key: cord-0945475-noa5devk
authors: Boppana, Sushma; Qin, Kai; Files, Jacob K; Russell, Ronnie M.; Stoltz, Regina; Bibollet-Ruche, Frederic; Bansal, Anju; Erdmann, Nathan; Hahn, Beatrice H.; Goepfert, Paul
title: SARS-CoV-2-specific peripheral T follicular helper cells correlate with neutralizing antibodies and increase during convalescence.
date: 2020-10-12
journal: medRxiv
DOI: 10.1101/2020.10.07.20208488
sha: 215541663cb4516a5fc3e30dc869617a983dd3a0
doc_id: 945475
cord_uid: noa5devk

T-cell immunity is likely to play a role in protection against SARS-CoV-2 by helping generate neutralizing antibodies. We longitudinally studied CD4 T-cell responses to the M, N, and S structural proteins of SARS-CoV-2 in 21 convalescent individuals. Within the first two months following symptom onset, a majority of individuals (81%) mount at least one CD4 T-cell response, and 48% of individuals mount detectable SARS-CoV-2-specific peripheral T follicular helper cells (pTfh, defined as CXCR5(+)PD1(+) CD4 T cells). SARS-CoV-2-specific pTfh responses across all three protein specificities correlate with antibody neutralization with the strongest correlation observed for S protein-specific responses. When examined over time, pTfh responses increase in frequency and magnitude in convalescence, and robust responses with magnitudes greater than 5% were detected only at the second convalescent visit, an average of 38 days post-symptom onset. These data deepen our understanding of antigen-specific pTfh responses in SARS-CoV-2 infection, suggesting that M and N protein-specific pTfh may also assist in the development of neutralizing antibodies and that pTfh response formation may be delayed in SARS-CoV-2 infection.

Cases of COVID-19, caused by the novel severe acute respiratory syndrome coronavirus 45 2 (SARS-CoV-2), were first reported in Wuhan, China at the end of 2019 (1). Since then, the 46 COVID-19 pandemic has caused significant morbidity, mortality, and economic disruption 47 worldwide (2). In SARS-CoV-2 infection, initial studies reported significant lymphopenia in 48 hospitalized patients (3). An elevation of both activation and exhaustion markers on T cells in both 49 severe and mild disease has also been described (4-6). More recently, data on antigen-specific 50 T-cell responses in individuals recovered from SARS-CoV-2 infection has emerged. These 51 studies have reported CD4 T-cell responses to SARS-CoV-2 in 80-100% of convalescent 52 individuals, with most publications focusing on the Spike (S) protein (7-10). 53

Several SARS-CoV-2 vaccine efficacy trials are in progress, and recent Phase I/II trial 54 data have highlighted the presence of neutralizing antibodies as evidence of plausible vaccine 55 efficacy (11) (12) (13) . Although the key components of a protective immune response against SARS-56

CoV-2 remain unclear, studies in non-human primates have found that neutralizing antibodies 57 (nAb) are a correlate of protection in infection and vaccination (14, 15) . With this in mind, a better 58 understanding of how T-cell responses contribute to the formation of nAb is critical to optimizing 59 future vaccine design. 60

Because direct study of lymphoid tissues in humans is difficult, peripheral T follicular cells 61 (pTfh), or T follicular helper cells (Tfh) circulating in the blood, serve as an important surrogate for 62 understanding Tfh responses within germinal centers. While there is some controversy regarding 63 how to best identify these cells, there is general consensus that these cells express CXCR5, a 64 lymph node homing receptor, and many groups use PD1 expression in conjunction with CXCR5 65 to define pTfh (16) (17) (18) . While frequencies of circulating CXCR5 + PD1 + CD4 T cells are typically 66 low, these cells are closely linked to Tfh in lymphoid tissue (19) and have been shown to support 67 humoral responses (20, 21) . Antigen-specific pTfh have been shown to correlate with neutralizing 68 antibodies in the context of infection and vaccination of several pathogens (17, (22) (23) (24) (25) (26) . Although 69 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020 . . https://doi.org/10.1101 pTfh responses have not been described in the context of SARS-CoV or MERS-CoV infection, 70

CD4 T-cell responses have been shown to be important in controlling SARS-CoV in mouse 71 models (27), and a recent study of a MERS-CoV vaccine in mice found that Tfh frequencies in 72 draining lymph nodes correlated with neutralizing antibodies (28) . 73

