key: cord-0857069-av5txnbf
authors: Lee, Hye Kyung; Knabl, Ludwig; Moliva, Juan I.; Werner, Anne P.; Boyoglu-Barnum, Seyhan; Kapferer, Sebastian; Pateter, Birgit; Walter, Mary; Sullivan, Nancy J.; Furth, Priscilla A.; Hennighausen, Lothar
title: mRNA vaccination in octogenarians 15 and 20 months after recovery from COVID-19 elicits robust immune and antibody responses that include Omicron
date: 2022-03-25
journal: Cell Rep
DOI: 10.1016/j.celrep.2022.110680
sha: efeca00e9dd1ba2638f4571cfc24a32f59001048
doc_id: 857069
cord_uid: av5txnbf

Knowledge about the impact of prior SARS-CoV-2 infection of the elderly on mRNA vaccination response is needed to appropriately address the demand for additional vaccinations in this vulnerable population. Here we show that octogenarians, a high-risk population, mount a sustained SARS-CoV-2 spike-specific IgG antibody response for 15 months following infection. This response boosts antibody levels 35-fold upon receiving a single dose of BNT162b2 mRNA vaccine 15 months after recovery from COVID-19. In contrast, antibody responses in naïve individuals boost only 6-fold after a second vaccine. Spike-specific ACE2 antibody binding responses in the previously infected octogenarians following two vaccine doses exceed those found in a naïve cohort after two doses. RNA-seq demonstrates activation of interferon-induced genetic programs, which persist only in the previously infected. A preferential increase of specific IGHV clonal transcripts that are the basis of neutralizing antibodies is observed only in the previously infected nuns.

In contrast, antibody responses in naïve individuals boost only 6-fold after a second 32 vaccine. Spike-specific ACE2 antibody binding responses in the previously infected 33 octogenarians following two vaccine doses exceed those found in a naïve cohort after 34 two doses. RNA-seq demonstrates activation of interferon-induced genetic programs, 35 which persist only in the previously infected. A preferential increase of specific IGHV 36 clonal transcripts that are the basis of neutralizing antibodies is observed only in the 37 previously infected nuns. 38

With the persistence of the COVID-19 epidemic the question of when and who to offer 41 additional vaccinations emerged as a discussion point across geographic areas (Dolgin, 42 2021a, b) . A point of agreement is the vulnerability of the aged to COVID-19 morbidity 43 and mortality (Covino et al., 2021) . There is a body of information on vaccination response 44 both before and after COVID-19 disease in younger, and generally healthy, populations, 45 with median ages ranging from 32 to 47 years old (Ebinger et al., 2021; Goel et al., 2021; 46 Krammer et al., 2021; Lozano-Ojalvo et al., 2021; Sokal et al., 2021; Wang et al., 2021) . 47

While a large-scale study has revealed high and comparable efficacy of the BNT162b2 48 mRNA COVID-19 vaccine in young and older adults (Polack et al., 2020) , at least in the 49 short term, there is limited data available for those in the eighth and higher decades of 50 life (Hyams et al., 2021) . 51

The need for a data-driven approach for optimizing vaccination strategies in the 52 very old population is four-fold. One, this is a group that has disproportionately 53 experienced death due to COVID-19, two, aging of the immune system can be associated 54 with functional declines and poor vaccine responses titers (Bartholomeus et al., 2018; 55 Gonçalves et al., 2019; Sokal et al., 2021) , three, this population has a higher prevalence 56 of co-morbidities, many of which are specifically linked to COVID-19 morbidity and 57 mortality, than younger populations and four, this population is not infrequently domiciled 58 in close living quarters facilitating disease transmission. 59

Tyrol, Austria was an epicenter early in the COVID-19 pandemic, whereby an 60 outbreak in the ski resort of Ischgl in February 2020 led to a seroprevalence of 42% 61 (Borena et al., 2021; Lee et al., 2021b) . Another outbreak took place in a nearby convent 62 neutralizing antibody at the time of receiving the second vaccine dose (5wk (D0)) 155

