection of VRC01, which resolved less than 4 hours after applying inert lotion and was deemed related to the study product. Another participant had a brief syncopal episode several hours after receiving a VRC01 infusion, deemed not related to the study product. A third participant had mild, brief chest tightness following the first SC injection that resolved spontaneously; this was deemed related to the study product; however, this participant was determined after unblinding to be a placebo recipient.

Regarding adherence, 35 out of 44 (80%) participants in groups T2, T4, and T5 received all expected 8-weekly infusions, with 90% complete adherence in T2, 75% in T4, and 83% in T5. Eleven of 20 participants (55%) received all expected 4-weekly infusions in group T1, while 13 out of 24 (54%) participants received all biweekly injections in group T3/P3 (Table 1). Twenty-nine participants (33%) missed at least 1 infusion or injection during the study. The most commonly cited reasons for missed visits were travel, relocation, or loss to follow-up. The study product was not administered for 11 study visits, affecting 9 participants in the first 3 groups, due to the observation of frozen study product at the end of a thawing period of less than 1 hour, leading to revision of pharmacy instructions to request a thawing period of greater than 1 hour, without further abnormalities observed. The 59 participants (57 VRC01 recipients and 2 placebo recipients) who received all study product administrations comprise the PP cohort and, unless otherwise stated, were the focus of PK and immunological assessments.

Pharmacokinetics

Fig 3 and Table 2 depict the group-level geometric mean serum VRC01 concentrations measured by ELISA and neutralization assays for participants in the PP cohort. Individual-level VRC01 concentrations measured by ELISA are shown in S1 Fig for participants in the MITT (S1 Fig, panel A) and PP (S1 Fig, panel B) cohorts; summaries of group-level serum concentrations by time point are provided in S1 Table (MITT cohort) and S2 Table (PP cohort). Detailed population pharmacokinetics modeling of serum concentration data in the MITT cohort has been previously described [35]. In the PP cohort (S2 Table), as measured by ELISA, average (geometric mean) serum VRC01 levels in T1 3 days after the first (40 mg/kg) and second (20 mg/kg) IV infusion were 422 and 260 μg/ml, respectively. Average peak levels 1 hour after the last infusion were 796 μg/ml. Average trough levels 28 days post infusion were 69 μg/ml after the first infusion and 51, 46, 45, 43, and 46 μg/ml after infusions 2–6, respectively. VRC01 remained detectable (≥1.1 μg/ml) in 6/9 (67%) of PP participants 12 weeks after the last IV infusion in T1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. VRC01 serum concentrations over time in the PP cohort, measured by ELISA and TZM-bl assays.

ELISA measured anti-idiotypic binding. TZM-bl measured neutralization against 3 Env-pseudotyped virus strains (MN.3, MW965.26, and PVO.4). Each dot (bar) indicates the geometric mean (± standard error) concentrations across participants in each group. A double tick on the x-axis denotes a time interval of 3 days, representing visit days 3, 17, 31, 59, 115, and/or 157. Peak concentrations were measured at 1 hour after the last IV infusion in treatment groups T1, T2, T4, and T5 and 3 days after the first and last SC injections in group T3/P3. Concentrations were truncated at the limit of quantification of the least sensitive assay. T1: 20 mg/kg IV q 4 weeks with 40 mg/kg IV loading; T2: 40 mg/kg IV q 8 weeks; T3 (P3): 5 mg/kg SC q 2 weeks with 40 mg/kg IV loading; T4: 10 mg/kg IV q 8 weeks; T5: 30 mg/kg IV q 8 weeks. Env, HIV-1 envelope glycoprotein; PP, per-protocol; P3, placebo group 3; q, quodque; SC, subcutaneous; T1, treatment group 1; T2, treatment group 2; T3, treatment group 3; T4, treatment group 4; T5, treatment group 5.

https://doi.org/10.1371/journal.pmed.1002435.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Summary of serum concentration levels (μg/ml) in the PP cohort.

Shown are geometric mean, 95% CI around the mean, and proportion of observations above the LLoQ of the ELISA assay.

https://doi.org/10.1371/journal.pmed.1002435.t002

In T2, average serum VRC01 levels 3 days after each of the 3 bimonthly IV infusions of 40 mg/kg were 441, 414, and 486 μg/ml, respectively. The corresponding day 3 post-infusion levels in T5 (3 bimonthly IV infusions of 30 mg/kg) were 153, 183, and 242 μg/ml and in T4 (3 bimonthly IV infusions of 10 mg/kg) were 84, 87, and 113 μg/ml. Average peak levels 1 hour after the last infusion were 1,549; 1,177; and 420 μg/ml in T2, T5, and T4, respectively. Average trough levels 56 days after the first, second, and third IV infusion in T2 were 20, 24, and 27 μg/ml, respectively. In T5, these trough values were 12, 13, and 16 μg/ml. In T4, these trough values were 4, 4, and 6 μg/ml. VRC01 remained detectable (≥1.1 μg/ml) 16 weeks after the last infusion in 80%, 38%, and 0% of participants in T2, T5, and T4, respectively. The estimated mean ± standard deviation (SD) serum half-life of VRC01 in T4 and T5 was 11.4 ± 4.3 days (median = 11.0 days) and 13.7 ± 6.5 days (median = 10.2 days), respectively. Limited levels of VRC01 accumulation were observed in T4 and T5 (S3 Table). These results are consistent with previously reported half-life and drug accumulation estimates for VRC01 [33] and are independent of dose levels [35].

