key: cord-0716897-wwg3q917
authors: Westmeier, Jaana; Paniskaki, Krystallenia; Karaköse, Zehra; Werner, Tanja; Sutter, Kathrin; Dolff, Sebastian; Overbeck, Marvin; Limmer, Andreas; Liu, Jia; Zheng, Xin; Brenner, Thorsten; Berger, Marc M.; Witzke, Oliver; Trilling, Mirko; Lu, Mengji; Yang, Dongliang; Babel, Nina; Westhoff, Timm; Dittmer, Ulf; Zelinskyy, Gennadiy
title: Impaired cytotoxic CD8+ T cell response in elderly COVID-19 patients
date: 2020-08-23
journal: bioRxiv
DOI: 10.1101/2020.08.21.262329
sha: 5605469d2077f0a257a95d86d98eca538a624139
doc_id: 716897
cord_uid: wwg3q917

SARS-CoV-2 infection induces a T cell response that most likely contributes to virus control in COVID-19 patients, but may also induce immunopathology. Until now, the cytotoxic T cell response has not been very well characterized in COVID-19 patients. Here, we analyzed the differentiation and cytotoxic profile of T cells in 30 cases of mild COVID-19 during acute infection. SARS-CoV-2 infection induced a cytotoxic response of CD8+ T cells, but not CD4+ T cells, characterized by the simultaneous production of granzyme A and B, as well as perforin within different effector CD8+ T cell subsets. PD-1 expressing CD8+ T cells also produced cytotoxic molecules during acute infection indicating that they were not functionally exhausted. However, in COVID-19 patients over the age of 80 years the cytotoxic T cell potential was diminished, especially in effector memory and terminally differentiated effector CD8+ cells, showing that elderly patients have impaired cellular immunity against SARS-CoV-2. Our data provides valuable information about T cell responses in COVID-19 patients that may also have important implications for vaccine development. Importance Cytotoxic T cells are responsible for the elimination of infected cells and are key players for the control of viruses. CD8+ T cells with an effector phenotype express cytotoxic molecules and are able to perform target cell killing. COVID-19 patients with a mild disease course were analyzed for the differentiation status and cytotoxic profile of CD8+ T cells. SARS-CoV-2 infection induced a vigorous cytotoxic CD8+ T cell response. However, this cytotoxic profile of T cells was not detected in COVID-19 patients over the age of 80 years. Thus, the absence of a cytotoxic response in elderly patients might be a possible reason for the more frequent severity of COVID-19 in this age group in comparison to younger patients.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a highly virulent 63 sarbecovirus currently causing a global pandemic with millions of cases and hundred 64 thousands of fatalities. Virus replication in the lung epithelium and the corresponding 65 pneumonia are the main reasons for symptomatic COVID-19 cases, although other tissues 66 and organs such as the kidney are also affected (1). Elderly people are predisposed to 67 severe COVID-19 and the mortality increases dramatically with age (1-3) . In particular, 68 individuals over 80 years of age show the highest hazard ratio (8.93-13.77 ) in terms of 69 hospital admissions (1) and have the highest case fatality rate (4). 70

There is growing evidence that adaptive immune responses are necessary for the control 71 and subsequent elimination of the virus (5). Cytotoxic T lymphocytes (CTL) are a specialized 72 population of immune cells which is able to selectively kill infected cells and consequently 73 eliminate viruses. Usually, CD8+ T lymphocytes mediate adaptive cytotoxic T cell responses. 74

