SARS-CoV-2 mRNA vaccine BNT162b2 elicited a robust humoral and cellular response against SARS-CoV-2 variants.

BNT162b2 vaccine was introduced in Italy on 27th December 2020 and healthcare workers were rapidly vaccinated. In this study, we demonstrated that one vaccine dose was sucient for eliciting a sustained humoral and cell-mediated response in SARS-CoV-2 experienced healthcare workers but had a lower effect in SARS-CoV-2 naïve subjects. However, 98% naïve subjects developed both neutralizing antibodies and Spike-specic T-cells after the second dose. Moreover, the antibody and T-cell responses were effective against viral variants since a partial reduction in antibody response was observed only against the South-African variant in SARS-CoV-2 naïve individuals, while the T-cell response was less affected. magnitude of T cells specic for SARS-CoV2 S protein, RBD-binding and NT antibodies after the rst and the second dose of BNT162b2 vaccine; iii) the correlation between T-cell response, RBD-binding antibodies and NT antibodies; iv) the changes in total CD4+ and CD8+ T cells, B cells and NK cells after the rst and second dose of BNT162b2 vaccine; 5) the frequency of T-cell and NT antibody responders against SARS-CoV-2 variant strains; 6) the persistence of T-cell response, RBD-binding antibodies and NT antibodies. This last end-point will be analyzed at a later time.


Introduction
After the emergence of the Coronavirus Disease 2019  pandemic caused by the betacoronavirus named severe acute respiratory syndrome virus 2 (SARS-CoV-2), a strong effort has been made for the rapid development of protective vaccines. The BNT162b2 vaccine [1], a lipid nanoparticleformulated, nucleoside-modi ed RNA encoding the SARS-CoV-2 full-length Spike (S) protein derived from the rst SARS-CoV-2 isolated in Wuhan, China, modi ed by two proline mutations to lock it in the prefusion conformation, was the rst authorized for active immunization.
The vaccine showed a 95% protection against SARS-CoV-2 infection in a phase 2/3 trial [2]. Another mRNA-based vaccine, mRNA-1273 [3] showed similar effect and was subsequently authorized. However, data on the kinetics of immune response elicited by the vaccines are limited to low numbers of subjects analyzed and mainly to the antibody response [1,[4][5][6][7][8].
Moreover, the T-cell response elicited by the vaccine may have a crucial role in the protection from SARS-CoV-2 infection and disease as well as in the establishment of a long-term recall memory response.
Previous reports related to SARS-CoV-1 infection suggested a protective role of both humoral and cellmediated immunity. Moreover, while SARS-CoV-1 T-cell response may confer long-term protection, the antibody response was found to be relatively short-lived in convalescent individuals [9]. On the other hand, T-and B-cell memory speci c for SARS-CoV-2 was found to persist for at least 6-8 months [10][11][12].
The emergence of new SARS-CoV-2 variants with mutations in the Spike protein has raising signi cant concerns about vaccines e cacy and reinfection risk in subjects who experienced a natural SARS-CoV-2 infection. The rst emerging spike protein variant, D614G, rapidly became the dominant SARS-CoV-2 clade in Europe by May 2020 (EU variant). Recently, the new variant 501Y.V1 lineage B.1.1.7 (UK variant) including multiple mutations in both the receptor binding domain (RBD) and N-terminal domain of S protein [13] as well as the 501Y.V3 lineage P.1 (Brazilian, BZ) [14] and 501Y.V2 lineage B.1.351 (South African, SAF) variants mutations are present in S protein and, especially, in the RBD [15].
Preliminary data suggested that convalescent and sera from vaccinated individuals e ciently neutralized UK variant, while a reduction in neutralizing (NT) antibody titres has been observed against SAF variant [16,17].
In addition, recent data from adenoviral vaccine ChAdOx1 nCoV-19 studies reported a lower e cacy against SAF variant [18], whereas the Ad26.COV2.S vaccine appeared equally effective in USA, Brazil and South Africa, where different variants circulate [19]. The question of whether both SARS-CoV-2 experienced patients and vaccinated subjects might be protected against emerging variants needs to be urgently investigated. So far, in depth analysis of the level of vaccine-induced B-and T-cell immune response against SARS-CoV-2 variants in vaccinated subjects is still lacking.
