Circulating effector cd T cell populations are associated with acute coronavirus disease 19 in unvaccinated individuals

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection causes severe coronavirus disease 2019 (COVID-19) in a small proportion of infected individuals. The immune system plays an important role in the defense against SARS-CoV-2, but our understanding of the cellular immune parameters that contribute to severe COVID-19 disease is incomplete. Here, we show that populations of effector cd T cells are associated with COVID-19 in unvaccinated patients with acute disease. We found that circulating CD27 neg CD45RA + CX3CR1 + V d 1 effector cells expressing Granzymes (Gzms) were enriched in COVID-19 patients with acute disease. Moreover, higher frequencies of GzmB + V d 2 + T cells were observed in acute COVID-19 patients. SARS-CoV-2 infection did not alter the cd T cell receptor repertoire of either V d 1 + or V d 2 + subsets. Our work demonstrates an association between effector populations of cd T cells and acute COVID-19 in unvaccinated individuals.


INTRODUCTION
Coronavirus disease 2019 (COVID-19) remains a global burden with over 663 million confirmed cases and approximately 6.7 million deaths have been reported to the WHO to date. 1 Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection can result in severe lung inflammation (i.e.acute respiratory distress syndrome (ARDS)). 2 However, we do not completely understand the immune parameters involved in the disease pathogenesis.SARS-CoV-2 enters respiratory epithelial cells via the ACE2 receptor upon inhalation and initiates replication. 2SARS-CoV-2 antigens (Ags) are presented by major histocompatibility complex (MHC) I that are recognized by SARS-CoV-2-specific cytotoxic CD8 + T cells. 3,4Further Ag presentation induces both cellular and humoral immune cells.In COVID-19 patients with ARDS, a cytokine storm occurs as a result of an uncontrolled systemic immune response, which is the most common cause of death. 5It is composed of large amounts of interferon (IFN)-a, IFN-c, tumor necrosis factor (TNF) a, tumor growth factor (TGF) b, and interleukin (IL)-1b, -6, -12, -18, -33 as well as many chemokines produced by effector immune cells. 4While both innate and adaptive immune responses play important roles in the defense against SARS-CoV-2, the role of cd T cells is poorly understood.
cd T cells are a unique population of unconventional lymphocytes.Circulating cd T cells represent 5-10% of all CD3 + T cells and express a T cell receptor (TCR), mostly consisting of Vc9 paired to Vd2, 6 termed Vd2 + T cells.Vd2 + T cells activate rapidly upon exposure to the phosphoAg (pAg) (E)-4-Hydroxy-3-methyl-but-2-enyl pyrophosphate (HMBPP). 7This pAg is produced by an array of microbes including Plasmodium falciparum and Mycobacterium tuberculosis.The Vd2 + T cell subset possesses a cdTCR repertoire that contains public (i.e.shared) complementarity determining region 3 (CDR3) sequences. 6Vd2 À cdT cells on the other hand, are present at high frequencies in tissues where they are thought to play an important role in immune surveillance. 8The majority of tissue-infiltrating cells express a Vd1 TCR paired to various Vc chains. 9Vd1 + T cells show private (i.e.not shared) clonotypic cd TCR expansions and differentiate from na€ ıve into effector cells upon microbial challenges such as repeated P. falciparum 10 and cytomegalovirus (CMV) 9 infection.cd T cells have been implicated in the immune response to SARS-CoV-2 and are associated with COVID-19 severity. 112][13] Importantly, activated cd T cells correlated with COVID-19 severity, with increased frequencies 11 and counts 14 found in severe vs. mild infection. 11In addition, IL-6 and IL-18 levels correlated with activated cd T cells, indicating that these cells may contribute to the cytokine storm that occurs in severe acute COVID-19.Only a few studies investigated the involvement of specific cd T cell subsets in SARS-CoV-2 infection.Reduced circulating Vd2 + T cell frequencies 12 and numbers have been found in COVID-19 patients correlating with disease severity. 15While no differences in Vd1 + T cells have been identified, a recent study showed that TCRd1 + (TRDV1) CDR3 sequences showed evidence of clonal focusing in COVID-19 patients aged > 50 years. 16d T cell subsets have been implicated in controlling other viral infections such as CMV and severe acute respiratory syndrome (SARS).In CMV, Vd1 + T cells expanded rapidly and undergo cd TCR clonotype expansion upon infection in both stem cell 17 and kidney [18][19][20][21] transplantation settings and have been implicated in the resolution of CMV infection. 22Moreover, Vd1 + T cells expressed an effector phenotype (i.e.CD27 lo CD45RA + or CD45RA + CD28 À ) in CMV-seropositive patients. 9,18In SARS patients, Vd2 + T cells expanded during convalescent disease. 23Moreover, Vd2 + T cells lines mounted IFN-cdependent responses upon in vitro stimulation with IPP and IL-2.Supernatants from stimulated Vd2 + T cells lines decreased both viral load and infectious units from SARS-infected Vero cells.In line with this, Vd2 + T cells derived from controls co-cultured with SARS-infected THP-1 cells had increased IFN-c production and cytotoxicity.Importantly, it remains to be elucidated whether and how cd T cell subsets and cd TCR repertoires change in response to SARS-CoV-2 infections and whether this associates with COVID-19 severity.
To gain a deeper understanding of cd T cell repertoire responses in SARS-CoV-2 infection and how each subset associates with disease severity in SARS-CoV-2 infection, we analyzed peripheral blood mononuclear cells (PBMCs) from acute and convalescent COVID-19 patients using multi-color flow cytometry and cd TCR sequencing.An increased activated Vd1 effector population was observed in acute COVID-19 patients compared with convalescent patients and healthy controls (HCs).This population had increased cytotoxic potential in acute COVID-19 patients.In addition, an expanded cytotoxic Vd2 effector population was found in patients with acute COVID-19.However, no cd TCR selection was identified for either subset and the repertoire did not change over time.Together, this study suggests that circulating effector populations of cd T cells may be associated with acute COVID-19.

Activated effector cd T cell frequencies are increased in acute COVID-19
To explore the relationship between circulating cd T cells and acute COVID-19, we examined a cohort of 23 acute (median age of 53 years, range 27-73 years; 11 females) and 49 convalescent (median age of 57 years, range 22-76 years; 17 females) COVID-19 patients from VIC, Australia (Figure 1a) and as described before. 11We compared this cohort of COVID-19 patients with 21 healthy controls (HCs) with no history of COVID-19.We observed significantly decreased total cd T cell frequencies in convalescent patients compared with HCs (Figure 1b), and in line with previous findings. 12We then examined T cell memory phenotypes of cd T cells in our cohorts.We divided cd T cells into na€ ıve-like, central memory (CM), effector memory (EM), and terminally differentiated CD45RA-expressing EM (TEMRA) subsets based on CD27 and CD45RA expression using high dimensional FFT-accelerated interpolation-based t-SNE (FItSNE) dimensional reduction analyses (Figure 1c).Acute COVID-19 patients presented with significantly lower frequencies of na€ ıve-like cells and CM cells compared with HCs and convalescent patients, respectively, but increased TEMRA frequencies in comparison with both HCs and convalescent patients (Figure 1d).We then examined cd T cell activation status by measuring CD38 and HLA-DR cell surface expression (Supplementary figure 1a).CD38 + HLA-DR + cd T cells were significantly increased in acute COVID-19 patients compared with both HCs and convalescent patients (Figure 1e).This activated cd T cell phenotype was not restricted to a single memory population but increased across all subsets (Supplementary figure 1b).Together, these data demonstrate that acute SARS-CoV-2 infection results in an increased proportion of activated cd T cells with a TEMRA phenotype.
Individuals with acute SARS-CoV-2 infection have increased circulating effector Vd1 + and Vd2 + T cells Circulating cd T cells comprise adaptive-like Vd1 + and innate-like Vd2 + T cell subsets.We examined whether acute COVID-19 disease was associated with specific subsets and phenotypes of cd T cells.This was investigated using a multi-parameter flow cytometry panel in a sub-cohort of 18 COVID-19 patients, comprising nine acute (median age of 58 years, range 27-72 years; five females) and nine convalescent donors (median age of 40 years, range 19-74 years; four females; Table 1; Figure 2a).Absolute counts of cd, Vd1, and Vd2 T cells were not different between patient groups (Supplementary figure 1c).As we observed an increase in TEMRA cd T cell frequencies in acute COVID-19 patients, we investigated whether there was a skewing of effector populations in the cd T cell subsets between patient groups.In the sub-cohort of acute COVID-19 patients the TEMRA population had significantly higher proportions of Vd1 + T cells, while Vd2 + T cell frequencies were not statistically different (Figure 2b).We then focused on effector populations of Vd1 + and Vd2 + cd T cells that we have previously phenotyped in healthy adults. 9,24CD27 lo CX3CR1 + Vd1 effector T cell proportions were significantly increased in acute COVID-19 patients in comparison with the convalescent group (Figure 2c).Conversely, convalescent patients displayed higher frequencies of CD28 + CD27 hi Vd1 naive T cells compared with acute patients (Supplementary figure 1d).Within the Vd2 + T cell population, a significantly reduced CD27 hi and increased CD27 lo cell population was observed in acute COVID-19 patients compared with convalescent patients (Figure 2d).We also assessed the total ab and CD8 + T cell frequencies within CD3 + T cells and they were not significantly different between both acute and convalescent groups (Supplementary figure 1e).Finally, we correlated cd T cells and their subsets to age.Vd1 effector frequencies correlated with age, while total cd T cells, total Vd2 + and Vd1 + T cells did not (Supplementary figure 1f).Together, our data suggest that COVID-19 does not alter or drive changes in cd T cell populations but that the frequency of effector cd T cell populations was elevated in patients that developed acute COVID-19.

Increased proportion of cd T cells with cytolytic potential in COVID-19 patients with acute infection
cd T cells are known to possess a pre-programmed cytolytic potential which increases after microbial exposure. 10Cytolytic Vd1 + T cell frequencies, as a proportion of the entire GzmA + GzmB + T cell population, were significantly higher in acute COVID-19 patients compared with convalescent patients, while cytolytic Vd2 + T cells did not differ (Figure 3a).We observed that CD27 lo Vd2 + T cell proportions are increased in acute COVID-19 patients (Figure 2); thus, we gated on cytolytic cells within CD27 À Vd2 + T cells to investigate differences between convalescent and acute COVID-19 patients.GzmB + CD27 À Vd2 + T cell proportions were significantly higher in acute COVID-19 patients, while GzmA + cells were not different (Figure 3b).Together these data show that cd T cell subsets have increased cytolytic potential during acute SARS-CoV-2 infection.

cd TCR repertoires remain stable after SARS-CoV-2 infection
Acute viral infection with CMV has been reported to drive the clonotypic selection within the cd TCR repertoire. 9We investigated whether acute SARS-CoV-2 infection drives changes in the cd TCR repertoire.Vd chain usage was skewed between acute and convalescent COVID-19 patients, with acute patient's cd T cells expressing more Vd1 chain sequences (Figure 4a).The Vd1 + (Figure 4b) and Vd2 + (Figure 4c) cd TCR repertoires of acute patients showed no difference compared with convalescent patients.This was evident across visual (treeplots) and repertoire metrics (accumulated frequencies of the top 20 clonotypes and D75 index).Subsequently, we longitudinally mapped the top 20 cd TCR clonotypes in a single acute patient and follow-up convalescent sample 20 days later (Figure 4d).The top 20 cd TCR clonotypes in this individual were stable and did not change after SARS-CoV-2 infection.Taken together, these data suggest that COVID-19 may not drive the acute selection of specific cd TCR clonotypes.

DISCUSSION
cd T cells and their cd TCR repertoires have been implicated in protective immune responses to viral infections such as CMV and SARS.However, the role of cd T cell subsets (i.e.Vd1 and Vd2 + T cells) in COVID-19 has remained unclear.It is also unknown if and how these subsets and their cd TCR repertoires change after SARS-CoV-2 infection and associate with COVID-19 severity in unvaccinated patients.Here, we describe an association of increased circulating activated Vd1 effector and Vd2 effector T cells with disease severity in unvaccinated COVID-19 patients.While increased cytotoxic potential within both cd T cell subsets was found in acute SARS-CoV-2 infection, no cd TCR selection nor repertoire changes were observed over time.
COVID-19 patients with acute infection were overall lymphopenic and this finding is in line with previously published data. 14,25Similarly, cd T cell counts were decreased in patients with acute compared with mild disease. 14,25Although our results deviate from this finding, likely due to smaller patient cohorts, increased cd T cell frequencies were found in patients with acute COVID-19 compared with those with mild disease.While cd T cell frequencies were not restored in convalescent patients, further longitudinal studies should be performed to investigate changes in their frequencies long-term post-COVID-19.cd T cells, particularly Vd1 + T cells, are mostly tissue resident and infiltrate tissues such as the lungs.Two studies showed abundant cd T cell numbers in the lungs and BALF from patients with SARS-CoV-2 infection. 26,27While these findings should be confirmed in more patients, it is likely that cd T cells are recruited to and implicated in local immune responses in the lung.Acute COVID-19 correlated with an expanded Vd1 + T cell frequency 25 ; however, the Vd1 + T cells' phenotype was not characterized.In this study, we found an increased TEMRA cd T cell population in unvaccinated patients with acute COVID-19, which were mainly Vd1 + T cells.Another study found increased na€ ıve-like and decreased effector cd T cells, while the TEMRA subset was not different. 14Although this seems opposite to our findings, this study did not investigate the different cd T cell subsets and designated na€ ıve and effector cells based on CD45RA and CD62L expression.We observed increased frequencies of CX3CR1 + Vd1 effector cells in patients with acute COVID-19.CX3CR1 plays a role in cell migration as its ligand (CX3CL1) is highly expressed in tissues, including lung tissue. 28,29In addition, in vitro chemotaxis assays have shown that CX3CL1 induced migration of CX3CR1 + cells with cytotoxic potential included cd T cells. 30However, there are no in vivo experiments demonstrating cd T cell migration from the blood into tissues specifically in SARS-CoV-2 infection.Moreover, an increased presence of this subset could also predispose COVID-19 patients for acute disease.In our cohorts, Vd1 effector cells correlated with age, supporting earlier findings that Vd1 + T cells in the cord blood 9 and children 10 were mainly na€ ıve (i.e.CD27 hi ), while HCs had a mixture of na€ ıve and CD27 lo effector cells. 9In addition, only Vd1 effector T cells acquire cytotoxic potential, regardless of age. 10 Interestingly, children (who have a na€ ıve Vd1 + T cell repertoire) are at lower risk of contracting COVID-19. 31Alternatively, an increased Vd1 effector population could also be the result of acute infection.It has been shown that Vd1 + T cells do not respond in vitro to SARS-CoV-2 nucleocapsid or spike proteins. 32Therefore, they may become activated and respond to as yet unknown SARS-CoV-2 related Ags that drives their expansion.Further studies should investigate pre-COVID-19 samples to determine whether cd T cell subsets predispose patients for acute disease.
There is emerging evidence of the adaptive functions of Vd1 + T cells in viral infections, which are similar to conventional ab T cells. 33While clonally expanded cd TCRs were present in COVID-19 patients, we did not detect differences between acute and convalescent patient groups.Moreover, transition into convalescent disease in a single patient did not result in notable changes in their top 20 clonotypes.However, this may be due to the limited patient samples studied, as a recently published study did find clonotypic focusing of TRDV1 in a larger cohort of COVID-19 patients. 16Here, TRDV1 cd TCR clonotypic focusing was sustained over ~9 days after the first positive SARS-CoV-2 test, albeit only in patients aged > 50 years.No association with disease severity was found.The TRDV1 TCR repertoire was private and no specific CDR3 sequences were associated with SARS-CoV-2 exposure.The authors speculated that expanded TRDV1 TCRs are not specific for pathogenderived Ags but respond to stress induced ligands. 34nother option could be a response to cytokine stimulation, as conventional CD8 ab T cells can expand clonally in response to IL-12 and IL-18. 34Acute COVID-19 is associated with cytokine storms of mainly IL-6 and TNFa, but patients also showed elevated IL-12 levels. 35It has been shown previously that in vitro stimulation of CD27 lo Vd1 + T cells (which contained a set of focused  cd TCR clones) with IL-15 resulted in proliferation of several clones. 9While no differences in IL-15 levels have been found in COVID-19, increased levels were reported in patients with Middle East respiratory syndrome coronavirus (MERS-CoV) compared with controls. 36uture studies should sequence Vd1 + TCR repertoires from COVID-19 patients with varying disease intensities and pre-COVID-19 samples to elucidate whether and how cytokines and/or SARS-CoV-2 drives changes herein.Vd2 + T cells are involved in immune responses, particularly against microbes mainly via recognition of their Ag HMBPP.While no associations of Vd2 + T cells with disease severity in COVID-19 were found in our study, reduced cell counts were found in acute COVID-19. 25evertheless, they have been implicated in protective immune responses in COVID-19 as patients who did not survive the infection had lower Vd2 + T cell counts than those who survived. 37Vd2 + T cells may have direct antiviral potential in COVID-19 as they rapidly activated (i.e.increased CD69 expression and IFN-c production) in PBMC cultures stimulated with influenza A. 38 Inhibition of the mevalonate pathway and isopentyl pyrophosphate (IPP; i.e. mammalian equivalent of HMBPP) synthesis in PBMCs significantly reduced IFN-c production by Vd2 + T cells in response to influenza A. In addition, another study confirmed these results and also showed that these cells had enhanced cytolytic potential, killed influenzainfected cells (A and B) and utilized a public Vc9 TCR CDR3 paired to various Vd2 CDR3s. 39Transfected Vd2 TCRs from COVID-19 patients activated in response to epitopes that were present on non-structural protein 8 (NSP8), a protein encoded in the SARS-CoV-2 genome. 40owever, we did not find any evidence of disease severity being associated with expansions of the similar TCRs in COVID-19.We showed that CD27 À Vd2 effector T cell proportions correlated with disease severity.They are composed of a larger GzmB + population, which is in line with previously published data. 41These cells were also more responsive to HMBPP stimulation in this study.Together this suggests that while Vd2 + T cells may not directly recognize virus infected cells, infection does have the potential to induce IPP accumulation and to drive the Vd2 + T cell response.
A limitation of this study is that we were unable to include a pre-COVID-19 sample, and thus cannot compare with a baseline measurement how COVID-19 changes the cd T cell population.Future studies should aim to investigate cd T cell subsets in a pre-COVID-19 sample and also longitudinally in patients with both acute and mild acute COVID-19 to study whether disease severity is associated with long-term changes in the cd T cell compartment.Another limitation is that tissueresident cd T cells were not analyzed.As Vd2 À cd T cell subsets are mostly tissue resident, and SARS-CoV-2 initially infects the mucosal lining of the respiratory tract, 2 tissue infiltrating cd T cells should be studied in the context of COVID-19 to detect tissue-specific effects of the infection and association with disease severity.Other avenues that could also be explored are the potentially long-term impacts of one or multiple vaccinations against COVID-19, and infection with different SARS-CoV-2 variants, on the cd T cell population.
In summary, an association was found between disease severity and activated TEMRA cd T cells.These cells consisted mainly of cytotoxic Vd1 effector T cells in patients with acute SARS-CoV-2 infection.Moreover, Vd2 effector T cells with cytotoxic potential were increased in acute COVID-19.Our study suggests that cd T cell subsets may predispose patients for acute disease; however, further research should be performed to investigate whether vaccination against COVID-19 changes the cd T cell compartment to promote mild disease.