Data on SARS-CoV-2-specific T follicular helper cell responses are also limited. 74 Thevarajan et al. was the first to report on pTfh frequencies in SARS-CoV-2, and found that 75 frequencies of total pTfh increased during acute infection (29). Since then, a few studies have 76 drawn a correlation between total CD4 T cell or total Tfh-like cell frequencies and antibody levels 77 (30, 31). Another study found increased expression of CXCR5 and ICOS, two Tfh markers, on 78 SARS-CoV-2-specific CD4 T-cells but did not examine pTfh responses directly (32). In deceased 79 donors with COVID-19, Kaneko et al. recently reported that BCL6-expression in germinal center 80

Tfh was lost within thoracic lymph nodes. This study suggests that Tfh response formation may 81 be impaired in severe SARS-CoV-2 infection (33), but how this affects the formation of antigen 82 specific Tfh responses is unclear. 83

The most direct examination of pTfh to date was conducted by Juno et al, where circulating 84 Tfh in the blood were defined as CD45RA -CXCR5 + CD4 T cells. They demonstrated a correlation 85 between S protein-specific pTfh and nAb, suggesting that Tfh responses are formed in mild 86 SARS-CoV-2 infection (34). However, these data leave several questions unanswered, including 87 at what point in convalescence these responses evolve and whether Tfh specific for other SARS-88

CoV-2 proteins contribute to the formation of neutralizing antibodies. While this study was a useful 89 first glimpse at antigen-specific Tfh responses, it did not examine PD1 expression, a canonical 90

Tfh marker, and used the activation markers, Ox40 and CD25, to identify antigen-specific 91 responses, which have previously been shown include a high percentage of T regulatory cells 92 (35). It is also important to note that pTfh specificity does not necessarily correspond with 93 neutralizing antibody specificity. For example, in HIV infection and vaccination, intrastructural help 94 occurs, where CD4 T-cell responses to internal, structural proteins correlated with neutralizing 95 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 12, 2020 . . https://doi.org/10.1101 antibodies against the exterior, envelope protein (36, 37) . These studies underscore the 96 importance of examining pTfh responses across the SARS-CoV-2 proteome. 97

Here, we report on SARS-CoV-2-specific CD4 T-cell responses to the membrane (M), 98 nucleocapsid (N), and spike (S) proteins studied longitudinally in 21 convalescent individuals. We 99 directly examined antigen-specific pTfh (CXCR5 + PD1 + CD4 T cells) and observed correlations 100 between antigen-specific pTfh responses across all protein specificities and antibody 101 neutralization, with the strongest correlation observed for S protein-specific pTfh frequencies. 102

High magnitude SARS-CoV-2-specific pTfh responses (>5% activation of total pTfh population) 103

were only detected at the second convalescent visit, more than 30 days following symptom onset. 104

These data are the first to examine the kinetics of pTfh responses that arise after SARS-CoV-2 105 infection as well as the relationship between neutralizing antibodies and pTfh responses to the 106 SARS-CoV-2 M and N proteins. These results also suggest that pTfh formation may be delayed 107 in SARS-CoV-2 infection. Table 1) . 121 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020. . https://doi.org/10. 1101 We utilized two flow cytometry-based strategies: 1) upregulation of activation-induced markers 122 (AIM), and 2) production of effector molecules by intracellular cytokine staining (ICS). Gating 123 strategies for AIM and ICS in an unstimulated, negative control are shown in S1 Fig. 124 responses against all three SARS-CoV-2 proteins. At the first convalescent visit, we found that 129 57% (12/21) of individuals mounted a SARS-CoV-2-specific CD4 T-cell response by AIM and that 130 these CD4 responses targeted all three tested proteins with similar frequencies (Fig 1C) . 131

Meanwhile, by ICS, 47% (10/21) of individuals had at least one SARS-CoV-2-specific CD4 132 response at this visit, with a similar distribution across the tested proteins (Fig 1D) . As a control, 133

