key: cord-0862083-wln66iw8
authors: Foti, James J.; Lema, Kevin; Strickland, Justin; Tjon, Emily; Li, Adrienne; Rivera, Amalia; Cabral, Crystal; Cormier, Laura; Dowal, Louisa; Rao, Sudhir; Vemulapalli, Vijetha; Flechtner, Jessica B.
title: The ATLAS™ screening assay reveals distinct CD4+ and CD8+ SARS-CoV-2 antigen response profiles which have implications to Omicron cellular immunity
date: 2022-05-17
journal: bioRxiv
DOI: 10.1101/2022.05.17.491668
sha: 4f405f2dfa1d88a0be5a511652322e776a15af28
doc_id: 862083
cord_uid: wln66iw8

The emergence of SARS-CoV-2 variants are a persistent threat to the efficacy of currently developed prophylactic vaccines and therapeutic antibodies. These variants accumulate mutations in the spike protein which encodes the epitopes necessary for neutralizing antibody binding. Moreover, emerging evidence suggest that robust antibody responses are insufficient to prevent severe disease and long-lasting viral immunity requires T cells. Thus, understanding how the T cell antigen landscape evolves in the context of these emerging variants remains crucial. T cells responses are durable and recognize a wider breadth of epitopes reducing the possibility of immune escape through mutation. Here, we deploy the ATLAS™ assay which identifies CD4+ and CD8+ T cell antigens by utilizing the endogenous HLA class-I and class-II peptide processing pathways. Profiling of T cells from exposed and unexposed donors revealed rich and complex patterns which highlighted the breadth of antigenic potential encoded in SARS-CoV-2. ATLAS revealed several common or frequent antigenic regions as well as an abundance of responses in the unexposed cohort potentially the result of pre-exposure to related coronaviruses. ORF10 was a common CD4+ response in the unexposed cohort while spike was identified as a common and frequent target in both cohorts. Moreover, the spike response profiles allowed us to accurately predict the impact of Omicron spike mutations. This analysis could thus be applied to study the impact of future emerging VOCs.

CoV-1 and MERS-CoV viruses could quickly be adopted to guide the development of SARS-CoV-2 64

prophylactic vaccines and recombinant antibody treatments (Salvatori et al., 2020) . The current 65 generation of vaccines and therapies have thus far proven to be extremely effective in limiting the 66 incidence and severity of the disease ( The in-depth profiling of SARS-CoV-2 recognition presented here identifies underappreciated antigenic 108 hot-spots which may inform the design of next-generation COVID-19 vaccines. In addition, ATLAS could 109 be used as an Immunomonitoring tool during the development of these new vaccines. Finally, these 110 studies might help predict the impact of mutations encoded in subsequent VOCs which will likely continue 111 to emerge in the future. 112 Patient enrollment, sample collection, and processing 114 Peripheral blood mononuclear cells (PBMCs) were enriched and frozen from consented donors and 115 classified into three cohorts: unexposed, severe, and mild. For the exposed cohorts, subjects were PCR 116 confirmed donors and samples were collected 1 week to 9 months post-disease resolution. Severe and 117 mild donors were recruited in roughly equal numbers and categorized as severe if they required 118 hospitalization and mild if they did not. Unexposed donor PBMCs were collected before mid-2019. Donor 119 demographic information is summarized in S1 Table. All samples were purchased from vendors who 120 adhered to ethical practices for donor consent, compensation, and sample collection. 121

To generate a SARS-CoV-2 plasmid library for use in the ATLAS assay (described below and Fig 1) , a 123 consensus sequence was derived by aligning 1676 sequences available on NCBI as of May 2020 to the 124 SARS-CoV-2 reference sequence (NC_045512) using the Geneious Prime software (Biomatters). A panel 125

of DNA fragments were designed to profile the entire SARS-CoV-2 proteome along with several control 126 polypeptides (S2 Table) . were codon optimized prior to synthesis and cloning (Twist Biosciences) in-frame with a red fluorescent 136 protein (Fresno RFP; Atum Bio). 137

The ATLAS assay was previously described ( 

Where: 167 = normalized cytokine value for a fragment 168 = natural log of the raw cytokine value for a fragment 169

Median absolute deviation of (NG) = (� − � �) 170

Profiling of the exposed and unexposed cohorts using the ATLAS library depicted in Fig 1B revealed a wide  173 breadth of responses throughout the SARS-CoV-2 proteome with every fragment producing a CD4 + and/or 174 CD8 + response (Fig 2) . Responses to ATLAS fragments were also observed for both T cell types in all orfs 175 except for orf7b to which no donor had a CD8 + response. However, there was a skew towards S, M, and 176

N as the number of responses per fragment for these orfs was larger (CD4 + = 11 and CD8 + = 11) than the 177 rest of the proteome (CD4 + = 7 and CD8 + = 7). Biases were also observed when comparing the response 178

profiles of the T cell subsets. For example, several orfs (orf3a, E, M, ORF7b, ORF8, and ORF10) and the C-179