(reciprocal ID50 GMT of 11; 5 samples not collected), with a single exception that had low 156 levels of neutralizing antibody ( Figure 2E ). By day 7 after the second dose (6wk (D7)), 157 the reciprocal ID50 GMT in this group was 763, a ~68-fold increase from a week prior (5wk 158 (D0)). However, in contrast to the previously infected octogenarians, the naïve vaccinated 159 group experienced a ~3.5-fold decrease (reciprocal ID50 GMT of 223) in neutralizing 160 antibody levels from 6 weeks to 4 months (4mo) ( Figure 2E ). By 4 months, 4/14 naïve 161 vaccinated individuals had no detectable neutralizing antibody against Delta. Overall, the 162 fold change elicited by the second dose, from 5 weeks (D0) to 4 months, in the naïve 163 individuals was significantly higher, ~20-fold, than in the previously infected cohort after 164 the first dose, ~16-fold. However, at 17.5 months (2.5 months after the single vaccination) 165 the previously infected group had a larger magnitude of neutralizing antibody than the 166 naïve group after receiving their second dose. 167

Next, we determined the degree to which antibodies induced in the previously 168 infected octogenarians after one or two vaccinations can prevent binding of the Omicron 169 spike protein to the ACE2 receptor. For this we used the angiotensin-converting enzyme 170 2 (ACE2) binding inhibition assay (Ebinger et al., 2021) . ACE2 binding inhibition 171 approached 100% for the ancestral variant and Alpha, Beta, Gamma and Delta after the 172 second vaccination ( Figure 1I ). A highly significant increase of ACE2 binding inhibition 173 was also observed for the Omicron variant after the second vaccination ( Figure 1I ). In 174 contrast, little ACE2 binding inhibition for Omicron was obtained in the naïve group after 175 the second vaccination ( Figure 2F ). 176

Early responses to vaccination are elevated levels of interferons and other cytokines. To 179 gauge the early vaccine response, we measured serum levels of a panel of cytokines in 180 the previously infected nuns prior to and after the single vaccination, and prior to and after 181 the second vaccination in the naïve group (Figure S4A-D; Data S1, Table S3 ). Out of the 182 10 cytokines measured, a significant increase of circulating IFN- and CXCL10 was 183 observed in both groups within one day following vaccination. CXCL10 (IP-10) is 184 regulated by IFN- (Lee et al., 2021a) and is rapidly induced following vaccination and 185 viral infections and it has been identified as a biomarker reflecting COVID-19 severity 186 (Huang et al., 2005; Laing et al., 2020; Sobolev et al., 2016) . While IFN- levels in both 187 groups returned to baseline levels at day 7, CXCL10 levels had returned to baseline in 188 the naïve group but remained significantly elevated in the COVID-19 recovered group, 189

suggesting an extended inflammatory response. Levels of the other cytokines tested were unchanged in the 191 COVID-19 recovered group (Data S1, Table S3 ). In contrast, in the naïve group, IL-16 192 levels declined post vaccination and IL-8 levels increased at day 7 post vaccination 193 ( Figure S4E ). 194

To further understand the molecular differences to the vaccine response between 195 the two cohorts, we performed bulk RNA-seq on buffy coats from the 16 previously 196 infected nuns prior to and after the first vaccination and from the 14 naïve subjects prior 197 to and after the first and second dose ( Figure 4 ; Data S1, Tables S4-10). RNA-seq was 198 conducted on 115 samples with an average sequencing depth of 240 million reads per 199 sample (Data S1, Table S1 ). The greatest transcriptome differences in the previously 200 infected nuns were observed at day 7 post vaccination ( Figure 4A ). A total of 161 genes 201 were induced at least two-fold within one day of the vaccination ( Figure 4B ; Data S1, 202 Table S4 ), 894 genes were induced by day 7 (Data S1, Table S5 ) and 652 genes were 203 activated between day 1 and day 7 (Data S1, Table S6 ). Gene-set enrichment analysis 204 (GSEA) demonstrated that the genes activated within one day after the vaccination were 205 enriched in immune-response, interferon, and JAK-STAT pathways ( Figure Figure 4D ; Data S1, Table S7 ), only 32 genes were elevated at day 7 as compared to 209 day 0 ( Figure 4D ; Data S1, Table S8 ). A total of 77 genes were activated between days 210 1 and 7 (Data S1, Table S9 ). As expected, the genes activated at day 1 were part of 211 interferon and cytokine pathways ( Figure S5D ). We also investigated the immune 212 transcriptomes prior to and after the primary vaccination of the naïve group and very few 213 induced genes related to chemokine and cytokine signaling were identified ( Figure S6 ; 214 Data S1, Table S10 ). 215