we also measured T cell responses to SARS-CoV-2 peptide pools in COVID-19 negative 134 individuals by assaying samples collected from healthy individuals before the COVID-19 135 pandemic. In the healthy controls tested, we detected three low magnitude (≤0.17%), presumably 136 cross-reactive memory CD4 T-cell responses in two of the ten tested individuals (20%) in line with 137 previously published reports (8). Representative staining of an AIM-detected and an ICS-detected 138 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020. pTfh responses in 4 of the 21 individuals tested (19%) and equally spread across each of the 157 three proteins (Fig 2B) . These data indicate that only a minority of individuals have mounted 158 detectable SARS-CoV-2-specific pTfh responses early in convalescence. However, previous 159 studies on pTfh responses have rarely calculated responder rates, and, therefore, it is difficult to 160 conclude whether this responder frequency is atypical. 161 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020. . https://doi.org/10. 1101 Meanwhile, none of the healthy controls tested had detectable SARS-CoV-2-specific pTfh 162 responses. This lack of pTfh responses in COVID negative individuals is not surprising, as pTfh 163 compose a minor population of the total CD4 T cells in the blood and pTfh responses induced by 164 other seasonal coronaviruses, if present, are likely to exist at very low, undetectable frequencies. 165 Additionally, the fact that these responses were only detected in convalescent individuals 166 bolsters our confidence that these pTfh responses were induced by recent SARS-CoV-2 infection 167 and do not represent cross-reactive, memory responses. Because pTfh are important for the development of an antibody response, we investigated 172 whether the frequency of SARS-CoV-2-specific pTfh correlated with antibody level and 173 neutralization at the first convalescent visit. We used two measurements of antibodies: The first 174 was the commercially available Abbott test that detects N protein-specific IgG. The second assay 175 measured antibody neutralization by luciferase expression and is likely a more biologically 176 relevant metric because neutralizing antibodies have been shown to correlate with protection in 177 preclinical studies (14, 15) . For all three proteins, we see a similar level of significant correlation 178 between the antigen-specific pTfh frequency and N protein IgG titer (Fig 3A) . However, we find 179 that pTfh frequencies across proteins differentially correlate with antibody neutralization (Fig 3B) : 180 S protein-specific pTfh responses most strongly correlate with nAb (p < 0.0001, r = 0.75), followed 181 by M protein-specific ones (p = 0.001, r = 0.66), and finally N protein-specific pTfh (p = 0.02, r = 182 0.52). To ensure these correlations were specific to SARS-CoV-2-induced responses, we 183 quantified the frequency of total pTfh (CXCR5 + PD1 + ). We did not see any correlation between the 184 overall frequency of pTfh and antibody levels or neutralization (Fig 3C) . Taken together, these 185 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020. protein where the CD4 T-cell response rate increased from 38% to 57% and the pTfh response 196 rate increased from 10% to 33% (Fig 4A) . Additionally, M protein-specific CD4 T-cell and pTfh 197 response magnitudes by AIM trended up from the first to second visit (p = 0.09 and p = 0.07, 198 respectively), while other antigen-specific subsets appeared at similar magnitudes at both 199 timepoints (Fig 4B) . is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020. . https://doi.org/10. 1101 In this study, we longitudinally examined the CD4 T-cell responses targeting the major 212 SARS-CoV-2 structural proteins, M, N, and S, in 21 convalescent individuals by measuring the 213 expression of activation marker and the production of effector cytokines. We found that at the first 214 convalescent visit, antigen-specific pTfh responses could be detected against all three proteins 215 and that the frequency of antigen-specific pTfh in these individuals correlated with nAb, albeit to 216 varying degrees. We also found that pTfh responses increase over time in convalescence and 217 that truly robust pTfh responses (>5% frequency) were only detected at a second, later visit. 218

The relative weakness of the correlation between N protein-specific pTfh frequency and 219

antibody neutralization compared to the M and S proteins may relate back to the structure of 220 SARS-CoV-2. Both the spike and membrane proteins have portions that are located exteriorly, 221 while the nucleocapsid protein is found exclusively internally. Collectively, these data suggest that is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020. . https://doi.org/10.1101/2020.10.07.20208488 doi: medRxiv preprint second visit for all individuals assessed in this study occurred an average of 38 days following 238 symptom onset. A delay in pTfh response formation could be due to the T-cell dysfunction that 239

acute SARS-CoV-2 infection (4, 5, 39), and our group has recently illustrated that this dysfunction 241 is sustained during convalescence, even in non-hospitalized individuals (6) frequency at the first visit was 48% (10/21), but if using three times the background, the CD4 260 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020. . https://doi.org/10.1101/2020.10.07.20208488 doi: medRxiv preprint responder rate is 76% (16/21). This strategy likely decreases our false positive rate but may also 261 contribute to the discrepancy between our data and previously published studies.