The present study was designed to evaluate the development and the persistence of the T-and B-cell responses in subjects receiving the BNT162b2 vaccine. In this article, we report the analysis of the immune response in 145 subjects after the rst and second dose of the vaccination schedule, whereas the long-term persistence of the immune response will be analyzed subsequently. Moreover, we aimed to de ne the level of both antibody and cell mediated response against the emerging SARS-CoV-2 variants after a complete vaccination schedule in comparison to convalescent patients.
SARS-CoV-2 speci c T-cell response ELISpot assay was used for detecting anti-S and, as control, anti-nucleocapsid (N) cell-mediated response in vaccinated subjects at baseline, at T1 and T2.
Levels of N-speci c T cells did not change with time in both experienced and naïve subjects (Fig. 2b). Of note, 30% of SARS-CoV-2 naïve subjects showed detectable SARS-CoV-2 speci c T-cell response against S and 18.2% against N at baseline.
The phenotype of S-speci c T cells elicited by the vaccine was analysed in a subgroup of 20 naïve subjects at T2 by ow-cytometry analysis of T-cells proliferating in response to S and N peptide pools and co-expressing CD25 and ICOS activation markers ( Fig. 3; Supplementary Fig. 2). The vaccine elicited CD4 + T-cell (median 7.0%, IQR 2.1-8.2% CD25 + ICOS + cells) and, particularly (p = 0.005), CD4 + T FH -cell response (median 7.8%, IQR 6.2-12.4% CD25 + ICOS + cells). In addition, also S-speci c CD8 + T cells (median 2.6%, IQR 1.4-6.1% CD25 + ICOS + cells) were detected in vaccinated subjects. This pattern is similar to that observed in 9 SARS-CoV-2-infected subjects with mild symptoms collected a median time of 38 (range 28-90) days after infection, who showed a slightly lower number of proliferating CD4 + T and CD4 + T FH cells than vaccinated subjects (p = 0.044 and 0.047, respectively). As observed also with the ELISpot assay, the response to N protein was detected in few vaccinated subjects.

Antibody and T-cell response against SARS-CoV-2 variants
Sera collected at T2 from 31 subjects for SARS-CoV-2 infection were tested against RBD of the wild type (WT) Chinese strain, the 501Y.V1 lineage B.  (Fig. 4a). A reduction of approximatively 50% and 99% of RBD-speci c antibody titres was observed from WT to UK and SAF strains, respectively (Fig. 4b-c).
Sera collected at T2 from 68 SARS-CoV-2 vaccinated subjects (61 naïve per COVID-19 and 7 COVID-19 positive) were tested against WT strain and challenged against EU, UK, SAF, and 501Y.V3 lineage P.1 (Brazilian; BZ) ( Fig. 4d). Among the 61 SARS-CoV-2 naïve subjects, an increase in four-fold dilution for SARS-CoV-2 NT antibodies was observed against the EU variant with respect to WT strain (p < 0.001) ( Fig. 4e), while the level of SARS-CoV-2 NT antibodies was not affected when sera were tested against UK and BZ strains ( Fig. 4f-g). A ve-fold and twenty-fold reduction in median titre of SARS-CoV-2 NT antibodies was observed in response to SAF strain in comparison to WT virus and EU strain, respectively (p < 0.001) (Fig. 4h). However, although at lower levels NT antibodies against the SAF variant were still detectable in in 55/68 (80.9%) subjects. Among the 7 SARS-CoV-2 experienced vaccinated subjects, a sustained NT level against the all ve variants was observed. The median reduction of SARS-CoV-2 NT against the SAF variant was only onefold than respect to WT (median 160 vs 320; respectively p = 0.603).
In addition, sera from six vaccinates subjects, were used to assess the NT level also against the C.36 Thailand (Thai) and B.1.525 Nigerian (NIG) variants. No signi cant difference was observed between SARS-CoV-2 NT titer against WT, UK, BZ and Thai variant (p = 0.999). Similarly to what observed for the SAF variant, a six-fold decrease of NT level was documented against the NIG variant (p = 0.081; Supplementary Fig. 3).