Additionally, a fraction of the CD4+ T cell population is able to differentiate into cells with 75 cytotoxic properties (6). Both populations of cytotoxic cells can contribute to virus control by 76 eliminating infected cells. T cells responding to viral antigens expand and differentiate from 77 cells with a naïve phenotype into subpopulations of terminally differentiated cytotoxic effector 78 T cells or cells with an effector memory phenotype. Both effector cell subpopulations are 79 abundant during the acute phase of antiviral immune responses (7). Accordingly, the number 80 of cells with these phenotypes rises during the acute immune responses against several 81 respiratory viral infections (8). The SARS-CoV-2 infection is associated with a reduction of 82 CD8+ and CD4+ T cells (9, 10). One prominent cause of lymphopenia may be an enhanced 83 migration of T cells into infected compartments (11, 12) . Despite the lymphopenia, expanded 84 virus-specific CD8+ and CD4+ T cells can be detected in 14) . The 85 CD4+ and CD8+ T cells are specific towards several proteins of SARS-CoV-2 as has been 86 recently shown (15) (16) (17) . During the early phase of the immune response, CD8+ and CD4+ T 87 cells reacted against the spike, membrane, and nucleocapsid proteins (15, 16) . The T 88 lymphocytes of convalescent patients responded to structural proteins or nonstructural 89 proteins which provides evidence of the development of memory to different viral proteins 90 after infection (17, 18) . Interestingly, some individuals who were not infected with SARS-91

CoV-2 also responded to the antigens of this virus which have a low homology with "common 92 cold" human coronaviruses (17, 18) . 93

The detection of these virus-specific cells was possible after the in vitro stimulation of T cells 94 with viral peptides. This method allows for the definition of the specificity of analyzed T cells, 95 but has a modulating impact on the T cell phenotype and functionality. Moreover, the 96 stimulation of activated effector T cells in vitro can lead to restimulation-induced cell death 97 pneumonia. Nine patients did not need oxygen supplementation, but all others received 134 oxygen during their stay. All patients received empiric antimicrobial treatment except four 135 patients. Four patients were treated additionally with oral Oseltamivir until a negative 136 influenza test was available. The median length of hospitalization was 8 (3-108) days. The 137 median age of patients was high (71 years), consistent with the fact that symptomatic 138 COVID-19 disease and hospital admissions are more prevalent in the elderly. Peripheral 139 blood was drawn immediately after hospitalization to analyze T cell responses during acute 140 infection. The laboratory parameters of the patients are depicted in Table 1 (Sup. Tab. 1). All 141 SARS-CoV-2 infections were unequivocally confirmed by certified diagnostic RT-PCRs. 142

CD4+ T cells usually function as helper cells, but have been shown to be capable of 145 cytotoxicity after several virus infections, including those with coronaviruses (6, 24). 146 Therefore, we analyzed the production of cytotoxic molecules in CD4+ T cells upon SARS-147

CoV-2 infection without any additional stimulation of lymphocytes. First, we determined the 148 numbers of CD4+ T cells in the blood of COVID-19 patients and stratified the patients into 149 age groups of 29-79 (median 62) and 80-96 (median 86) years. CD4+ T cell counts were 150 reduced compared to normal clinical references: in the 29-79 age group, the median was 333 151 CD4+ T cells per µl vs. 555-1460 CD4+ T cells per µl in healthy donors, and in the 80-96 152 age group, the median was 319 CD4+ T cells per µl vs 540-720 CD4+ T cells per µl in age-153 matched control individuals (25) (Fig. 1A) . No difference in CD4+ T cell counts between the 154 analyzed age groups was observed. Next, we determined the differentiation status of all 155 CD3+CD4+ T cells according to the expression of CD45RO, CCR7, and CD28 and stratified 156 CD4+ T cells into naïve (N, CD45RO-CCR7+ CD28+), central memory (CM, CD45RO+ 157 CCR7+ CD28+), transitional memory (TM, CD45RO+ CCR7-CD28+), effector memory (EM, 158 CD45RO+ CCR7-CD28-), and terminally differentiated effector (E, CD45RO-CCR7-CD28-) 159 subpopulations (Fig. 1B) . The gating strategy is shown in Figure S1 . Subsequently, we 160 compared the distribution of subpopulations between COVID-19 patients and age-matched 161 healthy controls, again stratified according to age. No obvious differences between COVID-162 19 patients and healthy controls were found for any of the CD4+ T cell subtypes (Fig. 1C) . To 163 characterize their cytotoxic profile, we stained total CD4+ T cells directly ex vivo without re-164 stimulation for the cytotoxic molecules GzmA, GzmB, and perforin and compared the two 165 age groups between COVID-19 patients and age-matched healthy controls ( Fig. 1C-F) . 166

Again, we did not find clear differences between groups, except that the frequency of GzmB-167 producing cells was slightly increased in the 29-79 year group of COVID-19 patients 168 compared to healthy controls, yet with largely overlapping confidence intervals (Fig. 1C, E) .