These data further our understanding of CD4 T-cell responses, particularly pTfh 263 responses, against SARS-CoV-2. Our study directly measures SARS-CoV-2-specific pTfh 264 responses to three major structural proteins, M, N, and S. We clearly demonstrate that SARS-265

CoV-2-specific pTfh responses that arise early in convalescence strongly correlate with antibody 266 neutralization and that S protein-specific responses most closely relate to antibody neutralization. Table 1 . Paired Visit 1 and Visit 2 PBMC samples from 21 individuals who 281 had recovered from COVID-19 were assessed in this study. Clinical data from these individuals 282 were retrieved from the Enterprise Biorepository REDCap database (43). All tested individuals 283 were symptomatic, but none were hospitalized during the course of their illness. Symptom severity 284 was quantified using a self-reported severity score on a scale of 1 to 3, where 1 represented no 285 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020 . . https://doi.org/10.1101 interference in daily life, 2 a moderate impact on daily life, and 3 a significant decrease in quality 286 of life due to symptoms. A majority of individuals reported moderate severity (71%, 15/21), and a 287 minority reported mild severity of symptoms (29%, 6/21). None reported severe symptoms. 288

Additionally, all but two had a positive SARS-CoV-2 nasopharyngeal swab. The two individuals 289 who did not have a PCR test completed had a known COVID contact, were symptomatic, and with the co-stimulatory antibody mix; Monensin and Brefeldin A (BD Bioscience) were added after 309 1 hour. Cells were incubated for 12 hours in total, instead of 18. Staining was conducted in three 310 steps: 1) Surface marker staining for 30 min at 4 o C with Dead cell dye-UV, CD3-A780, CD4-311 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020. . https://doi.org/10.1101/2020.10.07.20208488 doi: medRxiv preprint BV785, CD8-V500, CD14-PercpCy5.5, and CD19-PercpCy5.5. 2) Permeabilization with 312

at 4 o C with IFNγ-A700, TNFα-PeCy7, and CD154-APC. CD154 was plotted against IFNγ. 314

Additional details regarding the antibodies used in both the AIM and ICS assays can be found in 315 S1 Table, and the gating strategies for AIM and ICS in an unstimulated, negative control are 316 shown in S1 Fig.  317 Antibody assays: Plasma samples from the first time point for all 21 individuals were tested for 318 SARS-CosV-2-specific antibodies. The Abbott Architect assay was used to detect immunoglobin 319 G (IgG) reactivity to the SARS-CoV-2 nucleocapsid protein (46). The IgG quantity is reported as 320 a calculated index specimen/calibrator ratio, and values over 1.4 were considered positive for N 321 protein IgG. Manufacturer-reported specificity of this assay is 99.6% (99.1%-99.9%). 322

Antibody neutralization assays were conducted as previously described (47). Briefly, the 323 SARS-CoV-2 Spike (Wuhan 1, with a 19 amino acid cytoplasmic tail deletion) was pseudotyped 324 onto an HV-1 nanoluciferase reporter backbone by co-transfection in HEC 293T cells. 325

Pseudovirus was incubated with five-fold serial dilutions of patient plasma and then used to infect 326 1.5x10 4 293T clone 13 cells expressing ACE2. Two days post-infection, cells were washed with 327 PBS, lysed, and nanoluciferase activity was determined according to manufacturer's instructions 328 (Nano-Glo® Luciferase Assay System). Luciferase activity in wells with virus and no patient 329 plasma were set to 100%, and the dilution of plasma at which luminescence was reduced to 50% 330 (ID50) was calculated. 331

Statistical analysis: Comparisons between paired visit 1 and visit 2 magnitudes were conducted 332 by Wilcoxon signed-rank tests. All correlations were determined by Spearman Rank tests, with 333 the exception of Supplemental Figure 2A , where multiple measurements were plotted for each 334 individual (across the three proteins) and therefore a generalized linear mixed effect model 335 accounting for multiple measurements per individual was employed. In Figure 3 , all axes were 336 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020 . . https://doi.org/10.1101 transformed using log10(x+1) to allow for visualization of zeros, and correlations were determined 337 with untransformed data. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020 . . https://doi.org/10.1101 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020. . https://doi.org/10.1101/2020.10.07.20208488 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 12, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 12, 2020. . 