Poor correlation between RBD-speci c and NT antibody titres for the different SARS-CoV-2 variants was observed. In 36 of the 68 vaccinated subjects (53%), residual PBMC isolated at T2 were challenged against inactivated virus preparations from supernatant of Vero E6 cells infected with the ve SARS-CoV-2 strains (WT, EU, UK, BZ and SAF) and cell-mediated response was analysed by ELISpot assay (Fig. 6a).
The ELISpot assay was less sensitive in measuring T-cell response using WT virus than using peptide pools from the homologous S protein: median IFNγ SFU/million PBMC were 112.5 (IQR 71.3-233.8) and 47.5 (IQR 23-91), respectively. However, even if a slightly lower level of cell-mediated T-cell response was observed against UK, BZ and SAF variants (Fig. 6a), SARS-CoV-2 speci c T cells were detected in the large majority of the subjects, with no signi cant difference in the frequency of responders. In detail, 34/36 (94.4%) were positive for WT SARS-CoV-2 T-cell response and 33/36 (91.7%) were positive for EU SARS-CoV-2 T-cell response. UK, BZ and SAF-speci c T-cell response was detectable in 30/36 (83.3%), 29/36 (80.5%) and 31/36 (86.1%) vaccinated subjects (Fig. 6b), indicating that the BNT162b2 vaccine was able to elicit T-cells able to recognize conserved epitopes from any virus variant tested.

Discussion
Here we report a prospective and comprehensive evaluation of the development of antibody and T-cell responses elicited by the mRNA vaccine BNT162b2 for SARS-CoV-2 in a cohort of 145 immunocompetent individuals (18 SARS-CoV-2 experienced and 127 SARS-CoV-2 naïve). The major results of the study show that: i) all vaccinated subjects tested developed a NT antibody response, and 98.6% a speci c T-cell response after complete vaccination schedule; both antibody and T-cell response were not affected by age after the second dose; ii) antibody and T-cell responses were not correlated; iii) the antibody and T-cell responses are effective against most viral variants (a reduction in antibody response was observed only against the SAF variant, while the T-cell response was less affected).
The mRNA-based vaccine was highly effective in eliciting both antibody and T-cell responses. Regarding the T-cell subsets induced, we observed both CD4 + and CD8 + T cells proliferative responses. This pattern is similar to that observed in a preclinical study for another mRNA-based vaccine against human cytomegalovirus [20]. Among CD4 + T cells, we observed an effective response of the T FH subset, which is involved in providing help for B-cell differentiation and antibody a nity maturation. A reduction in total T and B cells after vaccination was observed; this event may be a consequence of lymphocyte redistribution to lymphoid organs during the development of the immune response to the vaccine.
In this study, we observed that all the SARS-CoV-2 naïve individuals developed a humoral response higher than that observed after natural infection after two vaccine doses but not after one vaccine dose, when about 6% of individuals was still negative for SARS-CoV-2 NT Abs. On the other side, the cell-mediated response was detectable in 98.4% naïve individuals after two doses of vaccine, whereas one vaccine dose was not su cient for eliciting detectable SARS-CoV-2 T-cell response in 32.1% of subjects.
Conversely, both humoral and cellular responses were strongly induced after the rst dose of the BNT162b2 vaccine in previously SARS-CoV-2 infected subjects: therefore, it appears that prime vaccine dose acts as a boost in these subjects. Moreover, both antibody and T-cell responses did not appear to increase after the second dose in these individuals. According to previous observation [21,22], this evidence strengthens the concept that just one vaccine dose may be su cient in subjects with previous SARS-CoV-2 infection [23].
On the other side, we observed that level of antibody and, especially, cell-mediated response after one vaccine dose was weak or undetectable in SARS-CoV-2 naive subjects. This is in contrast to what was observed in a previous report analysing the antibody response in a small number of subjects [20], and suggests that only one dose could not be su cient to elicit a complete protection in naive subjects.