To further understand the stark differences to the first vaccination in the previously 218 infected group and the second vaccination in the naïve cohort, we dug deeper and 219 analyzed the longitudinal expression of the genes activated at day one post vaccination. 220

Out of the 161 genes activated in the previously infected population, expression of 108 221 genes was still significantly activated at day 7 ( Figure 5A ; Data S1, Table S11 ). Forty 222 percent of these genes are part of interferon and virus-response pathways. In contrast, 223 out of the 173 genes induced in the naïve population at day 1, only five genes (IFI44, 224 IFI44L, RSAD2, IFIT1 and GBP1P1) were expressed at elevated levels at day 7 ( Figure  225 5A; Data S1, Table S12 ). These findings suggest a prolonged vaccine-induced 226 transcriptomic response in the previously infected individuals. 227

A direct comparison of genes induced in both groups at day 1 post vaccination 228 identified 58 immune relevant genes shared between the previously infected and SARS-229

CoV-2 naïve population ( Figure 5B ; Figure S7 ; Data S1, Table S11 Figure S8 ; Data S1, Table S11 ). While a total 236 of 548 genes were induced at least two-fold in the previously infected cohort, only 21 237 genes were induced in the naïve population between day 1 and 7 post vaccination. 238

For an initial exploration of T cell immunity, a standard clinical laboratory assay of T cell 241 stimulation was performed in serum four months after vaccination. Data showed that the 242 previously infected cohort with a single vaccine dose demonstrated a significantly higher 243

proportion of definitively positive tests to COVID-19 spike protein (13 out of 16) than the 244 naïve group (1 out of 14) (22%) ( Figure S9 ; Data S1, Table S13 ). Three out of the five S11; Data S1, Table S14 ). RNA-seq was conducted prior to the vaccination (D0) and at 256 seven days (D7) and four months after the vaccination. Immunoglobulin G heavy chain 257 variable (IGHV) genes used in rearrangements of high level CDR3 revealed the use of a 258 broad range of germlines in both cohorts, with a larger breadth in the previously infected 259

persons. In addition, a preferential expansion of transcripts from specific germline genes 260 occurred in the previously infected individuals ( Figure 6A ; Data S1, Table S15 ). Most 261 notably, the IGHV1-2, IGHV1-24, IGHV2-5, IGHV3-13, IGHV3-30, IGHV3-33, IGHV5-51, 262 IGHV3-53/3-66, IGHV4-31 and IGHV3-7 clonal transcripts increased more than 15-fold 263 within one week after the single vaccination of the previously infected group, exceeding 264 that observed in the naïve group ( Figure 6B ; Data S1, Table S15 ). While similar patterns 265 have been observed by others (Andreano et al., 2021b) there are also differences that 266 could be explained by the sequencing depth. Notably, while elevated usage of IGHV2-5 267 in a recent study was restricted to previously infected individuals (Andreano et al., 2021b) 268 we observed some induction also in the naïve group. At four months post vaccination, 269 expression of all transcript classes that were elevated at day 7 had returned to levels 270 seen prior to the vaccination. previously infected nuns after a single vaccine has also been identified in extremely 279 potent monoclonal antibodies from COVID-19 recovered patients (Andreano et al., 280 2021a) . BCR diversity was also measured using the widely used Chao1 biodiversity index 281 that is sensitive to changes in rare species (Chao, 1984) . It significantly increased at D7, 282 both in the recently infected group after the first vaccination and in the naïve group after 283 the second dose ( Figure S12 ). 284

Lastly, we analyzed the CDR3 sequences for shared characteristics in both 285 cohorts. We found nine CDR3 sequences induced in both groups between days 0, 7 and 286 46 in the previously infected group, and at day 39 in the naïve group (Data S1, Table  287 S16). The induction levels were higher in the previously infected group compared to the 288 naïve group. Although median of CDR3 amino acid length are similar in both groups, the 289 previously infected group shows more diversity compared to the naïve group and longer 290 ones were detected distinctively in individual's repertoire (Data S1, Table S16 ). 291