In T3, the average serum VRC01 level 3 days after IV infusion of 40 mg/kg was 416 μg/ml. The 3-day levels after the first and last SC injection of 5 mg/kg were 170 and 60 μg/ml, respectively, the latter concentration being a more accurate reflection of levels achieved by low-dose SC injection, with minimal contribution from the high-dose IV loading. Similarly, average trough levels measured 14 days after each subsequent 2-weekly SC injection decreased steadily from 85 μg/ml after the first injection to 34 μg/ml after the final injection. VRC01 remained detectable (≥1.1 mg/ml) in 60% of participants 10 weeks after the last injection. No VRC01 was detected in the placebo group (P3). The relatively stable pharmacokinetics after multiple IV and SC dosing suggests the absence of anti-VRC01 antibodies, which was confirmed via ADA assay in a highly sensitive MSD platform (S4 Table).

Immunologic activity of VRC01 post infusion

To test whether VRC01 retained expected neutralizing activity post infusion, serum samples from the same time points used for the ELISA-based PK analysis were assayed against 3 Env-pseudotyped viruses (tier 1A: MN.3, MW965.26; tier 2: PVO.4) in the TZM-bl neutralization assay. As shown in Fig 3, serum concentrations of VRC01 measured by neutralizing activity closely approximated the concentrations measured by ELISA, with even the lowest IV dose (10 mg/kg) demonstrating expected magnitude and breadth of neutralizing activity, regardless of whether neutralization was measured with the tier 1 viruses or the tier 2 virus. Similar results have been reported previously over a shorter duration and lower frequency of VRC01 administration [33].

To examine post-infusion neutralizing activity in greater detail, serum samples from 6 T5 (30 mg/kg) participants obtained 4 weeks after each of the 3 IV infusions every 8 weeks and 1 hour and 8 weeks after the third infusion were assessed for neutralizing activity against a multiclade panel of 11 additional tier 2-circulating strains of HIV-1 that represent globally circulating strains and exhibit a range of known sensitivities to VRC01. As shown in Fig 4A, very potent broadly neutralizing activity was seen at 1 hour post infusion in all 6 participants that subsequently declined in magnitude and breadth by 4 weeks post infusion. Notably, equivalent magnitude and breadth of neutralizing activity was seen 1 month post each infusion in the 4 participants who received all 3 infusions and at 4 weeks post first and third infusion in the 2 participants (104–27 and 104–88) who missed the second infusion. Results for these latter 2 participants also illustrate that tier 2 virus-neutralizing activity was detectable against >80% of the viruses 12 weeks after a single infusion. As expected, the level of activity present at 8 weeks post third infusion in all 6 participants was intermediate between the levels seen after 4 and 12 weeks. Fig 4B shows that the geometric mean serum neutralization ID50 titer against tier 2 viruses was approximately 1:100 at 4 weeks post each infusion and that this titer was approximately 1:30 at the trough time point (8 weeks post infusion, measured post final infusion only). S5 Table contains the ID50 titer data for each participant. Fig 4C illustrates the consistency of measuring VRC01 concentration in these 6 T5 participants via ELISA or via TZM-bl ID50 neutralization assay estimation using either tier 1A or a global panel of tier 2 HIV-1 Env-pseudoviruses. Overall, these results indicate no loss of expected VRC01 neutralizing activity after multiple IV infusions over a period of at least 24 weeks. Table 3 shows that midpoint and trough serum VRC01 levels achieved in this study following each group’s final dose are consistent with concentrations that are known to neutralize a majority of circulating strains of clade B and clade C HIV-1 in vitro.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Serum neutralizing activity against multiple isolates at 4 weeks after the first, second, and third infusions and at 1 hour and 8 weeks following the third infusion in 6 T5 participants.

(A) MB curves based on TZM-bl neutralization assay results against 11 tier 2 viruses for each of the 6 individuals in T5. Two individuals who missed their second infusion are indicated by an * at the red line (4 weeks post infusion 2). MB curves based on results with HIV-1 Env-pseudotyped viruses PVO.4, 398F1, CNE8, X2278, 246-F3, TRO.11, CNE55, CH119, 25710, X1632, and Ce0217. Response rate on the y-axis is the percent of the 11 viruses neutralized at serum dilutions (Log10 ID50 titer) shown on the x-axis. (B) Serum ID50 neutralization titers against the 11 tier 2 and 2 tier 1 isolates. Shown are the geometric mean and standard error of the mean over the ID50 titers of 4 T5 individuals, excluding the 2 individuals who missed their second infusion. (C) VRC01 serum concentration measured by ELISA and TZM-bl against tier 1 and tier 2 isolates. Shown are the geometric mean and standard error of the mean over the estimated concentration of VRC01 in 4 T5 individuals, excluding the 2 individuals who missed their second infusion. Individual level neutralization data for participants can be found in S5 Table. Env, HIV-1 envelope glycoprotein; ID50, 50% infectious dose; MB, magnitude-breadth; T5, treatment group 5.

