key: cord-0750047-jgmkltbk
authors: Wu, Jianghua; Tang, Lu; Ma, Yanling; Li, Yu; Zhang, Dongmei; Li, Qian; Mei, Heng; Hu, Yu
title: Immunological Profiling of COVID-19 Patients with Pulmonary Sequelae
date: 2021-09-07
journal: mBio
DOI: 10.1128/mbio.01599-21
sha: 3967521e9cf8da5b7bb717f0b902e26a7703be35
doc_id: 750047
cord_uid: jgmkltbk

Cellular immunity may be involved in organ damage and rehabilitation in patients with coronavirus disease 2019 (COVID-19). We aimed to delineate immunological features of COVID-19 patients with pulmonary sequelae (PS) 1 year after discharge. Fifty COVID-19 survivors were recruited and classified according to radiological characteristics, including 24 patients with PS and 26 patients without PS. Phenotypic and functional characteristics of immune cells were evaluated by multiparametric flow cytometry. Patients with PS had an increased proportion of natural killer (NK) cells and a lower percentage of B cells than patients without PS. Phenotypic and functional features of T cells in patients with PS were predominated by the accumulation of CD4-positive (CD4(+)) T cells secreting interleukin 17A (IL-17A), short-lived effector-like CD8(+) T cells (CD27-negative [CD27(−)] CD62L(−)), and senescent T cells with excessive secretion of granzyme B/perforin/interferon gamma (IFN-γ). NK cells were characterized by the excessive secretion of granzyme B and perforin and the downregulation of NKP30 and NKP46; highly activated NKT and γδ T cells exhibited NKP30 and TIM-3 upregulation and NKB1 downregulation in patients with PS. However, immunosuppressive cells were comparable between the two groups. The interrelationship of immune cells in COVID-19 was intrinsically identified, whereby T cells secreting IL-2, IL-4, and IL-17A were enriched among CD28(+) and CD57(−) cells and cells secreting perforin/granzyme B/IFN-γ/tumor necrosis factor alpha (TNF-α)-expressed markers of terminal differentiation. CD57(+) NK cells, CD4(+)Perforin(+) T cells, and CD8(+) CD27(+) CD62L(+) T cells were identified as the independent predictors for residual lesions. Overall, our findings unveil the profound imbalance of immune landscape that may correlate with organ damage and rehabilitation in COVID-19.

87.50%), and tractive bronchiectasis (8, 33 .33%) in PPSs (Table 1) . Representative chest computed tomography (CT) scans longitudinally exhibited the change of lung lesions (Fig. 1A) . Although the lung lesions gradually resolved in all patients, ground-glass opacity, (GGO), fiber streak shadow, tractive bronchiectasis, reticulation, and bronchovascular bundle distortion could be observed in the representative patients ( Fig. 1A to C). Importantly, there was a statistically significant increment in hemoglobin in PPSs in comparison to NPSs ( Table 1 ), suggesting that residual pulmonary lesions may influence diffusion capacity and induce a compensatory increase in hemoglobin in those with PS.

NK cells and short-lived effector-like CD8 + T cells accumulate in PPSs. Immunologic disturbance induced by SARS-CoV-2 infection is characterized by lymphopenia in those with acute COVID-19 (16) . Strikingly, white blood cell and lymphocyte counts were higher in PPSs than NPSs 1 year after discharge, whereas no significant differences in neutrophil, monocyte, and platelet counts were found between the two groups ( Table 1) . We next analyzed the presence of lymphocyte subsets to obtain an overview of the general distribution in the peripheral blood. The NK cell percentage was significantly higher in PPSs than in NPSs, but there were no significant differences in CD3 1 T, CD4 1 T, CD8 1 T, NKT, and gd T cell percentages ( Fig. 2A) . A decreased proportion of B cells was detected in PSSs, though the memory B cell frequency was not significantly different ( Fig. 2A and B) .

We next performed immunophenotypic analyses of circulating CD4 1 and CD8 1 T cells to identify their state of differentiation, exhaustion, and senescence. CD27 and CD62L were used to distinguish maturation and memory subsets in CD4 1 and CD8 1 T cells. We found no significant differences in CD4 1 T cell populations based on CD27 and CD62L expression between PPSs and NPSs (Fig. 2C ). CD27-negative (CD27 2 ) CD62L 2 T cells represent short-lived effector-like T cells characterized by the enrichment for antigen-experienced and senescent T cells, while CD27 1 CD62L 1 T cells consist of naive T cells and central memory T cells (17) . Importantly, there was a statistically significant increase in the CD8 1 CD27 2 CD62L 2 T cells percentage in PPSs, yet the CD8 1 CD27 1 CD62L 1 T cell frequency was significantly reduced (Fig. 2C ). CD4 1 CD57 1 T cells percentage tended to be higher in PPSs than in NPSs, though there was no statistical difference (Fig. 2D) . A higher percentage of CD8 1 CD57 1 T cells was observed in PPSs (Fig. 2D) . Moreover, increases of both CD4 1 T and CD8 1 T cells expressing KLRG-1 and TIM-3 were detected in PPSs, with no changes in the frequencies of CD28and PD-1-expressing T cells between PPSs and NPSs (Fig. 2D ).

Circulating NK cells and innate-like lymphocytes exhibit distinct phenotypes between PPSs and NPSs. NK, NKT, and gd T cells, large granular lymphocytes with the ability to lyse virally infected cells, are important components of antiviral immunity (18) (19) (20) . Having established that PPSs displayed an increase in the NK cells percentage, we further analyzed the NK cell phenotype. Importantly, there was a statistically significant increment in NK cells expressing CD57, whereas the frequencies of NKP30-and NKP46-expressing NK cells were significantly reduced in PPSs (Fig. 3A) . No changes in the frequencies of CD27-, KLRG1-, PD-1-, TIM-3-, NKB1-, NKG2A-, and NKG2D-expressing NK cells were noted between the two groups (Fig. 3A) . Focusing on NKT cells, we observed an increase in the fraction of those expressing TIM-3 and NKP30, but the fraction of NKB1 1 NKT cells was markedly decreased in PPSs (Fig. 3B) , suggesting that NKT cells exhibit an activated or exhausted state. However, there were no significant differences in the frequencies of CD27-, CD57-, KLRG1-, PD-1-, NKG2A-, NKG2D-and NKP46expressing NKT cells between the two groups (Fig. 3B) . Moreover, gd T cells from PPSs overexpressed CD57, KLRG1, TIM-3, and NKP30 (Fig. 3C) , which is compatible with a Relative proportions of memory and naive B cells between the two groups. (C) Relative proportions of CD4/ CD8 1 CD27 2 CD62 2 , CD4/CD8 1 CD27 1 CD62 2 , CD4/CD8 1 CD27 1 CD62 1 , and CD4/CD8 1 CD27 2 CD62 1 T cells between the two groups. (D) Relative proportions of CD4/CD8 1 CD28 1 , CD4/CD8 1 CD57 1 , CD4/CD8 1 KLRG1 1 , CD4/CD8 1 PD-1 1 T cells, and CD4/CD8 1 TIM-3 1 T cells between the two groups. Data are mean 6 SD. The Mann-Whitney U test or unpaired two-tailed Student's t tests were used to compare the two groups. *, P , 0.05; **, P , 0.01; ***, P , 0.001; ****, P , 0.0001.

Immunological Profiling of COVID-19 Patients ® senescent, hyperactivated exhaustion profile. Downregulation of the inhibitory receptor NKB1 and CD27 in gd T cells was found in PPSs, but PD-1, NKG2A, and NKP46 expressions were unaffected (Fig. 3C) .

