An Omicron-specific, self-amplifying mRNA booster vaccine for COVID-19: a phase 2/3 randomized trial

As of 3 March 2024, 774,834,251 confirmed cases of COVID-19 and 7,037,007 deaths had been reported worldwide. Vaccines have mitigated the pandemic, and mRNA vaccines in particular offer advantages: they are nonintegrating, synthetically produced in a cell-free, scalable, and rapid manner, making them ideal for variant updates. However, cold-chain requirements limit access in low- and middle-income countries. The Omicron BA.1 (B.1.1.529) variant rapidly became dominant and evaded immunity induced by vaccines targeting ancestral strains, creating an unmet need for variant-adapted boosters. GEMCOVAC-OM was developed as a monovalent, Omicron-specific (BA.1) self-amplifying mRNA (samRNA) booster that differs from other mRNA vaccines by encoding a replicase for intracellular RNA amplification (enabling lower doses), being lyophilized and thermostable at 2–8 °C for 12 months, and being delivered intradermally via a needle-free Tropis device. The study aimed to assess the safety and immunogenicity of GEMCOVAC-OM as a heterologous booster in adults previously vaccinated with BBV152 or ChAdOx1 nCoV-19.

The paper situates the work within prior evidence: (1) Omicron variants show substantial escape from vaccine-induced neutralizing antibodies; (2) initial Omicron-adapted boosters by Moderna and Pfizer were bivalent, but immune imprinting limited additional benefit from ancestral antigen, leading WHO TAG-CO-VAC to recommend monovalent variant-specific boosters; (3) neutralizing antibodies measured by live-virus assays correlate with protection and serve as immunobridging endpoints; (4) self-amplifying mRNA (samRNA) platforms are emerging, with ARCT-154 obtaining approval in Japan and showing noninferiority/superiority to BNT162b2 as a booster; (5) T cell cross-reactivity to Omicron is generally preserved, and CD8+ T cell responses contribute to protection and durability; (6) real-world data indicate higher reactogenicity among females after mRNA COVID-19 vaccines.

Design: Prospective, multicenter, open-label, randomized phase 2 study seamlessly followed by phase 3 across 20 hospitals in 13 Indian cities, conducted under ICH-GCP and the Declaration of Helsinki. Trial registration: CTRI/2022/10/046475. An interim phase 3 analysis at day 29 supported Emergency Use Authorization in India (June 19, 2023). Participants: Healthy adults ≥18 years, having received two doses of BBV152 or ChAdOx1 nCoV-19 ≥4 months before screening, and no known COVID-19 infection within 3 months. Key exclusions: pregnancy/lactation, immunocompromised states, or investigator-determined unsafe conditions. Screening included medical history, exam, vitals, and SARS-CoV-2 RT-PCR on day 1 (RT-PCR–positive individuals excluded from immunogenicity analyses). Randomization and masking: Phase 2 randomized 1:1 to GEMCOVAC-OM vs GEMCOVAC-19. Phase 3 included a safety cohort (n=3,140; 3,000 GEMCOVAC-OM; 140 ChAdOx1 nCoV-19) and within it an immunogenicity cohort (n=420) randomized 2:1 to GEMCOVAC-OM (n=280) vs ChAdOx1 nCoV-19 (n=140) using IWRS and stratified block randomization. Open-label; laboratory assays were blinded. Interventions: GEMCOVAC-OM (10 µg in 0.1 ml) intradermally via Tropis needle-free injector; GEMCOVAC-19 (10 µg in 0.5 ml) intramuscularly; ChAdOx1 nCoV-19 (COVISHIELD; 5×10^10 viral particles) intramuscularly. Both GEMCOVAC vaccines used in vitro transcribed mRNA in a cationic lipid nano-emulsion (buffer: 10% sucrose in 10 mM sodium citrate, pH 6.5). Antigen sequences: GEMCOVAC-OM (DDBJ LC769018), GEMCOVAC-19 (DDBJ LC776732.1). Visit schedule: Day 1 vaccination; follow-up visits at day 29 (±7), day 90 (±14), and day 180 (±14). Safety monitoring: Solicited AEs (local: pain, redness, swelling, warmth, pruritus, bruising; systemic: fever, headache, myalgia, arthralgia, fatigue, malaise, nausea, chills) captured via diary for 7 days; unsolicited AEs through day 180; myocarditis was an AE of special interest. AE grading per DAIDS criteria. Immunogenicity assessments: - Live-virus neutralization PRNT50 against BA.1 at IRSHA using Vero E6 cells; titers determined via logistic regression. - Surrogate neutralization (cPass, GenScript) against BA.1. - Anti-spike (BA.1) IgG by in-house indirect ELISA (full-length spike-coated plates; titers interpolated from 5-parameter logistic fits). - Cellular immunity via intracellular cytokine staining on PBMCs stimulated with BA.1 spike peptide pool (PepTivator); markers included IFNγ, TNF, and IL-2 in CD4+ and CD8+ T cells; Omicron spike-specific B cells quantified using biotinylated BA.1 spike and B-cell markers (CD19, CD20). Cellular analyses in subsets: phase 2 (~20%; 14/arm); phase 3 (~25%; 71 GEMCOVAC-OM, 35 ChAdOx1). Endpoints: Phase 2 primary—safety and anti-spike IgG at day 29; secondary—seroconversion (≥2-fold rise), cPass, cellular immunity at day 29; exploratory—humoral and cellular responses at day 90. Phase 3 primary—noninferiority at day 29 of PRNT50 LSGMR (GEMCOVAC-OM/ChAdOx1) with NI margin >0.67, and difference in seroconversion (≥2-fold rise) with NI margin >−10%; superiority assessed if lower bound 95% CI for LSGMR >1 and seroconversion difference >0%. Secondary—anti-spike IgG LSGMR and seroconversion, cPass, cellular immunity at day 29, and safety to day 180; exploratory—day 90 humoral and cellular responses. Statistics: Full analysis (intention-to-treat) for immunogenicity. GMTs with 95% CIs on log-transformed data; within-arm changes by paired tests; LSGMR by ANCOVA with baseline titers as covariates; seroconversion CIs by Clopper–Pearson; between-arm seroconversion differences by Miettinen–Nurminen; cPass change by ANCOVA; cellular changes by paired t-test or Wilcoxon tests; between-group cellular comparisons by t-test or Wilcoxon rank-sum; normality by Shapiro–Wilk. Sex-disaggregated analyses used ANCOVA (immunogenicity) and unadjusted odds ratios (safety).

