Safety and immunogenicity of a parenteral P2-VP8-P[8] subunit rotavirus vaccine in toddlers and infants in South Africa: a randomised, double-blind, placebo-controlled trial

Background Efficacy of live oral rotavirus vaccines is reduced in low-income compared with high-income settings. Parenteral non-replicating rotavirus vaccines might offer benefits over oral vaccines. We assessed the safety and immunogenicity of the P2-VP8-P[8] subunit rotavirus vaccine at different doses in South African toddlers and infants. Methods This double-blind, randomised, placebo-controlled, dose-escalation trial was done at a single research unit based at a hospital in South Africa in healthy HIV-uninfected toddlers (aged 2 to <3 years) and term infants (aged 6 to <8 weeks, without previous rotavirus vaccination). Block randomisation (computer-generated, electronic allocation) was used to assign eligible toddlers (in a 6:1 ratio) and infants (in a 3:1 ratio) in each dose cohort (10 μg, followed by 30 μg, then 60 μg if doses tolerated) to parenteral P2-VP8-P[8] subunit rotavirus or placebo injection. The two highest tolerated doses were then assessed in an expanded cohort (in a 1:1:1 ratio). Parents of participants and clinical, data, and laboratory staff were masked to treatment assignment. P2-VP8-P[8] vaccine versus placebo was assessed first in toddlers (single injection) and then in infants (three injections 4 weeks apart). The primary safety endpoints were local and systemic reactions within 7 days after each injection, adverse events within 28 days after each injection, and all serious adverse events, assessed in toddlers and infants who received at least one dose. In infants receiving all study injections, primary immunogenicity endpoints were anti-P2-VP8-P[8] IgA and IgG and neutralising antibody seroresponses and geometric mean titres 4 weeks after the third injection. This trial is registered at ClinicalTrials.gov, number NCT02109484. Findings Between March 17, 2014, and Sept 29, 2014, 42 toddlers (36 to vaccine and six to placebo) and 48 infants (36 to vaccine and 12 to placebo) were enrolled in the dose-escalation phase, in which the 30 μg and 60 μg doses where found to be the highest tolerated doses. A further 114 infants were enrolled in the expanded cohort between Nov 3, 2014, and March 20, 2015, and all 162 infants (12 assigned to 10 μg, 50 to 30 μg, 50 to 60 μg, and 50 to placebo) were included in the safety analysis. Serum IgA seroresponses were observed in 38 (81%, 95% CI 67–91) of 47 infants in the 30 μg group and 32 (68%, 53–81) of 47 in the 60 μg group, compared with nine (20%, 10–35) of 45 in the placebo group; adjusted IgG seroresponses were seen in 46 (98%, 89–100) of 47 infants in the 30 μg group and 47 (100%; 92–100) of 47 in the 60 μg group, compared with four (9%, 2·5–21) of 45 in the placebo group; and adjusted neutralising antibody seroresponses against the homologous Wa-strain were seen in 40 (85%, 72–94) of 47 infants in both the 30 μg and 60 μg groups, compared with three (7%, 1·4–18) of 45 participants in the placebo group. Solicited reactions following any injection occurred with similar frequency and severity in participants receiving vaccine and those receiving placebo. Unsolicited adverse events were mostly mild and occurred at a similar frequency between groups. Eight serious adverse events (one with placebo, two with 30 μg, and five with 60 μg) occurred in seven infants within 28 days of any study injection, none of which were deemed related to study treatment. Interpretation The parenteral P2-VP8-P[8] vaccine was well tolerated and immunogenic in infants, providing a novel approach to vaccination against rotavirus disease. On the basis of these results, a phase 1/2 trial of a trivalent P2-VP8 (P[4], P[6], and P[8]) subunit vaccine is underway at three sites in South Africa. Funding Bill & Melinda Gates Foundation. This single-centre, double-blind, randomised, placebo-controlled, dose-escalation trial assessed participants at the Respiratory and Meningeal Pathogens Research Unit based at the Chris Hani Baragwanath Academic Hospital (Johannesburg, South Africa), which serves the urban population of Soweto. The protocol was approved by the Human Research Ethics Committee, University of the Witwatersrand (Johannesburg, South Africa), the Western Institutional Review Board (Puyallup, WA, USA), and the Medicines Control Council (Pretoria, South Africa). We also