Conversely, perforin responses were reduced in the older age group of COVID-19 patients 170 ( Fig. 1C , F). The overall data failed to reveal a meaningful cytotoxic response of CD4+ T 171 cells early after SARS-CoV-2 infection, and we did not further analyze CD4+ cells in the 172 current study. can kill virus-infected cells. Therefore, we analyzed the production of cytotoxic molecules in 177 CD8+ T cells in a cohort of SARS-CoV-2-infected individuals. 178

We first determined CD8+ T cell numbers in the blood of COVID-19 patients in the two age 179

groups. CD8+ T cell counts were clearly reduced in both groups compared to the numbers 180 reported in the literature (25), and an additional significant reduction was found for COVID-19 181 patients over 80 years of age ( Fig. 2A) . When calculating the Pearson correlation, we found 182 an inverse association between CD8+ T cell counts in peripheral blood and patient age (Fig.  183 2B). Next, we determined the distribution of different CD8+ T cell subsets (defined parallel to 184 the criteria for CD4+ T cells described above) and compared COVID-19 patients with age-185 matched healthy controls, again in two age groups. The gating strategy is shown in Figure  186 S2. Differences were found for the 29-79 year group in which the frequency of naïve CD8+ T 187 cells was clearly reduced in COVID-19 patients, whereas percentages of effector, effector 188 memory, and transitional memory cells were enhanced compared to healthy individuals, 189

suggesting an ongoing CD8+ T cell response in COVID-19 patients (Fig. 2C) . Interestingly, 190 this difference was almost absent in the older age group, most likely because the pool of 191 naïve CD8+ T cells largely disappears in elderly individuals (23). To characterize the profile 192 of CD8+ T cells, we stained CD8+ cells for cytotoxic molecules and compared the two age 193 groups from COVID-19 patients and healthy controls (Fig. 3A ). Cells were analyzed directly 194 ex vivo without any re-stimulation. We found a significant difference between the patients 195 and controls in the younger age group. COVID-19 patients had higher frequencies of CD8+ T Some studies on T cell responses in COVID-19 patients reported that CD8+ T cells may 203 already become functionally exhausted during acute infection (26). This hypothesis was based on the analysis of PD-1 expression by T cells during early COVID-19. In functional 205 terms, PD-1, in conjunction with its ligands PD-L1 and PD-L2, exerts potent immune-206 inhibitory activities. However, its expression is induced by T cell receptor (TCR) activation 207

(27) and TCR downstream NFAT signaling (28). PD-1 expression is a hallmark of recent 208 TCR-based recognition of MHC-presented antigens that is often up-regulated on cytotoxic 209 effector T cells during acute infections (29, 30) . We found that about 20% of total CD8+ T 210 cells expressed PD-1 in healthy controls, as well as in COVID-19 patients (Fig. 4A ). Most of 211 these cells expressed GzmA with no apparent differences between the groups (Fig. 4B) . 212

However, for GzmB and perforin, we found a higher frequency of positive cells among PD-1+ 213 CD8+ T cells in the group of younger COVID-19 patients compared to healthy controls. This 214 difference was absent for the older age group (Fig. 4C , D). Our data indicate that PD-1+ 215

Here, we clearly demonstrate a cytotoxic profile in CD8+ T cells upon SARS-CoV-2 infection, 218 which was also found in CD8+ T cells expressing PD-1. 219 220

Individual subpopulations of CD8+ T cells differ in their ability to produce cytotoxic molecules, 223 with the highest potency for effector T cell populations. To investigate which CD8+ T cell 224 subpopulation dominates the cytotoxic profile of CD8+ T cells in mild COVID-19 patients, we 225 analyzed the expression of Gzms and perforin in all five T cell subpopulations. 226