Furthermore, in these subjects the emergence of viral resistance or escape might be increased if a suboptimal immune response is elicited. Thus far, while a single vaccine dose may be su cient in SARS-CoV-2 experienced subjects, it can be unfavourable in SARS-CoV-2 naive subjects. Also, while the antibody levels observed after the rst dose were inversely correlated with age, thus resulting in a potentially ineffective immunization in subjects at higher risk for developing the more severe consequences of SARS-CoV-2 infection, the second dose was effective in boosting the response in all subjects, with no correlation with their age.
Among the 5 SARS-CoV-2 variants tested, only SAF (501Y.V2 lineage B.1.351), signi cantly impacted neutralizing antibody levels in vaccinated subjects, while SARS-CoV-2 T-cell response was not signi cantly affected. Since the beginning of the pandemic, about 100 million individuals were infected by SARS-CoV-2 and neutralizing antibody titres are extremely variable [24]. Those with low titre could be re-infected, and due to the presence of incomplete protective neutralizing antibodies might develop escape variants [25,26]. The accumulation of mutations in RBD and N-terminal domain of Spike protein might be associated to increased escape from neutralization. Mutation at 484 site has been related to the reduction in neutralization sensitivity [27,28]. Spike mutations in the South African 501Y.V2 lineage B1.351 are linked to immune escape from several classes of anti-SARS-CoV-2 monoclonal antibodies, and a signi cant decrease in neutralization titer of plasma from convalescent subjects was observed when sera were challenged with South African variant [29]. Usually, after natural infection or vaccination a polyclonal immune responses arise against multiple antigenic epitopes. Consequently, small numbers of variations in antigen sequence should have little impact on recognition by the immune system, including both NT antibodies and T cells.
Our results show that vaccinated individuals developed an equally effective NT antibodies response against the WT and UK variant, as well as against BZ variant, while the NT response against the EU variant was increased. On the other side, according to previous observation, sera from vaccinated subjects showed a signi cant reduction in SARS-CoV-2 NT antibodies when challenged against the SAF variant. Nevertheless, serum from the majority (almost 80%) of vaccinated subjects maintained a neutralizing activity against this variant that could be effective in preventing the development of severe disease. These results are in contrast with another report showing a higher reduction in neutralization titres against pseudoviruses with S protein from UK, BZ and SAF variants, with respect to WT, and no increase in NT against European variant [30]. Although the binding titer to SAF RBD was highly reduced with respect to WT (about 100 folds), the NT titer was less affected, showing a ve-fold reduction. In addition, we observed that anti-RBD antibody levels, measured in vaccinated subjects with either commercial or in house assays, have a poor correlation with NT titers. This may suggest that anti-S antibodies binding to sites outside the RBD, which are less involved by the mutations reported, could contribute to virus neutralization. Importantly, our studies rely on the use of wild type viruses rather than pseudoviruses. So far, the use of pseudoviruses could give discrepant results with respect to natural strains, as reported [31], likely because the arti cial lentiviral particles cannot resemble the complete biology of actual clinical isolates.
Interestingly, the level of NT antibodies against SARS-CoV-2 variants in SARS-CoV-experienced vaccinated subjects was robust. This preliminary observation might be helpful in the design of vaccination strategies, suggesting that three exposures to SARS-CoV-2 Spike antigen could increase the level of neutralizing antibodies against SARS-CoV-2 variants.
However, and most considerably, our studies on cell-mediated response revealed that SARS-CoV-2 T-cell response is minimally affected by the mutations occurring in SARS-CoV-2 variants, as reported [32], and most subjects tested showed a detectable T-cell response against each virus strain. Differently from Tarke and colleagues, we used inactivated preparations of different SARS-CoV-2 variant isolates in order to recapitulate realistically antigen presenting mechanisms and natural epitope recognition from T cells. The levels of T-cell response and NT antibodies elicited by the vaccine were not correlated; therefore, a low NT titer does not exclude the presence of a protective T-cell response able to control SARS-CoV-2 infection and avoid the development of a severe disease. Furthermore, potential cross-reactive T-cells originally elicited by the common cold coronaviruses may be boosted by the vaccine and contribute to protection. In our study, about 20-30% SARS-CoV-2 naïve subjects showed a T-cell response speci c for S or N proteins. This might be related to the presence of cross-reactive epitopes from common cold coronaviruses that seasonally circulates [12,[33][34][35][36].