The spike (S) protein is the major surface antigen of SARS-CoV-2 and it uses its 292 RBD to engage the host receptor angiotensin I converting enzyme 2 (ACE2) for viral entry 293 (Zhou et al., 2020) . RBD-targeting antibodies can neutralize SARS-CoV-2 by blocking 294 ACE2 binding. 294 SARS-CoV-2 RBD-targeting antibodies with information on 295 immunoglobulin G heavy-chain variable (IGHV) gene usage have been described (Yuan 296 et al., 2020) . Here we have identified isoforms of IGHV3-30 and IGHV3-33 as well as 297 isoforms of IGHV3-53 and IGHV3-66 that are frequently used in these antibodies ( Figure  298 6). The prevalence of IGHV3-53 has been recognized in COVID-19 patients (Yan et al., 299 2021; Yuan et al., 2020) . 300

In this real-world study, we provide evidence that a single dose of the mRNA vaccine 303

BNT162b2 elicits a strong immune response in an octogenarian population after receiving 304 a single vaccination 15 months after a documented infection with SARS-CoV-2 and 305 recovery from COVID-19. Aging is associated with a decline of the immune system, 306 commonly referred to as immunosenescence, and increased chronic low-grade systemic 307

inflammation, also referred to as inflammaging (Zost et al., 2020) , have been associated 308 with a poor vaccine response (Lozano-Ojalvo et al., 2021) . However, our data 309 demonstrate that the immune response to BNT162b2 in this previously infected elderly 310 population (median 81 years old) exceeds that of a younger naïve cohort (median 59 311 years old) receiving a two-dose regimen. 312

The optimal window for providing the booster vaccine to individuals previously 313 infected with SARS-CoV-2 has not been precisely defined and may be age-314 dependent. Recent studies have investigated the immune response in younger 315 populations recovered from COVID-19 (Andreano et al., 2021a; Hyams et al., 2021; Wang 316 et al., 2021) . In general, the immune responses, including spike-specific IgG antibody 317 levels, in individuals younger than 50 years having received booster doses within one to 318 six months after the original SARS-CoV-2 infection were similar to those seen after two 319 doses of vaccine in individuals of similar age without prior infection (Ebinger et al., 2021; 320 Goel et al., 2021; Hyams et al., 2021) . While our previously infected population has a 321 median age of 81 years, other studies use different definitions of elderly, ranging from 61 322 years (Abu Jabal et al., 2021) to >71 years (Anderson et al., 2020) . A large-scale clinical 323 study provided evidence that natural immune protection that develops after a SARS-CoV- The rise in antibody titers with vaccinations was preceded by a robust induction of 337 IFN- pathway genes that exceeded the response induced by a two-vaccination course 338 in the uninfected slightly younger comparative community group. The strong, 339 prolonged immune response in the previously infected elderly that had received a single 340 vaccination more than one year after infection greatly exceeded that observed after the 341 booster vaccination in naïve individuals. This was seen both in the antibody evolution and 342 in transcriptomes. Exploring the genomic immune responses after vaccination through 343 RNA-seq approaches can identify transcriptional signatures associated with 344 effective antibody production but published data are limited to younger age individuals 345 (Andreano and Rappuoli, 2021; Arunachalam et al., 2021; Lee et al., 2021b) . Sequencing 346 depth is a consideration for interpretation of such studies and a sequencing depth of more 347 than 240 million reads per sample permitted the identification of specific gene signatures 348 in the elderly after a single vaccine dose. This also allowed the identification of specific 349 IGHV germline classes, including IGHV1-69, IGHV1-24, IGHV1-2 and IGHV3-53 that are 350 preferentially expressed in some of the previously infected octogenarians after a single 351 vaccine but induced less in the naïve individuals. These HCV genes are used by several 352 potent neutralizing antibodies (Andreano et al., 2021a) . While the expression of some 353 IGHV germline classes, such as IGHV3-30 and IGHV3-53, is specifically elevated in our 354 octogenarian cohort after the single dose, expression in younger COVID-19 recovered 355 individuals with a median age of 41 years was independent of the vaccination status 356 (Wang et al., 2021) . Longitudinal antibody measurements in our elderly cohort, especially 357 after receiving additional BNT162b2 doses, will provide a better understanding of the 358 need, timing and value of specific vaccination regimen in this vulnerable population. 359

