nature immunology

Article https://doi.org/10.1038/s41590-024-02072-9

Robust mucosal SARS-CoV-2-specific

T cells effectively combat COVID-19 and
establish polyfunctional resident memory
in patient lungs

Received: 4 June 2024 A list of authors and their affiliations appears at the end of the paper

Accepted: 17 December 2024

Published online: xx xx xxxx

Mucosal antigen-specific T cells are pivotal for pathogen clearance and
immune modulation in respiratory infections. Dysregulated T cell responses
Check for updates exacerbate coronavirus disease 2019 severity, marked by cytokine storms
and respiratory failure. Despite extensive description in peripheral blood,
the characteristics of severe acute respiratory syndrome coronavirus 2
(SARS-CoV-2)-specific T cells in the lungs remain elusive. Here we
conducted integrated single-cell profiling of SARS-CoV-2-specific T cells in
122 bronchoalveolar lavage fluid (BALF) and 280 blood samples from 159
patients, including 27 paired BALF and blood samples from 24 patients.
SARS-CoV-2-specific T cells were robustly elicited in BALF irrespective
of prior vaccination, correlating with diminished viral loads, lessened
systemic inflammation and improved respiratory function. SARS-CoV-2-
specific T cells in BALF exhibited profound activation, along with
proliferative and multi-cytokine-producing capabilities and a glycolysis-
driven metabolic signature, which were distinct from those observed in
peripheral blood mononuclear cells. After viral clearance, these specific
T cells maintained a polyfunctional tissue-resident memory phenotype,
highlighting their critical roles in infection control and long-term
protection.

The global coronavirus disease 2019 (COVID-19) pandemic, caused by SARS-CoV-2-specific T cell responses have been thoroughly
severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), per- investigated in peripheral blood4–8. These T cells generally display a
sists worldwide. Respiratory failure and severe pneumonia are major cytotoxic, type 1 helper T (TH1)-dominant effector phenotype during
contributors to morbidity and mortality in COVID-19, emphasizing the the acute phase4,9. While interferon-γ (IFNγ)-producing CD8+ T cells
importance of understanding the immune responses in the lower res- in particular have been linked to lower disease severity10, another
piratory tract1. However, research on specific T cell responses in lungs study showed that IFNγ-producing SARS-CoV-2-specific CD4+ and CD8+
has been limited because of challenges in obtaining bronchoalveolar T cell numbers in the blood did not correlate with clinical severity6.
lavage fluid (BALF) and lung tissue samples, and the low number of Another contentious issue is whether virus-specific T cells expressing
T cells in these samples. While previous studies identified dysregulated inhibitory markers such as programmed cell death protein 1 (PD-1)
T cells2 and inflammatory circuits involving alveolar macrophages undergo exhaustion11–13. In addition, peripheral memory T (TM) cells
(AMs)3 in the lungs of patients with COVID-19, detailed characterization in convalescents persist and retain antiviral functionality14,15, but it is
of SARS-CoV-2-specific T cells in the alveolar space is lacking. unclear how these responses in blood correlate with those in the lungs,

e-mail: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

or whether they have a protective or pathogenic role in the pulmonary BALF virus-specific T cells are linked to improved outcomes
environment. Additionally, studies on vaccine-induced T cell responses To investigate factors influencing airway T cell responses, we strati-
have shown rapid recall and enhanced immunity during breakthrough fied samples based on vaccination history, disease severity, sampling
infections (BIs), suggesting T cell immune imprinting16–18. However, time after symptom onset, and viral RNA shedding. Unlike periph-
the impact of vaccination on pulmonary T cells is not well understood. eral T cells17–19, airway virus-specific T cell frequencies showed no
In this study, we systematically mapped the phenotypic and func- significant differences between BIs and NIs when analyzed at simi-
tional profiles of SARS-CoV-2-specific T cells in the lower airway of lar time points (Fig. 2h,i). Furthermore, no correlation was observed
patients with COVID-19 throughout the disease course. A total of 122 between T cell responses and the number of vaccine doses received,
BALF samples and 280 blood samples from 159 patients were character- suggesting that prior intramuscular COVID-19 vaccination had lim-
ized using molecular, transcriptomic and T cell receptor (TCR) profiling ited impact on mucosal T cell responses (Fig. 2k,l). Although airway
at the single-cell level. Our analysis revealed that BALF virus-specific SARS-CoV-2-specific T cell levels tended to be higher in severe cases
T cells were highly activated, proliferative and exhibited multi-cytokine later in the disease course (>2 weeks post symptom onset (wpo)), this
production, along with a glycolysis-drive metabolic signature. These did not reach statistical significance (Fig. 2h,i).
features were notably distinct from those of peripheral T cells, with Moreover, a strong positive association was observed between
differences in magnitude, functionality, phenotype and metabolism. airway-specific CD8+ and CD4+ T cells (Fig. 2j), indicating a potential
Importantly, robust airway virus-specific T cells were associated synergistic relationship. Neither CD8+ nor CD4+ T cell responses were
with reduced viral loads and improved respiratory function, under- associated with neutralizing antibody levels in BALF or the age of
scoring their protective roles during infection. Notably, pulmonary patients in this cohort (Fig. 2k,l). To examine the functional role of
SARS-CoV-2-specific T cell responses were comparable between natural airway virus-specific T cells, we correlated their frequencies with viral
and BIs, suggesting limited induction of mucosal T cell memory by prior loads and respiratory function. Enhanced T cell responses were associ-
intramuscular vaccination. After viral clearance, these specific T cells ated with lower viral gene copies, suggesting a role in viral clearance
retained a resident memory phenotype, potentially contributing to (Fig. 2k,l). Moreover, higher frequencies of specific CD8+ T cells in
future immune protection. These findings emphasize the critical role BALF correlated with improved oxygenation (PaO2:FiO2 ratio), oxy-
of airway virus-specific T cells in viral clearance and clinical outcomes. gen delivery, and blood pH levels, indicating a beneficial effect on
respiratory function (Fig. 2l). These T cells were also inversely corre-
Results lated with systemic inflammatory markers (C-reactive protein (CRP),
Demographic characteristics interleukin-8 (IL-8), interleukin-6 (IL-6)), and corticosteroid use (Fig. 2l).
In this study, we enrolled 159 patients with COVID-19 from the First Affili- In contrast, higher specific CD4+ T cell levels in BALF were associated
ated Hospital of Guangzhou Medical University between January 2020 with markers of severe disease, such as elevated carboxyhemoglobin
and January 2023. The cohort included individuals infected with the (COHb), thrombomodulin (TM), neutrophilia, lymphopenia, and
wild-type, Delta, and BA.5 strains, all of whom were experiencing their higher neutrophil:lymphocyte ratios (NLRs) (Fig. 2k), indicating a
primary SARS-CoV-2 infection at the time of enrollment. The median distinct functional role compared to CD8+ T cells.
age of participants was 64 years (interquartile range (IQR) = 52–76),
including 98 males (61.6%) and 61 females (38.4%) (Fig. 1a and Extended Enhanced activation and functionality of BALF-specific T cells
Data Table 1). Among these patients, 55 patients (34.6%) had BIs after In addition to the higher frequencies, specific T cells in BALF exhibited
receiving inactivated COVID-19 vaccines (CoronaVac or BBIBP-CorV), higher levels of IFNγ and tumor necrosis factor (TNF) compared to
while 104 patients (65.4%) were unvaccinated at the time of natural PBMCs (Fig. 3a), thus reflecting stronger activation. We further char-
infection (NI) (Fig. 1a and Extended Data Table 1). The cohort included acterized these T cells using a 29-color flow cytometry panel, assessing
81 mild (51%) and 78 severe (49%) cases, with nine fatalities (5.7%) (Fig. 1a viability, surface markers, intracellular cytokines, chemokine recep-
and Extended Data Table 1). Over 70% of patients had comorbidities, tors and immune checkpoint molecules. Virus-specific T cells in BALF
with hypertension being the most prevalent (50.4%) (Extended Data exhibited superior multi-cytokine production, with polyfunctionality
Table 1). Twenty-four patients (PT1–PT24) provided 27 paired BALF increasing over the course of the disease (Fig. 3b). Correlation analyses
and peripheral blood mononuclear cells (PBMC) samples, with PT1 showed a positive association between disease duration and polyfunc-
contributing samples at three time points and PT7 at two time points; tional specific T cells in BALF, but not in PBMCs (Fig. 3c), highlighting
95 patients (PT25–PT119) provided only BALF samples, each at a single the dynamic nature of T cell functionality in the lungs.
time point; 44 patients (PT1, PT4, PT5, PT8 and PT120–PT159) contrib- Dimensionality reduction analysis of paired BALF and PBMC sam-
uted a total of 253 longitudinal blood samples (Fig. 1). ples revealed distinct segregation between the two compartments
(Fig. 3d), identifying nine T cell clusters (Fig. 3e). Notably, clusters
Robustly elicited SARS-CoV-2-specific T cells in BALF enriched in BALF included C1 (CD103+CD8+ tissue-resident TM cells
We assessed SARS-CoV-2-specific T cell responses in alveoli and peri­ (TRM)), C5 (CXCR3hiCCR5hiCD4+ T cells), C6 (PD-1+CD4+ T cells) and C7
pheral blood using intracellular cytokine staining (ICS) after stimulation (polyfunctional CD4+ T cells), while PBMCs were enriched in less acti-
with a SARS-CoV-2 peptide library (S/N/M/E/ORF) (Extended Data Fig. 1b). vated subsets (IFNγlo) (Fig. 3e and Extended Data Fig. 2a). These T cells
In paired BALF and PBMC samples from patient PT1 at three sequential in BALF displayed potent cytokine production (IFNγ, TNF), activation
time points, virus-specific T cells were detected in the lungs as early as markers (CD69, HLA-DR) and tissue-resident markers (CD103, CCR5),
16 days post symptom onset (dpo), increasing notably by 1 month after while displaying lower levels of immune inhibitory molecules (T cell
disease onset, while remaining at low levels in peripheral blood (Fig. 2a). immunoglobulin and mucin domain-containing protein 3 (TIM-3),
Stronger virus-specific T cell responses were observed in BALF compared PD-1) (Fig. 3f,g). Memory subset analysis showed that virus-specific
to paired PBMC samples (Fig. 2b,c), which was extended by the analysis T cells in BALF, along with the total T cell population, were predomi-
of 114 BALF and 272 PBMC samples from 111 and 55 patients, respectively nantly effector memory T (TEM) cells (CCR7−CD45RA−), in contrast
(Fig. 2d,e). Moreover, the time since symptom onset positively correlated to the memory profiles seen in PBMCs (Fig. 3h and Extended Data
with the abundance of virus-specific CD4+ and CD8+ T cells in BALF, but Fig. 2b,c). These findings suggest that airway virus-specific T cells
not in PBMCs. This finding was further confirmed in the correlation possess enhanced activation, broad functionality and a tissue-resident
analyses (Fig. 2k,l). Notably, no correlation was observed between the phenotype, probably driven by the antigen-rich inflammatory lung
proportion of specific T cells in BALF and PBMCs (Fig. 2f,g), suggesting environment. This hypothesis is further supported by elevated levels
distinct immune dynamics in the lung and peripheral compartments. of IFNs (IFNα, IFNγ), pro-inflammatory cytokines (IL-1β, IL-6, TNF,

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

a PT1
PT2
PT3
PT4
PT5
PT6
PT7
PT8
PT9
PT10
PT11
PT12
PT13
PT14
PT15
PT16
PT17
PT18
PT19
PT20
PT21
PT22
PT23
PT24
PT25
PT26
PT27
PT28
PT29
PT30
PT31
PT32
PT33
PT34
PT35
PT36 BALF cytokine Age Severity
PT37
PT38 No 17−19 Mild
PT39 Yes 20−29 Severe
PT40
PT41 30−39
PT42 40−49
PT43
PT44
Plasma cytokine 50−59 Clinical outcomes
PT45 No 60−69 Discharge
PT46
PT47 Yes 70−79 Expired
PT48 80−89
PT49
PT50 90−99
PT51 ICS Sample type
PT52
PT53 No BALF
PT54
Yes Sex
PT55 BALF and PBMC
PT56 Female
PT57 PBMC
PT58 Male
PT59 scRNA-seq
PT60
PT61 No
Comorbidity ViralRNA
PT62 Yes +
PT63 RNA
PT64 No −
PT65 RNA
PT66 Smart-seq2 Yes
PT67 Unknown
PT68
PT69 No
PT70 Yes Vaccination
PT71
PT72
Sample type
PT73 No BALF
PT74 Yes
PT75 BALF and Blood
PT76
PT77 Blood
PT78
PT79
PT80 Time to endpoint event
PT81
PT82
PT83
PT84
PT85
PT86
PT87
PT88
PT89
PT90
PT91
PT92
PT93
PT94
PT95
PT96
PT97
PT98
PT99
b
PT100
PT101 PT1 3 sequential
PT102
PT103 27 paired
PT104
PT105 24 PTs PT7 2 sequential BALF and blood
PT106
PT107
PT108
PT109 PT2–PT6, PT8–PT24 22 single
PT110
PT111
PT112
PT113
PT114
95 PTs PT25–PT119 95 single BALF
PT115
PT116
PT117
PT118
PT119 PT1, PT4, PT5, PT8
PT120
PT121 44 PTs 253 longitudinal blood
PT122
PT123 PT120–PT159
PT124
PT125
PT126
PT127
PT128
PT129
PT130
PT131
PT132
PT133
PT134
PT135
PT136
PT137
PT138
PT139
PT140
PT141
PT142
PT143
PT144
PT145
PT146
PT147
PT148
PT149
PT150
PT151
PT152
PT153
PT154
PT155
PT156
PT157
PT158
PT159
BALF cytokine
Plasma cytokine
ICS
scRNA-seq
Smart-seq2
Age
Sex
Comorbidity
Vaccination
Severity
Clinical outcomes
Sample type

0 14 28 42 56 70 84 98 112 126
dpo

Fig. 1 | Study design and patient cohort. a, Schematic overview of patient Comorbidities depicted include hypertension, diabetes, chronic pulmonary
demographics, sample details, and omics assays used in this study. The x axis diseases, cancer, coronary artery disease, stroke, history of organ transplantation,
represents the days after disease onset, while the y axis lists the patient IDs and chronic renal disease. b, Summary of patients who provided BALF and PBMC
(PT1–PT159). Patient IDs are prominently displayed in the central column. samples in this study. For additional details, see Extended Data Table 1.