The representative histogram shows that GzmA was produced by transitional memory, 227 effector memory, and effector cells, whereas GzmB and perforin were only found in the latter 228 two populations in our ex vivo analysis (Fig. 5A ). Next, we assessed whether the production 229 of cytotoxic molecules by effector CD8+ T cell subpopulations is influenced by the age of 230 COVID-19 patients. For a precise analysis of age effects on the expression of cytotoxic 231 molecules we stratified the COVID-19 patients into 3 age groups (29-66 (median 56); 70-76 232 (median 73); 80-96 years (median 86)). Interestingly, for effector and effector memory cells, 233 the percentages of GzmA-as well as perforin-positive cells were significantly reduced in the 234 80-96 age group compared to the 29-69 age group (Fig. 5E , G, H, G). For transitional 235 memory cells, this was only the case for GzmA (Fig. 5B ). This suggests a functional 236 impairment of the cytotoxic program in the CD8+ T cells of elderly COVID-19 patients. 237

The simultaneous expression of different cytotoxic molecules is a feature of effector cells 238 with a strong cytolytic potential. Therefore, we also performed single-cell analysis of CD8+ T 239 cells from COVID-19 patients to determine the expression profiles of cytotoxic molecules for 240

cytotoxic profile produced only GzmA and there was no obvious difference between the age 242 groups ( Fig. 6A-C) . Surprisingly, the vast majority of effector and effector memory cells 243 produced all three cytotoxic molecules simultaneously (Fig. 6D-I) . While all patients from the 244 youngest age group had multifunctional effector cells, some individual patients from the older 245 age groups showed reduced multifunctional responses (Fig. 6E, F For an aged immune system, the reduction of T lymphocytes and processes of immune 291 senescence are characteristic (37). However, the nature of the progressive loss of circulating 292 CD8+ T cells in elderly COVID-19 patients is not completely understood. One possible 293 explanation may be an enhanced migration of T cells from the blood into the infected tissue. 294

Usually, the accumulation of T lymphocytes leads to a progressive inflammation in the 295 infected organs. SARS-CoV-2 infects lung epithelial cells, which might recruit cytotoxic T 296 cells into the lung. In the early phase of infection, which we analyzed here, they most like 297 contribute to virus control in the lung. However, sustained T cell cytotoxicity might also 298 contribute to organ damage. Thus, the precise recognition and elimination of infected cells 299 without the induction of too much inflammation and tissue destruction is necessary for the 300 survival of infected patients. This delicate balance of two opposing processes is very 301 important for survival. Here, multifunctional T cells, producing both perforin and Gzms at the 302 same time may be very important, as it has been shown that perforin is a critical enabler of 303 the apoptotic effects mediated by Gzms. Cells producing perforin and Gzms are necessary 304 for the efficient control of virus infections (20). We found many T cells producing Gzms and 305 perforin in our COVID-19 patient cohort, although their frequencies were reduced in elderly 306 patients. They might have contributed to efficient virus control, since all our patients showed 307 only mild symptoms and fully recovered from COVID-19. Multifunctional cytotoxic T cells 308 often express PD-1 (38), not because they are functionally impaired during acute infection, 309 but as an important negative switch to shut them down when responses are either too strong 310 or maintained for too long. 311

The cytotoxic molecules analyzed here share some overlapping functions but also elicit non-312 redundant features. The critical effector molecule for target cell killing is perforin, as it 313 promotes the entry of Gzms into target cells and in this way enables the cytotoxic 314 functionality of Gzms (20). Thus, CTLs producing Gzms without perforin can induce severe 315 inflammation triggered by the aimless release of Gzms and inflammatory cytokines. Once 316

Gzms enter infected target cells in the presence of perforin, they mediate apoptosis of these 317 cells. Thus, it is tempting to speculate that the herein identified age-associated reduction of 318

CTLs expressing perforin may be an additional factor in COVID-19 progression, as it might 319 support lung inflammation. 320