Terminus of N were almost exclusively CD4 + targets. Additional biases were observed for orf1ab, CD4 + 180 responses were localized to a few fragments in the orf1a region (NSPs 1-10) but were spread throughout 181 the orf1b region (NSPs 12 -16) while the reciprocal pattern was observed for CD8 + cells. 182

Several common and frequent CD4 + and/or CD8 + SARS-CoV-2 antigens were identified in both cohorts ( Fig  183  2 and S3 Table) . Common and frequent antigens were identified in orf1ab, N, and spike whereas M and 184

ORF10 only encoded common CD4 + antigens. Profiling of orf1ab identified a common CD4 + antigen in both 185 orf1a and orf1b regions whereas frequent CD8 + antigens were identified in both regions along with a 186 common antigen in orf1b. 187

When analyzing the response pattern of the unexposed cohort, multiple CD4 + (median = 6) and CD8 + 189

(median = 8) responses were detected with all donors having at least one response (Fig 2) . A wide breadth 190 of responses was also observed in this cohort, with 72% and 90% of the fragments inducing at least one 191 CD4 + and CD8 + response, respectively. Unexpectedly, a comparison of unexposed and exposed profiles 192 revealed both cohorts responding to similar regions of the proteome at comparable frequencies (S3 193  Table) . The strong similarities in response profiles resulted in the inability of our statistical analyses to 194 stratify donors based on SARS-CoV-2 exposure or disease severity (data not shown). Although multiple 195 studies observed the separation of exposed and unexposed cohorts, some unexposed donors in these 196 studies responded to SARS-CoV-2 antigens potentially because of pre-exposure to related coronaviruses 197 (Bacher et al., 2020; Le Bert et al., 2020; Sette and Crotty, 2020). As discussed in more detail below, ATLAS 198 may be fine-tuned to identify otherwise infrequent cross-reactive antigens which might be missed by 199 other methods, and which resulted in the absence of stratification in this study. 200 the C-terminus (Fig 3A) . When mapping the position of the Alpha and Delta mutations onto these 205 fragments, the polymorphisms are primarily located in the first half of the NTD and the central portion of 206 the protein (i.e., the region encompassing the RBM and S1/S2 domains), however, neither variant had 207 more than 3 mutations in any fragment (S3 Table) . Conversely, the number of Omicron mutations 208 mapped to each fragment outnumbered mutations from the other VOCs except for the last fragment 209 which had none. In addition, 12 Omicron mutations mapped to fragments which encode the RBD (S_316-210 492 and S_473-649) with a particularly striking density in the second half of the RBM, the region critical 211 for ACE2 binding. All eight of these mutated RBM residues are encoded in S_473-649 whereas 3 are in 212 S_316-492 ( Fig 3B) . 213

As observed in the rest of genome, the ATLAS recognition profile of unexposed and exposed individuals 214 was overlapping (Fig 3C) 

Most of the studies reported so far used in silico tools to predict potential antigenic peptides or used 225 overlapping peptides. Using these approaches, approximately 1500 SARS-CoV-2 derived T cell epitopes 226

have been described (reviewed in (Grifoni et al., 2021) ). To further enhance our understanding of the 227 cellular responses to SARS-CoV-2, we deployed ATLAS, which is an ex vivo assay capable of identifying 228

CD4 + and CD8 + T cell antigens without prior knowledge of the donors' HLA repertoire or putative antigenic 229 sequences. ATLAS capitalizes on natural processing and presentation pathways (reviewed (Rock et al., 230 2016)) and can use peptides hundreds of amino acids long. As a result, it maximizes the ability of the 231 immunoproteasome to generate and present optimally processed peptides compared to assays which 232 rely exclusively on exogenous loading of synthetic peptides. Together, these factors could have 233

contributed to the identification of additional ross-reactive antigens in the same regions which resulted 234

in the homogenizing of response profiles and the absence of patterns which differentiated donors based 235 on exposure or disease severity (refer to Fig 2) . The work presented here allowed us to make several useful observations that may guide understanding 258 the impact that Omicron and possibly future VOCs mutations have on the ability of CD4 + and CD8 + T cells 259

to recognize spike. The S_316-492 fragment contains 3 of Omicron's 8 RBM mutations, and the reference 260 strain did not elicit any significant CD8 + T cell reactivity. Hence, it may be safe to assume that, unless they 261 introduce novel antigens, mutations in this area will bear no functional significance for CD8 + T cells. In 262 contrast, the CD8 + T cell immunoprevalent S_473-649 fragment contains all the Omicron RBM mutations, 263

with 5 being uniquely encoded within a span of 13 amino acids. In the region spanning those 5 mutations 264

(refer to Fig 3b) handful of potential HLA combinations. It is also possible that mutations could produce sequences with 275 enhanced peptide processing and presentation resulting in the generation of beneficial epitopes. 276

Therefore, the overall impact to both CD8 + and CD4 + recognition of spike, and the ability to mount an 277 immune response to the RBM/RBD domain might be minimal. 278

Taken together, the data presented here suggests that, at the population level, the natural cellular 279

immunity to Omicron, especially as it relates to the ability to recognize spike and contribute to the 280 generation of neutralizing antibodies against the RBM, should not be significantly different between 281

Omicron and the reference strain. Mutations from the Omicron (ο), Delta (δ), and Alpha (α) VOCs are shown as red, cyan, and black barcodes, respectively.

