Immunogenicity and safety of a recombinant adenovirus type-5-vectored COVID-19 vaccine in healthy adults aged 18 years or older: a randomised, double-blind, placebo-controlled, phase 2 trial

Introduction

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused more than 12·1 million cases of COVID-19 worldwide, resulting in 551 000 deaths and severe economic disruption.1, 2 After the initial outbreak, with more than 80 000 cases and 3000 deaths in China, COVID-19 has now spread to 216 countries and territories. Large numbers of cases and deaths are reported daily from Europe, the USA, Brazil, Russia, India, and many other countries.3, 4 The current pandemic has highlighted the need for effective preventive solutions to reduce burden and spread of the disease. As long as there is a COVID-19 epidemic in one area in the world, there is a risk of a pandemic.

Unlike typical vaccine development, which often takes decades, developing a vaccine to prevent COVID-19 has become a race between humans and the virus.⁵ Many countries have accelerated the process of clinical trials to determine an effective and safe vaccine to prevent COVID-19 and influence the course of the current pandemic.6, 7 Currently, about 250 candidate vaccines against SARS-CoV-2 are in development worldwide, including mRNA vaccines, replicating or non-replicating viral vectored vaccines, DNA vaccines, autologous dendritic cell-based vaccine, and inactive virus vaccines.⁸ To date, at least 17 of these vaccine candidates are under evaluation in clinical trials.

In March 2020, we did a single-centre, open-label, non-randomised, first-in-human phase 1 trial⁹ with CanSino Biologics' (Tianjin, China) adenovirus type-5 (Ad5)-vectored COVID-19 vaccine in a dose-escalating manner (5 × 10¹⁰, 1 × 10¹¹, and 1·5 × 10¹¹ viral particles). Generally, the candidate vaccines had acceptable safety and tolerability profiles and promising immunogenicity in healthy Chinese adults; however, the high-dose vaccine was associated with an increased risk of severe adverse reactions. We therefore carried forward the phase 2 trial with only the 5 × 10¹⁰ and 1 × 10¹¹ viral particles doses of the candidate vaccine, aiming to further evaluate the immunogenicity and safety in a larger population, and to determine an appropriate dose for the efficacy study.

Methods

Results

603 volunteers were recruited and screened for eligibility between April 11 and 16, 2020 (figure 1 ). 95 individuals were excluded, leaving 508 eligible participants who consented to participate in the trial and were randomly assigned to vaccine or placebo. 253 were randomly assigned to the 1 × 10¹¹ viral particles dose group, 129 to the 5 × 10¹⁰ viral particles dose group, and 126 to the placebo group. The mean age of the participants was 39·7 years (SD 12·5; range 18–83), with 309 (61%) individuals aged 18–44 years, 134 (26%) aged 45–54 years, and 65 (13%) aged 55 years or older across the treatment groups (table 1 ). 254 (50%) of 508 participants were male. Baseline characteristics of the participants and the pre-existing Ad5 neutralising antibody titres were largely similar across the treatment groups. Among the 508 participants, 266 (52%) had high pre-existing immunity and 242 (48%) had low pre-existing immunity to the Ad5 vector. All the participants completed the scheduled safety visits within 28 days post vaccination and gave blood samples at days 0 and 28, and 506 (>99%) donated blood samples at day 14.

The baseline antibody titres of the participants are given in the appendix (p 1). RBD-specific ELISA antibody responses induced by the Ad5-vectored COVID-19 vaccine were detected from day 14 onwards, with GMTs of 94·5 (95% CI 80·5–110·8) and 85·1 (66·0–109·7) in the 1 × 10¹¹ and 5 × 10¹⁰ viral particles dose groups, respectively (figure 2 ). At day 28, the RBD-specific ELISA antibodies peaked at 656·5 (575·2–749·2) in the 1 × 10¹¹ viral particles dose group and 571·0 (467·6–697·3) in the 5 × 10¹⁰ viral particles dose group. 244 (96%, 95% CI 93–98) of 253 participants in the 1 × 10¹¹ viral particles dose group and 125 (97%, 92–99) of 129 participants in the 5 × 10¹⁰ viral particles dose group showed seroconversion of RBD-specific ELISA antibodies at day 28, whereas the participants in the placebo group showed no antibody increase from baseline.

Both vaccine doses induced significant neutralising antibody responses to live SARS-CoV-2, with GMTs of 19·5 (95% CI 16·8–22·7) and 18·3 (14·4–23·3) in participants in the 1 × 10¹¹ and 5 × 10¹⁰ viral particles dose groups, respectively, at day 28 post vaccination (figure 2). Seroconversion of the neutralising antibody responses to live SARS-CoV-2 occurred in 148 (59%, 95% CI 52–65) of 253 participants receiving the 1 × 10¹¹ viral particles dose, and in 61 (47%, 39–56) of 129 participants receiving the 5 × 10¹⁰ viral particles dose 28 days post vaccination. The GMTs of neutralising antibody responses to pseudovirus were 61·4 (95% CI 53·0–71·0) in the 1 × 10¹¹ viral particles dose group and 55·3 (45·3–67·5) in the 5 × 10¹⁰ viral particles dose group, with seroconversion of 214 (85%, 95% CI 80–89) and 107 (83%, 76–88), respectively. No significant differences were observed between the two dose groups in the neutralising antibody responses to live virus and pseudovirus.

Before vaccination, 266 (52%) of 508 participants had high pre-existing anti-Ad5 neutralising antibodies (table 1). Participants with low pre-existing anti-Ad5 immunity had RBD-specific ELISA antibody and neutralising antibody levels that were approximately two-times higher than the participants with high pre-existing anti-Ad5 immunity (appendix pp 2–3). Increasing age was found to be another independent negative impact factor on the RBD-specific ELISA antibody (p=0·0018), and neutralising antibody responses to live virus (p<0·0001) or pseudovirus (p=0·046; appendix pp 4–6). The stratified analysis based on age found that participants aged 55 years or older were associated with relative low antibody responses in both dose groups post vaccination, particularly in terms of neutralising antibodies to live virus (appendix pp 7–8). Nevertheless, the ELISA antibodies to RBD and neutralising antibodies at day 28 in vaccine recipients were still significantly higher than in placebo recipients in this population. Male and female participants who received the vaccine showed similar RBD-specific ELISA antibody and neutralising antibody responses post vaccination (appendix pp 9–10). Both the ELISA antibody titres to RBD and neutralising antibody titres to pseudovirus were significantly correlated with the neutralising antibody titres to live virus, with a correlation coefficient of 0·75, and 0·72 (p<0·0001), respectively.

Baseline ELISpot T-cell responses were negative in 506 (>99%) of 508 participants. Ad5-vectored COVID-19 vaccine induced significant SARS-CoV-2 spike glycoprotein-specific IFNγ-ELISpot responses in 227 (90%, 95% CI 85–93) of 253 participants receiving the 1 × 10¹¹ viral particles dose, and 113 (88%, 81–92) of 129 participants receiving the 5 × 10¹⁰ viral particles dose at day 28 (figure 3 ). A median of 11·0 spot-forming cells (IQR 5·0–25·0) and 10·0 spot-forming cells (6·0–21·0) per 1 × 10⁵ peripheral blood mononuclear cells in participants in the 1 × 10¹¹ viral particles and 5 × 10¹⁰ viral particles dose groups, respectively, were observed at day 28, with increases of more than ten-times in both dose groups. The IFNγ-ELISpot responses were not significantly different between the dose groups at day 28. No positive IFNγ-ELISpot T-cell responses were detected in the placebo group post vaccination. Significant increases of post-vaccination T-cell responses in terms of spot-forming cells were observed both in participants with high and low pre-existing neutralising antibody at day 28. In the participants with high pre-existing immunity against Ad5, 88% of participants across both groups (111 of 126 in the 1 × 10¹¹ viral particles dose group, and 66 of 75 in the 5 × 10¹⁰ viral particles dose group) showed positive IFNγ-ELISpot T-cell responses post vaccination. Sex and age of the participants did not differ their IFNγ-ELISpot T-cell responses induced by vaccination (appendix p 11). In addition, 241 (95%, 95% CI 92–97) of 253 participants in the 1 × 10¹¹ viral particles dose group, and 118 (91%, 85–95) of 129 participants in the 5 × 10¹⁰ viral particles dose group showed either positive IFNγ-ELISpot T-cell response or seroconversion of neutralising antibody to live SARS-CoV-2 at day 28 post vaccination (appendix p 12).

Within 14 days after the vaccination, 183 (72%) of 253 participants in the 1 × 10¹¹ viral particles dose group, and 96 (74%) of 129 participants in the 5 × 10¹⁰ viral particles dose group reported at least one solicited adverse reaction, both of which were significantly higher than the 46 (37%) of 126 participants in the placebo group (p<0·0001; table 2 ). The most common systemic solicited reactions in the 5 × 10¹⁰ and 1 × 10¹¹ viral particles dose groups were fatigue, reported by 42% and 34%; fever, reported by 32% and 16%; and headache, reported by 29% and 28%, respectively. The most common injection site solicited reaction was pain, reported by 57% of the 1 × 10¹¹ viral particles dose group and 56% of the 5 × 10¹⁰ viral particles dose group. Although most adverse reactions were reported as either mild or moderate, 24 (9%) of 253 participants receiving the vaccine at 1 × 10¹¹ viral particles had severe (grade 3) adverse reactions, which was significantly higher than those receiving the vaccine at 5 × 10¹⁰ viral particles (p=0·0011) or placebo (p=0·0004). The most commonly reported grade 3 adverse reactions was fever, in 20 (8%) of 253 participants in the 1 × 10¹¹ viral particles dose group, and one (1%) of 129 participants in the 5 × 10¹⁰ viral particles dose group. High pre-existing Ad5 immunity, increasing age, and male sex were associated with significantly lower occurrence of fever post vaccination (appendix p 13). The grade 3 reactions were self-limited and resolved within 72–96 h without medication (appendix pp 14–15). The unsolicited adverse reactions within 14 days post vaccination were reported by 19 (8%) participants in the 1 × 10¹¹ viral particles dose group, seven (5%) in the 5 × 10¹⁰ viral particles dose group, and seven (6%) in the placebo group, showing no difference across the groups. Overall, 196 (77%) participants in the 1 × 10¹¹ viral particles dose group, 98 (76%) in the 5 × 10¹⁰ viral particles dose group, and 61 (48%) in the placebo group experienced at least one or more adverse event within 28 days after the vaccination. No serious adverse events were documented within 28 days.

All participants tested were seronegative for the antibodies to nucleocapsid protein of SARS-CoV-2 at day 28, by IgG and IgM rapid test kit (Vazyme Biotech, number CD101, Nanjing, China), suggesting none had exposure to SARS-CoV-2 during this study period.

Discussion

This study is the first randomised controlled trial for evaluation of the immunogenicity and safety of a candidate non-replicating Ad5-vectored COVID-19 vaccine. The phase 2 trial is a necessary and crucial step to turn an early-stage experimental vaccine into a promising vaccine candidate in an efficacy trial in a large population. In this study, a single injection of the Ad5-vectored COVID-19 vaccine at 1 × 10¹¹ viral particles and 5 × 10¹⁰ viral particles induced comparable specific immune responses to the spike glycoprotein at day 28, with no significant differences noted between the two groups. The vaccine induced seroconversion of the neutralising antibodies in 59% and 47% of participants, and seroconversion of binding antibody in 96% and 97% of participants, in the 1 × 10¹¹ and 5 × 10¹⁰ viral particles dose groups, respectively. Positive specific T-cell responses measured by IFNγ-ELISpot were found in 90% and 88% of participants receiving the vaccine at 1 × 10¹¹ and 5 × 10¹⁰ viral particles, respectively. 95% of participants in the 1 × 10¹¹ viral particles dose group and 91% of the recipients in the 5 × 10¹⁰ viral particles dose group showed either cellular or humoral immune responses at day 28 post vaccination (appendix p 12). Pre-existing immunity to the Ad5 vector and increasing age could partially hamper the specific immune responses to vaccination, particularly for the humoral immune responses.

In this study, most reactions reported post vaccination were mild or moderate. Although the proportions of participants who had adverse reactions such as fever, fatigue, and injection site pain were significantly higher in vaccine recipients than those in placebo recipients, adverse reactions within 28 days were generally not severe, and resolved within a short period of time (no more than 48 h). Among recipients of either dose of the Ad5-vectored COVID-19 vaccines, all grade 3 adverse reactions were reported from the dose group of 1 × 10¹¹ viral particles, with the exception of one from the 5 × 10¹⁰ viral particles dose group. Although this study was powered only to capture common adverse events after immunisation, the results suggest that the experimental Ad5-vectored COVID-19 vaccine has a good safety profile, which was in line with the results of our phase 1 trial in healthy adults.⁹