submitted a US Food and Drug Administration Investigational New Drug Application, which was approved. The study was undertaken in accordance with South African Good Clinical Practice Guidelines.18 The dose-escalation phase was designed to test three dose levels (10 μg, 30 μg, and 60 μg) of vaccine, first in toddlers and then in infants. Cohorts of 14 toddlers (12 vaccine, two placebo) per dose level were to receive a single injection. Cohorts of 16 infants (12 vaccine, four placebo) per dose level were to receive three injections at 4-week intervals. The first participants were toddlers at 10 μg dose, followed by toddlers at 30 μg and infants at 10 μg dose, then toddlers at 60 μg and infants at 30 μg dose, and then infants at 60 μg dose. Progression from one dose to the next and from toddlers to infants required review by a safety review committee (SRC) of safety data for 7 days after the first injection in the respective dose or age cohort. The two highest tolerated dose levels were then assessed in an expanded cohort of 114 infants (38 at each dose level and 38 placebo). Toddlers and infants were identified from hospital birth registers and postnatal wards, and invited for screening 1–7 days before randomisation. Healthy HIV-uninfected toddlers (aged ≥2 and <3 years) and infants (aged ≥6 and <8 weeks, ≥37 weeks gestation, and without previous receipt of rotavirus vaccination) were eligible for enrolment. Eligibility criteria were assessed through medical history, physical examination, and screening laboratory tests. Exclusion criteria included acute illness at time of enrolment, presence of malnutrition or any systemic disorder, congenital defects, known or suspected impaired immunological function, immunoglobulin therapy or chronic immunosuppressant medications, and concurrent participation in another clinical trial. Full inclusion and exclusion criteria are listed in the appendix p 1. Participants were only enrolled if their parents were literate and provided written informed consent, and intended to stay in the area with the child during the study. Toddlers were randomly assigned to receive vaccine or placebo in groups of 14, beginning with 10 μg, then followed by 30 μg, then 60 μg. Within each group, randomisation was done in two blocks of seven toddlers (six were randomised to vaccine and one to placebo), in a random order within the block. Infants were randomly assigned in groups of 16 beginning with 10 μg, then followed by 30 μg, then 60 μg. Within each group, randomisation was done in four blocks of four infants (three to vaccine and one to placebo), in a random order within the block. Infants in the expanded cohort were allocated either to one of the two highest doses tolerated in the dose escalation phase, or to placebo, in blocks of three or six infants, with block size chosen at random and infants randomly ordered within the block in a 1:1:1 ratio. The randomisation sequence was computer generated by the Emmes Corporation (Rockville, MD, USA). A study investigator enrolled and randomly assigned participants electronically, and was provided with a blinded treatment number. This number was given to the unmasked pharmacist who prepared and dispensed the injection on the basis of the treatment number, in a masked syringe (taped to conceal the colour of the liquid), which was then administered intramuscularly into the thigh by a masked study investigator. Parents of participants and clinical, data, and laboratory staff were all masked to treatment assignment. Toddlers in the dose-escalation phase of the trial received a single intramuscular injection of vaccine or placebo in the anterolateral thigh on the day of randomisation (day 0). Infants in both the dose-escalation phase and the expanded cohort received three intramuscular injections of vaccine or placebo in the anterolateral thigh on the day of randomisation (day 0) at age 6–7 weeks, day 28 at age 10–13 weeks, and day 56 at age 14–17 weeks (appendix p 6). The P2-VP8-P[8] protein was produced at the Walter Reed Army Institute of Research, Pilot Bioproduction Facility (Silver Spring, MD, USA).17 The protein was diluted using sterile saline and formulated with aluminium hydroxide (Alhydrogel, Brenntag Biosector, Frederikssund, Denmark) by the study pharmacist within 6 h of administration to yield dose concentrations of 10 μg, 30 μg, and 60 μg per 0·5 mL containing 0·56 mg of aluminium. An injection of sterile saline was used