Cohort and ethics statement
Acute and convalescent COVID-19 patients were recruited with written consent via the University of Melbourne (#2057366, #2056901, #2056689, #2056761, #1442952, #1955465, and #1443389), Austin Hospital (HREC/63201/Austin-2020) and Alfred Hospital (#280/14), as previously described, 3 in accordance with the Declaration of Helsinki Principles and the Australian National Health and Medical Research Council Code of Practice (Table 1).Healthy pre-pandemic buffy packs were obtained from the Australian Red Cross LifeBlood (2015#8, West Melbourne, Australia; Table 1).PBMCs or flow-through fractions following tetramer-associated magnetic enrichment (TAME) from our previous study 3 were used for the assessment of cd T cell subsets in this study and were selected based on available PBMCs (MUHREC #26340).

Peripheral blood mononuclear cell isolation
Peripheral blood was layered over Ficoll-Paque (GE Healthcare, Chicago, IL, USA) or lymphoprep (STEMCELL Technologies, Vancouver, Canada) and centrifuged for 20 min at 600-8009 g at room temperature without brake.The PBMC layer was removed and washed twice in Roswell Memorial Park Institute (RPMI) 1640 medium for 10 min at 4009 g at 4°C.Cells were frozen at À80°C in fetal calf serum (FCS) and 10% dimethyl sulfoxide (both Sigma-Aldrich, St Louis, MO, USA) and stored in liquid nitrogen.

Computational flow cytometry analysis
Computational analysis of cd T cell data was performed on previously published dataset of healthy and COVID-19 acute and convalescent blood samples 11 using FlowSOM and the Specter R package 42 (https://github.com/ImmuneDynamics/Spectre), as described previously in detail. 11A gating strategy is provided in Supplementary figure 2a.Cellular expression of CD27, CD45RA, CD38 and HLA-DR was determined on the cd T cell cluster and plotted by FltSNE plots for Figure 1.Subsets of cd T cells included na€ ıve-like CD45RA + CD27 + , TEMRA CD45RA + CD27 À , CM CD45RA À CD27 + and EM CD45RA À CD27 À subsets were evaluated for CD38 and HLA-DR expression using manual gating in FlowJo.Volcano plots were created in R, where comparisons were performed using a Wilcoxon rank-sum test (equivalent to the Mann-Whitney U-test) with the wilcox.testfunction in R. Statistics displayed in volcano plots were corrected with a False Discovery Rate (FDR) adjustment.

Antibodies and flow cytometry
Staining for cd T cell subsets was performed as described previously. 43In short, frozen PBMCs were thawed and washed twice in PBS.The cells were first stained with Zombie Aqua (1:500 in PBS; BioLegend, San Diego, CA, USA), washed, and then split and stained with both the cd T cell and Gzm panel (Supplementary table 2).Intracellular staining continued by fixing and permeabilizing PBMCs using the Foxp3/ Transcription factor staining buffer set (eBioscience, San Diego, CA, USA) and incubation with intracellular antibodies diluted in permeabilization buffer.An acute COVID-19 patient had to be excluded from subsequent cd T cell subset and Gzm analyses as no anti-CD27 antibody was added and no Gzm panel staining was performed.Absolute cell counts were quantified in a selection of samples (n = 2 acute and n = 5 convalescent COVID-19 patients) using accucheck counting beads (Thermo Fisher Scientific, Waltham, MA, USA).A representative gating strategy is shown in Supplementary figure 2b.All samples were acquired on a Fortessa X20 flow cytometer and resulting FCS files were analyzed using FlowJo v10 (both BD Biosciences, Franklin Lakes, NJ, USA).