The fast-spreading SARS-CoV-2 Omicron variant has a propensity of immune 360 evasion and breakthrough infections are common (Kuhlmann et al., 2022) . While a third 361 dose of BNT162b2 augments the magnitude of the antibody response to Omicron (Kotaki 362 et al., 2022; Muik et al., 2022; Sievers et al., 2022) , questions remained about the efficacy 363 of previous SARS-CoV-2 infection followed by a two-dose mRNA vaccination in the 364 elderly. Our study demonstrated previous infection followed by two doses of BNT162b2 365 about 1.5 years later resulted in a strong Omicron neutralization based on an ACE2 366 binding inhibition assay, far exceeding that seen in naïve individuals receiving two 367

BNT162b2 doses. A recent study demonstrated that previously infected and vaccinated 368 persons display residual neutralization of Omicron (Cele et al., 2022) , possibly account 369

for the milder disease seen in many individuals. 370

Results from this real-world study are encouraging for vaccine efficacy in 371 previously infected individuals in their 80s and beyond. While the optimal window between 372 previous infection and a booster shot is not known, our study demonstrates that a 15-373 months gap between infection and the first vaccination did not negatively interfere with 374 the immune response but resulted in robust production of antibodies, qualitatively and 375 quantitatively exceeding that of naïve individuals who received two doses. Similarly, a 376 larger interval between vaccination followed by a breakthrough infection correlated with 377 increased neutralization activity against SARS-CoV-2 variants (Miyamoto et al., 2022; 378 Sidik, 2022) . This has practical implications for health care professionals making 379 decisions on the need for booster vaccinations. 

The authors declare not competing interests. 410

We worked to ensure ethnic or other types of diversity in the recruitment of human 413 subjects. We worked to ensure that the study questionnaires were prepared in an 414 inclusive way. While citing references scientifically relevant for this work, we also actively 415 worked to promote gender balance in our reference list. The author list of this paper 416 includes contributors from the location where the research was conducted who 417 participated in the data collection, design, analysis, and/or interpretation of the work. 

• RNA-seq data from this study were generated in the laboratory of the last author 434

and deposited under the accession GSE190747 in the Gene Expression Omnibus 435 (GEO). 436

• Cytokine data displayed in Figure S4 and antibody data shown in Figure 1C -I, 437 cell activation data shown in Figure S9 are listed in Data S1, Table S13. Analyzed 440 immunoglobulin genes in Figure 6 and Figures S10-11 are listed in Data S1, 441

Tables S14-16. 442

• This paper does not report original code. 443

• Additional Supplemental Items are available from Mendeley Data at 444 https://data.mendeley.com/datasets/smhwct443j/1 . 445

• Any additional information required to reanalyze the data reported in this paper is 446 available from the lead contact upon request. 447

Sixteen COVID-19 recovered volunteers who were infected with SARS-CoV-2 and 452 developed COVID-19 in the spring of 2020 and 14 SARS-CoV-2 naïve healthy volunteers 453 and (Data S1, Table S1) were recruited for the study under informed consent. 454

Recruitment and blood sample collection took place between January and August 2021. 

The V-PLEX COVID-19 ACE2 Neutralization kit (Meso Scale Discovery, K15570U) was 497 used to quantitatively measure antibodies that block the binding of ACE2 to its cognate 498 ligands (SARS-CoV-2 and variant spike subdomains). Plates were coated with the 499 specific antigen on spots in the 96 well plate and the bound antibodies in the samples 500

(1:10 dilution) were then detected by Human ACE2 protein conjugated with the MSD 501 SULPHO-TAG which is then read on the MSD instrument which measures the light 502 emitted from the tag. 503

Pseudotyped lentiviral reporter viruses were produced as previously described (Corbett 506 et al., 2020; Wu et al., 2021) . Briefly, HEK293T/17 cells (ATCC CRL-11268) were 507 transfected with the following: [1] a plasmid encoding S protein from Wuhan-Hu-1 strain 508 (GenBank no. MN908947.3) with a p.Asp614Gly mutation (D614G) or a plasmid encoding 509 B.1.617.2 S protein that was altered via site-directed mutagenesis (Delta), [2] a plasmid 510 encoding luciferase reporter, [3] a plasmid encoding the lentivirus backbone, and [4] a 511 plasmid encoding the human transmembrane protease serine 2 (TMPRSS2) gene. 512