as placebo. Vaccine or placebo in infants was given in the same thigh as hepatitis B vaccine (Heberbiovac-HB; The Biovac Institute, Cape Town, South Africa), which is given routinely to infants at ages 6, 10, and 14 weeks, whereas 13-valent pneumococcal conjugate vaccine (Prevenar13; Pfizer Laboratories, New York, NY, USA), routinely given at ages 6 and 14 weeks only, and the combination vaccine for diphtheria, tetanus, pertussis, poliomyelitis, and Haemophilus influenzae type b (Pentaxim; Sanofi Pasteur, Paris, France), routinely given at ages 6, 10, and 14 weeks, were given in the opposite thigh. Infants also received three doses of the oral Rotarix rotavirus vaccine (GlaxoSmithKline, Rixensart, Belgium) as part of this study, at 4, 8, and 12 weeks after the third study injection. Participants were observed for 30 min after administration of each injection. Parents were provided with, and trained to correctly use, a thermometer, a tool to assess the size of injection site redness and swelling, and a memory aid booklet to assess and record local (injection site pain or tenderness, redness, swelling, and itching) and systemic (fever, vomiting, decreased appetite, irritability, and decreased activity) symptoms daily for 7 days after each injection. Clinic visits were done 3 and 7 days after each injection. Haemoglobin, white blood cell count with differential, platelet count, total bilirubin, albumin, creatinine, and alanine transaminase were assessed at baseline and 7 days after the first injection. Unsolicited adverse events were recorded from randomisation until the final study visit, 6 months after the last injection. Parents were advised to contact study staff if the child developed an adverse event during the course of the study. Grading and causality of adverse events were determined by the investigator using a grading scale developed for this study, and all safety data, including all adverse events, were reviewed by the SRC throughout the study. Study stopping rules are described in the appendix p 1. Serum was collected at baseline, and at 4 weeks (all participants) and 24 weeks (infants only) after the final study injection (appendix p 6). Anti-P2-VP[8] IgG and IgA were quantified using standard ELISA assay techniques.17 Neutralising antibodies to Wa (G1P[8]), 89-12 (G1P[8]), DS-1 (G2P[4]), and 1076 (G2P[6]) were established as previously described.19 Rotavirus serum IgA was measured by an ELISA assay using whole virus lysate.20 Testing was done at the Division of Infectious Diseases, Cincinnati Children’s Hospital Medical Center (Cincinnati, OH, USA). Stool samples were collected from infants at 5, 7, and 9 days after the first dose of Rotarix and tested for the presence of rotavirus using the commercially available ProsPecT Rotavirus Microplate Assay (Oxoid Ltd, Ely, UK), according to the manufacturer’s instructions. ELISA-positive specimens were confirmed and genotyped by PCR amplification of the genes encoding VP7 and VP4 (appendix p 2). Testing was done at the National Institute for Communicable Diseases (Johannesburg, South Africa). The primary objectives were to assess the safety and reactogenicity of the P2-VP8-P[8] vaccine at escalating doses in toddlers and infants, and to investigate the immunogenicity at different doses in infants. The primary safety endpoints were the number of serious adverse events, the number of adverse events within 28 days after each injection, and the number of vaccine-induced local and systemic reactions within 7 days after each injection and overall for the three combined injections. The primary immunogenicity endpoints were the proportion of infants with anti-P2-VP8 IgG and IgA seroresponses, the proportion of infants with neutralising antibody responses against rotavirus, anti-VP8 IgG and IgA geometric mean titres (GMTs) 4 weeks after the third injection in infants, and neutralising antibody GMTs 4 weeks after the third injection in infants. An unadjusted seroresponse was defined as an increase of four times or more in titre between baseline and 4 weeks after the third injection. Adjusted IgG and neutralising antibody post-injection titres accounted for the decay in maternal antibodies using the half-life calculated from participants in the placebo group who had detectable baseline titres that were higher than at the post-injection visit. This adjustment value was established for each assay separately. An adjusted seroresponse was