Our current data support the concept that cytotoxic CD8+ T cells play an important role in the 321 control of early SARS-CoV-2 infections, but may also be a factor of immune pathogenesis 322 and COVID-19 progression during later periods of infection. Thus, it will be important to 323 carefully balance therapeutic measures either supporting or suppressing T cell responses in 324 future COVID-19 therapy. Recent suggestions to therapeutically administer checkpoint 325 inhibitors, which are efficiently used for tumor immune therapy, for the treatment of COVID-326

19 patients (39) should be reevaluated, since we did not find functionally exhausted CD8+ T 327 cells in our patients. In agreement with our previous findings based on acute virus infection 328 models in mice (30, 38), our clinical study suggests that a checkpoint therapy might enhance 329 the functionality of the PD-1-expressing cytotoxic CD8+ T cells in COVID-19 patients and 330 improve virus control, but with a potential to exaggerate the immunopathology in the lung and 331 other organs, which might actually accelerate decompensation. 332 were reported for the control group.

peripheral blood by gradient centrifugation. Blood was collected in EDTA S-monovettes 347 (Sarstedt). Collected blood was pre-diluted with RPMI-1640 (Thermo Fisher Scientific) and 348 overlaid onto 15 mL Pancoll separating solution (PAN-Biotech). The tubes were centrifuged 349 at 1600 rcf for 15 min at room temperature with deceleration of the centrifuge set to low. The PD-1/PD-L1 Pathway Affects the Expansion and Function of Cytotoxic CD8(+) T Cells 508

During an Acute Retroviral Infection. Front Immunol 10:54. 

The simultaneous production of GzmA, GzmB, and perforin by CD8 

Clinical Characterisation Protocol: prospective observational cohort study

Zhu 388 SY, Zhong NS, China Medical Treatment Expert Group for C. 2020

Coronavirus Disease 2019 in China

Clinical features of COVID-19 in elderly patients: A 391 comparison with young and middle-aged patients

Estimates of the severity of coronavirus disease 2019: a model-based 397 analysis

Immunology of COVID-19: Current State of 403 the Science

Fas Ligand-mediated cytotoxicity of CD4+ T cells 406 during chronic retrovirus infection

Virus-specific CD8 T cells: activation, 408 differentiation and memory formation

The CD8 T Cell Response to Respiratory Virus Infections

Clinical and 413 immunological features of severe and moderate coronavirus disease 2019

Metabolic disturbances and 416 inflammatory dysfunction predict severity of coronavirus disease 2019 (COVID-19): a 417 retrospective study

Zhang 419 S, Zhang Z. 2020. Single-cell landscape of bronchoalveolar immune cells in patients with 420 COVID-19

Pathological findings of COVID-19 associated 423 with acute respiratory distress syndrome

COVID-19 progression is potentially driven by T cell immunopathogenesis

CoV-2-Specific Humoral and Cellular Immunity in COVID-19 Convalescent Individuals

The SARS-CoV-2 T-cell immunity is directed against the spike, 436 membrane, and nucleocapsid protein and associated with COVID 19 severity

Phenotype and kinetics of SARS-CoV-2-specific T cells in COVID-19 patients with 441 acute respiratory distress syndrome

Targets of T Cell 445 Responses to SARS-CoV-2 Coronavirus in Humans with COVID-19 Disease and Unexposed 446 Individuals

SARS-CoV-2-specific T cell immunity in cases of COVID-19 and SARS, and 450 uninfected controls

The power and the promise 452 of restimulation-induced cell death in human immune diseases

Perforin and granzymes: function, dysfunction 455 and human pathology

Combination immunotherapy with 460 anti-PD-L1 antibody and depletion of regulatory T cells during acute viral infections results in 461 improved virus control but lethal immunopathology

Pathological findings of COVID-19 associated 464 with acute respiratory distress syndrome

Cellular 466 senescence impact on immune cell fate and function

Cytotoxic CD4+ T Cells in Viral 468 Infections

Aging and 470 lymphocyte subpopulations: whole-blood analysis of immune markers in a large population 471 sample of healthy elderly individuals

Reduction and Functional Exhaustion of T Cells in Patients With 474

Front Immunol 11:827

PD-1-mediated suppression of IL-2 production induces CD8+ T cell anergy in vivo

GzmB perforin