Serum, in duplicate, were tested for neutralizing activity against the pseudoviruses by 513 quantification of luciferase activity in relative light units. The percentage of neutralization 514 was normalized, with luciferase activity in uninfected cells defined as 100% neutralization 515 and luciferase activity in cells infected with pseudovirus alone as 0% neutralization. Titers 516 were calculated using a log (agonist) versus normalized-response (variable slope) 517 nonlinear regression model in GraphPad and are reported as the serum dilution required 518 to achieve 50% (50% inhibitory dilution [ID50]) neutralization. The input dilution of serum 519 is 1:20, thus, 20 is the lower limit of quantification. Samples that do not neutralize at the 520 limit of detection at 50% are plotted at 10, and that value was used for geometric mean 521 calculations. 522

T-cell reactivity to SARS-CoV-2 peptides was measured by an ELISPOT assay using a 525 human IFN- kit (Mabtech, Nacka Strand, Sweden) according to manufacturer's 526 instructions. After washing of ELISPOT plate with sterile PBS, conditioning of plate with 527 medium for 30 min and subsequent removal of medium, buffy coat cells were added to 528 wells. Buffy coat cells were stimulated by adding each, SARS-CoV-2 S1 peptide and 529 SARS-CoV-2 N, M, O-peptide mix, followed by incubation in a humified incubator at 37°C 530 and 5% CO2. Anti-CD-28 was added to each well in a concentration of 1µg mL-1 to 531 enhance stimulation. Anti CD3 mAbs served as positive control. Detection of stimulated 532 T-cells was done by adding PBS plus 0.5 fetal calf serum (PBS/0.5% FCS) -containing 533 detection antibodies to each well and incubation for 2 hrs. at room temperature after 534 removal of cells and washing with sterile PBS. Subsequent incubation and washing using 535 PBS, streptavidin-ALP diluted in PBS/0.5% FCS was added. After another incubation of 536 1h at room temperature and washing with PBS, substrate solution was added to each 537 well. After incubation at room temperature till emergence of distinct spots, color 538 development was stopped by thorough washing with tap water. The plate was dried 539 before counting spots by using an AID ELISPOT reader system. Normalized reads were 540 obtained by subtraction of the negative control wells. Results were presented as spot 541

forming colonies per million immune cells in percent. 542

Whole blood was collected, and total RNA was extracted from the buffy coat and purified 545 using the Maxwell RSC simply RNA Blood Kit (Promega) according to the manufacturer's 546 instructions. The concentration and quality of RNA were assessed by an Agilent 547

Bioanalyzer 2100 (Agilent Technologies, CA). 548 mRNA sequencing (mRNA-seq) and data analysis. 550

The Poly-A containing mRNA was purified by poly-T oligo hybridization from 1 mg of total 551 RNA and cDNA was synthesized using SuperScript III (Invitrogen, MA). Libraries for 552 sequencing were prepared according to the manufacturer's instructions with TruSeq 553 The raw data were subjected to QC analyses using the FastQC tool (version 557 0.11.9) (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). mRNA-seq read 558 quality control was done using Trimmomatic (Bolger et al., 2014 ) (version 0.36) and STAR 559 RNA-seq (Dobin et al., 2013) (version STAR 2.5.4a) using 150 bp paired-end mode was 560 used to align the reads (hg19). HTSeq (Anders et al., 2015) (version 0.9.1) was to retrieve 561 the raw counts and subsequently, Bioconductor package DESeq2 (Love et al., 2014) in 562 R (https://www.R-project.org/) was used to normalize the counts across samples and 563 perform differential expression gene analysis. Additionally, the RUVSeq (Risso et al., 564 2014) package was applied to remove confounding factors. The data were pre-filtered 565 keeping only genes with at least ten reads in total. The visualization was done using dplyr 566 (https://CRAN.R-project.org/package=dplyr) and ggplot2 (Wickham, 2009) . The genes 567 with log2 fold change >1 or <-1 and adjusted p-value (pAdj) <0.05 corrected for multiple 568 testing using the Benjamini-Hochberg method were considered significant and then 569 conducted gene enrichment analysis (GSEA, https://www.gsea-570 msigdb.org/gsea/msigdb). 571

For T-or B-cell receptor repertoire sequencing analysis, trimmed fastq files from 572 bulk RNA-seq were aligned against human V, D and J gene sequences using the default 573 settings with MiXCR (Bolotin et al., 2017; Bolotin et al., 2015) . CDR3 sequence and the 574 rearranged BCR/TCR genes were identified. The diversity of BCR/TCR genes was 575 investigated by the Chao1 index (Chao, 1984) . Document S1. Figures S1-S12. 598 599 Data S1. Tables S1-16. 600

Tables S1, participant demographics, related to Figure 1 . 601 Table S2 . Anti-SARS-CoV-2 Spike/Nucleocapsid protein profiles and Neutralizing 603 antibody to variants of all individuals in all cohorts, related to Figures 1-3 . 604 Table S14 . List and number of clones of V genes of heavy and light chains, related to 649 Figure 6 . 650 Table S15 . List and number of IGHV genes rearranged to highly expressed CDR3 652 sequence, related to Figure 6 . 653 654   Table S16 . List and number of induced CDR3 between day0 and day7, related to Figure  655 6. 656 Comorbidities. The comorbidities of the individuals are listed in Data S1, Table S1 . (C-D) 848

Plasma IgG antibody binding the SARS-CoV-2 RBD (spike) from different variants and 849 the N protein. p-value are from 2-way ANOVA with Tukey's multiple comparisons test (C) 850

and unpaired t-test with Welch's correction (D). *p < 0.05, ***p < 0.001, ****p < 0.0001. 851 (5wk (D0), n = 9; 6wk (D7), n = 14; 4 mo, n = 14; 5.5 mo, n = 11) (E) Neutralizing antibody 852 response to the Delta variant (B.1.617.2). p-value are from 1-way ANOVA with Tukey's 853 multiple comparisons test. *p < 0.05. The individual's data are listed in Data S1, Table 854 individual's data are listed in Data S1, Table S2 . Colors in dot plots were matched to ones 867 in timetable in A and C. (Naïve 2 doses (median 73yrs.), n = 5; Naïve 2 doses (median 868 58yrs.), n = 14) 869 genes significantly activated at day7 in COVID-19 recovered (left) and SARS-CoV-2 898 naïve (right) cohorts. The genes induced between day0 and day7 fall into different 899 categories based on their activation pattern (Data S1, Tables S11-12). Group A, genes 900 whose activation progresses between day0 and day7; Group B, genes highly activated 901 preferentially between day0 and day1; Group C, genes highly activated between day1 902 and day7. to the vaccination (Day0) and after 7 days (Day7) and 4 months (Data S1, Table S15 ). 907

The number of sequences analyzed for each individual are shown in the inner circle. 908

Sizes of pie slices are proportional to the number of clonally related sequences. Persisting 909 clones (same IGV genes) in both time points are shown as colored slices. White indicates 910 sequences isolated at single time point. (B) Induction fold activity of specific variable gene 911 classes identified in the three convalescent and three naïve individuals prior to the 912 vaccination and after 7 days (Data S1, Table S15 ). The color code in panels a and b are 913 identical.

Impact of age, ethnicity, sex and prior infection status on 660 immunogenicity following a single dose of the BNT162b2 mRNA COVID-19 vaccine: 661 real-world evidence from healthcare workers

HTSeq--a Python framework to work with 664 high-throughput sequencing data

Safety 667 and Immunogenicity of SARS-CoV-2 mRNA-1273 Vaccine in Older Adults

Extremely potent human 671 monoclonal antibodies from COVID-19 convalescent patients

Hybrid immunity improves B 675 cells and antibodies against SARS-CoV-2 variants

Immunodominant antibody germlines in COVID-677 19

Systems vaccinology of the BNT162b2 680 mRNA vaccine in humans

BNT162b2 vaccine booster 683 dose protection: A nationwide study from Israel. medRxiv

Transcriptome 686 profiling in blood before and after hepatitis B vaccination shows significant differences in 687 gene expression between responders and non-responders

Trimmomatic: a flexible trimmer for 689

Illumina sequence data

Antigen 692 receptor repertoire profiling from RNA-seq data

MiXCR: software for comprehensive adaptive 695 immunity profiling

Persistence of immunity to SARS-698

CoV-2 over time in the ski resort Ischgl

Omicron extensively but 701 incompletely escapes Pfizer BNT162b2 neutralization

Nonparametric estimation of the number of classes in a population

Evaluation of the 706 mRNA-1273 Vaccine against SARS-CoV-2 in Nonhuman Primates

Frailty Assessment in the 710

Emergency Department for Risk Stratification of COVID-19 Patients Aged ≥80 Years

STAR: ultrafast universal RNA-seq aligner

COVID vaccine immunity is waning -how much does that matter? 716

Is one vaccine dose enough if you've had COVID? What the science 718 says

BNT162b2 mRNA vaccine in individuals previously infected with SARS-CoV-2

Distinct antibody and 725 memory B cell responses in SARS-CoV-2 naïve and recovered individuals following 726 mRNA vaccination