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

granulocyte colony-stimulating factor), chemokines (CCL2, CXCL8, public repertoire, while 19 CD4+ and 33 CD8+ clones in PBMCs (Fig. 4i).
CCL3) and T cell-related cytokines (interleukin-2 (IL-2)) in BALF com- This confirms that aligning single-cell TCR sequences with autologous
pared to plasma (Extended Data Fig. 2d). short-term T cell line sequences enhances SARS-CoV-2-specific T cell
identification.
Heightened clonal expansion of BALF-specific T cells
While the transcriptomic landscape of total T cells in the BALF of Activated antiviral transcriptome of BALF-specific T cells
patients with COVID-19 has been characterized2, the features of Unsupervised clustering of CD4+ and CD8+ T cells in the transcriptomic
SARS-CoV-2-specific T cells in this compartment are poorly understood. dataset (Extended Data Fig. 3) revealed 12 distinct clusters (Fig. 5a),
To map the transcriptional profile of virus-specific T cells in BALF, we distributed across BALF and PBMCs (Fig. 5b). These clusters were
performed single-cell RNA sequencing (scRNA-seq) and single-cell TCR annotated as cytotoxic (C0; NKG7, GZMB, GZMH, GNLY, PRF1, CCL4
sequencing (scTCR-seq) on 16 paired BALF and PBMC samples from 13 and CCL5), type I IFN response (C7; IFIH1 and IFIT1–IFIT3), cycling (C6;
patients, generating a dataset of 186,451 cells (Fig. 4a and Extended CENPU, MCM4 and MKI67), regulatory T (Treg) cells (C8; CTLA4, FOXP3
Data Fig. 3a,b). This dataset included epithelial cells, tissue-resident and IL2RA), naive T (TN) cells (C1 and C9; EEF1B2, EEF1G, RPS3A and
AMs (TRAMs), monocyte-derived macrophages, monocytes, dendritic RPL22), central memory T (TCM) cells (C5 and C10; CCR7, LEF1, IL7R and
cells (DCs), B cells, T cells, plasma cells, natural killer (NK) cells and TCF7) and TRM cells (C3 and C4; ZEB2, ITGA1, ITGAE, RBPJ and CXCR6)
NKT cells. Consistent with previous studies20–22, the cellular distribu- (Fig. 5a–d). BALF-specific T cells were predominantly CXCR3+CD4+
tion in BALF differed substantially from that in the peripheral blood. TEM cells (C2), RBPJ+CD4+ TEM/TRM cells (C4) and NKG7+CD8+ TEM cells
We identified 12,093 T cells in BALF, distributed across 5,290 unique (C0), ITGA1+CD8 TEM/TRM cells (C3) and proliferative T cells (C6), with
clones, compared to 22,750 T cells in PBMCs, corresponding to 10,236 higher expression of effector and antiviral genes compared to PBMCs
distinct clones (Fig. 4b). Notably, 300 clonotypes were shared between (Fig. 5e–g). Gene set enrichment analysis (GSEA) identified significant
the two compartments, indicating a degree of clonal overlap (Fig. 4b). enrichment of pathways related to IFNγ response, T cell activation,
To identify SARS-CoV-2-specific T cells, autologous PBMCs were inflammation, tissue migration, proliferation and metabolism in BALF
labeled with Cell Trace Violet (CTV), stimulated with a SARS-CoV-2 pep- virus-specific T cells (Fig. 5h).
tide pool and cultured ex vivo. After 12 days, proliferated antigen-specific Functional gene modules highlighted higher expression of core
T cells (CTVlo) were sorted and subjected to bulk TCR-seq (Fig. 4a and genes linked to activation (HLA-DRB5, ICOS), proliferation (CENPM),
Extended Data Fig. 4a)23. TCR sequences from scRNA-seq were matched survival (BCL2), cytokine (IFNG, CCL5), cytotoxicity (GZMB, PRF1), tis-
to these TCR sequences to identify virus-specific T cells. A T cell clone sue residency (RBPJ, ITGA1, ZNF683), TCR signaling pathways (LCP2,
was classified as SARS-CoV-2-specific if its TCR alpha and beta chains JUN) and anti-ferroptosis (GPX4) markers in BALF-specific T cells, while
matched 100% with the bulk TCR sequences; otherwise, it was considered peripheral markers and genes related to apoptosis (CASP3), pyroptosis
undefined (Fig. 4a). Under these criteria, we identified 430 specific T cell (NLRP3), peripheral circulation (CCR7, S1PR1), terminal differentiation
clones in BALF and 333 in PBMCs, including 1,188 and 690 specific T cells, (KLRG1) and oxidative phosphorylation (MT-CO3, COX7C) genes were
respectively (Fig. 4c). Of these, 49 clones were shared between BALF and downregulated (Fig. 5i). Metabolic analysis revealed a shift toward
PBMCs (Fig. 4c). The proportion of specific T cells identified using this mammalian target of rapamycin (mTOR) signaling (SLC7A5, CFL1),
approach showed strong concordance with those identified using the glycolysis (TPI1, LDHA), fatty acid metabolism (FABP5) and amino
ICS assay (Extended Data Fig. 4b). BALF T cells displayed greater clonal acid metabolism (SLC3A2) in BALF-specific T cells (Fig. 5i), reflecting
expansion (defined as a clone size >1) compared to those in PBMCs, with a robust response to viral infection26. These results suggest hyper-
CD8+ T cells showing more pronounced expansion than CD4+ T cells activation of T cells rather than exhaustion, as indicated by positive
(Fig. 4d,e). Moreover, the proportion of clonally expanded specific CD4+ correlations between activation, cytokine production, cytotoxicity
and CD8+ T cells was remarkably higher in BALF (Fig. 4f), with reduced and mTOR signaling (Fig. 5j and Extended Data Fig. 5a,b).
TCR diversity, indicating clonal focusing (Fig. 4g and Extended Data Further comparison of T cells bearing identical TCRs in BALF and
Fig. 4c). Notably, virus-specific T cells were predominantly represented PBMCs revealed that 49 overlapping virus-specific clones were distrib-
in the top 200 most expanded clones (Fig. 4h). Longitudinal analysis of uted across distinct clusters in each compartment (Fig. 6a). In BALF,
patient PT1 showed that virus-specific T cell clones were among the top these clones were predominantly found in clusters C0, C2, C3, C4 and
100 clones across all BALF samples (Extended Data Fig. 4d). C6; in PBMCs, they appeared in clusters C0, C2, C3 and C6 (Fig. 6b).
To further validate these findings, we cross-referenced the identi- Notably, the dominant clones in BALF did not consistently align with
fied TCR sequences with public SARS-CoV-2-specific TCR repertoires the largest clones in PBMCs (Fig. 6c), suggesting a potential difference
from the Immune Epitope Database (IEDB)24,25. As most published TCR in TCR usage between airway and peripheral compartments. Addition-
sequences are derived from single-chain data, we used 9,978 TRBV ally, shared clones in the airway, but not in the periphery, exhibited
sequences for CD4+ and 133,118 for CD8+ T cells for alignment. In BALF, significantly higher expression of genes linked to a metabolically active
25 CD4+ and 31 CD8+ T cell clones shared TRBV sequences with the and highly proliferative state (Fig. 6d).

Fig. 2 | Robust specific T cell responses in the BALF of patients with COVID-19. PBMC samples from 15 patients. h,i, Comparison of specific CD4+ (h) and CD8+
a, Representative flow plotting of specific CD4+ and CD8+ T cells expressing IFNγ (i) T cell responses in BALF between mild and severe patients with NI and BI.
and TNF upon SARS-CoV-2 peptide stimulation in paired samples from human Data are shown as mean ± s.e.m. The number of samples (ns) and patients (np) is
alveoli (BALF) and blood (PBMCs) collected over three sequential days post indicated in each panel. Statistical analysis was conducted using two-way analysis
symptom onset from patient PT1. b,c, Summary plots comparing the frequencies of variance. j, Two-tailed Spearman correlation analysis between specific CD4+
of IFNγ+CD4+ (b) and IFNγ+CD8+ T (c) cells between 18 paired BALF and PBMC and CD8+ T cell responses in 114 BALF samples from 111 patients. The shaded
samples from 15 patients. Each dot represents an individual sample. Statistical area represents the 95% confidence interval (CI). k,l, Spearman correlation
significance was assessed using a two-tailed Wilcoxon matched-pairs signed-rank analysis of specific CD4+ (k) and CD8+ (l) T cells in BALF with the indicated clinical
test. d,e, Kinetics of antigen-specific T cell responses in BALF (d) and PBMCs (e) parameters. The number of samples and patients for each analysis is indicated
after symptom onset. The number of patients and samples (ns/np) is indicated (ns/np). Two-tailed multiple comparisons were corrected using the Benjamini–
in the top left corner. Analysis was performed using a linear mixed model fit by Hochberg method. Spearman correlation coefficients (rs) and adjusted P values
restricted maximum likelihood (REML) (βdpo, fixed effects estimate), with t-tests (Padj) are shown. The shaded area represents the 95% CI. Clinical parameters
based on Satterthwaite’s method. f,g, Two-tailed Spearman correlation analysis include COHb, TM, lymphocyte count (lym), neutrophil count (neu), NLR,
of specific CD4+ (f) and CD8+ (g) T cell responses between 18 paired BALF and PaO2:FiO2 ratio and CRP.

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

a PT1 Dpo 16 Dpo 31 Dpo 34 b c

PBMCs BALF PBMCs BALF PBMCs BALF 50 –5 10 0.0010
5.30 × 10

Percentage of IFNγ+CD4+ T

Percentage of IFNγ CD8 T

0.005 1.132 0.015 13.928 0.026 16.699

+
40 8

CD4+ T

+
30 6

0.056 1.037 0.096 8.786 0.081 8.447

20 4
0.037 0.516 0.033 2.057 0.012 1.767
TNF-BUV737

10 2

CD8+ T
0 0

0.847 1.033 0.301 6.557 0.294 5.565 BALF PBMCs BALF PBMCs

IFNγ-BV421
d e h
100
100.00 100.00 NI-mild

specific CD4+ T (%)

ns/np: 114/111 ns/np: 272/55
NI-severe
specific CD4+ T

specific CD4+ T
10
–16
βdpo < 2 × 10
10.00 10.00
P = 0.271 BI-mild

BALF-
PBMC-
BALF-

1.00
1 BI-severe
1.00

0.10 βdpo = 0.03474 0.10 0.1

ns: 11 5 9 8 18 26 24 13
P = 0.0151 np: 11 5 9 8 18 23 24 13
0.01 0.01 0.01
0–2 >2
0 20 40 60 0 30 60 90
wpo
100.00 ns/np: 114/111 100.00 ns/np: 272/55 βdpo = 7.83 × 10
–9
i 100

specific CD8 T (%)

NI-mild
specific CD8+ T

specific CD8+ T

P = 0.505
10.00 10.00 10 NI-severe

+
PBMC-
BALF-

BI-mild

BALF-
1.00 1.00 1 BI-severe
0.10 βdpo = 0.03186 0.10 0.1
ns: 11 5 9 8 18 26 24 13
P = 0.039
0.01 0.01 np: 11 5 9 8 18 23 24 13
0.01
0 20 40 60 0 30 60 90 0–2 >2
dpo dpo wpo
f k
Specific CD4+ T 100.00 100.00 100.00 100.00 100.00
BALF specific CD4+ T

50
10.00 10.00 10.00 10.00 10.00
40
rs = –0.1455
30 P = 0.5645 1.00 1.00 1.00 1.00 1.00
BALF

20 rs = 0.047 rs = –0.112 rs = 0.210 rs = 0.216 rs = 0.032

0.10 0.10 0.10 0.10 0.10
Padj = 0.688 Padj = 0.286 Padj = 0.045 Padj = 0.045 Padj == 0.735
10 ns/np: 114/111 ns/np: 114/111 ns/np: 113/110 ns/np: 114/111
ns/np: 114/111
0.01 0.01 0.01 0.01 0.01
0
4 8
0 0.3 0.6 0.9 1.2 30 60 90 0 1 2 3 0 20 40 60 10 10 10 100 1,000
PBMC Age (year) Vaccine dose Sampling dpo Viral gene copies BALF FRNT50
100.00 100.00 100.00 100.00 100.00
BALF-specific CD4+ T

g Specific CD8+ T
10.00 10.00 10.00 10.00 10.00

10
1.00 1.00 1.00 1.00 1.00
8
0.10 rs = 0.214 0.10 rs = –0.222 0.10 rs = 0.260 0.10 rs = 0.232 rs = 0.291
0.10
6 rs = –0.0878 Padj = 0.045 Padj = 0.045 Padj = 0.045 Padj = 0.045
BALF

Padj = 0.045
P = 0.7291 ns/np: 102/100 ns/np: 103/101 ns/np: 103/101 ns/np: 103/101 ns/np: 55/55
4 0.01 0.01 0.01 0.01 0.01

2 0 1 0 20 40 60 30 60 90 0.1 1.0 10.0 100.0 10 20

0 COHb (%) Percentage of lym in blood Percentage of neu in blood NLR Highest TM ± 24 h
0 0.3 0.6 0.9 1.2 l 100.00
100.00 100.00 100.00 100.00
PBMCs
BALF-specific CD8 T
+

10.00 10.00 10.00 10.00 10.00

1.00 1.00 1.00 1.00 1.00

j rs = 0.201
15 0.10 rs = 0.140 0.10 rs = –0.128 0.10 rs = –0.251 0.10
rs = 0.223 0.10
BALF-specific CD8+ T

Padj = 0.150 Padj = 0.174 Padj = 0.040 Padj = 0.052

Padj = 0.047
ns/np: 114/111 ns/np: 114/111 ns/np: 113/110 ns/np: 103/101
ns/np: 114/111
0.01 0.01 0.01 0.01 0.01
10
4 8
30 60 90 0 1 2 3 0 20 40 60 10 10 250 500
Age (year) Vaccine dose Sampling dpo Viral gene copies Lowest PaO2/
5
rs = 0.676 FiO2 ± 24 h
–15
Padj: 1.68 × 10 100.00 100.00 100.00 100.00 100.00
BALF-specific CD8+ T

0 ns/np: 114/111

10.00 10.00 10.00 10.00 10.00

0 10 20 30 40
BALF-specific CD4+ T 1.00 1.00 1.00 1.00 1.00

0.10 rs = 0.352 0.10 rs = 0.239 0.10 rs = –0.204 0.10 rs = 0.328 0.10 rs = –0.268
Padj = 0.050 Padj = 0.047 Padj = 0.050 Padj = 0.050 Padj = 0.050
ns/np: 36/34 ns/np: 102/100 ns/np: 114/111 ns/np: 41/41 ns/np: 69/69
0.01 0.01 0.01 0.01 0.01
2 4 6 7.2 7.3 7.4 7.5 0 10 20 30 40 0 100 200 0 200 400
Oxygen delivery pH Corticosteroid use Highest CRP ± 24 h Highest IL-8 ± 24 h

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

a b PT2 PT3 PT1 PT1 PT4 PT5

Specific CD4+ T Specific CD8+ T d15 d19 d31 d34 d67 D70 n = 18
5 0.0151 6 P = 0.0088

BALF
P = 1.74 × 10–8

Specific CD4+ T
4
IFNγ MFI (×103)

IFNγ MFI (×103)

4 P = 0.0065
3

PBMCs
2
2
P = 0.0011
1
Cytokines

BALF
0

Specific CD8+ T
0 1
BALF PBMCs BALF PBMCs 2
3
20 8
0.0494 4

PBMCs
5
15 6
TNF MFI (×102)
TNF MFI (×102)

10 4
IFNγ + + + + + + + + + + + + + + + +
5 2
TNF – + – – – + + + – – – + + + – +
IL-2 – – + – – + – – + + – – + + + +
GM-CSF – – – + – – + – + – + + + – + +
0 0
BALF PBMCs BALF PBMCs IL-10 – – – – + – – + – + + + – + + +

c d Gated on specific T e Gated on specific T

Specific CD4+ T Specific CD8+ T
100 80 (BALF + PBMCs) (BALF + PBMCs)
BALF BALF
Polyfunctional (%)
Polyfunctional (%)

80 rs = 0.4417 rs = 0.3947
60
P = 0.0348 P = 0.0766
60 ns/np: 23/21 ns/np: 21/19
CD4+ T
40 3: IFNγ
lo

40 PBMC PBMC hi
rs = –0.0945 rs = 0.0423 4: CD38
20
TriMap y
hi hi
20 P = 0.7091 P = 0.8675 5: CXCR3 CCR5
ns/np: 18/15 ns/np: 18/15 6: PD-1
+

0 0
7: Polycytokines
0 20 40 60 80 0 20 40 60 80 +
TriMap x 8: IL-10
dpo dpo 100
9: TRM
BALF
f 80

Proportion (%)
IFNγ TNF IL-2 GM-CSF +
CD4 T CD8+ T
+ 60
CD8 T 1: TRM
lo
40 2: IFNγ
Max.
PBMCs
+ 20
CD4 T
+
CD8 T 0

