Immune Responses to the ChAdOx1 nCoV-19 and BNT162b2 Vaccines and to Natural Coronavirus Disease 2019 Infections Over a 3-Month Period

Abstract

Background

There are limited data directly comparing immune responses to vaccines and to natural infections with coronavirus disease 2019 (COVID-19). This study assessed the immunogenicity of the BNT162b2 and ChAdOx1 nCoV-19 vaccines over a 3-month period and compared the immune responses with those to natural infections.

Method

We enrolled healthcare workers who received BNT162b2 or ChAdOx1 nCoV-19 vaccines and patients with confirmed COVID-19 and then measured S1 immunoglobulin (Ig) G and neutralizing antibodies and T-cell responses.

Results

A total of 121 vaccinees and 26 patients with confirmed COVID-19 were analyzed. After the second dose, the BNT162b2 vaccine yielded S1 IgG antibody responses similar to those achieved with natural infections (mean IgG titer [standard deviation], 2241 [899] vs 2601 [5039]; P = .68) but significantly stronger than responses to the ChAdOx1 vaccine (174 [96]; P < .001). The neutralizing antibody titer generated by BNT162b2 was 6-fold higher than that generated by ChAdOx1 but lower than that by natural infection. T-cell responses persisted for 3 months with BNT162b2 and natural infection but decreased with ChAdOx1.

Conclusions

Antibody responses after the second dose of BNT162b2 are higher than after the second dose of ChAdOx1 and like those occurring after natural infection. T-cell responses are maintained longer in BNT162b2 vaccinees than in ChAdOx1 vaccinees.

Vaccines against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) are expected to end the coronavirus disease 19 (COVID-19) pandemic. The lipid nanoparticle–formulated messenger RNA (mRNA)–based vaccine, BNT162b2, developed by BioNTech/Pfizer, has been reported to be 95% effective in preventing COVID-19 [1]. The ChAdOx1 nCoV-19 vaccine, developed by Oxford University/AstraZeneca, a replication-deficient chimpanzee adenovirus vector-based vaccine, has been reported to be 70.4% effective [2]. Separate studies reported that BNT162b2 and ChaAdOx1 nCoV-19 elicited strong antibody responses and robust spike protein specific CD8⁺ and CD4⁺ T-cell responses [3–8].

In our group’s previous study [9], antibody responses to BNT162b2 after the first vaccination of a homogeneous population were faster and stronger than responses to ChAdOx1 nCoV-19, but spike-specific T-cell responses were similar. However, there are limited comparisons of the detailed kinetic immunogenicity across differently platformed, second-dose-completed ChAdOx1 and BNT162b2. Many studies have reported rapid waning of immunoglobulin (Ig) G antibodies within 3–4 months in patients naturally infected with COVID-19 [10–13], and our group’s previous findings [14] were consistent with the findings of those studies. However, studies directly comparing immune responses to natural infection and vaccines have also been limited. In the current study, we examined immune responses, both humoral and cellular, to the BNT162b2 and ChAdOx1 nCoV-19 vaccines over a 3-month period and compared the antibody and T-cell responses induced by the vaccines with those generated by natural COVID-19 infection.

MATERIALS AND METHODS

Study Participants and the Specimens

A nationwide vaccination program against COVID-19 is ongoing in South Korea. Our study enrolled healthcare workers who received the ChAdOx1 nCoV-19 vaccine or the BNT162b2 vaccine at a tertiary care hospital in Seoul, South Korea, between 5 and 25 March 2021. In accord with the policy of the Korean government, the BNT162b2 vaccine was assigned to high-risk healthcare workers in direct contact with patients who had COVID-19, and the ChAdOx1 vaccine was assigned to those involved in general patient care. All participants agreed to peripheral blood sampling, and blood sampling was carried out once before vaccination, for baseline serology. Thereafter, blood samples were taken from the participants vaccinated with NT162b2 at 3 weeks after the first dose, 2 weeks after the second dose (5 weeks after the first dose), and 12 weeks after the first dose. The blood samples collected from the participants vaccinated with the ChAdOx1 nCoV-19 vaccine were collected 3, 8, and 12 weeks after the first dose and 2 weeks after the second dose (14 weeks after the first dose).

Patients naturally infected with COVID-19 and admitted to Asan Medical Center were enrolled between March 2020 and February 2021. COVID-19 infection was confirmed by real-time reverse transcription-polymerase chain reaction (RT-PCR) for the RdRp, N, and E genes of SARS-CoV-2 (Allplex 2019-nCoV assay; Seegene). All participants agreed to peripheral blood sampling, which was carried out on the day of hospital admission and 1 month, 2 months, and either 3 or 4 months after symptom onset. The study was reviewed and approved by the Institutional Review Board of Asan Medical Center (IRB nos. 2020-0297 and 2021-0170). Informed consents were obtained from all participants.

Measurement of Antibody Responses

SARS-CoV-2 S1-specific IgG and IgM antibody titers were measured using an enzyme-linked immunosorbent assay developed in-house, details of which are described elsewhere report [8]. The data are presented as international units per milliliter, which is standardized with reference pooled sera from International Vaccine Institute (Seoul, South Korea).To determine cutoff values for the enzyme-linked immunosorbent assay, the mean and standard deviation (SD) of negative control plasma were measured, and cutoff values were defined as mean plus 3-fold the SD value; the cutoff value was 10 IU/mL for IgG, as reported elsewhere [15, 16].

We also measured plasma levels of neutralizing antibodies using a microneutralization assay. Briefly, 100 times the median tissue culture infective dose of SARS-CoV-2 (βCoV/Korea/KCDC/2020 NCCP43326) was mixed with an equal volume of diluted plasma specimen, incubated at 37°C for 30 minutes, and added to Vero cells. After 96 hours, the cytopathic effect of SARS-CoV-2 on the infected cells was measured. Neutralizing antibody titer was calculated as the reciprocal of the highest dilution of test plasma giving 50% neutralization. The microneutralization assay was performed in a biosafety level 3 laboratory in Institut Pasteur Korea (Seongnam, Republic of Korea).

Measurement of T-Cell Responses

An interferon (IFN) γ enzyme-linked immunospot assay was performed to measure the SARS-CoV-2–specific T-cell response of peripheral blood mononuclear cells (PBMCs) isolated from participants’ blood samples. T cells were stimulated with overlapping peptides of SARS-CoV-2 spike protein (Miltenyi Biotec), and numbers of spot-forming cells per 5.0 × 10⁵ PBMCs were counted using an automated enzyme-linked immunospot reader (AID iSPOT; Autoimmun Diagnostika).

Statistical Analysis

Statistical analyses were performed with SPSS Statistics for Windows software, version 24.0 (IBM), and graphs were plotted with GraphPad Prism 8 software. Depending on normality of the data, we used the χ² test or Fisher exact test to analyze categorical variables and Student t test or the Mann-Whitney U test for continuous variables. All tests of significance were 2-tailed, and differences were considered statistically significant at P < .05.

RESULTS

Baseline Characteristics of Study Participants

Of a total of 129 healthcare workers enrolled in this study, 93 (72%) were vaccinated with ChAdOx1 and 36 (28%) with BNT162b2; after 8 of the ChAdOx1 vaccinees dropped out, 85 ChAdOx1 and 36 BNT162b2 vaccinees were included in the final analysis. Baseline characteristics of the participants are shown in Table 1. The ChAdOx1 vaccinees were older than the BNT162b2 vaccinees (median age, 36 vs 32 years, respectively; P = .006) (Table 1). After the second dose of vaccine, local and systemic reactogenicity, based on severity grade, were significantly higher in the ChAdOx1 than in the BNT162b2 group (P = .04 and P < .001, respectively) (Table 1). Twenty-six patients with confirmed COVID-19 were classified according to disease severity, which was asymptomatic in 1, mild in 1, moderate in 8, severe in 12, and critical in 4. Detailed baseline characteristics by disease severity are shown in Table 2.

S1-IgG Antibody Responses

SARS-CoV-2 S1 protein-specific IgG antibody (S1-IgG) titers were measured in 36 of the plasma samples from the participants injected with the BNT162b2 vaccine, at baseline and at 5 and 12 weeks and in 35 of the plasma samples at 3 weeks after the first vaccination. S1-IgG titers were significantly higher 2 weeks after the second dose (5 weeks after the first dose) than at 3 weeks after the first dose (mean [SD], 2241 [899] vs 351 [180] IU/mL; P < .001). They then decreased by 12 weeks after the first dose (mean [SD], 834 [467] IU/mL; P < .001) (Figure 1A).