We initiated this phase 2 trial before the full analysis of the data from the phase 1 study was available. The vaccine doses chosen were mainly based on the safety data from the dose-escalating phase 1 trial: a higher proportion of participants reported grade 3 adverse reactions in the high-dose group (1·5 × 10¹¹ viral particles) compared with low-dose (5 × 10¹⁰ viral particles) or middle-dose (1 × 10¹¹ viral particles) groups (17% vs 6% and 6%, respectively).⁹ Therefore, we assumed that the vaccine doses at 1 × 10¹¹ and 5 × 10¹⁰ viral particles might have similar safety profiles. In addition, an increasing antigen dose is often associated with increasing immunogenicity; thus, we expected the vaccine dose at 1 × 10¹¹ viral particles to be the better of the two doses. Therefore, we designed the randomisation of this study in a ratio of 2:1:1 for the dose groups of 1 × 10¹¹ and 5 × 10¹⁰ viral particles, and placebo group, respectively, giving more weight to the first group. We found, in contrast to our expectations, that the vaccine at 5 × 10¹⁰ viral particles had a better safety profile than, and comparable immunogenicity to, the vaccine at 1 × 10¹¹ viral particles.

Age and pre-existing anti-Ad5 immunity of the participants could have affected the candidate vaccine's safety and immunogenicity. We noted that the occurrence of fever was associated with decreasing age and low pre-existing immunity to the vaccine vector Ad5 virus. 19 (90%) of the 21 participants who experienced grade 3 fever had no pre-existing immunity to Ad5, with a neutralising antibody titre below detection. Increasing age and high pre-existing anti-Ad5 immunity were found to be able to significantly reduce the immune responses to the vaccine. In some participants with high pre-existing anti-Ad5 immunity, one injection of the vaccine might be inadequate to induce a high level of humoral immune responses, particularly for people aged 55 years or older. These results align with the finding that older people are more likely to have exposure history to Ad5, with higher baseline neutralising antibody to Ad5, which indicates that this population might be more tolerant of higher dose or a booster dose regimen of the Ad5-vectored COVID-19 vaccine than people who are young and naive to Ad5. Pre-existing anti-Ad5 immunity is considered to be the biggest obstacle for the candidate Ad5-vectored COVID-19 vaccine to overcome. A flexible additional dose (between months 3 and 6) might be a potential solution to provide enhancement of immune responses, according to our previous experience with an Ad5 vector-based Ebola vaccine in a homologous prime-boost immunisation study.¹³ More evidence about the immunogenicity and feasibility of additional dose immunisation in the older population will be evaluated in a phase 2b trial. However, the vaccine recipients in this study showed increased anti-Ad5 neutralising antibodies, by 5·0-times and 3·8-times in the 1 × 10¹¹ viral particles and 5 × 10¹⁰ viral particles dose groups, respectively, at day 28 post vaccination (appendix p 17). The high level of anti-Ad5 immunity could affect the boosting effect of the vaccine; therefore, we planned to follow the dynamic change of the Ad5-specific antibodies in participants until month 6 to determine the timing of booster administration.

We determined the participants' serostatus before and after immunisation by ELISA, neutralising tests to live SARS-CoV-2 or pseudovirus, and ELISpot, providing evidence of humoral and cell-mediated immunity for the candidate vaccine. Because the live virus neutralising antibody test needs to be done in biosafety level three laboratories, a pseudovirus neutralising antibody test was developed to serve as an alternative.¹⁰ However, we found the magnitudes of neutralising antibody responses to pseudovirus were greater than those to the live virus, which might be associated with the different methodological principles of the two tests. In the pseudovirus neutralising test, when the specific antibody in serum binds to the pseudovirus, it inhibits the pseudovirus from entering the cells, reducing the expression of luciferase on the cell surface. Thus, we can calculate the amount of neutralising antibodies to pseudovirus by detecting the total fluorescence. The neutralising antibody is detected by measuring the cytopathic effect after the viral infection. Although, generally, the output values of the two methods are correlated, the two methods have different detection sensitivities and the detection values do not always have a one-to-one corresponding relationship.