Coexistence of senescence and exhaustion phenotypic signatures in cytokinesecreting cells in PPSs. We next assessed cytokine production in CD3 1 T cells after stimulation with PMA and ionomycin and observed that, in comparison to NPSs, CD4 1 T cells from PPSs overexpressed IL-17A and interferon gamma (IFN-g) but not IL-2, IL-4, or tumor , and NKP46 expression on gd T cells between the two groups. Data are mean 6 SD. Mann-Whitney U test or unpaired two-tailed Student's t tests were used to compare the two groups. *, P , 0.05; **, P , 0.01; ***, P , 0.001; ****, P , 0.0001. necrosis factor alpha (TNF-a) (Fig. 4A ). The frequency of IFN-g expression was higher among CD8 1 T cells in PPSs than NPSs, with no significant differences in IL-4, IL-17A, and TNF-a production in CD8 1 T cells between the two groups ( Fig. 4B) . Nonetheless, CD8 1 T cells from PPSs showed downregulated IL-2 expression compared with NPSs (Fig. 4B ). We next investigated the degranulation capacity and cytotoxic molecule expression in CD4 1 T, CD8 1 T, NK, NKT, and gd T cells. Based on functional characterization, CD4 1 T, CD8 1 T, NK, and gd T cells exhibited upregulation of granzyme B (GZMB) and perforin in PPSs compared with NPSs ( Fig. 4C and D) . However, no significant differences in GZMB and perforin expression in NKT cells were observed ( Fig. 4C and D) .

We next used a clustering and visualization strategy to investigate the coexpression of CD27, CD28, PD-1, and CD57, together with cytokines (IL-2, IL-4, IL-17A, TNF-a, and IFN-g) and cytotoxic molecules (GZMB and perforin) ( Fig. 5A and B and Fig. S1H to I in the supplemental material). The advantage of this analysis lies in its integration of surface and functional markers at a single-cell level, providing an improved understanding of their high-dimensional relationship. CD4 1 IL-2 1 , CD4 1 IL-4 1 , and CD4 1 IL-17A 1 T cells hardly expressed CD57 (Fig. 5A ). Phenotypic characterization further demonstrated that CD4 1 T cells expressing IL-2, IL-4, and IL-17A were prominent among the FIG 4 Functional characterization of CD4 1 T, CD8 1 T, NK, NKT, and gd T cells. Circulating CD4 1 T, CD8 1 T, NK, NKT, and gd T cells from COVID-19 patients with pulmonary sequelae (PPSs; n = 24) and patients without pulmonary sequelae (NPSs; n = 26) were analyzed by multiparameter flow cytometry. (A) Relative proportions of CD4 1 IL-2 1 , CD4 1 IL-4 1 , CD4 1 IL-17A 1 , CD4 1 IFN-g 1 , and CD4 1 TNF-a 1 T cells between the two groups. (B) Relative proportions of CD8 1 IL-2 1 , CD8 1 IL-4 1 , CD8 1 IL-17A 1 , CD8 1 IFN-g 1 , and CD8 1 TNF-a 1 T cells between the two groups. (C) Relative proportions of perforin in circulating CD4 1 T, CD8 1 T, NK, NKT, and gd T cells between the two groups. (D) Relative proportions of GZMB in circulating CD4 1 T, CD8 1 T, NK, NKT, and gd T cells between the two groups. Data are mean 6 SD. The Mann-Whitney U test or unpaired two-tailed Student's t tests were used to compare the two groups. *, P , 0.05; **, P , 0.01; ***, P , 0.001; ****, P , 0.0001.