In conclusion, the BNT162b2 vaccine appears able to induce roust antibody and T-cell responses in all immunocompetent individuals and pave the way for the clinical evaluation of other mRNA vaccines currently in development [37]. Although the effectiveness of the antibody response may be partially reduced by some mutations naturally arising in virus strains, the T-cell response appears to be marginally affected. Although we analysed the T-cell response induced by BNT162b2 mRNA vaccine, it is likely the other SARS-CoV-2 mRNA vaccine licensed, the mRNA-1273 show a similar broadly reactive T-cell response. An important issue in the future evaluation of SARS-CoV-2 vaccines is the persistence of the memory response induced. This aspect will be prospectively addressed by our ongoing study.

Study design
We designed an observational, longitudinal prospective study to evaluate the immune response elicited by the BNT162b2 vaccine against SARS-CoV-2. The e cacy end-point were the development of SARS-CoV-2-speci c neutralizing antibody and T-cell response. The e cacy evaluation are conducted at shortterm (after the complete vaccination schedule), medium-term (6 months after vaccination); long-term (12 months after vaccination). Here we report results of the short-term evaluation.
Heparin-treated and untreated blood samples were collected before vaccination (T0), 21 days after the rst dose (T1, before second dose administration) and 21 days after the second dose (T2). Blood lymphocyte counts, SARS-CoV-2 S-speci c T-cells, RBD-binding antibody levels and SARS-CoV-2 NT serum titers were determined at each time point.
The primary end-point of the study was the frequency of "full responders" (i.e. subjects developing both T cells speci c for SARS-CoV2 S protein and NT antibody) after the second dose of BNT162b2 vaccine. Secondary end-points were the following: i) the frequency of "full responders" after the rst dose of BNT162b2 vaccine; ii) the magnitude of T cells speci c for SARS-CoV2 S protein, RBD-binding and NT antibodies after the rst and the second dose of BNT162b2 vaccine; iii) the correlation between T-cell response, RBD-binding antibodies and NT antibodies; iv) the changes in total CD4+ and CD8+ T cells, B cells and NK cells after the rst and second dose of BNT162b2 vaccine; 5) the frequency of T-cell and NT antibody responders against SARS-CoV-2 variant strains; 6) the persistence of T-cell response, RBDbinding antibodies and NT antibodies. This last end-point will be analyzed at a later time.
Data about the frequency of subjects developing T cells and NT antibodies in response to the BNT162b2 vaccine were limited, thus this was an exploratory study. The phase 2/3 clinical trial [2] showed a 95% (95% CI: 90-98%) e cacy of the vaccine in protection from SARS-CoV-2 infection. Thus, adopting a conservative approach, we hypothesized a 90% rate of "full responders" developing both T cells speci c for S protein and NT antibody. With a sample size of 150 subjects, the 95% con dence interval of the expected frequency is 84-94%. This correspond to a minimum number 126 "full responders" which should have been detected.
The study (CoVax) was approved by the local Ethics Committee (Comitato Etico Area Pavia) and Institutional Review Board (P-20210000232). All the subjects signed informed written consent.

Lymphocyte counts
Total lymphocytes subpopulation count was determined on whole blood by ow cytometry using BD Lyric ow cytometer, (BD Biosciences, San Jose,CA, USA). After lysis of red blood cells, CD3 + CD4 + , and CD3 + CD8 + T cells, CD16 + /56 + NK and CD19 + B cells/μl were determined by ow cytometry, using BD Multitest™ 6-color TBNK reagent and TruCOUNT tubes (BD Biosciences). Gating strategy was set up on CD45 + and side scatter (SSC). After overnight (or 1 h) blocking with 5% (wt/vol) skimmed milk, the plates were washed and incubated for 1 h with human serum four fold serial dilutions (starting from 1:50), then for 45 min with horseradish peroxidase-labeled goat IgG to human IgG and, nally, for 25 min with 5mg/ml orthophenylendiamine before the addition of 4 N sulphuric acid. The optical density (OD) value of the serum incubated without RBD was subtracted from the OD value of the serum incubated with RBD. Cutoff of 0.100 net OD was calculated based on mean+2SD results of SARS-CoV-2 naive subjects at the serum dilution 1:50. Serum dilution yielding 0.100 net OD value was considered as the RBD-binding serum titer.

T-cell response
Peripheral blood mononuclear cells (PBMC) were isolated from heparin-treated blood by standard density gradient centrifugation. The number of IFNγ-producing spot forming cells (SFC) was determined by ELISpot as previously reported [12]. Brie y, PBMC (2x10 5 /100μl culture medium per well) were stimulated in duplicate for 24 h in 96-well plates (coated with anti-IFN-γ monoclonal capture antibody) with peptide pools (15mers, overlapping by 10 aminoacids, Pepscan, Lelystad The Netherlands) representative of the S and peptide pool (15 mers, overlapping 11 aminoacids; JPT Peptide Technologies GmbH, Germany) representative of Nucleocapsid (N) proteins, at the nal concentration of 0.25 µg/ml. Phytoheamagglutinin (PHA; 5 µg/mL) was used as positive control, and medium alone as negative control. For the evaluation of cell-mediated response against the SARS-CoV-2 variants, supernatant of Vero E6 cells cultured in 25cm 2 asks and infected for 72 h with 100 TCID 50 /ml of different SARS-CoV-2 strains was UV-inactivated for 1 hour and used as antigen formulation. PBMC (4x10 5 /100μl culture medium per well) were cultured in duplicate in the presence of inactivated SARS-CoV-2-infected (or mockinfected as negative control) cell supernatant diluted 1:10 in culture medium. Culture medium was RPMI 1640 supplemented with 2mM L-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin, and 10% of heat inactivated fetal bovine serum (FBS, Euroclone). Spots were counted using an automated AID ELISPOT reader system (AutoImmun Diagnostika GmbH, Strasburg, Germany). The mean number of spots from duplicate cultures were adjusted to 1 x 10 6 PBMC. The net spots per million PBMC was calculated by subtracting the number of spots in response to negative control from the number of spots in response to the S or N antigen. Responses ≥10 net spots/million PBMC were considered positive based on background results obtained with negative control (mean SFC+2SD).
Characterization of CD4 + , CD4 + follicular helper (T FH ), and CD8 + T-cell proliferative response To evaluate T-cell subsets proliferation, PBMC (600,000/200μl culture medium per well) collected from 19 vaccinated subjects, 9 subjects who experienced mild SARS-CoV-2 infection and 5 unexposed control subjects were stimulated in triplicate in 96-well round-bottom plates with peptide pools representative of the S and N proteins, at the nal concentration of 0.1 µg/ml for 7 days. Peptide pool from human actin, was used as negative control antigen. Culture medium was RPMI 1640 supplemented with 2mM Lglutamine, 100 U/ml penicillin and 100 μg/ml streptomycin, 5% of heat inactivated human serum AB, 1 mM Sodium Pyruvate, 100 μM non-essential amino acids and 50µM 2-Mercaptoethanol. After culture, cells were washed with PBS 0,5µM EDTA and stained in PBS with Live/Dead Fixable Violet Dye (Invitrogen) at 4°C. After washing, cells were stained at room temperature in PBS 5% FCS with anti-CXCR5, followed by anti-IgG2b (biotinylated) and, subsequently, with Streptavidin BV421, CD3 PerCP 5.5, CD4 APC Cy7, CD8 FITC, CD25 PECy7, CD278 (ICOS) APC antibodies. Finally, cells were washed and suspended in 1% paraformaldehyde. The frequency of CD25 + ICOS + expanded CD3 + CD4 + , CD3 + CD4 + CXCR5 + and CD3 + CD8 + T-cells was determined by subtracting the frequency of PBMC incubated with actin peptides from the frequency of PBMC incubated with SARS-CoV-2 S and N peptides.