CXCR3 CCR4 CCR5 CCR6

C
LF

M
BA

PB
g BALF PBMCs h
Min.
MFI CD4+ T CD8+ T
IFNγ
Max. 100
Percentage of specific

TNF
CD4+ or CD8+ T cells

CD103 HLA-DR CD38 CD69 TN

CD69 80
TCM
CD103 60 TEM
CCR5 TEMRA
40
HLA-DR Min.
20
CD38
PD-1 TIM-3 LAG-3 IL-10 PD-1 0

TIM-3
s
PB LF

PB LF
s
C

C
BA

BA
M

CD4 CD8 CD4 CD8

Fig. 3 | Activated and polyfunctional phenotypes of specific T cells in the BALF Two-tailed Spearman correlation results are shown, with sample and patient
of patients with COVID-19. a, Summary plots comparing the mean fluorescence numbers indicated (ns/np). d,e, TriMap projections of specific (IFNγ+) T cells in
intensity (MFI) of IFNγ and TNF in the IFNγ+CD4+ and IFNγ+CD8+ T cells between BALF (n = 10) and PBMC (n = 11) (d), grouped into nine clusters based on FlowSOM
18 paired BALF and PBMC samples from 15 patients. A two-tailed Wilcoxon clustering (e). The analysis included ten paired BALF and PBMC samples, plus
matched-pairs signed-rank test was used. b, Multi-cytokine expression of specific an additional PBMC sample, from eight patients. Cluster characteristics are
T cells was assessed using flow cytometry after SARS-CoV-2 peptide stimulation. displayed alongside the map, and cluster distributions and proportions are
Each 10 × 10 dot plot represents the coexpression of multiple cytokines in shown below (e, see also Extended Data Fig. 2a). f, TriMap projection showing
a representative sample, with the dots indicating 1% of cells and the colors the distribution of the indicated proteins. g, Heatmap of z-scores calculated per
reflecting cytokine profiles (legend below). Cumulative cytokine patterns from 18 marker, based on the MFI of protein expression on specific CD4+ and CD8+ T cells
paired BALF and PBMC samples of 15 patients are shown as pie charts, color-coded in BALF (n = 10) and PBMC (n = 11) samples. h, Memory phenotypes of specific T
using the number of cytokines expressed: gray (IFNγ), orange (two cytokines), cells in BALF (n = 23) and PBMCs (n = 22) were summarized based on CD45RA and
green (three cytokines), pink (four cytokines) and purple (five cytokines). CCR7 expression: TN cells (CD45RA+CCR7+), TCM cells (CD45RA−CCR7+), TEM cells
Analysis was performed using a mixed model fit by REML, with Šídák’s multiple (CD45RA−CCR7−) and TEM cells re-expressing CD45RA, CD45RA+CCR7− (TEMRA). Data
comparisons test. c, Polyfunctional specific T cells, producing more than one are presented as the mean ± s.e.m. GM-CSF, granulocyte-macrophage colony-
cytokine, were analyzed for their correlation with the sampling time points (dpo). stimulating factor. max., maximum; min., minimum.

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

a Alveoli
16 paired
BALF and PBMC

scRNA-seq T cells T cell transcriptome

Patients with scTCR-seq
P Au Single-cell transcriptomes and
COVID-19 Blood BM tolo TCR repertoire
(n = 13) Cs g
(n ou TCR alignment
= s
16
) Specific T cells
SARS-CoV-2 CTVlo T Specific T cells
85.1
peptide pool cell Bulk TCR
TCRα
stimulation sorting sequencing
12-day culture TCRβ
CTV-labeled
Short-term Proliferated Specific TCR repertoire
PBMCs
T cell lines specific T cells

b Total T c Specific T d e f g
BALF CD4+ T PBMC CD4+ T
Clone
100% 100% 100%
size BALF CD8+ T PBMC CD8+ T
4,990 clones 381 clones
75% 75% 75% x=1 Abundance Diversity
BALF 11,073 cells BALF 906 cells 15%
2≤x≤4 200
300 clones 49 clones 50% 50% 50%
4<x≤8

Abundance
BALF: 1,020 cells BALF: 282 cells 10% 150
25% 25% 25% x>8
PBMCs: 1,270 cells PBMCs: 222 cells

D
PBMCs PBMCs 0% 0% 0% 100

q
9,936 clones 284 clones 5%

PBMC CD4+ T

PBMC CD8+ T
BALF CD4+ T

BALF CD8+ T
CD4+ T
CD8+ T
PBMCs
BALF

21,480 cells 468 cells 50

0%

h
0 1 2 3
10 10 10 10 0 1 2 3 4
100
Specific Rank q
Undefine
i 10,000
9,936
Intersection

6,488
Set size
size

5,000

10,000
5,000
2,497
75 1 21 1 6 5 13 1 14 177 73 11 257 9

0
0
PBMC total CD4+ T
Percentage of top 200 T cell clones

PBMC-specific CD4+ T

BALF total CD4+ T

BALF-specific CD4+ T
50
Published specific CD4+ T

13,3056
Intersection

–5
1 × 10
Set size
size

50,000

50,000
5
1 × 10
25
27 4 5 2 2 3 89 158 26 105 26 2,370 1,388
0 0
+
PBMC total CD8 T
PBMC-specific CD8+ T

BALF total CD8+ T

BALF-specific CD8+ T
0
Published specific CD8+ T
PBMC CD4+ T

PBMC CD8+ T
BALF CD4+ T

BALF CD8+ T

Fig. 4 | Clonality and diversity of the specific T cell repertoire in BALF and clonality based on tissue (BALF and PBMC) (d) and cell types (CD4+ and CD8+ T
PBMCs. a, Experimental workflow for single-cell analysis (Chromium 5′ scRNA- cells) (e). f, Summary of specific T cell clonality. g, Abundance and diversity of
seq and scTCR-seq) and identification of SARS-CoV-2-specific T cells. Briefly, 16 specific T cell repertoire in BALF and PBMCs. Comparison of the Hill diversity
paired BALF and PBMC samples from 13 patients were subjected to single-cell index (qD, y-axis) over varying diversity orders (q, x-axis) was shown. The shaded
sequencing. The remaining cells from these PBMC samples were labeled with area represents the 95% CI. h, Distribution of specific T cell clones among the top
CTV and cultured ex vivo under stimulation with SARS-CoV-2 peptide pools; 200 ranked T cell clones in the entire T cell repertoire from BALF and PBMCs.
the proliferated short-term T cell lines (CTVlo population) were sorted for The red rectangles indicate specific T cell clones and the gray ones stand for
subsequent bulk TCR-seq, identifying specific TCRs. scTCR-seq sequences were undefined clones. i, Upset plots illustrating the TRBV clonotypes shared among
then aligned with the bulk TCR sequences to identify antigen-specific T cells and virus-specific and total T cells identified from single-cell sequencing data in
obtain their single-cell transcriptomic profiles. A cell was defined as antigen- the BALF and PBMC samples, along with published SARS-CoV-2-specific TCR
specific if both its TRAV and TRBV sequences matched those in the specific TCR sequences. Colored dots connected by lines represent shared clonotypes. The
repertoire. b,c, Venn diagrams illustrating the number of overlapping and unique horizontal bar graph shows the total number of TCR clonotypes for each cluster
TCR clones, along with the corresponding total T cell (b) and SARS-CoV-2-specific intersection, while the vertical bar graph indicates the number of clonotypes
T cell (c) counts between BALF and PBMC. d,e, Comprehensive summary of T cell shared across overlapping clusters.

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

a b c e f
BALF 100%
C3 PBMCs 100% C0
Specific BALF PBMCs C1
C0
C4 Undefined C2
C7 C0 C1 75%
C5 75% C2 C3
C11
C10 C3 C4
C4 50% C5
C2
C6 C5 C6
C1 50%
C9 C6 C7
UMAP2

C8 C7 25% C8
C8 C9
25% C10
UMAP1 C6 Proliferating T C9
C0 NKG7 TEM
C10 0% C11
+
C1 Naive CD4 C7 ISGhi TEM + T + T + T + T
C11
0% 4 4 8 8
C2 CXCR3 CD4 TEM C8 Treg CD CD CD CD
LF C LF C
C3 ITGA1 CD8 TEM/TRM C9 Naive CD8 + BA PBM BA PBM

s
hi

PB LF
C
BA

M
C4 RBPJ CD4+ TEM/TRM C10 CCR7int TCM

C5 IL7R CD4 TCM +

C11 MAIT

d Naive Cytotoxic
Cytokines/
activation TRM:TEM Tcm
IFN
Cycling response Treg MAIT
MAPK and NF-κB
signaling
C0
C1
C2 Average Percent
C3 expression expression
C4 2 0
C5 25
C6 1
50
C7 0
75
C8
C9 −1

C10
C11

REL
TNF

JUN
IL7R

FOS
CD4

LEF1

IFIT1
PRF1

RBPJ

SELL

IFIH1
IFNG

ZEB2

TCF7

IFIT3
IFIT2

RELB
GNLY

JUND
CCL4
CCL5

CCR7

S1PR1
CD8A
CD8B

NKG7

ITGA1

IL2RA
RPL22
EEF1G

GZMK

GZMB

ITGAE

MKI67
GZMH

MCM4
RPS3A

FASLG

CTLA4
FOXP3
EEF1B2

CXCR3
CXCR6

CENPU

NFKBIA
ZNF683

TRAV1−2
HLA−DPB1
HLA−DQB1

HLA−DRB5

g Specific CD4+ T (BALF versus PBMCs)

h j R = 0.27, P = 1.2 × 10–8
−14
R = 0.18, P = 9.5 × 10
−7
R = 0.35, P < 2.2 × 10
−16

Specific CD4+ T (BALF versus PBMC) R = 0.27, P = 1.7 × 10 R = 0.09, P = 0.063 R = 0.14, P = 0.0031
Upregulated: 166
Downregulated: 379 Activated 4 4 4
Padj

LAG3

LAG3

LAG3
No difference
60 Cytokine–cytokine receptor interaction
Response to IFNγ 2
TCF7 SRGN
TNF signaling via NF-κB 0.04 2 2
mTORC1 signaling 0.03
−log10(Padj)

IL7R PABPC1
B2M
40
RASA3 KLF3 IL-2−STAT5 signaling 0.02 0
LEF1 CCL5 Inflammatory response 0 0
SORL1 HLA−A 0.01
CD55
RPS23
Cell activation 0 1 2 3 0 1 2 3 4 0 1 2 3 4
PLAC8
RPL7 GAPDH GZMA
Biological adhesion
20
KLF6 DUSP4 Leukocyte migration PRF1 GZMB CCL3
S100A11 RGS1 CCL4
T cell proliferation Count
IFNG Cholesterol homeostasis 25 R = 0.3, P < 2.2 × 10
−16
R = 0.3, P < 2.2 × 10
−16
R = 0.31, P < 2.2 × 10
−16

Amino acid import R = 0.15, P = 0.0025

4
R = 0.14, P = 0.0045
4
R = 0.18, P = 0.00015
0
MCL1
Glycolysis 50
Positive regulation of protein metabolic process 4 3 3

HAVCR2

HAVCR2
75
LAG3

−2 0 2 0 1 2 2 2
2
log2(fold change) NES 1 1

Specific CD8+ T (BALF versus PBMCs) 0 0 0

Specific CD8+ T (BALF versus PBMCs) Padj 0 2 4 6 0 1 2 3 0 1 2 3 4
60 Upregulated: 486 Activated
Downregulated: 89 IL-2−STAT5 signaling 0.010 CCL4 PRF1 GZMB
No signaling
Response to IFNγ R = 0.34, P < 2.2 × 10
−16
R = 0.26, P = 4.9 × 10
−13
0.005 4 4
MT−CYB
MT−CO3
Cytokine–cytokine receptor interaction R = 0.12, P = 0.014 R = −0.0045, P = 0.93
40
Leukocyte migration
−log10(Padj)

3 3
HAVCR2

HAVCR2

VIM
DUSP4
NKG7
MT−CO2 PLP2 Biological adhesion BALF-specific CD4+ T
ITGA1
Count 2 2
FGFBP2 RPS6 LGALS3 CKLF Inflammatory response BALF-specific CD8+ T
20
MT−ND5 RPS27A KLF6 S100A11 Cell activation 20 1 1
KLF3 RPL23A LDLRAD4 RBPJ
MAPK cascade 30
MCL1 40 0 0
CCL4 G-protein-coupled receptor signaling pathway
IFNG 50 0 1 2 3 4 0 2 4 6
0 0 1 2 60 CCL3 CCL4
−2 −1 0 1 NES
o
m in

log2(fold change)
is m

i
ol a
d

ab nd
is n

io
en
os s a

at

et a
ph ive
in
id
ty

yl

m cid
pt si

al
s

os at
s
ci
ro to

or
n

ry
es

re

si
gn
r
io

id a
ph id
i
l

py op

to

to
ox

ly
va

in

e-

R
g

ac tty
at

x
si

co
bi

ec
ok

TO
in

O
su
ot
Ap
vi
iv

Fa
hi
l

ly
Eff
yt
yt

s
t

yc

TC
Su

m
Ac

In

Ti

G
C
C
C

BALF CD4+ T Average Percent

expression expression
1.5
PBMC CD4+ T 1.0
0
0.5 25
BALF CD8+ T 0 50
−0.5 75
−1.0 100
PBMC CD8 T +
IL2
TK1

LTA

REL
JUN

TPI1
TNF

TOX

FOS

PKM
PRF1

RBPJ

CFL1
IFNG

ZEB2

LCP2

RELB

PGK1
ENO1
ICOS

BCL2

CSF2
MCL1

LAG3

S1PR1
CCR7
NKG7

KRAS

JUND

LDHA

GPX4
CCL3
CCL4
CCL5

ITGA1
CD69

RRM2

GZMK

ITGAE

RHOA

IL2RG
MKI67

GZMA
GZMB

KLRG1

NFAT5

HSPE1

FABP5
PRDX1
CASP1

NLRP3

FASLG

INSIG1
PDCD1

CD247

HSPD1
CENPF

PGAM1
CASP3
CASP4
CASP5

CXCR3
CXCR6

COX7C
NFKBIA

NAMPT

UQCRB
CENPM

BCL2L1

NFATC2
PHLDA1

ZNF683

SLC7A5

SLC3A2
HAVCR2

MAP2K3

SQSTM1

CACYBP

NDUFB9

ARL6IP5
MT-CO3
CD40LG

NDUFA12
TNFRSF4
TNFRSF9
HLA-DPB1
HLA-DQB1
HLA-DRB5

Fig. 5 | Transcriptional features of specific T cells in BALF and PBMCs. a, BALF and PBMCs. The names of the indicated DEGs are shown. Horizontal black
Uniform manifold approximation and projection (UMAP) plot displaying the dashed line: Benjamini–Hochberg Padj = 0.05. h, Enriched immune pathways in
transcriptomic clusters of all T cells from the 16 paired BALF and PBMC samples specific T cells in BALF versus PBMCs were identified based on the normalized
from 13 patients. b, Distribution of clusters in BALF (red) and PBMCs (blue). enrichment score (NES), with statistical significance defined as P < 0.05 after
c, Summary of cluster proportions on T cells in BALF and PBMCs. d, Average Benjamini–Hochberg correction. The count denotes the pathway size reflecting
expression (color scale) and percentage of expressing cells (size scale) of the number of genes detected in the expression dataset. i, Dot plots illustrating
selected genes across the indicated clusters of total T cells. e, Distribution of key genes in the indicated functional modules of specific T cells in BALF and
specific T cells (red dots) in BALF (left) and PBMCs (right). f, Summary of cluster PBMCs. j, Two-tailed Spearman correlation analysis between the indicated genes
proportions of specific T cells in BALF and PBMCs. g, Volcano plots showing expressed in specific CD4+ (pink) and CD8+ (aqua) T cells in BALF.
differentially expressed genes (DEGs) in specific CD4+ and CD8+ T cells between

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

a BALF PBMCs b c
BALF

C0
C1 200

C2

Clone size
C3
C4
C5
PBMCs C6 100

C7
C8
Overlapped clones C9
cl. 13 cl. 63 cl. 245 cl. 433 cl. 679 cl. 1324 cl. 2690 Other C10
cl. 16 cl. 86 cl. 290 cl. 473 cl. 709 cl. 1394 cl. 2701
cl. 33 cl. 116 cl. 291 cl. 474 cl. 870 cl. 1404 cl. 2719 C11 0
cl. 38 cl. 137 cl. 320 cl. 529 cl. 884 cl. 1440 cl. 2924 BALF PBMCs
cl. 52 cl. 138 cl. 323 cl. 632 cl. 1080 cl. 2487 cl. 3429
cl. 57 cl. 204 cl. 412 cl. 642 cl. 1166 cl. 2597 cl. 11708
cl. 68 cl. 220 cl. 419 cl. 675 cl. 1197 cl. 2622 cl. 15777

d Specific CD4+ T Specific CD8+ T

IFNG CCL4 GZMA ICOS IFNG CCL4 GZMA ICOS

5 4 4 6
0.0023 0.0089 0.00066 0.027 0.039 4e−05 4 0.024 3 0.043
4 4 3 3 3
3 3 4 3 2
2 2 2
2 2 2
1 2 1
1 1 1 1 1
0 0 0 0 0 0 0 0

PDCD1 LAG3 RBPJ NFKBIA LAG3 RBPJ ITGAE JUND

0.03 3 0.0046 0.033 0.0094 0.0032 0.0018 2.0 0.049 3 0.0013
2.0 3 4
2 2 1.5
1.5 2 3 2
2
1.0 2 1.0
1 1 1
1 1
0.5 1 0.5
0 0 0 0 0 0 0 0

PHLDA1 CXCR6 CCR7 NDUFA12 BCL2 CXCR3 MT-CO3 ARL6IP5

3 0.027 2.5 0.0021 0.0054 3 0.0039 2.0 0.0046 0.027 6 0.0001 0.014
3
2.0 1.5 1.5 5 2
2 2
1.5 2 4
1.0 1.0
1 1.0 1 1
0.5 1 3
0.5 0.5
2
0 0 0 0 0 0 0
1
BALF PBMCs BALF PBMCs BALF PBMCs BALF PBMCs BALF PBMCs BALF PBMCs BALF PBMCs BALF PBMCs

Fig. 6 | Comparison of overlapping specific T cell clones in BALF and PBMCs. clones colored as in a. d, Comparison of expression levels of selected genes in
a, Distribution of specific T cell clones (cl.) overlapping between BALF and the 49 overlapped T cell clones between BALF (red) and PBMCs (blue). Each dot
PBMCs, mapped onto T cell clusters colored according to the indicated clones. represents one T cell clone, with consistent clones connected by a gray line. The
b, Pie charts summarizing the cumulative cluster proportions of specific T cells box plots depict the median and IQRs, with the whiskers extending to 1.5× the IQR
covering the 49 overlapped T cell clones, with T cell clustering corresponding or maximum value (two-tailed pairwise Wilcoxon rank-sum tests with Benjamini–
to Fig. 5a. c, Variation in clone size for each clone from BALF and PBMCs, with Hochberg correction).

Smart-seq2 analysis of epitope-specific T cells in BALF population, highlighting a highly activated, potent local T cell response
To extend our findings, we performed targeted analysis on CD8+ T cells distinct from that observed in PBMCs.
specific for the B40/N322 (MEVTPSGTWL) epitope in four paired BALF
and PBMC samples from four patients (PT6, PT17, PT18 and PT19) bear- Enhanced BALF- specific T cell function after viral clearance
ing the HLA-B*40:01 alleles. In line with previous findings (Fig. 2b), the The important role of antigen-specific T cells in the lung has been well
frequency of B40/N322-specific CD8+ T cells was significantly higher established in animal models27,28. Our results revealed a negative corre-
in BALF compared to PBMCs (Extended Data Fig. 5c,d). Smart-seq2 lation between the abundance of virus-specific T cells in BALF and viral
transcriptional profiling showed elevated expression of genes linked loads (Fig. 2k,l), highlighting their potential role in viral clearance. In
to cytotoxicity (FASLG), chemotaxis (CCL3), tissue residency (ITGA1) samples collected after viral clearance (RNA−), the frequency of specific
and cell survival (PHLDA1) in BALF compared to PBMCs (Extended T cells was significantly higher compared to those collected before viral
Data Fig. 5e). GSEA further confirmed pathway enrichment involved clearance (RNA+), particularly during the 14–28 dpo window, where
in T cell activation, chemotaxis, proliferation and cytotoxicity in BALF sample numbers in the two groups were comparable (Fig. 7a,b). This
(Extended Data Fig. 5f). These B40/N322-specific CD8+ T cells mirrored increase was accompanied by the expansion of specific T cell clones in
the transcriptional characteristics of the broader antigen-specific T cell BALF (Fig. 7c). Notably, not all T cell clones showed uniform contraction

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

a + – b + – c BALF-specific T d BALF-specific T cells from PT1

RNA RNA RNA RNA
100.00 100.00
P = 0.145 BALF +
Persistent clones

BALF-specific CD4 T
BALF-specific CD4 T

P = 0.0475 100% RNA

+
+

cl. 38
10.00 10.00 cl. 52
75%
cl. 63 75
1.00 1.00 cl. 86
50% cl. 116
0.10 0.10 cl. 137

Clone size
ns: 61 52 37 31 cl. 138
25% cl. 189
np: 60 51 36 31 50
0.01 0.01 cl. 190

UMAP2
0% cl. 192
0 14 28 42 56 All data 14−28 days cl. 287
cl. 320

CD4 T RNA+
CD4 T RNA−
CD8 T RNA+
CD8 T RNA−
dpo UMAP1
cl. 326 25
100.00 100.00 P = 0.0797 P = 0.0135 − cl. 416
BALF-specific CD8 T

BALF-specific CD8 T
RNA
cl. 419
+

+
10.00 10.00 cl. 474
cl. 481
Clone size cl. 523 0
1.00 1.00 cl. 630
x=1 cl. 859 RNA
+
RNA
−

0.10 0.10 ns: 61 52 37 31 2≤x≤4 cl. 879

cl. 1166
np: 60 51 36 31 4<x≤8 cl. 2511
0.01 0.01
x>8 cl. 2517
0 14 28 42 56 All data 14−28 days Other
dpo

e BALF-specific CD4 T (RNA versus RNA )

+ − + f + +
g BALF-specific T
BALF specific CD4 T BALF specific CD4 T
CD4 CD8

Running enrichment score

Running enrichment score

Upregulated: 107 0.5

Downregulated: 981 0 NES FDR
2.36 <0.001 6.1 × 10
−15 2 2 × 10
−14
No signaling 0.4

Activation

Activation
60 2.24 <0.001 1 1
−0.2 1.83 0.0034

score

score
RPL13A RPL31 RPL39 0.3
−log (Padj)

RPS17 RPL27A UBC

0 0
40 MT−CO3 RPS20 EEF1A1 RGCC
0.2
RPL41 PPP1R15A FOS −0.4 −1 −1
10

SMAP2
IFI16 MT−CO1 HLA−B MT−ATP8 0.1 + − −
KLF6
NES FDR RNA RNA RNA+ RNA
20 −0.6 0
–2.90 <0.001 P < 2.22 × 10
−16
1.1 × 10
−11

Cytokines

Cytokines
–2.54 <0.001
−0.1 0.5 0.5

score

score
IFNG
0
0 0
+ +
BALF specific CD8 T BALF specific CD8 T
Running enrichment score

−3 −2 −1 0 1 2 3
Running enrichment score

0 −0.5 −0.5
log2(fold change) NES FDR RNA RNA
+ −
RNA RNA
+ −

2.05 0.0097
+ − + 0.4 2.18 0.0012
−12 1.5 P < 2.22 × 10−16
BALF-specific CD8 T (RNA versus RNA ) 1.0
3.4 × 10
1.83 0.0034 1.0

Resident

Resident
Upregulated: 365 −0.2

score

score
0.5 0.5
Downregulated: 299 0 0
0.2
30 No difference −0.5 −0.5
RPS15 −0.4
NES FDR RNA RNA
+ −
RNA RNA
+ −
CORO1A RACK1
−3.37 <0.001
0
−log (Padj)

20 RPS14
−2.22 0.0069 2 P < 2.22 × 10−16 −16
−0.6 P < 2.22 × 10
ACTB RPL35 RPS25 CD69 IFNG
RPL6 1
10

MBNL1 RGCC
IFITM2 RPS2 1

score

score
NAMPT
STK17B DDX3X
IFNα response T cell activation

IFN

IFN
GLS PTPN22
10 LINC01578 0 0
Response to IFNβ Tissue migration
Cytokine production −1 −1
+ − + −
RNA RNA RNA RNA
0 −16
1.5 8.4 × 10 1.0 0.0011

Inhibitory

Inhibitory
−2 −1 0 1 2 1.0
0.5

score

score
0.5
log2(fold change) 0
0
h 100% i Flow j −0.5 −0.5
DC
CD14+ monocyte Ep RNA RNA
+ −
RNA RNA
+ −

T ith
g el
M

CD16+ monocyte
CD27

ERAP1
ICAM

in ial
FLT3

cl
on rive
R1

-d AM
IL18

100
SE

Cy
IL18

TRAM log fold change (RNA– versus RNA+)

oc d
3

0.0145
e
MA

D1
SE

XN

yt
4A

Monocyte-derived AM
M
Percentage of AM in BALF

PL

High
A3

CD
75% MF 86 2
C

+ 80 GB Ligand
CD4 T AD
NG IT Low
CD8+ T
AM
CD 10 1 High
T RAM

70 GR Receptor
IFN
Cycling T 60 TIMP
2 CD4
0 Low
NK/NKT CXCL13 ITGAM

50% B Gain of interaction

40 CCR10
CD8 T

Plasma FASLG Ligand and receptor

DC IGF2
+

R
20 R3 AX
Red blood cell CXC NO L Loss of interaction
S TC
FA 4 NR H2
Platelet LA
CT B1
HA P1
Ligand and receptor
25% 0 G N
VC
Epithelial cell IT RP R2
Receptor ligand
e

+ – 2
2

RNA RNA
yt
AM

Monocyte/platelet
IT

oc
AL 2

GA
CH 48
IC

Ligand receptor
S9
CB

3LG

on
IFNG

L
CD

Monocyte/B doublet
RM
NU

GZMB
LTB

CD
IFNG

m
LG

3
FLT

+
R2

T/B doublet 4+
14

T
CD

0%
+ −
RNA RNA

Fig. 7 | Potently activated specific T cells contribute to viral clearance. the RNA+ (n = 8) and RNA− (n = 8) stages. The scoring methods are detailed in
a,b, Kinetics (a) and summary (b) of specific T cells in BALF over the sampling the Methods. Data are presented as the mean ± s.d.; a two-tailed Mann–Whitney
time (dpo) during viral shedding (RNA+, red, ns/np: 61/60) and after viral clearance U-test was used. h, Proportion of cell subclusters (clustered in Extended Data
(RNA−, blue, ns/np: 52/51). Data are presented as the mean ± s.e.m., with the Fig. 3a) in BALF during the RNA+ and RNA− stages. i, Proportion of the AM
number of samples (ns) and patients (np) indicated. Statistical comparisons population in leukocytes (CD45+) in BALF samples between the RNA+ (n = 56)
were performed using a linear mixed-effects model fit by REML, with two-tailed and RNA− (n = 50) stages based on flow data. Each dot represents one sample.
t-tests based on Satterthwaite’s method. c, Clonality analysis of specific T cells in Data are shown as the mean ± s.e.m.; a two-tailed Mann–Whitney U-test was
BALF at the RNA+ (n = 8) and RNA− (n = 8) stages. d, Distribution (left) and clone used. j, Circos plot depicting prioritized ligand–receptor interactions between
size (right) of persistent specific T cell clones from RNA+ to RNA− stages in the T cells and other cell types in BALF. The outer ring shows the color-coded cell
BALF samples. e, Volcano plot of DEGs in specific CD4+ and CD8+ T cells in BALF types, while the inner ring highlights the ligand–receptor pairs. The line and
comparing the RNA+ and RNA− stages. Horizontal black dashed line: Benjamini– arrow widths are proportional to the log fold change between the RNA− and RNA+
Hochberg Padj = 0.05. f, GSEA analysis of specific CD4+ (top) and CD8+ (bottom) progression stages in ligands and receptors, respectively. Different colors and
T cells in BALF comparing RNA+ (n = 8) versus RNA− (n = 8) stages. g, Comparison line types denote several types of interactions, as outlined in the legend. FDR,
of immune function-associated scoring modules between specific T cells during false discovery rate.

Nature Immunology
 Article https://doi.org/10.1038/s41590-024-02072-9

a BALF-specific T f BALF-specific CD4+ T g BALF-specific CD8+ T

S1PR1
RNA+ CD4+ T C6 C4 C2 TCF7
S1PR1
CCR7 TNF
RNA– CD4+ T C4 C2 CCL4
CCL5
CXCR3 CCR7
RNA+ CD8+ T C6 C3 CXCR3

RNA– CD8+ T C6 C3 ITGAE

HLA-DRB5
FABP5
0% 25% 50% 75% 100% ITGAE
+
C0 NKG7 TEM C4 RBPJ CD4 TEM/TRM C8 Treg MKI67
ITGA1 CENPF
+ +
C1 Naive CD4+ C5 IL7R CD4 TCM C9 Naive CD8
MKI67
C2 CXCR3 CD4 TEM C6 Proliferating T C10 CCR7
int
TCM CENPF z-score z-score
CENPU
hi GZMB 2 2
hi
C3 ITGA1 CD8 TEM/TRM C7 ISG TEM C11 MAIT 1 1
0 0
−1 −1
−2 NFAT5 −2

b Specific CD4+ T c Specific CD8+ T

IL7R BCL2

2 BCL2
2 3 2 2
Component 2

Component 2

LCP2
0 0
LCP2
−2 −2 RBPJ ZEB2
NFAT5
−4
0 −4 0
ITGA1
1 1 ZEB2
−6 −6 ZNF683
CD247
FASLG
−10 0 −10 −5 0 5 10 JUND JUND
Component 1 Component 1 Pseudotime Pseudotime

0 10 20 0 10 20

d BALF PBMCs e BALF PBMCs h BALF-specific CD4+ T i BALF-specific CD8+ T

2 2 0.9
1.0
0 0

RNA+

RNA+
0.6
RNA+

RNA+

−2 −2 0.5
Cell type
Component 2

0.3
Component 2

−4 −4 Cell type

Density
C2
Density

0 0 C0
−6 −6 C4
C3
2 2 C6 0.9
1.0 C6
0 0 C10
RNA−

RNA−
RNA−

RNA−

0.6
−2 −2 0.5
0.3
−4 −4
−6 −6 0 0
−10 0 −10 0 −10 −5 0 5 10 −10 −5 0 5 10 0 10 20 0 10 20
Component 1 Component 1 Pseudotime Pseudotime

j Flow k Flow l BALF

Gated on specific T RNA+ RNA– RNA+ RNA– Cluster 5 Cluster 7 Cluster 9
(BALF and PBMCs) P = 0.156 6.00 × 10
–6
CXCR3hiCCR5hi Polycytokines TRM

40 0.0190 40 0.0381 50 0.0714

BALF
BALF
Specific CD8+ T

Specific CD4+ T

specific CD4+ T

specific CD4+ T

specific CD4+ T
Percentage of

Percentage of

Percentage of

30 40
30
NS NS 30
20 20
20
10 10
PBMCs

PBMCs

10
TriMap y

0 0 0
RNA + – + – + –
+ +
CD8 T CD4 T
lo +
TriMap x 1: TRM 3: IFNγ 6: PD-1
lo hi
2: IFNγ 4: CD38 7: Polycytokines
+ – +
RNA RNA 5: CXCR3 CCR5
hi hi
8: IL-10
9: TRM

Fig. 8 | Antigen-specific T cells in BALF exhibit a differentiation bias toward marker genes shown. h,i, Density plots reflecting the number of specific CD4+ (h)
multifunctional TRM cells after viral clearance. a, The proportion of indicated and CD8+ (i) T cells among the indicated T cell subclusters stratified according
specific T cell subclusters (related to Fig. 5a) in BALF varied from viral shedding to RNA+ versus RNA− in BALF. The x axis indicates pseudotime (related to f,g)
stage (RNA+, n = 8) to viral clearance (RNA−, n = 8). b,c, Pseudotime trajectories and the y axis indicates probability density distribution of each T cell lineage at a
for specific CD4+ (b) and CD8+ (c) T cells from 16 paired BALF and PBMC samples different pseudotime. j, Distribution of specific T cells from BALF and PBMCs on
based on Monocle, showing three lineages on CD4+ and two lineages on CD8+ TriMap according to RNA+ (n = 4, 5) and RNA− (n = 6, 6) stages based on the flow
T cells. Cells are color-coded based on their phenotypes (related to a), with the cytometry dataset. k, Cumulative cluster proportions of specific T cells from
arrows indicating the pseudotime trajectories. Each dot represents one cell. BALF and PBMCs at the RNA+ (n = 4, 5) and RNA− (n = 6, 6) stages were compared
d,e, Trajectory inference analyses depicting the transition potential of specific using pie charts. Each segment is color-coded to represent different functional
CD4+ (d) and CD8+ (e) T cells from viral shedding stage (RNA+) to viral clearance clusters, as shown at the bottom (related to Fig. 3e). A two-tailed chi-squared test
(RNA−) in BALF and PBMCs. f,g, Gene expression dynamics in specific CD4+ (f) and was used. l, Clusters that significantly increased in BALF from the RNA+ to the
CD8+ (g) T cells along the pseudotime in BALF. Genes were clustered into eight RNA− stages are shown. Each dot represents one sample. Data are presented as
and 12 gene sets based on specific expression profiles, respectively, with selected the mean ± s.e.m.; a two-tailed Mann–Whitney U-test was used.

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

from RNA+ to RNA− stages. Longitudinal tracking of persistent clones high-dimensional flow cytometry, scRNA-seq and TCR repertoire analy-
in PT1 revealed shifts in dominant clonal populations after viral clear- ses. We observed a significantly higher proportion of virus-specific
ance in BALF (Fig. 7d) and PBMCs (Extended Data Fig. 6a). These clonal T cells in BALF compared to PBMCs, with no direct correlation between
dynamics suggest that differences in affinity, survival signaling and the two compartments. This discrepancy highlights the distinct nature
adaptation to the local antigen environment could drive variability of local versus systemic T cell immunity, consistent with findings in
among clones. Additionally, gene expression profiling of BALF-specific SARS-CoV-2 infection models28, where epitope-specific T cells are
T cells showed significantly elevated expression of IFNG, CD69 and detected earlier in the lungs than in the blood.
RGCC, alongside reduced expression of IFI16, IFITM2 and MT-CO3 after Activated T cells rely on aerobic glycolysis to sustain their effec-
viral clearance, whereas those in PBMCs displayed minimal changes tor functions and enhance IFNγ production, a hallmark of TH1 dif-
(Fig. 7e and Extended Data Fig. 6b). These findings underscore a ferentiation31. Our data demonstrate that glycolytic metabolism in
tight relationship between local viral elimination and the activity of lung-resident CD4+ T cells supports a TH1-biased response, facilitated by
airway-specific T cells. upregulated TCR signaling and mTOR pathway activation26,32. This met-
GSEA and functional module analysis of BALF virus-specific T cells abolic reprogramming probably provides the energy required for the
after viral clearance revealed enrichment in pathways related to acti- proliferation and heightened cytotoxic activity of SARS-CoV-2-specific
vation, cytokine production and tissue migration, with a reduction in T cells in the airways. Correspondingly, we observed enhanced cytokine
type I IFN responses (Fig. 7f,g). These functional shifts in BALF-specific secretion and increased cytotoxic potential in airway T cells compared
T cells from RNA+ to RNA− were consistently observed in longitudinal to their peripheral counterparts in the context of an inflammatory lung
samples from patient PT1 (Extended Data Fig. 6c–f). Additionally, environment marked by elevated MCP-1, MIP-1α, IL-8 and IL-2 levels.
immune cell subset distribution shifted after viral clearance character- Accompanied by robust activation, airway virus-specific T cells
ized by a decrease in inflammatory monocytes (CD14+ monocyte and exhibited increased expression of inhibitory markers such as PDCD1,
monocyte-derived AM subpopulations) and plasma cells, alongside HAVCR2 and LAG3, which are commonly associated with CD8+ T cell
an increase in TRAM cells, the predominant immune cells in the BALF exhaustion in chronic infections and cancer33,34. They have been linked
of healthy donors29 (Fig. 7h). Flow cytometry corroborated this to an exhaustion-like status in peripheral specific T cells of patients
finding, revealing an increase in alveolar macrophage numbers (Fig. 7i). with COVID-19 (refs. 11–13). However, in this study, these markers
Ligand–receptor interaction analysis suggested reduced T cell inter- were positively correlated with functional modules linked to mTOR
actions with alveolar cells, indicating a regulated immune environ- signaling, activation and cytotoxicity, suggesting a regulatory role
ment after clearance (Fig. 7j). Consistently, inflammatory cytokine rather than exhaustion. The upregulation of inhibitory markers during
levels tended to reduce, albeit not significantly in BALF after clearance early infection, followed by their downregulation after viral clearance,
(Extended Data Fig. 6g). implies their involvement in modulating hyperactivation rather than
indicating functional impairment.
Post-clearance TRM differentiation of BALF-specific T cells Our findings support the critical role of activated specific T cells in
Cluster analysis of BALF virus-specific T cells revealed a substantial viral clearance, with a strong inverse correlation between alveolar T cell
increase in CXCR3+CD4+ TEM cells (C2) and ITGA1+CD8+ TEM/TRM cells (C3), abundance and viral load. This observation aligns with studies in animal
coupled with a reduction in proliferating T cells (C6) after viral clear- models of SARS-CoV-2 infection28,35,36, where CD8+ T cells contribute
ance (Fig. 8a). Pseudotime trajectory analysis30 indicated three distinct to viral control and improved lung lesions. Intriguingly, while airway
differentiation pathways for specific CD4+ T cells: early CD4+ TEM (C1, virus-specific CD8+ T cells correlated with improved clinical outcomes,
C2) cells were linked to proliferating T cells (C6), which then diverged specific CD4+ T cells were paradoxically associated with indicators of
into TRM-driven (C4) and TH1-polarized TEM (C2) cell lineages (Fig. 8b). severe disease, indicating a complex, multifaceted role for these cells
Similarly, specific CD8+ T cells followed a more linear trajectory from in immune regulation.
cytotoxic CD8+ TEM (C0) to proliferative T cells (C6) and then to TRM-like Given the challenges posed by antibody evasion, inducing long-
CD8+ cells (C3) (Fig. 8c). Notably, these pseudotime trajectories were lived mucosal T cell memory has become a key target for next-
predominantly observed in BALF virus-specific T cells, but not in PBMCs, generation vaccines. Peripheral T cell responses to COVID-19 vaccina-
underscoring the localized tissue-specific differentiation occurring in tion are more rapid and extensive during BIs16–18,37,38, but our findings
lungs during viral clearance (Fig. 8d,e). Marker gene profiling along these indicate that intramuscular vaccination does not establish robust
trajectories—covering key genes involved in tissue migration (CXCR3), pulmonary T cell immunity. We observed no significant differences
proliferation (MKI67, CENPF), cytokines (GZMB, TNF, CCL4, CCL5), tissue in the frequency or kinetics of specific T cells in BALF between vac-
residency (ITGAE, ITGA1, ZEB2) and survival (BCL2, IL7R)—confirmed cinated and unvaccinated individuals, which is consistent with previ-
the functional specialization of these clusters (Fig. 8f,g). Density plots ous reports demonstrating the absence of mucosal T cell memory in
illustrating the distribution of T cells in these dominant subclusters uninfected vaccinees after intramuscular vaccination39–41. These results
across RNA+ and RNA− phases further supported these observations, highlight a critical gap in the current vaccination strategy, emphasiz-
revealing a shift from highly proliferative TEM to TH1-polarized TEM and ing the need for mucosal vaccines to elicit potent local immunity in
TRM cells in BALF, but not in PBMCs (Fig. 8h,i and Extended Data Fig. 6h). the respiratory tract.
Flow cytometry profiles from 10 BALF and 11 PBMC samples In summary, airway-specific T cells in patients with COVID-19
(related to Fig. 3e) confirmed a predominance of TEM (cluster 7, IFNγhi exhibit enhanced activation, proliferation and survival, without evi-
TNFhi IL-2hi; cluster 5, CXCR3hi CCR5hi HLA-DRhi) and TRM (cluster 9, dence of exhaustion despite high antigenic stimulation. These T cells
CD103hi CCR5hi CXCR3hi) phenotypes in BALF after clearance (Fig. 8j–l, might have a crucial role in viral clearance and symptom resolution,
Extended Data Fig. 6i). In contrast, undefined T cells displayed minimal establishing a durable tissue-resident memory population after
changes in cellular composition across viral clearance stages (Extended infection. Our findings provide valuable insights into mucosal T cell
Data Fig. 6j). These results highlight the establishment of a poised immunity in COVID-19, with implications for vaccine development and
TM cell population in the lungs, primed for rapid response upon infectious disease control.
reinfection. There are several limitations to this study. First, detailed anti-
gen repertoire profiling and cross-reactivity analyses were limited by
Discussion the number of cells available for ex vivo peptide stimulation assays.
In this study, we comprehensively profiled virus-specific T cell Second, comprehensive analyses of antigen-specific T cell transcrip-
responses in the airway and blood of patients with COVID-19 using tomes and TCR profiles were restricted to patients with sufficient

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

paired BALF and PBMC samples. Third, longitudinal BALF samples 18. Painter, M. M. et al. Prior vaccination promotes early activation of
were only available for two patients, limiting the generalizability of memory T cells and enhances immune responses during
these findings. Lastly, the study included 27 matched BALF with PBMC SARS-CoV-2 breakthrough infection. Nat. Immunol. 24, 1711–1724
samples, and pairwise comparisons were conducted within this limited (2023).
dataset. Further studies with larger cohorts and more extensive lon- 19. Yu, S. et al. Systemic immune profiling of Omicron-infected
gitudinal sampling are needed to fully elucidate the dynamics of local subjects inoculated with different doses of inactivated virus
and systemic T cell responses in COVID-19. vaccine. Cell 186, 4615–4631 (2023).
20. Xu, G. et al. The differential immune responses to COVID-19 in
Online content peripheral and lung revealed by single-cell RNA sequencing. Cell
Any methods, additional references, Nature Portfolio reporting sum- Discov. 6, 73 (2020).
maries, source data, extended data, supplementary information, 21. Saris, A. et al. Distinct cellular immune profiles in the airways and
acknowledgements, peer review information; details of author contri- blood of critically ill patients with COVID-19. Thorax 76, 1010–1019
butions and competing interests; and statements of data and code avail- (2021).
ability are available at https://doi.org/10.1038/s41590-024-02072-9. 22. Szabo, P. A. et al. Longitudinal profiling of respiratory and
systemic immune responses reveals myeloid cell-driven lung
References inflammation in severe COVID-19. Immunity 54, 797–814 (2021).
1. Roukens, A. H. E. et al. Prolonged activation of nasal immune cell 23. Low, J. S. et al. Clonal analysis of immunodominance and
populations and development of tissue-resident SARS-CoV-2- cross-reactivity of the CD4 T cell response to SARS-CoV-2.
specific CD8+ T cell responses following COVID-19. Nat. Immunol. Science 372, 1336–1341 (2021).
23, 23–32 (2022). 24. Vita, R. et al. The Immune Epitope Database (IEDB): 2018 update.
2. Wauters, E. et al. Discriminating mild from critical COVID-19 Nucleic Acids Res. 47, D339–D343 (2019).
by innate and adaptive immune single-cell profiling of 25. Maekawa, Y. et al. Notch controls the survival of memory CD4+
bronchoalveolar lavages. Cell Res. 31, 272–290 (2021). T cells by regulating glucose uptake. Nat. Med. 21, 55–61 (2015).
3. Grant, R. A. et al. Circuits between infected macrophages and 26. O’Neill, L. A., Kishton, R. J. & Rathmell, J. A guide to immunometa­
T cells in SARS-CoV-2 pneumonia. Nature 590, 635–641 (2021). bolism for immunologists. Nat. Rev. Immunol. 16, 553–565 (2016).
4. Weiskopf, D. et al. Phenotype and kinetics of SARS-CoV-2-specific 27. Zhao, J. et al. Airway memory CD4+ T cells mediate protective
T cells in COVID-19 patients with acute respiratory distress immunity against emerging respiratory coronaviruses. Immunity
syndrome. Sci. Immunol. 5, eabd2071 (2020). 44, 1379–1391 (2016).
5. Qiu, S. et al. Successful clearance of persistent SARS-CoV-2 28. Zhuang, Z. et al. Mapping and role of T cell response in
asymptomatic infection following a single dose of Ad5-nCoV SARS-CoV-2-infected mice. J. Exp. Med. 218, e20202187 (2021).
vaccine. Signal Transduct. Target. Ther. 8, 123 (2023). 29. Travaglini, K. J. et al. A molecular cell atlas of the human lung
6. Olea, B. et al. SARS-CoV-2-reactive IFN-γ-producing CD4+ and from single-cell RNA sequencing. Nature 587, 619–625 (2020).
CD8+ T cells in blood do not correlate with clinical severity in 30. Qiu, X. et al. Reversed graph embedding resolves complex
unvaccinated critically ill COVID-19 patients. Sci. Rep. 12, 14271 single-cell trajectories. Nat. Methods 14, 979–982 (2017).
(2022). 31. Peng, M. et al. Aerobic glycolysis promotes T helper 1 cell
7. Tan, A. T. et al. Early induction of functional SARS-CoV-2-specific differenti­ation through an epigenetic mechanism. Science 354,
T cells associates with rapid viral clearance and mild disease in 481–484 (2016).
COVID-19 patients. Cell Rep. 34, 108728 (2021). 32. Kishton, R. J., Sukumar, M. & Restifo, N. P. Metabolic regulation
8. Lafon, E. et al. Potent SARS-CoV-2-specific T cell immunity and of T cell longevity and function in tumor immunotherapy. Cell
low anaphylatoxin levels correlate with mild disease progression Metab. 26, 94–109 (2017).
in COVID-19 patients. Front. Immunol. 12, 684014 (2021). 33. McLane, L. M., Abdel-Hakeem, M. S. & Wherry, E. J. CD8 T cell
9. Sekine, T. et al. Robust T cell immunity in convalescent individuals exhaustion during chronic viral infection and cancer. Annu. Rev.
with asymptomatic or mild COVID-19. Cell 183, 158–168 (2020). Immunol. 37, 457–495 (2019).
10. Rydyznski Moderbacher, C. et al. Antigen-specific adaptive 34. Wherry, E. J. & Kurachi, M. Molecular and cellular insights into
immunity to SARS-CoV-2 in acute COVID-19 and associations with T cell exhaustion. Nat. Rev. Immunol. 15, 486–499 (2015).
age and disease severity. Cell 183, 996–1012 (2020). 35. Sun, J. et al. Generation of a broadly useful model for COVID-19
11. Rha, M.-S. & Shin, E.-C. Activation or exhaustion of CD8+ T cells in pathogenesis, vaccination, and treatment. Cell 182, 734–743 (2020).
patients with COVID-19. Cell. Mol. Immunol. 18, 2325–2333 (2021). 36. Pruner, K. B. & Pepper, M. Local memory CD4 T cell niches in
12. Kusnadi, A. et al. Severely ill COVID-19 patients display impaired respiratory viral infection. J. Exp. Med. 218, e20201733 (2021).
exhaustion features in SARS-CoV-2-reactive CD8+ T cells. Sci. 37. Guo, L. et al. Omicron BA.1 breakthrough infections in inactivated
Immunol. 6, eabe4782 (2021). COVID-19 vaccine recipients induced distinct pattern of antibody
13. Rha, M.-S. et al. PD-1-expressing SARS-CoV-2-specific CD8+ T cells and T cell responses to different Omicron sublineages. Emerg.
are not exhausted, but functional in patients with COVID-19. Microbes Infect. 12, 2202263 (2023).
Immunity 54, 44–52 (2021). 38. Ahmed, M. I. M. et al. Enhanced Spike-specific, but attenuated
14. Peng, Y. et al. Broad and strong memory CD4+ and CD8+ T cells Nucleocapsid-specific T cell responses upon SARS-CoV-2
induced by SARS-CoV-2 in UK convalescent individuals following breakthrough versus non-breakthrough infections. Front.
COVID-19. Nat. Immunol. 21, 1336–1345 (2020). Immunol. 13, 1026473 (2022).
15. Rodda, L. B. et al. Functional SARS-CoV-2-specific immune 39. Lim, J. M. E. et al. SARS-CoV-2 breakthrough infection in vaccinees
memory persists after mild COVID-19. Cell 184, 169–183 (2021). induces virus-specific nasal-resident CD8+ and CD4+ T cells of
16. Koutsakos, M. et al. SARS-CoV-2 breakthrough infection induces broad specificity. J. Exp. Med. 219, e20220780 (2022).
rapid memory and de novo T cell responses. Immunity 56, 40. Tang, J. et al. Respiratory mucosal immunity against SARS-CoV-2
879–892 (2023). after mRNA vaccination. Sci. Immunol. 7, eadd4853 (2022).
17. Yu, E. D. et al. Development of a T cell-based immunodiagnostic 41. Mitsi, E. et al. Respiratory mucosal immune memory to
system to effectively distinguish SARS-CoV-2 infection and SARS-CoV-2 after infection and vaccination. Nat. Commun. 14,
COVID-19 vaccination status. Cell Host Microbe 30, 388–399 (2022). 6815 (2023).

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

Publisher’s note Springer Nature remains neutral with regard to adapted material derived from this article or parts of it. The images
jurisdictional claims in published maps and institutional affiliations. or other third party material in this article are included in the article’s
Creative Commons licence, unless indicated otherwise in a credit
Open Access This article is licensed under a Creative Commons line to the material. If material is not included in the article’s Creative
Attribution-NonCommercial-NoDerivatives 4.0 International License, Commons licence and your intended use is not permitted by statutory
which permits any non-commercial use, sharing, distribution and regulation or exceeds the permitted use, you will need to obtain
reproduction in any medium or format, as long as you give appropriate permission directly from the copyright holder. To view a copy of this
credit to the original author(s) and the source, provide a link to the licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Creative Commons licence, and indicate if you modified the licensed
material. You do not have permission under this licence to share © The Author(s) 2025

Airu Zhu1,11, Zhao Chen1,11, Qihong Yan 1,11, Mei Jiang1,11, Xuesong Liu1,2,11, Zhengtu Li1,11, Na Li3,11, Chunli Tang1,11,
Wenhua Jian1,11, Jiangping He4, Lan Chen1,4, Jinling Cheng1, Canjie Chen1, Tian Tang1, Zhiwei Xu1, Qingtao Hu 1,5, Fang Li1,
Yanqun Wang1, Jing Sun1, Zhen Zhuang1, Liyan Wen1, Jianfen Zhuo1, Donglan Liu1, Yanjun Zhang1, Xiaofang Huang1,
Suxiang Li1, Qiuhui Zeng1, Fangli Chen1, Liang Zhou1,2, Dongdong Liu1,2, Changhao Zhong1, Yu Chen1, Shiyue Li1,
Kangli Liang6, Na Zhong7, Xinmei Zhang7, Jiekai Chen 8, Xiaobo Chen 1 , Yonghao Xu 1,2,4 , Nanshan Zhong1,4 ,
Jingxian Zhao 1,4 & Jincun Zhao 1,4,9,10

State Key Laboratory of Respiratory Disease, National Clinical Research Centre for Respiratory Disease, Guangzhou Institute of Respiratory Health, The
1

First Affiliated Hospital of Guangzhou Medical University, Guangzhou, China. 2Department of Critical Care Medicine, State Key Laboratory of Respiratory
Diseases, Guangzhou Institute of Respiratory Health, National Clinical Research Center for Respiratory Disease, The First Affiliated Hospital of Guangzhou
Medical University, Guangzhou, China. 3Southern Medical University Hospital of Integrated Traditional Chinese and Western Medicine, Southern
Medical University, Guangzhou, China. 4Guangzhou National Laboratory, Guangzhou, China. 5GMU-GIBH Joint School of Life Sciences, Guangzhou
Medical University, Guangzhou, China. 6Guangdong Hospital of Integrated Traditional Chinese and Western Medicine, Foshan, China. 7Shenzhen
Peacock Biotechnology Co. Ltd, Shenzhen, China. 8Center for Cell Lineage and Development, Guangdong Provincial Key Laboratory of Stem Cell and
Regenerative Medicine, Guangzhou Institutes of Biomedicine and Health, Chinese Academy of Sciences, Guangzhou, China. 9Shanghai Institute for
Advanced Immunochemical Studies, School of Life Science and Technology, ShanghaiTech University, Shanghai, China. 10National Clinical Research
Center for Infectious Disease, Shenzhen Third People’s Hospital; The Second Affiliated Hospital, School of Medicine, Southern University of Science and
Technology, Shenzhen, China. 11These authors contributed equally: Airu Zhu, Zhao Chen, Qihong Yan, Mei Jiang, Xuesong Liu, Zhengtu Li, Na Li,
Chunli Tang, Wenhua Jian. e-mail: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

Methods Flow cytometry data were acquired on an LSRFortessa X-20 or a

Patient enrollment and sample collection FACSymphony cell analyzer (BD Biosciences) with the FACSDiva soft-
From January 2020 to January 2023, we enrolled 159 patients with ware (v.9.0), and analyzed using FlowJo (v.10.6.2) (FlowJo LLC). Dimen-
COVID-19, confirmed using real-time quantitative PCR with reverse tran- sionality reduction and clustering analysis of flow cytometry data
scription (RT–qPCR), from the First Affiliated Hospital of Guangzhou were performed using the TriMap42 and FlowSOM43 plugins. Markers
Medical University (Ethics IDs: ES-2020-65, ES-2023-015-01). Written included in the cluster analysis were CD4, CD8, IFNγ, TNF, IL-2, IL-10,
informed consent was obtained from all participants. Detailed patient GM-CSF, CXCR3, CCR4, CCR5, CCR6, CD103, HLA-DR, CD38, CD69,
and sample information are provided in Fig. 1 and Extended Data Table 1. PD-1, TIM-3 and LAG-3.
Of the 159 patients (PT1–PT159), 44 (PT1, PT4, PT5, PT8, PT120–PT159)
were recruited during the Wuhan strain outbreak, 11 (PT2, PT3, PT6, Cytokine quantification
PT7, PT17–PT23) during the Delta variant wave and 104 (PT9–PT16, Cytokine quantification in the plasma and BALF supernatants was per-
PT24–PT119) during the Omicron BA.5 subvariant surge in China. formed using the BD Cytometric Bead Array (BD Biosciences) accord-
A total of 122 BALF samples were collected from 119 patients (PT1– ing to the manufacture’s protocol. Briefly, 50 μl of sample, 50 μl of
PT119), along with 280 longitudinal peripheral blood samples from 64 mixed capture beads and 50 μl of detection reagent were incubated
patients (PT1–PT24 and PT120–PT159). The median (IQR) time of BALF for 3 h at room temperature in the dark, followed by a washing step
collection was 18 (14–24) days after symptom onset. BALF samples were with 1 ml wash buffer and centrifugation at 300g for 5 min. The beads
centrifuged at 800g for 10 min at 4 °C to separate the supernatant from were then resuspended in BD Cytofix fixation buffer instead of assay
the cell pellet. PBMCs were isolated using Leucosep tubes (Greiner) and diluent. Standard curves for each analyte were generated using diluted
Ficoll-Paque PLUS (GE Healthcare) according to the manufacturer’s standards. Data acquisition was performed on a BD FACSVerse flow
instructions. BALF cell pellets and PBMCs were cryopreserved in liquid cytometer (BD Biosciences) with the FACSuite software (v.1.0.6), and
nitrogen, while supernatants and plasma samples were stored at −80 °C analyzed using the FCAP Array software v.3.0 (BD Biosciences).
until further analysis.
RT–qPCR detection of SARS-CoV-2
Peptide library Nucleic acids were extracted from the BALF samples using the Universal
A library of 20-mer peptides, overlapping by ten amino acids, covering RNA Purification Kit (cat. no. EZB-RN4, EZBioscience) according to
four SARS-CoV-2 structural proteins (Spike glycoprotein, nucleocapsid, the manufacturer’s protocol. RT–qPCR was conducted to quantify
membrane and envelope) and six accessory proteins (ORF3a, ORF6, SARS-CoV-2 viral RNA, targeting the open reading frame 1ab (ORF1ab)
ORF7a, ORF8, ORF9b and ORF9c), was synthesized by GL Biochem. and nucleocapsid (N) genes (cat no. DA0932, DaAnGene). RNA results
These peptides were used to stimulate BALF cells and PBMCs for ICS were defined as positive when the cycle threshold (Ct) values for both
and short-term T cell line expansion. genes were below 40.

Flow cytometry scRNA library preparation and sequencing

The following anti-human monoclonal antibodies were used: PE/ BALF cells and PBMCs were sorted for live cells using a FACSAria III cell
Dazzle594-CD103 (clone Ber-ACT8); APC-R700-CD127 (clone sorter (BD Biosciences), followed by scRNA-seq using microfluidic
HIL-7R-M21); BV570-CD14 (clone MφP9); BV711-CD14 (clone MφP9); single-cell sequencing platforms. Single-cell suspensions from the
APC-CD16 (clone B73.1); BB700-CD163 (clone GHI/61); PE/Cy7-CXCR3 PBMC or BALF sample were applied to the 10x Genomics or MobiNova
(clone 1C6/CXCR3); BV750-CCR4 (clone 1G1); BB790-CCR5 (clone workflow for cell capturing and scRNA gene expression and TCR
3A9); BB660-CCR6 (clone 11A9); PE-Cy7-CCR7 (clone G043H7); library preparation. Single-cell RNA libraries were prepared using the
PE-Cy5-CD206 (clone 19.2); BUV737-CD25 (clone 2A3); BV786-PD-1 Chromium Single Cell 5′ v2 Reagent (cat. no. 120237, 10x Genomics) or
(clone EH12.1); PE-Cy7-PD-1 (clone EH12.1); AF700-CD3 (clone HIT3a); MobiCube High-throughput Single Cell 5′ RNA-Seq Kit v2.0 (cat. no.
BUV395-CD3 (clone UCHT1); BV711-CD38 (clone HIT2); BV605-CD4 AS060200201, MobiNova) according to the manufacturer’s instruc-
(clone RPA-T4); BV421-CD4 (clone L200); BB515-CD4 (clone RPA-T4); tions. TCR libraries was prepared using the Chromium Single Cell V(D)J
BUV805-CD45 (clone HI30); BV510-CD45 (clone HI30); BUV737-CD45RA Enrichment Kit (cat. no. PN-1000005, 10x Genomics) or MobiCube TCR
(clone HI100); BUV395-CD56 (clone NCAM16.2); BV650-CD56 (clone Library Reagent Kit (Human) v2.0 (cat. no. AS060410301, MobiNova)
NCAM16.2); BV510-CD69 (clone FN50); BUV615-CD8 (clone SK1); according to the manufacturer’s instructions. Sequencing was per-
BV786-CD8 (clone RPA-T8); APC-R700-CD86 (clone 2331 FUN-1); formed on Illumina platforms using High Output v2 Kits (150 cycles)
BUV563-GM-CSF (clone BVD2-21C11); Percp-cy5.5-GM-CSF (clone with the recommended sequencing conditions for 5′ gene expression
BVD2-21C11); FITC-HLA-I (clone W6/32); BUV496-HLA-DR (clone G46- libraries and Mid Output v2 Kits for TCR libraries.
6); BV605-HLA-DR (clone G46-6); APC-IFNγ (clone B27), BV421-IFNγ
(clone B27); PE-Cy5-IL-10 (clone JES3-19F1); BV711-IL-10 (clone JES3- Antigen-specific T cell ex vivo expansion and RNA extraction
9D7); PE-IL-1β, BB515-IL-2 (clone MQ2-13A5); BV605-IL-2 (clone MQ1- Autologous PBMCs collected at the same time as BALF sampling were
17H12); BV480-IL-6 (clone MQ2-13A5); BV650-IL-8 (clone G265-8); revived, labeled with 5 μM CTV (Thermo Fisher Scientific), stimulated
APC-Cy7-LAG-3 (clone T47-530); BUV661-TIM-3 (clone 7D3); APC-TIM-3 with a SARS-CoV-2 peptide pool (200 nM for each peptide) and cultured
(clone FAB2365A); BUV737-TNF (clone MAb11); and PE-TNF (clone for 12 days ex vivo at 37 °C, 5% CO2. On days 3, 6 and 9 after stimulation,
MAb11). All antibodies were sourced from BD Biosciences, R&D Systems the culture medium was half-replaced with fresh complete Roswell
or BioLegend. Park Memorial Institute 1640 medium supplemented with 40 U ml−1
To detect specific T cells, cells were incubated in 96-well recombinant IL-2. The CTVloCD4+ T cells and CD8+ T cells that under-
round-bottom plates with 105–106 cells per well at 37 °C for 16–18 h, in went antigen-specific expansion were sorted using flow cytometry.
the presence of a SARS-CoV-2 peptide pool (final 200 nM for each pep- Subsequently, RNA extraction was performed according to the proto-
tide) and brefeldin A (final 1:1,000 dilution, BD Biosciences). Cells were col using the EZ-press RNA Purification Kit (EZBioscience).
first stained with the indicated surface antibodies at 4 °C for 15 min,
followed by viability dye staining (FVS440UV, BD Biosciences) at room Unbiased amplification of TCRs
temperature (20–25 °C) for 10 min in the dark. Cells were then fixed and The RNA of CTVlo SARS-CoV-2-specific CD4+ and CD8+ short-term T cell
permeabilized using Cytofix/Cytoperm solution (BD Biosciences) for lines was extracted and processed for TCR repertoire sequencing
intracellular marker staining. using iR-RepSeq-plus 7-Chain Cassette (cat. no. iR+7 chain-HLRI-C,

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

iRepertoire) (for detailed protocols, see the iRepertoire website; IL13, IL17A, IL21, IL22, TGFB1, IL1B, IL6, EBI3, IFNB1 and IFNA1; resident
https://irepertoire.com/). Briefly, NGS libraries covering TCRα and scores: CXCR3, CXCR6, ICAM1, ITGA1, ITGAE, RBPJ, ZEB2 and ZNF683; IFN
TCRβ chains were generated according to the manufacture’s proto- scores: IFIH1, IFIT1, IFIT2, IFIT3, FOS, JUN, JUND, REL, IRF1, IRF7, ISG15,
col. Reverse transcription was conducted with the OneStep RT–PCR MX1, STAT1, NFKBIA, TRIM21, OAS1 and OAS2; and inhibitory scores:
mix (QIAGEN). First-strand complementary DNA was purified with CTLA4, PDCD1, LAG3, HAVCR2, BTLA and TIGIT).
two rounds of SPRIselect beads binding (Beckman Coulter), followed For the Smart-seq2 data, reads were mapped to the GRCh38 refer-
by an extension using a V-gene primer mix, with beads purifying the ence genome using HISAT (v.2.1)53, and gene counts were quantified
first and second strand synthesis products. Library amplification with featureCounts in the SubReads package (v.1.5.3)54. Data integra-
used primers targeting communal sites engineered onto the 5′ end tion and visualization were conducted using ggplot2 (v.3.3.3), ggpubr
of the C and V primers. The final library included Illumina dual-index (v.0.6.0), ggSCvis (v.0.0.2) and ComplexHeatmap (v.2.14.0) in R.
adapters, ten-nucleotide unique molecular identifiers (UMIs) and
eight-nucleotide internal barcodes linked to the C gene primer. Ampli- Smart-seq2 for sensitive full-length transcriptome profiling in
fied libraries were multiplexed and sequenced on the Illumina NovaSeq single cells
platform (500-cycle kit, 250 bp paired-end reads) by a commercial ser- CD8 + T cells from the BALF and PBMC samples were stained
vice (Personal Biotechnology). The output immune receptor sequences with an B40-N322 pentamer (Pro5 MHC Class I Pentamer, B*40:
span from the first framework region to the beginning of the constant 01-MEVTPSGTWL-Pentamer-R-PE, cat. no. F4328-2B-C, ProImmune);
region, encompassing CDR1, CDR2 and CDR3. B40-N322-pentamer+CD3+CD8+ T cells were single-cell-sorted (FAC-
SAria III) into 96-well plates. scRNA-seq libraries were prepared using
scRNA-seq and TCR repertoire analysis the Smart-seq2 protocol as described in ref. 55. Briefly, cells were
Raw single-cell transcriptome sequencing data were processed using sorted into 96-well plates containing deoxynucleotide triphosphate
cellranger (v.7.2.0)44 or mobivision (v.3.0) with the GRCh38 reference mix, oligo(dT) primer and lysis buffer to lyse cells and processing to
genome. Downstream analysis was performed in R (v.4.2.3) with the the reverse transcription reaction immediately, followed by comple-
Seurat package (v.4.3.0)45. Quality control was applied by filtering mentary DNA synthesis and preamplification. DNA tagmentation and
cells based on 200–4,000 genes, a UMI count greater than 1,000 and adapter ligation were processed using TruePrep DNA Library Prep Kit
mitochondrial gene content of less than 20%. Filtered gene-barcode V2 for Illumina (cat. no. TD501-503, Vazyme) and TruePrep Index Kit V2
matrices were integrated across samples using harmony (v.0.1.1)46 to for Illumina (cat. no. TD202, Vazyme). Reads with a length of 300–700
correct for batch effects; data were normalized using LogNormalize in base pairs were purified for sequencing on the Illumina NovaSeq.
Seurat. Highly variable genes were identified (top 2,000, using the vst Gene expression per cell was calculated in fragments per kilobases
method) and scaling was performed while regressing out nCount_RNA per million mapped reads using RSEM (v.1.3.3). The RSEM output was
and percent.mito. Principal component analysis was conducted on further analyzed in R using the Seurat package45 according to our stand-
the top 2,000 variable genes, and UMAP embedding was generated ard procedure described in ‘scRNA-seq and TCR repertoire analysis’
using the top 30 principal components. Cell clusters were identified section. FindMarkers in Seurat was used to perform the DEG analysis.
using shared nearest neighbor modularity optimization, with a reso- GSEA was performed using clusterProfiler50. All the visualization was
lution of 1.2 and annotated by marker gene expression (PTPRC, CD14, conducted in R, using packages such as ggplot2.
FCGR3A, FCN1, CD68, CD1C, MARCO, FABP4, CD3D, IL7R, CCR7, CD4,
CD8A, CD8B, MKI67, NKG7, CD79A, MS4A1, MZB1, IGHG1, PPBP, HBA2, Statistical analysis
CAPS and TPPP3). T cells were reintegrated and reclustered at a resolu- Statistical analyses were conducted using R (v.4.4.0) and Prism v.9.5.1
tion of 0.8. Marker genes for T cell subtypes were determined with the (GraphPad Software), with specific statistical tests detailed in the
FindAllMarkers function. corresponding figure legends. A linear mixed-effects model fitted
scTCR-seq data were assembled, quantified and annotated follow- using REML was used to account for repeat sampling, with t-tests
ing the cellranger (v.7.2.0) or mobivision (v.3.0) vdj pipeline against based on Satterthwaite’s approximation for degrees of freedom. To
refdata-cellranger-vdj-GRCh38-alts-ensemble-7.1.0. T cells containing compare paired BALF and PBMC samples, a Wilcoxon matched-pairs
both the TCRα and TCRα sequences were retained, and clones were signed-rank test was used. Depending on the data type, group com-
defined by identical TRAV–TRBV pairs with matching CDR3 sequences. parisons were performed using a Kruskal–Wallis, Mann–Whitney U-test
Bulk TCR sequencing data from short-term T cell lines (regarded as or chi-squared test. Spearman correlation was applied to analyze the
specific T cells; see ‘Antigen-specific T cell ex vivo expansion and RNA associations among multiple variables. Multiple comparisons were
extraction’) were assembled, quantified and annotated using MiXCR corrected using the Benjamini–Hochberg method to derive Padj values,
(v.3.0.3) toolkit47. Specific T cells were identified by matching single-cell with statistical significance defined as P < 0.05.
TCR sequences to bulk TCR sequences for both α and β chains. For a
given T cell, if both α and β chains could be matched to the specific bulk Reporting summary
TCR sequences, it was defined as a specific T cell. Clonal diversity was Further information on research design is available in the Nature
analyzed using Alakazam (v.0.3.0)48. Portfolio Reporting Summary linked to this article.
To merge TCR and transcriptome data, cellular barcodes were
matched. DEGs were identified with FindMarkers with Padj < 0.05 and Data availability
a log2 fold change greater than 0.25. Functional enrichment was per- The RNA and TCR sequencing data first reported in this study are
formed using GSEA49 with clusterProfiler (v.4.6.2)50 and annotated deposited and accessible via the NCBI (https://www.ncbi.nlm.nih.
with the Kyoto Encyclopedia of Genes and Genomes51, Gene Ontology52 gov) under accession no. PRJNA1099166. Public SARS-CoV-2-specific
and Hallmark gene sets in the Molecular Signatures Database49. Trajec- TCR repertoires can be acquired from the IEDB database24. All data
tory inference analyses were performed separately for specific CD8+ supporting the findings of this study are available in the article, its
and CD4+ T cells using Monocle (v.2.26.0)30. Gene module scores (for Supplementary Information or from the corresponding author ( J.C.Z.)
example, activation, cytokines, resident and IFN score) were calculated upon reasonable request. Source data are provided with this paper.
using the AddModuleScore function in Seurat based on predefined
gene sets (activation scores: CD40LG, CD69, HLA-DPA1, HLA-DPB1, Code availability
HLA-DQB1, HLA-DRB5, ICOS, TNFRSF4, NFAT5, NFATC3, NFATC2, LCP2, All packages, functions and key parameters used in the study have been
LAT, ZAP70 and LCK; cytokines scores: IFNG, TNF, IL2, CSF2, IL10, IL4, IL5, described in Methods.

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

References the R&D Program of the Guangzhou Laboratory (SRPG22-006 to

42. Amid, E. & Warmuth, M. K. TriMap: large-scale dimensionality J.C.Z., SRPG23-001 to Y.X. and EKPG21-29 to A.Z.), the National
reduction using triplets. Preprint at https://arxiv.org/abs/ Natural Science Foundation of China (82495200, 82495203
1910.00204 (2022). and 82025001 to JC.Z., 82201933 to A.Z., 82101836 to Q.Z. and
43. Van Gassen, S. et al. FlowSOM: using self-organizing maps for 82201932 to Q.Y.), the Guangdong Basic and Applied Basic Research
visualization and interpretation of cytometry data. Cytometry A Foundation (2021B1515130005 to J.X.Z.), the China Postdoctoral
87, 636–645 (2015). Science Foundation (2022M710891 to Q.Y. and 2022M710892 to
44. Zheng, G. X. Y. et al. Massively parallel digital transcriptional A.Z.) and the State Key Laboratory of Respiratory Diseases Funded
profiling of single cells. Nat. Commun. 8, 14049 (2017). Program (SKLRD-Z-202330 to A.Z., SKLRD-Z-202324 to Q.Y. and
45. Butler, A., Hoffman, P., Smibert, P., Papalexi, E. & Satija, R. SKLRD-Z-202304 to Q.H.). We thank S. Perlman (University of Iowa) for
Integrating single-cell transcriptomic data across different his critical review of our paper.
conditions, technologies, and species. Nat. Biotechnol. 36,
411–420 (2018). Author contributions
46. Korsunsky, I. et al. Fast, sensitive and accurate integration of Conceptualization: J.X.Z. and J.C.Z. Data curation: A.Z., Z.C., Q.Y.,
single-cell data with Harmony. Nat. Methods 16, 1289–1296 (2019). J.X.Z. and J.C.Z. Formal analysis: A.Z., Z.C., Q.Y., M.J., J.H., J.X.Z. and
47. Bolotin, D. A. et al. MiXCR: software for comprehensive adaptive J.C.Z. Funding acquisition: A.Z., Q.Y., Y.X., N.S.Z., J.X.Z. and J.C.Z.
immunity profiling. Nat. Methods 12, 380–381 (2015). Investigation: A.Z., Z.C., Q.Y., X.L., Z.L., N.L., C.T., W.J., J.H., L.C., J.C.,
48. Gupta, N. T. et al. Change-O: a toolkit for analyzing large-scale C.C., T.T., Z.X., Q.H., F.L., Y.W., J.S., Z.Z., L.W., J.Z., D.L.L., C.H.Z., Y.C.,
B cell immunoglobulin repertoire sequencing data. Bioinformatics S.X.L., N.Z., X.Z., J.C., X.C., Y.X. and N.S.Z. Project administration: A.Z.
31, 3356–3358 (2015). Resources: W.J., L.C., J.C., C.C., T.T., Q.H., F.L., Y.W., J.S., Z.Z., L.W., N.Z.,
49. Subramanian, A. et al. Gene set enrichment analysis: a X.Z., J.C., N.S.Z., J.X.Z. and J.C.Z. Supervision: X.C., Y.X., J.X.Z. and J.C.Z.
knowledge-based approach for interpreting genome-wide Patient recruitment and sample collection: N.S.Z., X.L., Z.L., N.L., C.T.,
expression profiles. Proc. Natl Acad. Sci. USA 102, 15545–15550 Z.X., J.Z., D.L.L., Y.Z., X.H., S.Y.L., Q.Z., F.C., L.Z., D.D.L., C.H.Z., Y.C., S.L.,
(2005). K.L., X.C. and Y.X. Writing: A.Z., Z.C., Q.Y., J.X.Z. and J.C.Z.
50. Wu, T. et al. clusterProfiler 4.0: a universal enrichment tool for
interpreting omics data. Innovation 2, 100141 (2021). Competing interests
51. Kanehisa, M., Furumichi, M., Sato, Y., Kawashima, M. & The authors declare no competing interests.
Ishiguro-Watanabe, M. KEGG for taxonomy-based analysis of
pathways and genomes. Nucleic Acids Res. 51, D587–D592 (2023). Additional information
52. Ashburner, M. et al. Gene Ontology: tool for the unification of Extended data is available for this paper at
biology. Nat. Genet. 25, 25–29 (2000). https://doi.org/10.1038/s41590-024-02072-9.
53. Kim, D., Langmead, B. & Salzberg, S. L. HISAT: a fast spliced
aligner with low memory requirements. Nat. Methods 12, 357–360 Supplementary information The online version
(2015). contains supplementary material available at
54. Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general https://doi.org/10.1038/s41590-024-02072-9.
purpose program for assigning sequence reads to genomic
features. Bioinformatics 30, 923–930 (2014). Correspondence and requests for materials should be addressed
55. Picelli, S. et al. Full-length RNA-seq from single cells using to Xiaobo Chen, Yonghao Xu, Nanshan Zhong, Jingxian Zhao or
Smart-seq2. Nat. Protoc. 9, 171–181 (2014). Jincun Zhao.

Acknowledgements Peer review information Nature Immunology thanks Paul Moss and the
We thank all donors, healthcare personnel, study coordinators, other, anonymous, reviewer(s) for their contribution to the peer review
and laboratory managers involved in this study. We thank the of this work. Peer reviewer reports are available. Primary Handling
Biobank for Respiratory Disease at the National Clinical Research Editor: P. Jauregui, in collaboration with the Nature Immunology team.
Center for Respiratory Disease. This work was supported by grants
from the National Key R&D Program of China (2023YFC2306400 Reprints and permissions information is available at
and 2023YFC3041700 to J.C.Z., and 2021YFC0864500 to Y.X.), www.nature.com/reprints.

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

Extended Data Fig. 1 | Schematic overview of specific T cell analysis in BALF includes flow cytometry, 10x single-cell RNA and TCR sequencing, and Smart-
and PBMC samples from COVID-19 patients. (a) Schematic overview illustrating seq2 single-cell sequencing. (b) Flowcytometric gating strategy for identifying
the comprehensive analysis workflow of BALF and PBMC samples. The analysis SARS-CoV-2-specific T cells in PBMC and BALF samples.

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

Extended Data Fig. 2 | Highly activated T cells in the inflammatory airway expression. (c) Summary of cellular memory population of total CD4+ and CD8+
environment. (a) Heatmap displaying the expression of indicated proteins T cells in PBMC (n = 22) and BALF (n = 23). Data are presented as mean ± SEM. (d)
across 9 clusters mapped onto the specific T cell TriMap, derived from FlowSOM Wilcoxon matched-pairs comparison with signed rank test of indicated cytokine
data related to Fig. 3. (b) Representative flow plots of cellular memory population levels between 18 paired BALF supernatants and plasma samples.
of SARS-CoV-2-specific and total T cells according to CCR7 and CD45RA

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

Extended Data Fig. 3 | Transcriptional features of cells in BALF and PBMC. (a) Overview of the cell clusters in the integrated single cell transcriptomes of 186451 cells
derived from BALF and PBMC in COVID-19 patients. Proportion of cell clusters in BALF and PBMC were summarized. (b) Clusters were defined according to the cluster-
specific gene expression patterns.

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

Extended Data Fig. 4 | TCR repertoire features of T cells in BALF and PBMC. diversity and abundance of all the CD4+ and CD8+ T cells in PBMC and BALF. The
(a) Flow plots showing the gating strategy for specific T cells sorting from shaded area represents 95% CI. (d) Specific T cell clone distribution on the top
proliferated short-term T cell lines. (b) Spearman correlation (two-tailed) 100 T cell clones of the whole T cell repertoire in pairing BALF and PBMC samples
between frequencies of specific T cells in the BALF detected by intracellular from PT1 at three sequential time points. Rectangles in red represent specific T
cytokine staining and defined in single cell transcriptome based on the TCR cell clones and grey ones stand for undefined ones.
aligning to the proliferated T cells upon SARS-CoV-2 peptide stimulation. (c) TCR

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

Extended Data Fig. 5 | See next page for caption.

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

Extended Data Fig. 5 | SMARTseq2 analysis of HLA-B40:01-N322-330 epitope for single-cell sorting into 96-well plates following with sensitive full-length
(B40/N322 Pent+) specific CD8+ T cells in BALF and PBMC. (a-b) Corplots transcriptome sequencing. (d) Wilcoxon matched-pairs comparison of
showing spearman correlation of indicated functional gene modules among frequencies of B40/N322 Pent+CD8+ T cells between 4 paired BALF and PBMC
specific CD4+ (a) and CD8+ (b) T cells in BALF, related to Fig. 5. The colors of samples. (e)Volcano plot of DEGs in B40/N322-specific CD8+ T cells comparing
the ellipses indicate the values of correlation coefficient, and the ellipses have BALF versus PBMC by Smart-seq2. Horizontal black dashed line: Benjamini-
their eccentricity parametrically scales to the correlation values. (c) Flow plots Hochberg adjusted P value = 0.05. (f) GSEA analysis on B40/N322-specific CD8+
showing B40/N322 Pent+CD8+ T cells in paired BALF and PBMC samples from T cells comparing BALF versus PBMC. (g) Dot plots showing expression level of
patients carrying HLA-B40:01 subtype, which were sorted and then processed selected genes in B40/N322-specific CD8+ T cells from BALF and PBMC.

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

Extended Data Fig. 6 | See next page for caption.

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

Extended Data Fig. 6 | Comparison of specific T cells and the pulmonary and concentrations in BALF supernatant samples by RNA+ (n = 8) and RNA- (n = 10)
peripheral environments during viral shedding (RNA+) and viral clearance groups. Data are shown as mean ± SEM. (h) Density plots reflecting the number
(RNA−) stages. (a) Distribution (left) and summary (right) of specific T cell clone of T cells along the specific CD4+ and CD8+ T-cell lineages stratified by RNA+ and
size transitions from the viral shedding stage (RNA+) to the viral clearance (RNA-) RNA- stages in BALF and PBMC. The x-axis indicates pseudotime and the y-axis
stage among persistent clones in PBMC samples (n = 5, 6). (b) Volcano plots of indicates probability density distribution of each T-cell lineages at different
DEGs in specific CD4+ and CD8+ T cells in PBMC samples comparing RNA+ versus pseudotime. (i) T cell memory phenotyping according to the CCR7 and CD45RA
RNA- stages. Horizontal black dashed line: Benjamini-Hochberg adjusted P protein expression of total and specific T cells in PBMC (n = 5, 6) and BALF (n = 4,
value = 0.05. (c-f) Volcano plot displaying DEGs in specific CD4+ (c) and CD8+ (e) 6) at viral shedding (RNA+) and viral clearance (RNA-) stages, derived from
T cells from BALF samples collected from PT1 at RNA+ (16 days after onset) and flowcytometric data. Data are presented as mean ± SEM. (j) TriMap visualized
RNA- (31 and 34 days after onset) stages. Names of indicated DEGs are shown. undefined T cell distribution according to FlowSOM clusters. Proportions of
Comparison of immune function-associated scoring modules between specific indicated clusters in BALF (n = 4, 6) and PBMC (n = 5, 6) distribution at viral
CD4+ (d) and CD8+ (f) T cells at RNA+ and RNA- stages were shown. Details of the shedding (RNA+) and viral clearance (RNA-) stages were depicted by Pie charts.
scoring methods are introduced in the Methods section. Data are presented Chi-square test.
as mean ± SD, Mann–Whitney U-test. (g) Dot plots summarizing cytokine

Nature Immunology
Article https://doi.org/10.1038/s41590-024-02072-9

Extended Data Table 1 | Cohort demographics

Nature Immunology
 nature portfolio | reporting summary
Jincun Zhao, Jingxian Zhao, Nanshan Zhong,
Corresponding author(s): Yonghao Xu, Xiaobo Chen
Last updated by author(s): Dec 13, 2024

Reporting Summary
Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency
in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.
n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement
A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly
The statistical test(s) used AND whether they are one- or two-sided
Only common tests should be described solely by name; describe more complex techniques in the Methods section.

A description of all covariates tested

A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons
A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient)
AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals)

For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted
Give P values as exact values whenever suitable.

For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings
For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes
Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated
Our web collection on statistics for biologists contains articles on many of the points above.

Software and code

Policy information about availability of computer code
Data collection All flow cytometric data were acquired on BD LSR Fortessa X20 or FACSymphony with FACS Diva software (version 9.0) , or BD FACSVerse with
FACS Suite software (version 1.0.6).
The copy numbers of SARS-CoV-2 in BALF samples were measured by QuantStudio 6 (ABI).
Single cell sequencing data were collected by Illumina Novaseq.

Data analysis Flow data were analyzed with FlowJo software (version 10.6.2, BD Biosciences). Raw data were analyzed using R (version 4.2.3) and GraphPad
Prism (version 9.5.1). Statistical data analysis were performed with R (version 4.4.0).
Raw single-cell transcriptome and TCR sequencing data were processed following cellranger (version 7.2.0) or mobivision (version 3.0) count
protocol against GRCh38 reference genome. The cellranger or mobivision output was further analyzed in R (version 4.2.3) platform using the
Seurat package (version 4.3.0). A filtered gene-barcode matrix of all samples was integrated with harmony (version 0.1.1). Packages used were
MixCR (version 3.0.3), Alakazam (version 0.3.0), ClusterProfiler (version 4.6.2), Monocle (version 2.26.0), HISAT (version 2.1), SubReads
package (version 1.5.3), ggplot2 (version 3.3.3), ggpubr (version 0.6.0), ggSCvis (version 0.0.2), ComplexHeatmap (version 2.14.0), and RSEM
suit (version 1.3.3).
For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and
reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.
April 2023

1
 nature portfolio | reporting summary
Data
Policy information about availability of data
All manuscripts must include a data availability statement. This statement should provide the following information, where applicable:
- Accession codes, unique identifiers, or web links for publicly available datasets
- A description of any restrictions on data availability
- For clinical datasets or third party data, please ensure that the statement adheres to our policy

RNA and TCR sequencing data first reported in this study are deposited and accessible in the NCBI (https://www.ncbi.nlm.nih.gov) under accession number
PRJNA1099166. Public SARS-CoV-2-specific TCR repertoires are acquired from IEDB database (DOI: 10.1093/nar/gky1006). All data supporting the findings of this
study are available in the article, supplementary Information, or from the corresponding author JC.Z. on reasonable request. All statistical source data would be
provided with paper for publication.

Research involving human participants, their data, or biological material

Policy information about studies with human participants or human data. See also policy information about sex, gender (identity/presentation),
and sexual orientation and race, ethnicity and racism.
Reporting on sex and gender Participants were not excluded based on sex or gender.

Reporting on race, ethnicity, or No socially constructed or socially relevant categorization variables were used in this manuscript.
other socially relevant
groupings

Population characteristics Population characteristics are described in Figure 1a and Extended Data Table 1.

Recruitment In this prospective observational cohort study, adult patients were recruited with PCR-confirmed COVID-19. Participants were
recruited from a hospital located in Guangzhou, Guangdong, China, named the First Affiliated Hospital of Guangzhou Medical
University from January 2020 to January 2023. After informed consents were obtained, bronchoalveolar lavage fluid (BALF)
samples were obtained by trained clinicians from donors who had tested positive for SARS-CoV-2 and were diagnosed as
COVID-19 patients and agreed to participated the study. All methods were performed according to the relevant guidelines
and regulations.

Ethics oversight All human subjects research was approved by the First Affiliated Hospital of Guangzhou Medical University (Ethics ID:
ES-2020-65, ES-2023-015-01). All study participants have provided written informed consents and did not received
compensation for taking part in this study.

Note that full information on the approval of the study protocol must also be provided in the manuscript.

Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Life sciences study design

All studies must disclose on these points even when the disclosure is negative.
Sample size No statistical method was used to predetermine the sample size for our experiments, but our sample sizes are similar to or larger than those
reported in previous publications in the field. At least four replicates were included in the subgroup comparison. And no statistical tool was
used to predetermine sample sizes.

Data exclusions No data has been excluded for statistical analysis.

Replication Replications of experiments are specified in the according figure legends.

Randomization This was an observational study. Randomization was not applicable as this was not an intervention study.

Blinding Cell clustering was performed on all cells, resulting in a de facto blinded assignment of cells to distinct populations.
April 2023

Reporting for specific materials, systems and methods

2
We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material,
system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response.

nature portfolio | reporting summary

Materials & experimental systems Methods
n/a Involved in the study n/a Involved in the study
Antibodies ChIP-seq
Eukaryotic cell lines Flow cytometry
Palaeontology and archaeology MRI-based neuroimaging
Animals and other organisms
Clinical data
Dual use research of concern
Plants

Antibodies
Antibodies used Specificity Label Brand catalog# clone dilution
anti-CD103-PE-Dazzle594 (Biolegend, 350224, Ber-ACT8) 1/50
anti-CD127-APC-R700 (BD Horizon, 565185, HIL-7R-M21) 1/200
anti-CD14-BV570 (BD, 624298, MφP9) 1/100
anti-CD14-BV711 (BD Horizon, 563372, MφP9) 1/200
anti-CD16-APC (BD pharmingen, 561304, B73.1) 1/20
anti-CD163-BB700 (BD, 624381, GHI/61) 1/200
anti-CXCR3 -PE-Cy7 (BD, 560831, 1C6) 1/500
anti-CCR4 -BV750 (BD, 746980, 1G1) 1/100
anti-CCR5 -BB790 (BD, 624296, 3A9) 1/200
anti-CCR6 -BB660 (BD, 624295, 11A9) 1/200
anti-CCR7 -PE-Cy7 (Biolegend, 353226, G043H7) 1/100
anti-CD206 -PE-Cy5 (BD, 624294, 19.2) 1/25
anti-CD25 -BUV737 (BD Horizon, 612806, 2A3) 1/100
anti-PD-1 -BV786 (BD, 563789, EH12.1) 1/50
anti-PD-1-PE-Cy7 (BD pharmingen, 561272, EH12.1) 1/50
anti-CD3 -AF700 (BD, 624358, HIT3a) 1/100
anti-CD3 -BUV395 (BD Horizon, 563546, UCHT1) 1/400
anti-CD38 -BV711 (BD, 563965, HIT2) 1/100
anti-CD4 -BV605 (BD, 562658, RPA-T4) 1/20
anti-CD4 -BV421 (BD Horizon, 562842, L200) 1/200
anti-CD4 -BB515 (BD Horizon, 564419, RPA-T4) 1/100
anti-CD45 -BUV805 (BD, 612891, HI30) 1/500
anti-CD45 -BV510 (BD Horizon, 563204, HI30) 1/250
anti-CD45RA -BUV737 (BD Horizon, 612846, HI100) 1/200
anti-CD56 -BUV395 (BD, 563554, NCAM16.2) 1/100
anti-CD56 -BV650 (BD Horizon, 564057, NCAM16.2) 1/400
anti-CD69 -BV510 (BD, 747521, FN50) 1/100
anti-CD8 -BUV615 (BD, 612994, SK1) 1/200
anti-CD8 -BV786 (BD Horizon, 563823, RPA-T8) 1/250
anti-CD8 -Percp-cy5.5 (BD Horizon, 565310, SK1) 1/200
anti-CD86 -APC-R700 (BD Horizon, 565149, 2331 FUN-1) 1/100
anti-GM-CSF-BUV563 (BD, 624284, BVD2-21C11) 1/50
anti-GM-CSF-Percp-cy5.5 (Biolegend, 502312, BVD2-21C11) 1/50
anti-HLA-A,B,C-FITC (Biolegend, 311403, W6/32) 1/100
anti-HLA-DR-BUV496 (Biolegend, 569674, G46-6) 1/200
anti-HLA-DR-BV605 (BD Horizon, 562845, G46-6) 1/100
anti-IFN-γ-APC (BD pharmingen, 554702, B27) 1/200
anti-IFN-γ-BV421 (Biolegend, 506538, B27) 1/200
anti-IL-10-PE-Cy5 (BD pharmingen, 624350, JES3-19F1) 1/500
anti-IL-10-BV711 (BD Horizon, 564050, JES3-9D7) 1/50
anti-IL-1β-PE (R D, IC201P, # 8516) 1/100
anti-IL-2-BB515 (BD, 624279, MQ2-13A5) 1/250
anti-IL-2 -BV605 (BD Horizon, 564165, MQ1-17H12) 1/50
anti-IL-6-BV480 (BD, 624278, MQ2-13A5) 1/200
anti-IL-8 -BV650 (BD, 624280, G265-8) 1/500
anti-LAG-3-APC-Cy7 (BD, 624355, T47-530) 1/200
anti-TIM-3-BUV661 (BD, 624285, 7D3) 1/200
anti-TIM-3-APC (R D, 344823, FAB2365A) 1/25
April 2023

anti-TNF -BUV737 (BD, 624286, MAb11) 1/200

anti-TNF -PE (Biolegend, 502909, MAb11) 1/200
Fixable Viability Stain 440UV (BD, 566332) 1/1000
Pro5 MHC Class I Pentamer B*40:01-MEVTPSGTWL-Pentamer-R-PE (ProImmune, F4328-2B-C) 1/20

3
 Validation All antibodies used are commercial antibodies reported by the manufacturer to be validated for use. Likewise, we titrated the

nature portfolio | reporting summary

antibodies based on the staining condition in our study.

Plants
Seed stocks Report on the source of all seed stocks or other plant material used. If applicable, state the seed stock centre and catalogue number. If
plant specimens were collected from the field, describe the collection location, date and sampling procedures.

Novel plant genotypes Describe the methods by which all novel plant genotypes were produced. This includes those generated by transgenic approaches,
gene editing, chemical/radiation-based mutagenesis and hybridization. For transgenic lines, describe the transformation method, the
number of independent lines analyzed and the generation upon which experiments were performed. For gene-edited lines, describe
the editor used, the endogenous sequence targeted for editing, the targeting guide RNA sequence (if applicable) and how the editor
was applied.
Authentication Describe any authentication procedures for each seed stock used or novel genotype generated. Describe any experiments used to
assess the effect of a mutation and, where applicable, how potential secondary effects (e.g. second site T-DNA insertions, mosiacism,
off-target gene editing) were examined.

Flow Cytometry
Plots
Confirm that:
The axis labels state the marker and fluorochrome used (e.g. CD4-FITC).
The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a 'group' is an analysis of identical markers).
All plots are contour plots with outliers or pseudocolor plots.
A numerical value for number of cells or percentage (with statistics) is provided.

Methodology
Sample preparation BALF sample were centrifuged at 800 g, 4oC for 10 minus to separate supernatant and cell pellets. PBMC were isolated from
blood using Leucosep tubes (Greiner) and Ficoll-Paque PLUS (GE Healthcare) according to the manufacturer’s instructions.
BALF cell pellets and PBMC were stored in liquid nitrogen and supernatant and plasma were stored at -80oC before analysis.
To detect specific T cells, 1E5 to 1E6 cells were incubated in a well of 96-well round-bottom plates at 37 °C for 16-18 h in the
presence of SARS-CoV-2 peptide pool (200 nM for each peptide at final) and brefeldin A (BD Biosciences, 1:1000 at final). For
surface markers, cells were stained with the indicated antibodies at 4°C for 20-30 minutes, and then labeled with cell viability
staining dye (FVS440UV, BD Biosciences) at room temperature (20–25°C) in the dark for another 10 minutes. The cells were
subsequently fixed and permeabilized using Cytofix/Cytoperm Solution (BD Biosciences) to enable intracellular marker
labeling at 4°C for 30-40 minutes. Then cells were washed with 1X Perm/wash buffer twice and resuspended in FACS buffer
for flow cytometer detection.

Instrument LSR Fortessa, FACSymphony, FACSVerse and FACS-sort AriaIII

Software BD Diva software v9.0, FACS Suite software v1.0.6, and Flowjo v10.6.2

Cell population abundance Frequencies of cell populations are stated in dot plots and graphs.

Gating strategy Gating strategies are stated in the according Extended figures 1 and 4. In brief, all cells were gated on SSC-A/FSC-A and FSC-
H/FSC-A to identified as singlet, then as live (FVS440-), leukocytes (CD45+), T cells (CD3+CD56-), and further subdivided into
CD4+ and CD8+ T cells. SARS-CoV-2-specific T cells were identified as CD3+CD4+IFN-γ+, CD3+CD8+IFN-γ+, or B40/n322
pentamer+CD3+CD8+ T cells.

Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information.
April 2023