Both neutralising antibody and T-cell responses were important in eliminating the virus and controlling disease development in patients with COVID-19 who were naturally infected by SARS-CoV-2.14, 15 Antibodies are very likely to be effective against SARS-CoV-2, considering that convalescent serum samples have been applied with apparently good clinical results in COVID-19.¹⁶ But for the vaccine-induced immune responses, whether neutralising antibody alone is capable of preventing infection remains undetermined. Specific T-cell responses are essential for directly attacking and killing virus-infected cells.¹⁵ In addition, the CD4 T-cell responses are critical for the cytotoxic T-cell response and the maturating of neutralising antibodies.¹⁷ Thus, the evaluation of the cell-mediated responses, in addition to the neutralising antibodies, is important for a successful candidate vaccine.

Our trial has some limitations. First, this phase 2 trial started before the full analysis of the data from the phase 1 study was available, so we did not calculate the sample size based on study power in advance, which might lead to a lack of power to show the difference between dose groups. Second, the participants included in this study are all from Wuhan, China. The baseline anti-Ad5 immunity of the participants seemed to be representative of Chinese adults, according to the previously reported studies;¹⁸ however, anti-Ad5 immunity in adults varies from place to place globally, with reported proportions of adults with neutralising antibodies titres for Ad5 of more than 1:200 of about 80% in India, 78% in Kenya, 67% in Thailand, 64% in Uganda, around 60% in South Africa, 45% in Sierra Leone, and less than 30% in the USA.19, 20, 21 We might expect the candidate Ad5-vectored vaccine to have a superior immunogenicity in the population with a lower pre-existing anti-Ad5 immunity, but an inferior immunogenicity in people with a higher pre-existing anti-Ad5 immunity than observed in this phase 2 trial. Third, this trial did not include children. Although COVID-19 appears to have a more benign course in children, with almost no fatalities reported,22, 23 an ideal candidate vaccine for the ongoing pandemic should cover susceptible populations in all ages. Fourth, we report only data within 28 days of vaccination, and do not include data about the durability of the vaccine-induced immunity, which was not available at the time of publication. A subset of patients infected with SARS-CoV-2 might not develop long-lasting antibodies to the virus, and S-antibodies were reported to rapidly decline for people infected with seasonal coronaviruses and who recovered from COVID-19, especially those with mild symptoms or asymptomatic infection.²⁴ The ongoing phase 1 and 2 trials will enable the continued collection of safety data and assessment of antibody persistence over a 6-month period. Fifth, in this study, no participants had SARS-CoV-2 exposure after the vaccination, so we were unable to assess the efficacy of the candidate vaccine or any immunological risk associated with antibody induced by vaccination when exposed to the virus. However, the risk of COVID-19 and antibody-enhanced disease on exposure to the virus will be monitored long term. Finally, the clinical significance of these changes is difficult to assess because of the absence of an identified correlate of protective immunity, and reference standards for measuring neutralising antibodies against COVID-19. Future studies will need to establish an immune correlate of protection and a protective threshold to assess the feasibility of using the Ad5-vectored COVID-19 vaccine to provide protection for high-risk populations or for outbreak intervention.

WHO is facilitating collaboration and efforts to support vaccine development by defining the desired characteristics of promising candidate vaccines to combat COVID-19.25, 26 This situation emphasises the importance of being prepared to initiate an international multicentre, randomised, double-blind, controlled phase 3 effectiveness trial as soon as possible.²⁵ One immunisation of the Ad5-vectored COVID-19 vaccine at 5 × 10¹⁰ viral particles has a good safety profile (limited to common adverse reactions following immunisation) and could elicit significant specific immune responses to SARS-CoV-2, making it a potential candidate for emergency vaccination of acute protective response.

In conclusion, the results of this trial have extended our knowledge of the immunogenicity and safety of the Ad5-vectored COVID-19 vaccines. The results support testing of the Ad5-vectored COVID-19 vaccine at 5 × 10¹⁰ viral particles in a phase 3 effectiveness trial in healthy adults.

Data sharing

We support sharing of the individual participant data. The individual participant data that underlie the results reported in this Article, after de-identification (text, tables, figures, and appendices) will be shared. Individual participant data will be available beginning 3 months and ending 1 year after publication. Supporting clinical documents including study protocol, statistical analysis plan, and the informed consent form will be available immediately following publication for at least 1 year. Researchers who provide a scientifically sound proposal will be allowed access to the individual participant data. Proposals should be directed to jszfc@vip.sina.com or cw0226@foxmail.com. These proposals will be reviewed and approved by the funder, investigator, and collaborators on the basis of scientific merit. To gain access, data requesters will need to sign a data access agreement.

Supplementary Material