S1-IgG was measured in 84 plasma samples from ChAdOx1 nCoV-19 vaccinees at baseline, in 85 samples at 3 and 8 weeks, in 82 at 12 weeks, and in 83 at 14 weeks after the first vaccination. S1-IgG titers declined between 3 and 12 weeks after the first vaccination (mean [SD], 142 [139] at 3 vs 42 [29] at 12 weeks; P < .001) (Figure 1B). At 2 weeks after the second dose (14 weeks after the first dose), S1-IgG titers were higher than at 12 weeks but not significantly different from those at 3 weeks (mean [SD] at 14 weeks, 174 [96]; 12 vs 14 weeks, P < .001; 3 vs 14 weeks, P = .07) (Figure 1B).

For the naturally infected patients, S1-IgG was measured in 21 plasma samples at admission (median time after symptom onset [range] 6 [0–13] days), in 24 samples 4 weeks (30 [15–44] days) after symptom onset, in 18 samples 8 weeks (54 [47–74] days) after symptom onset, and 26 samples 15 weeks (105 [81–140] days) after symptom onset. The highest S1-IgG titers were measured at 4 weeks (mean [SD], 2601 [5039]) and declined by 15 weeks (457 [620]; P = .02) (Figure 1C).

At 4 weeks after symptom onset, naturally infected patients had almost the same antibody levels as second-dose BNT162b2 vaccinees at 5 weeks (mean [SD], 2601 [5039] and 2241 [899], respectively; P = .68), but these levels were significantly higher than at 2 weeks after the second dose of ChAdOx1 nCoV-19 (174 [96] at 14 weeks; P < .001). At 3 months after first vaccination or natural infection, S1-IgG antibody levels for ChAdOx1 nCoV-19 (mean [SD], 174 [96] at 14 weeks) were significantly lower than for BNT162b2 (834 [467] at 12 weeks; P < .001) or natural infection (457 [620] at 15 weeks; P = .001) (Figure 1D).

Virus Neutralizing Antibody Responses

SARS-CoV-2 virus neutralizing antibody titers were measured in plasma samples from 31 BNT162b2 and 37 ChAdOx1 nCoV-19 vaccinees 3 weeks after the first dose, and in samples from 36 BNT162b2 and 81 ChAdOx1 nCoV-19 vaccinees 2 weeks after the second dose. Neutralizing antibody titers were also measured in the plasma samples from 18 naturally infected patients at 4 weeks after symptom onset (peak response). After the first dose, the neutralizing antibody titers of the BNT161b2 vaccinees were about 1.6-fold higher than those in the ChAdOx1 nCoV-19 vaccinees (mean [SD], 183.1 [155.6] and 116.6 [116.2], respectively; P = .04) (Figure 1E). After the second dose, the titers were about 6-fold higher than in of the ChAdOx1 nCoV-19 vaccinees (mean [SD], 2544 [2547] vs 447 [341]; P < .001) but lower than those in naturally infected patients at 4 weeks (4708 [3270]; P = .01). Neutralizing antibody titers were significantly correlated with S1-IgG antibody titers (Pearson r = .747; P < .001) (Supplementary Figure 1).

IFN-γ–Producing T-Cell Responses

SARS-CoV-2 spike protein-specific IFN-γ–producing T-cell responses were measured in 35 PBMC samples from the participants injected with BNT162b2, at baseline and at 5 and 12 weeks, and in 33 PBMC samples 2 weeks after the first dose. T-cell responses had increased significantly by 2 weeks after the first dose, but by 2 weeks after the second dose they had not (mean [SD], 104.6 [97.8] at 2 vs 156.3 [113.6] at 5 weeks; P = .64) (Figure 2A). After that they remained roughly constant until 12 weeks after the first vaccination (5 vs 12 weeks, P = .64; 2 vs 12 weeks, P = .98) (Figure 2A).