defined as an increase of four times or more in titre between baseline and 4 weeks after the third injection (adjusted titre) in infants with an unadjusted post-injection titre greater than the limit of detection (the latter part of this definition was not included in the original statistical analysis plan but was added at the analysis stage). Neutralising antibody seroresponses were those against the strain from which the vaccine was based (homologous Wa strain) as well as against divergent rotavirus strains 89-12, DS-1, and 1079. The secondary objective was to assess the effect of P2-VP8-P[8] vaccination on shedding of Rotarix vaccine virus subsequently administered in infants, with the endpoint being the proportion of infants shedding rotavirus (determined by ELISA) at 5, 7, or 9 days after administration of the first dose of Rotarix (4 weeks after the third P2-VP8-P[8] or placebo injection). Exploratory objectives were to assess the immunogenicity of one dose of P2-VP8-P[8] vaccine at different doses in toddlers, and to characterise the serum IgA response in infants receiving Rotarix after receiving the P2-VP8-P[8] vaccine. Endpoints were the proportion of toddlers with anti-P2-VP8 IgG and IgA seroresponses; the proportion of toddlers with neutralising antibody responses against rotavirus; anti-VP8 IgG and IgA GMTs and neutralising antibody GMTs 4 weeks after the single injection in toddlers; and anti-rotavirus IgA, anti-P2-VP8 IgG and IgA GMTs, and neutralising antibody GMTs in infants before and after Rotarix vaccination. The effect of the vaccine on the shedding of Rotarix vaccine virus was a secondary objective, but was the primary factor used to calculate the sample size for the dose escalation and expanded cohorts of infants. Establishing sample size in this manner means that the study has the statistical power to address both primary and secondary objectives, because the sample size needed for the secondary objective exceeded that required for the primary objectives. We calculated that 50 infants per dose group, allowing for 10% of participants to drop out, would enable detection of an 80% reduction in Rotarix virus shedding in recipients of the vaccine compared with placebo recipients (>80% power assuming ≥30% shedding in placebo recipients). This sample size provided a 90% or greater chance of observing an adverse event that had a 4·5% risk of occurrence, and more than 95% power to detect a 40 percentage point or greater difference in seroresponses between a vaccine group and the placebo group. Sample size calculations were based on Fisher’s exact test using a two-sided α of 0·05. Toddlers and infants who received at least one dose of vaccine or placebo were assessed in the safety analysis. The unit of analysis was the proportion of participants with at least one event graded as moderate or worse (including solicited local reactions, systemic reactions, adverse events, and suspected adverse reactions), or any serious adverse events. The primary immunogenicity analysis included only infants who received all three study injections (per-protocol population). Logistic regression was used to detect differences in seroresponses between the treatment groups, including pair-wise comparisons if the overall difference was statistically significant. The Kruskal-Wallis test was used to compare magnitude of response between the treatment groups and Wilcoxon rank-sum test was used for pairwise comparisons. Assessment of shedding was done for each of the three specified post-Rotarix vaccination days, and for shedding on any of the 3 days combined. Infants for whom rotavirus was detected in any specimen by ELISA testing were considered to have undergone viral shedding, and the proportions of infants with shedding in the 30 μg and 60 μg dose groups and the combined 30 μg and 60 μg dose group were compared with the placebo group using logistic regression. The relative reduction in the proportion of participants with shedding of the Rotarix vaccine strain compared with the placebo group was analysed using the two-sample t distribution on log-transformed data. Data were analysed using SAS software (version 9.3) and statistical significance defined as a two-tailed p<0·05. The trial was registered on ClinicalTrials.gov ({"type":"clinical-trial","attrs":{"text":"NCT02109484","term_id":"NCT02109484"}}NCT02109484). The funder of the study had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.