SARS-CoV-2 T cell immunity: Specificity, function, durability, and 291 role in protection

Low-Avidity CD4+ T Cell Responses to SARS-CoV-2 in 294 Unexposed Individuals and Humans with Severe COVID-19

Update on and Future Directions for Use of Anti-SARS-CoV-2 Antibodies: National Institutes of 297 Health Summit on Treatment and Prevention of COVID-19

Coding potential and sequence conservation of 299 SARS-CoV-2 and related animal viruses

Broadly neutralizing antibodies overcome SARS-CoV-2 Omicron antigenic shift

SARS-CoV-2 Omicron has extensive but incomplete escape of Pfizer 305 BNT162b2 elicited neutralization and requires ACE2 for infection. medRxiv

Comparative tropism, replication kinetics, and cell damage profiling of SARS-CoV-2 and SARS-308

BNT162b2-elicited neutralization of B.1.617 and other SARS-CoV-2 variants

Striking Antibody Evasion Manifested by the Omicron Variant of SARS-CoV-2. bioRxiv

Are 363 pangolins the intermediate host of the 2019 novel coronavirus (SARS-CoV-2)?

Effectiveness and safety of SARS-CoV-2 vaccine in real-world 366 studies: a systematic review and meta-analysis

Identification of novel virus-specific antigens by CD4⁺ and CD8⁺ T 369 cells from asymptomatic HSV-2 seropositive and seronegative donors

Omicron-specific cytotoxic T-cell responses are boosted 372 following a third dose of mRNA COVID-19 vaccine in anti-CD20-treated multiple sclerosis patients. 373 medRxiv

Selective and cross-reactive SARS-CoV-2 T cell epitopes in unexposed 376 humans

Molecular basis of immune evasion by the Delta and Kappa

CoV-2 variants

Mutations of SARS-CoV-2 spike protein: Implications on immune evasion and vaccine-induced immunity. 382 Seminars in immunology

SARS-CoV-2 B.1.617.2 Delta variant replication and immune evasion

Improving Cancer Immunotherapies 387 through Empirical Neoantigen Selection

SARS-CoV-2 Variants: Mutations and Effective Changes

Resolution of Chlamydia trachomatis Infection

Minimal cross-over between mutations associated with Omicron variant 395 of SARS-CoV-2 and CD8+ T cell epitopes identified in COVID-19 convalescent individuals. bioRxiv

CD8+ T-Cell Responses in COVID-19 Convalescent Individuals Target

Conserved Epitopes From Multiple Prominent SARS-CoV-2 Circulating Variants. Open Forum Infectious 400 Diseases 8

Present Yourself! By MHC Class I and MHC Class II Molecules

Enhanced fusogenicity and pathogenicity of SARS-CoV-2 Delta P681R mutation. 405 Nature

CoV-2 SPIKE PROTEIN: an optimal immunological target for vaccines

An Effective COVID-19 Vaccine Needs to Engage T Cells

Pre-existing immunity to SARS-CoV-2: the knowns and unknowns

Adaptive immunity to SARS-CoV-2 and COVID-19

SARS-CoV-2 Delta VOC in Scotland: 415 demographics, risk of hospital admission, and vaccine effectiveness

SARS-CoV-2: Emergence of New Variants and 418 Effectiveness of Vaccines

Early induction of functional SARS-CoV-2-specific T cells associates with 421 rapid viral clearance and mild disease in COVID-19 patients

SARS-CoV-2 vaccination induces immunological memory able to cross-recognize 424 variants from Alpha to Omicron. bioRxiv

Impact of SARS-CoV-2 variants on the total CD4+ and CD8+ T cell reactivity in 427 infected or vaccinated individuals

Effectiveness of mRNA BNT162b2 COVID-19 vaccine up to 6 430 months in a large integrated health system in the USA: a retrospective cohort study

Detection of a SARS-CoV-2 variant of concern in South Africa

Waves 436 and variants of SARS-CoV-2: understanding the causes and effect of the COVID-19 catastrophe

Circulating SARS-CoV-2 spike N439K variants maintain 440 fitness while evading antibody-mediated immunity

Structural insights into coronavirus entry

Progress of the COVID-19 444 vaccine effort: viruses, vaccines and variants versus efficacy, effectiveness and escape

Immunology of COVID-19: Current State of the Science

Increased resistance of SARS-CoV-2 variant P.1 to antibody neutralization

Antibody resistance of SARS-CoV-2 variants B.1.351 and B.1.1.7

S2 Table ATLAS Fragment DNA and Protein Sequences 456 S3 Table Mutations, Percent Responses, and Antigen Classification

The authors are grateful for the donors who provided samples for our research study. They also thank 287 members of Genocea's ATLAS team including James Loizeaux, Madison Milaszewski, Oscar Cabrera, and 288Matt Rafferty. 289