IFN-γ–producing T-cell responses were measured in 29 PBMC samples from participants injected with ChAdOx1 nCoV-19, at baseline and at 12 and 14 weeks, and in 27 PBMC samples 2 weeks after the first dose. T-cell responses declined between 2 weeks and 12 weeks after the first dose (mean [SD], 121.2 [104.3] at 2 vs 34.3 [46.9] at 12 weeks; P < .001) (Figure 2B). By 2 weeks after the second dose (14 weeks after baseline), T-cell responses had not increased significantly (mean [SD], 44.93 [33.11]; P = .61).

T-cell responses in the patients with COVID-19 infection were measured in 18 PBMC samples at 4 weeks after symptom onset and 24 PBMC samples at 15 weeks. T-cell responses were maintained between 4 weeks and 15 weeks after symptom onset (mean [SD], 78.1 [83.8] at 4 vs 77.7 [119.1] at 15 weeks; P = .99) (Figure 2C). In vaccinees, 2 weeks after the first doses, T-cell responses to BNT162b2 and ChAdOx1 nCoV-19 were similar (P = .51) (Figure 2D), and they were like responses to natural infection at 4 weeks after symptom onset (natural infection vs BNT162b2, P = .52; natural infection vs ChAdOx1, P = .39). At 3 months after the first dose, the T-cell responses in the BNT162b2 vaccinees (at 12 weeks) were significantly higher than those in the ChAdOx1 nCoV-19 vaccinees at 14 weeks (mean [SD],10 8.7 [89.8] vs 44.93 [38.11]; P = .006), but comparable to those with the natural infection 15 weeks after symptom onset (77.7 [119.]; P = .08).

DISCUSSION

In the current study we assessed the immunogenicity of second-dose-completed BNT162b2 and ChAdOx1 nCoV-19 vaccines and compared the antibody and T-cell responses induced by the vaccines with those evoked by natural infection with COVID-19. After the second dose, S1-IgG antibody responses to BNT162b2 vaccine were similar to those of naturally infected patients and significantly higher than responses to ChAdOx1 nCoV-19 vaccine. Neutralizing antibody titers after the second dose were also significantly higher in BNT162b2 vaccinees than in ChAdOx1 nCoV-19 vaccinees but lower than in natural infection. S1-IgG antibody declined over 3 months in both BNT162b2 vaccinees and naturally infected patients. Three months after the first dose, the S1-IgG antibody response to BNT162b2 was highest, and the response to ChAdOx1 was lower than responses in the other groups. With both vaccines, the second doses did not elicit significant IFN-γ–producing T-cell responses. The T-cell response to BNT162b2 was more or less maintained, but the response to ChAdOx1 decreased by 3 months after the first dose. The T-cell responses to natural infection were constant for 3 months after infection.

In our group’s previous study, the first dose of BNT162b2 elicited a higher antibody response than ChAdOx1 nCoV-19 [9]. The differences between the S1-IgG and neutralizing antibody responses to the 2 vaccines increased after the second dose (Figure 1). After the second dose, BNT162b2b elicited an S1-IgG antibody response similar to that in the naturally infected patients, but the response to ChAdOx1 was significantly lower. The spike glycoprotein of SARS-CoV-2 binds a receptor, angiotensin-converting enzyme 2, through a domain that is part of S1, and it mediates cell entry [17, 18]. The spike protein is a key target for virus neutralizing antibodies [19] and a prime candidate for vaccine development. The ChAdOx1 nCoV-19 vaccine contains the full-length spike protein along with the leader sequence of tissue plasminogen activator [20]. The BNT162b2 vaccine also encodes the full-length spike protein, which is stabilized in the prefusion conformation by mutations of residues 986 and 987 to prolines [21, 22]. This structural difference between the spike proteins may affect antibody production, and the difference between the adenovirus-vector-based platform and the mRNA-based platform may also partly explain the discrepancy in antibody response between the 2 vaccines.