Bulk cell sorting and RNA-based TCR repertoire analysis
Part of the PBMCs from three acute (for one donor also a convalescent sample was analyzed) and two convalescent COVID-19 patients were stained for bulk sorting as described previously. 10In short, PBMCs were stained with Zombie Aqua dye (BioLegend) and then incubated with antibodies indicated in the bulk sort panel (Supplementary table 1) for 20 min on ice.The cells were then bulk sorted according to the sort strategy described in Supplementary figure 2c into RNAlater (Sigma Aldrich) using a FACS ARIA II Fusion (BD Biosciences).Total cell numbers sorted for each donor are given in Supplementary table 2. RNA was purified using a RNeasy plus micro kit (Qiagen, Hilden, Germany) and the human TCRd chain iR profile kits (iRepertoire Inc., Huntsville, AL, USA) were used to perform amplicon rescued multiplex (ARM)-PCR to generate complementarity determining region (CDR) 3 libraries for sequencing following the manufacturer's instructions.Sequencing was performed using an Illumina MiSeq (Monash Health Translation Precinct Medical Genomics Facility, Clayton, VIC, Australia).We used iRweb tools (iRepertoire Inc.) to assign CDR3 sequences, variable (V), diversity (D), and junction (J) gene usage, calculate diversity indexes (DI; this metric considers the clonal frequency to occupy 50% of the total repertoire (D50) and abundance of unique CDR3 sequences (Shannon entropy)) and plot tree maps.

Statistical analysis
Tabulated data were plotted and analyzed in GraphPad PRISM 9 (GraphPad Software, San Diego, CA, USA).Each data set was assessed for normality using Shapiro-Wilk normality test.Differences between cohorts were analyzed by an unpaired t-test for normally distributed data or a Mann-Whitney U-test for non-parametric data.Correlations were analyzed using a Pearson r-test.*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Graphical Abstract
The contents of this page will be used as part of the graphical abstract of html only.It will not be published as part of main.The immune system plays a role in controlling SARS-CoV-2 but our understanding of the immune compartments that contribute to COVID19 disease are incomplete.In this study we show that populations of effector gamma delta T cells are associated with acute COVID-19 in unvaccinated patients.

Figure 2 .
Figure 2. Vd1 + and Vc9/Vd2 + T cell subsets in COVID-19 patients.(a) Schematic overview of the number of acute (red) and convalescent (blue) COVID-19 patients included in the second part of the study (created in Biorender).(b) Representative plot of the CD27 lo CD45RA + TEMRA cd T cell population within an acute COVID-19 patient.Violin plots show frequencies of Vd1 + and Vd2 + T cells within the TEMRA population for acute (red; n = 9) and convalescent (blue; n = 9) patients.(c) Representative FACS plot gated on CD27 lo CX3CR1 + Vd1 effector cells for an acute COVID-19 patient.Violin plot show frequencies of CD27 lo CX3CR1 + Vd1 effector cells for acute (red; n = 8) and convalescent (blue; n = 9) patients.(d)Representative FACS plots gated on CD28 + CD27 + and CD28 À CD27 À Vc9/Vd2 + T cells for an acute (left) and convalescent (right) COVID-19 patient.Violin plots show frequencies of both subsets in acute (red; n = 8) and convalescent (blue; n = 9) patients.Normality was tested using a Shapiro-Wilk test.Normally distributed data were tested using an unpaired t-test and non-Gaussian distributed data were tested using a Mann-Whitney U-test.*P < 0.05, **P < 0.01.

Figure 4 .
Figure 4.The cd TCR repertoire in acute and convalescent COVID-19 patients.(a) Vd chain usage in acute (red; n = 3) and convalescent (blue; n = 2) COVID-19 patients.Treeplots, accumulated frequencies of the top 20 clonotypes, and D75 is shown in acute (red; n = 3 or 2) and convalescent (blue; n = 2) COVID-19 patients for TRD from (b) Vd1 + and (c) Vd2 + T cells.Each color in the treeplots represents a different cd TCR clonotype and the size represents the number of reads for each clonotype.(d) Sharing of the top 20 cd TCR clonotypes in an acute COVID-19 patient (n = 1) during acute and convalescent disease state.Blue scale = expanding clonotypes, Gray scale = contracting clonotypes, Orange-Yellow scale = new clonotypes.

Table 1 .
Demographic and clinical information a Not determined for 2 donors.