A Novel Coronavirus from Patients with 378 Pneumonia in China

Clinical Characteristics of Coronavirus 380 Disease 2019 in China

Clinical features of patients infected with 2019 382 novel coronavirus in Wuhan

Marked T cell 384 activation, senescence, exhaustion and skewing towards TH17 in patients with COVID-19 pneumonia

Immunological and inflammatory profiles in 387 mild and severe cases of COVID-19

Dysregulation in Individuals Recovering from SARS-CoV-2 Infection

Phenotype and 392 kinetics of SARS-CoV-2-specific T cells in COVID-19 patients with acute respiratory distress syndrome

SARS-CoV-2-reactive T cells in 395 healthy donors and patients with COVID-19

SARS-CoV-2-specific T cell 397 immunity in cases of COVID-19 and SARS, and uninfected controls

Targets of T Cell 399 Responses to SARS-CoV-2 Coronavirus in Humans with COVID-19 Disease and Unexposed Individuals

Safety and 402 immunogenicity of the ChAdOx1 nCoV-19 vaccine against SARS-CoV-2: a preliminary report of a phase 403 1/2, single-blind, randomised controlled trial

An mRNA 405 Vaccine against SARS-CoV-2 -Preliminary Report

Phase 1-2 Trial of a SARS

Recombinant Spike Protein Nanoparticle Vaccine

SARS-CoV-2 409 infection protects against rechallenge in rhesus macaques

DNA vaccine 411 protection against SARS-CoV-2 in rhesus macaques

Emerging concepts on T follicular helper cell dynamics in HIV infection

Vaccination establishes clonal 415 relatives of germinal center T cells in the blood of humans

Altered immune 417 cell follicular dynamics in HIV infection following influenza vaccination

T follicular helper cells in 420 human efferent lymph retain lymphoid characteristics

CXCR5 expressing human central 422 memory CD4 T cells and their relevance for humoral immune responses

Circulating precursor CCR7(lo)PD-1(hi)

CXCR5(+) CD4(+) T cells indicate Tfh cell activity and promote antibody responses upon antigen 426 reexposure

Activation of Peripheral T Follicular Helper Cells During Acute Dengue Virus Infection

Human 431 circulating PD-1+CXCR3-CXCR5+ memory Tfh cells are highly functional and correlate with broadly 432 neutralizing HIV antibody responses

Characteristics of the earliest 434 cross-neutralizing antibody response to HIV-1

Induction of 436 ICOS+CXCR3+CXCR5+ TH cells correlates with antibody responses to influenza vaccination

Peripheral CD4 T follicular 439 cells induced by a conjugated pneumococcal vaccine correlate with enhanced opsonophagocytic 440 antibody responses in younger individuals

Cellular immune responses 442 to severe acute respiratory syndrome coronavirus (SARS-CoV) infection in senescent BALB/c mice: CD4+ 443 T cells are important in control of SARS-CoV infection

The Potency of an Anti-MERS Coronavirus Subunit Vaccine 445 Depends on a Unique Combinatorial Adjuvant Formulation. Vaccines (Basel)

Breadth of 447 concomitant immune responses prior to patient recovery: a case report of non-severe COVID-19

Detection of SARS-CoV-2-Specific Humoral 450 and Cellular Immunity in COVID-19 Convalescent Individuals

SARS-CoV-2-specific T cells exhibit 454 phenotypic features of robust helper function, lack of terminal differentiation, and high proliferative 455 potential

Expressing T Follicular Helper Cells and the Absence of Germinal Centers in COVID-19

Humoral and circulating follicular 460 helper T cell responses in recovered patients with COVID-19

Comparative analysis of 462 activation induced marker (AIM) assays for sensitive identification of antigen-specific CD4 T cells. PLoS 463 One

Antibody Neutralization Breadth Is Associated with Enhanced HIV-Specific CD4+ T Cell Responses

Enhancing the Quality of Antibodies to HIV-1 Envelope by GagPol-Specific Th Cells

Safety and 471 immunogenicity of the ChAdOx1 nCoV-19 vaccine against SARS-CoV-2: a preliminary report of a phase 472 1/2, single-blind, randomised controlled trial

Deep immune profiling 474 of COVID-19 patients reveals distinct immunotypes with therapeutic implications

Prolonged Persistence of SARS-CoV-2 RNA in 477 Body Fluids

Coronavirus Disease 2019 Test Results After 479 Clinical Recovery and Hospital Discharge Among Patients in China

Optimization and validation 482 of an 8-color intracellular cytokine staining (ICS) assay to quantify antigen-specific T cells induced by 483 vaccination

A simple and 487 efficient method for the monitoring of antigen-specific T cell responses using peptide pool arrays in a 488 modified ELISpot assay

Comparison of 490 overlapping peptide sets for detection of antiviral CD8 and CD4 T cell responses

Performance 493 Characteristics of the Abbott Architect SARS-CoV-2 IgG Assay and Seroprevalence in

International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity