SARS-CoV-2 speci c T cells induced by both SARS- CoV-2 infection and mRNA vaccination broadly cross-recognize omicron


 The SARS-CoV-2 variant of concern (VOC) omicron (B1.1.529) is associated with high infectivity and efficient evasion from humoral immunity induced by previous infection or vaccination. In omicron-infected individuals who have been vaccinated or infected before, severe disease seems to be relatively infrequent pointing towards protection by previously primed SARS-CoV-2-specific T cells that cross-recognize omicron. By performing a comprehensive in-depth comparison of the SARS-CoV-2-specific T cell epitope repertoire after natural infection versus after mRNA vaccination, we here demonstrate that spike-derived epitopes are not dominantly targeted in convalescents compared to non-spike epitopes. In vaccinees, however, we detected a broader spike-specific T cell response compared to convalescents reflected by a more diverse repertoire of dominantly targeted spike-specific T cell epitopes. Booster mRNA vaccination induced a broader spike-specific T cell response in convalescents but not in vaccinees with complete initial vaccination. In convalescents and vaccinees, the targeted T cell epitopes are broadly conserved between ancestral and omicron SARS-CoV-2 variants. Hence, our data emphasize the relevance of mRNA vaccine-induced spike-specific CD8+ T cell responses in combating emerging SARS-CoV-2 VOC including omicron and support the benefit of also boosting convalescent individuals with mRNA vaccines.

convalescents reflected by a more diverse repertoire of dominantly targeted spike-specific T 48 cell epitopes. Booster mRNA vaccination induced a broader spike-specific T cell response in 49 convalescents but not in vaccinees with complete initial vaccination. In convalescents and 50 vaccinees, the targeted T cell epitopes are broadly conserved between ancestral and omicron 51 SARS-CoV-2 variants. Hence, our data emphasize the relevance of mRNA vaccine-induced 52 spike-specific CD8+ T cell responses in combating emerging SARS-CoV-2 VOC including 53 omicron and support the benefit of also boosting convalescent individuals with mRNA 54 vaccines. Continuously emerging variants of concern (VOC) sustain the pandemic state of SARS-CoV-61 2. Indeed, the current SARS-CoV-2 VOC omicron (B.1.1.529) is a global health care threat 62 due to its substantially increased infectivity and efficient ability to evade neutralisation by spike-63 specific antibodies induced by previous infection or vaccination [1][2][3][4][5][6] . Indeed, an approx. 5 to 25-64 fold higher concentration of neutralising antibodies is required to abolish infectivity of omicron 65 compared to the delta VOC. Both, increased infectivity and evasion from the humoral response 66 is at least partially based on mutations in the receptor binding domain (RBD) resulting in 67 enhanced affinity to ACE2 receptor and disruption of antibody binding sites 1 . Multiple additional 68 mutations in the spike protein of omicron confer the risk of a broad evasion from the immune 69 response that was readily established through vaccination and previous infection. This 70 assumption is in line with the reportedly higher risk of break through infection with omicron in 71 convalescents and vaccinees with complete initial immunization. Severe disease after omicron 72 infection, is, however, not common and may be prevented by broadly cross-reactive cellular 73 immunity 7 . Indeed, an important role of vaccine-and infection-induced T cells in preventing severe COVID-19 has been reported prior to the emergence of omicron [8][9][10][11] . Currently, very 75 little information is available about the evasion of omicron from cellular immunity although first 76 data indicate that this occurs to a lesser extend compared to evasion from humoral immunity. 77 The SARS-CoV-2-specific T cell epitope repertoire has been studied in some detail 11-16 78 (reviewed in 17 ), however, comparative in-depths studies of the epitope repertoire targeted by 79 infection-versus vaccine-induced T cell responses are so far lacking, hindering the precise 80 prediction of the immune escape potential of emerging VOC including omicron from the T cell 81 response in convalescents compared to vaccinees. 82 To address this important issue, we studied SARS-CoV-2-specific T cell responses in 83 convalescents recovered from natural SARS-CoV-2 infection (n=19) as well as individuals after 84 two (n=16) and three (n=7) doses of SARS-CoV-2 vaccination (Pfizer/BioNTech mRNA 85 vaccine) (Supplementary Table 1). We first mapped the overall SARS-CoV-2-specific CD8+ 86 T cell response and tested a set of 43 previously described immunodominant SARS-CoV-2-87 specific CD8+ T cell epitopes restricted by common HLA class I alleles 11-16 in epitope-specific 88 T cell cultures followed by cytokine staining (intracellular interferon-gamma production). 89 Convalescents displayed CD8+ T cell responses against the majority of epitopes that were 90 distributed over all viral proteins, with spike-specific epitopes definitively not being dominant 91 ( Figure 1A, left column). In vaccinees, in contrast and as expected, CD8+ T cell responses 92 were predominantly directed against spike epitopes ( Figure 1A, right column). Few CD8+ T 93 cell responses targeted non-spike epitopes, with the HLA-B*07/N105-113 epitope being the main 94 target. For this epitope, cross-recognition by T cells against common cold corona viruses has 95 been previously suggested [18][19][20] . Importantly, individual spike-specific CD8+ T cell epitopes 96 were more often targeted in vaccinees compared to convalescents, and the spike-specific 97 CD8+ T cell repertoire also appeared to be broader in vaccinees compared to convalescents. 98 When we compared the corresponding viral sequences between ancestral and omicron SARS-99 CoV-2 variants, only a single tested optimal CD8+ T cell epitope was affected by viral variation 100 ( Figure 1A, shown in red). 101 To further analyze the striking differences in spike-specific CD8+ T cell responses in 102 convalescents versus vaccinees and the nearly complete absence of viral variations in the 103 targeted epitopes in more detail, we analysed these responses using overlapping peptides 104 spanning the whole spike protein (180 18-mer peptides, sliding by 7 amino acids and thus 105 overlapping by 11 amino acids). For all positive responses, we evaluated the overlapping 106 peptide for described optimal epitopes restricted by the HLA class I alleles expressed by the 107 respective individual. If no matching optimal epitopes has been previously described, we 108 performed an in silico analysis to predict the most likely HLA class I restriction and optimal 109 epitope. Using this comprehensive approach, we identified an overall substantially broader 110 repertoire of spike-specific CD8+ T cell responses in vaccinees (Fig. 1B, Figure 1B). Indeed, in convalescents, no HLA class I allele restricted more 113 than two spike-specific CD8+ T cell epitopes, while several HLA class I alleles restricted five 114 or more spike-specific CD8+ T cell epitopes in vaccinees. In addition, we detected more spike-115 specific CD8+ T cell responses per individual in vaccinees compared to convalescents (Fig.  116   1C). Hence, the increased breadth of the spike-specific CD8+ T cell response in vaccinees 117 was evident on an individual and on a population level. 118 To test whether an increased breadth of the spike-specific response in vaccinees is also 119 evident for CD4+ T cells, we analysed the CD4+ T cell response applying the overlapping 120 peptides spanning the whole spike protein as described above. In contrast to the CD8+ T cell 121 response, the spike-specific CD4+ T cell response showed a more limited repertoire of 122 targeted epitopes after vaccination compared to natural infection ( Fig. 1D and Supplementary 123 Figure 1). In particular, fewer spike-specific CD4+ T cell epitopes were restricted by single 124 HLA class II alleles (Fig. 1D) and fewer CD4+ T cell responses were detectable per individual 125 ( Fig. 1E) in vaccinees compared to convalescents. The spike-specific CD4+ T cell repertoire 126 was therefore limited with respect to the individual and the population-based CD4+ T cell 127 response in vaccinees. Still, the fewer targeted spike-specific CD4+ T cell epitopes in 128 vaccinees exhibited high conservation between ancestral and omicron SARS-CoV-2 as it is 129 also the case for the majority of targeted epitopes in convalescents (Fig. 1D, variant epitopes 130 shown in red). Comparing CD8+ and CD4+ T cell response (Fig. 1F), mRNA vaccination 131 appears to particularly broaden and thus increase a cross-reactive spike-specific CD8+ T cell 132

response. 133
Next, to assess the effect of boosting vaccination-or infection-induced T cell responses by 134 mRNA vaccination on the spike-specific CD8+ T cell repertoire, we performed the very same 135 approach as described above using overlapping spike peptides to map spike-specific CD8+ 136 and CD4+ T cell responses in longitudinally followed vaccinees getting their 3 rd vaccine dose 137 mRNA vaccination, we observed a similarly broad and cross-reactive spike-specific CD8+ T 140 cell repertoire and similarly limited but still cross-reactive spike-specific CD4+ T cell repertoire 141 compared to the completed initial immunization with two vaccine doses ( Fig. 2A and  142 Supplementary Fig. 2A). Strikingly, however, we detected CD8+ T cell responses targeting 143 more overlapping peptides after the mRNA boost vaccination in convalescents, representing 144 a broader spike-specific CD8+ T cell repertoire ( Fig. 2B and Supplementary Fig. 2B). In 145 contrast, the CD4+ T cell repertoire was similar before and after boost vaccination of 146 convalescents ( Fig. 2B and Supplementary Fig. 2B). Again, the identified CD8+ and CD4+ 147 vaccination increased SARS-CoV-2-specific CD8+ T cell responses targeting conserved 149 regions within the spike protein of omicron in convalescent individuals. 150 To address the question whether the observed broader spike-specific CD8+ T cell repertoire 151 after mRNA vaccination may also be beneficial for potentially emerging future SARS-CoV-2 152 VOC beyond omicron, we analyzed the T cell response targeting highly conserved selective 153 sweep regions in SARS-CoV-2 21 in convalescents versus vaccinees. Selective sweep regions 154 mediate per definition an evolutionary advantage and therefore it is very likely that newly 155 emerging SARS-CoV-2 VOC also harbor high conservation within these regions. Four different 156 selective sweep regions have so far been described in the spike protein of SARS-CoV-2 21 that 157 also exhibit, as expected, a high degree of amino acid homology among the already evolved 158 SARS-CoV-2 VOC ( Fig. 3A/B). Importantly, compared to convalescents more vaccinated 159 individuals showed spike-specific CD8+ T cell responses targeting epitopes within the highly 160 conserved selective sweep regions indicating a spike-specific CD8+ T cell response with 161 focussed targeting of highly conserved regions after vaccination (Fig. 3C). A similarly focussed 162 spike-specific CD4+ T cell response was not evident after vaccination (Fig. 3D). Hence, a 163 broadly cross-recognizing spike-specific CD8+ T cell response is induced after mRNA 164 vaccination that may be also reactive towards emerging SARS-CoV-2 VOC in future beyond 165 omicron. 166

167
In conclusion, our data indicate that (1) convalescents target a variety of SARS-CoV-2-specific 168 CD8+ T cell epitopes over the complete SARS-CoV-2 proteome with spike-specific CD8+ T 169 cell responses being not dominant; (2) in contrast to the CD4+ T cell response, CD8+ T cell 170 responses in vaccinees are focussed on a broader repertoire of highly conserved spike-171 specific CD8+ T cell epitopes leading to an increased cross-recognizing potential; (3) boosting 172 convalescents with mRNA vaccination results in a broader spike-specific CD8+ T cell 173 response; and (4) CD8+ and CD4+ T cell responses in both, convalescents as well as 174 vaccinees, target epitopes that are highly conserved between ancestral, omicron and 175 potentially future emerging SARS-CoV-2 variants. Hence, our data emphasize the relevance 176 of mRNA vaccine-induced spike-specific CD8+ T cell responses in combating emerging SARS-177 CoV-2 VOC including omicron and support the benefit of also boosting convalescent 178 individuals with mRNA vaccines. 179 180

Data availability 181
All requests for raw and analyzed data and materials are promptly reviewed by the University 182 of Freiburg Center for Technology Transfer to verify if the request is subject to any intellectual 183 property or confidentiality obligations. Patient-related data not included in the paper were 184 generated as part of clinical examination and may be subject to patient confidentiality. Any 185 data and materials that can be shared will be released via a Material Transfer Agreement. 186 187

Acknowledgements 188
We thank all donors for participating in the current study and FREEZE-biobank-Center for 189 biobanking of the Freiburg University Medical Center and the Medical Faculty for support. This   acids with a free amine NH2 terminus and a free acid COOH terminus with standard Fmoc 31 chemistry and a purity of >70% (Genaxxon Bioscience). Similarly, 60 predescribed SARS-32 CoV-2 specific optimal CD8 + T cell epitopes were synthesized. 33

34
In vitro expansion and intracellular IFNγ staining with overlapping peptides or optimal 35 predescribed CD8 + T cell epitopes 36 In vitro expansion with OLPs or optimal epitopes was performed as follows: 20% of the PBMCs 37 were stimulated with a pool of all 181 SARS-CoV-2 spike OLPs or optimal epitopes (10 μg 38 ml −1 ) for 1 h at 37 °C, washed and co-cultured with the remaining PBMCs in RPMI medium 39 supplemented 20 U ml−1 with recombinant IL-2. On day 10, intracellular IFNγ staining was 40 performed with pooled OLPs (45 pools with 4 OLP each). Therefore, cells were re-stimulated 41 with OLP pools (50 μM), DMSO as negative control or PMA and ionomycin as positive control 42 in the presence of brefeldin A an IL-2. After 5 h of incubation at 37 °C, cells were stained for 43 surface markers (CD8+, CD4+; Viaprobe) and intracellular markers (IFNγ). Subsequently, on 44 day 12-14 the single overlapping peptides of positive pools and HLA-matched optimal CD8 + T 45 cell epitopes were tested by intracellular cytokine staining. Viral amino acid sequences of 46 positive individual OLPs were analysed for pre-described minimal epitopes 1-6 or the best HLA-47 matched predicted candidate using the Immune Epitope Database website (using two 48 prediction algorithms ANN 4.0 and NetMHCpan EL 4.123 for 8-mer, 9-mer and 10-mer 49 peptides with half-maximal inhibitory concentration (IC50) of <500 nM). Reference genome of human SARS-CoV-2 (MN908947.3) was downloaded from NCBI 67 database. SARS-CoV-2 epitopes were then mapped to the corresponding protein alignment. 68 Selective sweep regions were indicated as described by Kang et al. 10 . SARS-CoV-2 variants 69 of concern were identified via CoVariants (https://covariants.org/).