Enrollment and disposition: Phase 2 randomized 140 (70 per arm), all completed to day 180. Phase 3 enrolled 3,140 (3,000 GEMCOVAC-OM; 140 ChAdOx1); 14 participants excluded post hoc due to protocol deviations (prior booster). Day 29 immunogenicity cohort: 404 participants (271 GEMCOVAC-OM; 133 ChAdOx1). Phase 2 (GEMCOVAC-OM vs GEMCOVAC-19): - Anti-spike IgG GMT (BA.1): Baseline 30,048 (95% CI 23,910–37,763) → day 29 244,440 (229,122–260,782), P<0.0001 for GEMCOVAC-OM; Baseline 35,676 (29,401–43,289) → day 29 75,683 (61,687–92,853), P<0.0001 for GEMCOVAC-19. GMFR: 8.13 vs 2.12. LSGMR (OM/19): 3.40 (95% CI 2.79–4.13), P<0.0001. Seroconversion (≥2-fold): 92.9% vs 57.1%; difference 35.71% (95% CI 22.35–48.52), P<0.0001. Day 90 IgG GMTs remained higher than baseline and favored OM (LSGMR 3.31, 95% CI 2.72–4.02, P<0.0001). cPass mean percent neutralization increased more with OM to day 29 (change 19.3 vs 8.4; P<0.0001); day 90 means: 96.3% (s.d. 8.35) vs 90.0% (21.11). Cellular immunity: Significant increases from baseline to day 29 in IFNγ+ and IL-2+ CD4+ T cells and IFNγ+, IL-2+ CD8+ T cells in both arms; at day 90, OM showed higher TNF+CD4+ and IL-2+CD8+ than GEMCOVAC-19. Th2 cytokines (IL-4, IL-13) decreased from baseline at days 29 and 90. Phase 3 (GEMCOVAC-OM vs ChAdOx1): - PRNT50 (BA.1) GMTs: GEMCOVAC-OM baseline 623.9 (533.3–729.9) → day 29 1,099.9 (1,000.0–1,209.9), P<0.0001; ChAdOx1 baseline 775.3 (620.2–969.2) → day 29 754.9 (631.5–902.5), P=0.1490. GMFR: 1.76 vs 0.97. LSGMR at day 29: 1.58 (95% CI 1.36–1.84), P<0.0001—met noninferiority and superiority; seroconversion: 39.5% (33.6–45.5) vs 19.5% (13.1–27.3); difference 19.93% (10.5–28.4). Day 90 PRNT50 GMTs: 754.0 (682.1–833.6) vs 383.1 (319.4–459.6); LSGMR 2.09 (1.75–2.49), P<0.0001. - Anti-spike IgG (BA.1): Day 29 GMFR 7.25 (OM) vs 3.29 (ChAdOx1); LSGMR 2.15 (95% CI 1.83–2.52), P<0.0001; seroconversion 93.0% vs 76.7%; difference 16.30% (95% CI 9.02–24.64), P<0.0001. Day 90 LSGMR 2.23 (95% CI 1.87–2.66), P<0.0001. - cPass: Day 29 mean neutralization ~94% in both arms from ~68% at baseline (no between-group difference, P=0.8559). Day 90 mean neutralization higher with OM (91.7% vs 81.3%). - Cellular immunity (subset): At day 29, OM induced higher TNF+CD4+, IL-2+CD4+, and IL-2+CD8+ T cells than ChAdOx1 (all P<0.0001); both arms showed increases in certain IFNγ+ subsets. BA.1-specific B cells increased at day 29 with OM (P=0.0013) and were higher than ChAdOx1 (P<0.0001). Correlations: With OM, binding IgG correlated positively with neutralizing titers; strong positive correlations observed among IFNγ+ CD4+/CD8+ and IL-2+ subsets. Safety: Phase 2 reported 14 AEs total (5 OM; 9 GEMCOVAC-19), no serious or unsolicited AEs. Phase 3: Any AE—20.1% OM (602/2,990) vs 24.8% ChAdOx1 (33/133); mostly mild. Unsolicited AEs: 1.40% OM vs 2.26% ChAdOx1 to day 180. Serious AEs: OM (omphalitis, spontaneous abortion, pulmonary tuberculosis), ChAdOx1 (ligament rupture); all deemed unlikely related. No deaths. Sex-disaggregated analyses showed more local and systemic AEs reported by women in the OM arm; neutralizing titers were higher in women at day 90 in OM.