Innate gene 729 signature distinguishes humoral versus cytotoxic responses to influenza vaccination

An 732 interferon-gamma-related cytokine storm in SARS patients

Effectiveness of BNT162b2 and ChAdOx1 735 nCoV-19 COVID-19 vaccination at preventing hospitalisations in people aged at least 736 80 years: a test-negative, case-control study

memory B-cells are elicited by two doses of BNT162b2 mRNA vaccine

Antibody Responses in Seropositive Persons after a Single Dose of SARS-CoV-744 2 mRNA Vaccine

Breakthrough 747 infections with SARS-CoV-2 omicron despite mRNA vaccine booster dose

COVID-19 immune signature includes associations with poor prognosis

JAK inhibitors dampen activation of 754 interferon-stimulated transcription of ACE2 isoforms in human airway epithelial cells

Immune transcriptomes of 758 highly exposed SARS-CoV-2 asymptomatic seropositive versus seronegative 759 individuals from the Ischgl community

Moderated estimation of fold change and 761 dispersion for RNA-seq data with DESeq2

Differential effects of the second SARS-CoV-2 mRNA vaccine dose on T cell immunity 765 in naive and COVID-19 recovered individuals

Vaccination-infection interval determines 768 cross-neutralization potency to SARS-CoV-2 Omicron after breakthrough infection by 769 other variants. medRxiv

CoV-2 Omicron by BNT162b2 mRNA vaccine-elicited human sera

Reduced sensitivity of 776

SARS-CoV-2 variant Delta to antibody neutralization

BNT162b2 mRNA Covid-19 Vaccine

Normalization of RNA-seq data 781 using factor analysis of control genes or samples

Immunity against Omicron from breakthrough infection could be a 783 matter of timing

SARS-CoV-2 infection or mRNA vaccines have reduced neutralizing activity against 787 Beta and Omicron pseudoviruses

Adjuvanted influenza-H1N1 790 vaccination reveals lymphoid signatures of age-dependent early responses and of 791 clinical adverse events

mRNA vaccination 794 of naive and COVID-19-recovered individuals elicits potent memory B cells that 795 recognize SARS-CoV-2 variants

Naturally enhanced 798 neutralizing breadth against SARS-CoV-2 one year after infection

Ggplot2 : elegant graphics for data analysis

Serum Neutralizing 802

Activity Elicited by mRNA-1273 Vaccine

Germline IGHV3-53-encoded RBD-targeting neutralizing antibodies are 805 commonly present in the antibody repertoires of COVID-19 patients

Structural basis of a shared antibody response to SARS-CoV-2

Potent and protective IGHV3-53/3-66 public antibodies and 812 their shared escape mutant on the spike of SARS-CoV-2

A pneumonia outbreak associated with a new coronavirus of 815 probable bat origin

Rapid isolation and profiling of 818 a diverse panel of human monoclonal antibodies targeting the SARS-CoV-2 spike 819 protein

n = 15; 17.5 mo, n = 16) (I) Blocking of interactions between

Spike (RBD) interactions was tested by competition ELISA. p-value are from 2-way 841

ANOVA with Sidak's multiple comparisons test. *P < 0.05, ***P

Colors in dot plots were matched to ones

F) Blocking of interactions between ACE2 and SARS-CoV-2 Spike 856 (RBD) interactions was tested by competition ELISA

Figure 4. Immune transcriptomes following vaccination. (A) Principal-component 871

PCA) of transcriptomes generated prior to (Day0) and after the vaccination days 872 1 (Day1) and 7 (Day7) from the 16 previously infected octogenarians. The variation in the 873 global gene expression profiles across the three time points is shown

05) with a log2 fold change (FC) of more than 1 or less than -1 are indicated in 879 red and blue, respectively. Non-significant DEGs are indicated in gray. The numbers of 880 upregulated and downregulated genes are listed in Data S1, Tables S4-6. (C) PCA of 881 transcriptomes from the SARS-CoV-2 antigen

Day7 after the second vaccination

Red dots indicate significant 886 upregulation; blue dots indicate significant downregulation