To the best of our knowledge, there are very limited data directly comparing SARS-CoV-2–specific T-cell responses to the different COVID-19 vaccine platforms after the first and second doses of vaccine. Our study revealed that the spike protein–specific IFN-γ–producing T-cell response to the BNT162b2 vaccine was more stable over 3 months than that to the ChAdOx1 nCoV-19 vaccine. T cells may play a major role in the resolution of COVID-19 [23], but the effects on long-term memory T cells and their effects on long-lasting immunity are unclear. Nevertheless, current studies have shown that T-cell responses remain robust up to 6 or 12 months after exposure to SARS-CoV-2 [24, 25], even though they are accompanied by waning of antibody responses. Jung et al [26] also reported that SARS-CoV-2 spike, membrane, and nucleocapsid protein specific IFN-γ-producing T-cell responses were sustained over 200 days after infection. Thus, vaccine-induced T-cell responses may well contribute to long-term immunity. Although our findings indicate that the BNT162b2 vaccine elicits more durable SARS-CoV-2–specific T-cell responses than the ChAdOx1 nCoV-19 vaccine, further studies are needed to assess which vaccine platform may be superior in terms of long-term immunity against SARS-CoV-2.

In both BNT162b2 vaccinees and natural-infected patients, S1-IgG antibody declined significantly at 3 months after vaccination or infection. Other studies have reported results consistent with ours, namely, waning of spike-IgG antibody to the BNT162b2 vaccine [27, 28] and to natural COVID-19 infections [11, 12, 24]. However, a previous study reported that antibody waning dynamics were varied among naturally infected patients with rapid waning, slow waning, and persistent responses group [29]. Thus, further detailed kinetic studies are needed on the different waning patterns of vaccine-induced immunity depending on longevit.

In addition, S1-IgG antibody levels induced by the ChAdOx1 nCoV-19 vaccine declined between 3 weeks and 12 weeks after the first dose, although we did not observe a further decline after the second dose. However, this waning immunity was reversed after the second dose of ChAdOx1 nCoV-19 vaccine, which was given in accordance with Korean national vaccine policy based on evidence that a 3-month interval between doses may be beneficial for protection against COVID-19 [30]. In contrast, even though the second dose was delayed, the ChAdOx1 vaccine induced a lower S1-IgG antibody titer at 3 months after the first dose than either BNT162b2 or natural infection. Meanwhile, Goel et al [31] found that spike-specific IgG antibody responses decreased by half at 28–33 days after mRNA vaccinations, but spike-specific memory B-cell responses did not decline up to 6 months.

It is worth noting that various immunologic factors in addition to the antibody response may affect vaccine effectiveness against symptomatic or severe COVID-19. Hence, further studies of vaccine effectiveness and of the immunologic properties of the various COVID-19 vaccine platforms are needed.

The gradual reduction in antibody titers may affect the efficacy of the vaccines, especially in response to infections by variants. A current prime concern about COVID-19 is the emergence and spread of diverse variants of SARS-CoV-2. The B.1.617.2 (delta) variant, already dominant worldwide, has several spike protein mutations [32] that may affect immune responses to the key antigenic regions of this receptor-binding-protein [33]. Indeed, neutralizing antibody activity against the delta variant in BNT162b2 recipients is reported to be 5.8 times lower than that against wild-type SARS-CoV-2 [34]. Furthermore, the effectiveness of the mRNA vaccines in nursing home residents fell from 75% to 53% after the emergence of the delta variant [35]. Additional studies of the effect of the reduced efficacies of the different COVID-19 vaccine platforms against the delta variant are urgently needed.

Limitations of the present study include the relatively small number of enrolled participants and the differences in vaccine schedules and sampling times. Despite these limitations, our study provides clear comparative data on the immune responses to second-dose-completed BNT162b2 and ChAdOx1 nCoV-19 vaccines as well as to natural COVID-19 infections. The antibody responses induced by the BNT162b2 vaccine were much higher than those induced by the ChAdOx1 vaccine and similar in magnitude to the responses to natural infections. The antibody responses to the vaccines, like those to natural infections, waned after 3 months. T-cell responses were maintained in the BNT162b2 vaccinees and natural infected patients but not in the ChAdOx1 vaccinees. Further research into the induction of long-term immunity is needed, especially to manage emerging variants.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Financial support. This study was supported by the Korea Health Technology R&D Project through the Korea Health Industry Development Institute, funded by the Ministry of Health & Welfare, South Korea (grant HW20C2062) and the National Research Foundation of Korea, through the Ministry of Science and ICT, Government of Korea (grant 2017M3A9G6068254).

Potential conflict of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References

Author notes