Administered as a heterologous booster, the Omicron BA.1-specific samRNA vaccine GEMCOVAC-OM demonstrated robust humoral and cellular immunogenicity with a favorable safety profile. In phase 3, GEMCOVAC-OM met noninferiority criteria versus ChAdOx1 nCoV-19 for PRNT50 at day 29 and achieved superiority on both neutralizing titers and seroconversion; these advantages persisted to day 90. Anti-spike IgG responses were significantly higher with OM at days 29 and 90. The lack of between-arm differences in cPass at day 29 reflects assay ceiling effects (>90% neutralization and 100% cap), while live-virus PRNT50—considered the gold standard—captured meaningful differences and supports immunobridging. Cellular analyses indicated a TH1-skewed response, with OM eliciting higher TNF+ and IL-2+ CD4+ T cells and IL-2+ CD8+ T cells than ChAdOx1, and increased BA.1-specific B cells at day 29, aligning with durable cellular immunity. Safety was acceptable, with mostly mild AEs, no vaccine-related serious AEs or deaths, and sex-related patterns consistent with higher reactogenicity reported in females for mRNA vaccines. These findings are consistent with performance of other BA.1-adapted mRNA boosters, which showed titer ratios around 1.5–1.8 versus prototypes, and underscore the value of variant-updated, monovalent strategies to mitigate immune imprinting. GEMCOVAC-OM’s lyophilized, thermostable, intradermal, and needle-free delivery features address access barriers in LMICs.

GEMCOVAC-OM, a thermostable, intradermally delivered self-amplifying mRNA vaccine, was safe, well tolerated, and induced superior humoral and cellular immune responses against Omicron BA.1 compared with ChAdOx1 nCoV-19 when used as a heterologous booster. Anti-spike IgG and live-virus neutralization were significantly enhanced, with supportive T cell and B cell responses and acceptable reactogenicity. The platform’s dose-sparing samRNA design, lyophilization, and needle-free delivery may improve global vaccine access and equity. Future research should include head-to-head comparisons with other Omicron-specific mRNA vaccines, evaluation against contemporaneous Omicron sublineages, durability beyond 3 months, and efficacy/effectiveness endpoints.

- Sex imbalance: Participants were predominantly male, so findings may generalize more to men; sex-disaggregated analyses were underpowered. - Comparator: Lack of an Omicron-specific mRNA comparator in India necessitated use of ChAdOx1 nCoV-19 (Wuhan-based), limiting direct comparisons to other Omicron mRNA vaccines. - Open-label design: Different delivery routes precluded blinding, potentially biasing safety reporting. - Prior infection uncertainty: Multiple waves and asymptomatic infections impeded accurate ascertainment of prior COVID-19, complicating interpretation of boosting effects. - No placebo/efficacy arm: Given booster approvals, placebo was unethical; an immunobridging approach was used instead of clinical efficacy assessment.