https://doi.org/10.1371/journal.pmed.1002435.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Predicted coverage based on in vitro neutralizing activity of VRC01.

https://doi.org/10.1371/journal.pmed.1002435.t003

Non-neutralizing antibody effector functions were assessed in serum from study participants. ADCC, ADCP, and virion capture activities were assessed at 3 days post second administration. ADCC was assessed and found to be weak or absent (Fig 5A), which agrees with the low ADCC activity exhibited by non-infused VRC01 in this assay (S2 Fig). Despite this low ADCC effector function, 100% of serum samples from all 5 treatment groups exhibited ADCP activity against glycoprotein 140 (gp140)-coated microspheres (Fig 5B) and demonstrated capacity of the fragment crystallizable (Fc) portion of this mAb to engage monocyte effector cells. All specimens from PP cohort participants also demonstrated capacity to bind infectious virions (Fig 5C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. VRC01 mAb retains nonneutralizing Fc effector functions postintravenous or postsubcutaneous administration in the PP cohort.

(A) ADCC using the Luciferase-HIV CH0505.LucR T2A.ecto/293T/17 assay, (B) average phagocytosis score, and (C) average virus capture percentage are presented (average of 2 replicate experiments). Grey squares indicate baseline (preadministration) time points; red circles represent positive responses at postadministration time points; open blue triangles represent negative responses at postadministration time points. Horizontal dashed lines represent the positivity cutoff based on the 95th percentile of all baseline visits. Percent responders and the number of positive responders/total number are shown above each treatment group. T1: 20 mg/kg IV q 4 weeks with 40 mg/kg IV loading; T2: 40 mg/kg IV q 8 weeks; T3 (P3): 5 mg/kg SC q 2 weeks with 40 mg/kg IV loading; T4: 10 mg/kg IV q 8 weeks; T5: 30 mg/kg IV q 8 weeks. ADCC, antibody-dependent cellular cytotoxicity; Fc, fragment crystallizable; mAb, monoclonal antibody; PP, per-protocol; P3, placebo group 3; q, quodque; SC, subcutaneous; T1, treatment group 1; T2, treatment group 2; T3, treatment group 3; T4, treatment group 4; T5, treatment group 5.

https://doi.org/10.1371/journal.pmed.1002435.g005

Lastly, the correlations between the functional activities and VRC01 serum concentration measured at 3 days post the second infusion/injection in the PP cohort (Fig 6) were assessed. The strongest and most significant correlations were observed between ADCP and virion capture activities (Spearman ρ = 0.78) and between both ADCP and virus capture and VRC01 serum concentration (ρ = 0.72 and 0.76, respectively). Statistically significant but weaker correlations were observed between ADCC activity and ADCP activities (ρ = 0.29), ADCC and virion capture (ρ = 0.38), and ADCC and VRC01 serum concentration (ρ = 0.36).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Correlations between functional activities and VRC01 level in the PP cohort (all treatment arms).

Lower diagonal squares show scatterplots of each pair of assay variables. A Lowess smoother line using locally weighted polynomial regression with a span of 2/3 was added. Upper diagonal squares show rank-based Spearman correlation coefficients for each pair of assay variables. Diagonal squares show histograms of each variable. ADCC, antibody-dependent cellular cytotoxicity; PP, per-protocol.

https://doi.org/10.1371/journal.pmed.1002435.g006

Disclaimer

The opinions expressed in this article are those of the authors and do not necessarily represent the official views of the NIAID or the National Institutes of Health (NIH).

HVTN 104 protocol team

Kenneth Mayer and Charlie Gregor, The Fenway Institute, Boston, MA, USA; Kelly Seaton, Duke Human Vaccine Institute, Durham, NC, USA; Nicole Grunenberg, Yunda Huang, Eva Chung, Janet Quade, Marianne Hansen, Lisa Sunner, Gina Escamilla, Ryan Jensen, Gail Broder, Adi Ferrara, Jill Zeller, Abby Isaacs, Doug Grove, Rachael McClennen, On Ho, and Shelly Ramirez, Vaccine and Infectious Disease Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA; Chuka Anude, Scharla Estep, Mary Allen, and Chris Butler, Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institute of Health, Bethesda, MD, USA; Barney Graham, Julie Ledgerwood, and Mary Enama, Vaccine Research Center, National Institute of Allergy and Infectious Diseases, National Institute of Health, Bethesda, MD, USA; David Crawford, HVTN Community Advisory Board, Chicago, IL, USA; Steve Csipke, HVTN Community Advisory Board, Boston, MA, USA; Samuel Cohen, Brigham and Women’s Hospital, Boston, MA, USA; Debora Dunbar, University of Pennsylvania, Philadelphia, PA, USA.