Immunological Profiling of COVID-19 Patients CD28 1 T cells (Fig. 5A ). According to a representative scatterplot, IFN-g could be secreted by CD4 1 CD57 1 T cells and CD4 1 CD57 2 T cells; furthermore, CD4 1 CD57 1 IFN-g 1 T cells were enriched among CD4 1 CD28 2 T cells (Fig. 5A ). As observed in CD4 1 T cells, the coexpression of IL-2, IL-4, and IL-17A, together with CD28 and CD57, could be similarly observed in CD8 1 T cells (Fig. S1H ). Next, we investigated whether PD-1 expression correlates with cytokine production in the T cell population. We found that IL-2, IL-4, IL-17A, TNF-a, and IFN-g expression was mainly distributed in PD-1 2 T cells ( Fig. 5A and Fig. S1H ). We summarized these data in a heatmap and exhibited the FIG 5 Senescence and exhaustion phenotypes coexist in cytokine-secreting CD4 1 T, CD8 1 T, NK, NKT, and gd T cells. The coexpression of CD27, CD28, and PD-1, together with cytokines (IL-2, IL-4, IL-17A, TNF-a, and IFN-g), in CD4 1 T and CD8 1 T cells from COVID-19 patients with pulmonary sequelae (PPSs; n = 24) and patients without pulmonary sequelae (NPSs; n = 26) was analyzed by multiparameter flow cytometry. (A) Bidimensional map obtained from flow cytometric data displaying the coexpression of CD27, CD28, and PD-1, together with cytokines (IL-2, IL-4, IL-17A, TNF-a, and IFN-g), in CD4 1 T cell. Figure S1H in the supplemental material displays the coexpression of CD27, CD28, and PD-1, together with cytokines (IL-2, IL-4, IL-17A, TNF-a, and IFN-g), in CD8 1 T cell. The coexpression of CD27, CD28, and 57, together with granzyme B (GZMB) and perforin, in circulating CD4 1 T, CD8 1 T, NK, NKT, and gd T cells from COVID-19 PPSs (n = 24) and NPSs (n = 26) was analyzed by multiparameter flow cytometry. (B) Bidimensional map obtained from flow cytometric data displaying the coexpression of CD28 and CD57, together with GZMB and perforin, in CD4 1 T, CD8 1 T, NK, NKT, and gd T cells. Figure S1I displays the coexpression of CD27 and CD57, together with GZMB and perforin, in CD4 1 T, CD8 1 T, NK, NKT, and gd T cells. (C) Heatmap represents the percentage of cells in CD27 1 , CD27 2 , CD28 1 , CD28 2 , PD-1 1 , and PD-1-clusters that express IL-2, IL-4, IL-17A, TNF-a, and IFN-g in CD4 1 T and CD8 1 T cells for PPSs and NPSs. (D) Heatmap represents the percentage of cells in CD27 1 , CD27 2 , CD28 1 , CD28 2 , CD57 1 , and CD57 2 clusters that express GZMB and perforin in CD4 1 T, CD8 1 T, NK, NKT, and gd T cells for PSSs and NPSs. distributions of IL-2, IL-4, IL-17A, TNF-a, and IFN-g expression together with CD28, CD57, and PD-1 between the two groups (Fig. 5C) .

We next investigated the expression of CD27, CD28, and CD57 together with cytotoxic molecules in CD4 1 T, CD8 1 T, NK, NKT, andgd T cells (Fig. 5B) . According to the representative scatterplot, GZMB and perforin expression was higher among CD4 1 CD57 1 T cells than CD4 1 CD57 2 T cells in both groups; furthermore, CD4 1 CD57 1 GZMB 1 T cells and CD4 1 CD57 1 perforin-positive T cells were mainly enriched among CD27 2 cells and CD28 2 cells (Fig. 5B and Fig. S1I ). As observed in CD4 1 T cells, GZMB and perforin expression in CD8 1 T, NK, NKT, and gd T cells was similarly enriched among CD57 1 cells, CD28 2 cells, and CD27 2 cells (Fig. 5B and Fig. S1I ). We summarized the data in a heatmap and exhibited the distributions of GZMB and perforin expressions together with CD27, CD28, and CD57 between the two groups (Fig. 5D) .

Comparable immunosuppressive cells between PPSs and NPSs. Immunosuppressive cells served as the main mechanisms for maintaining immune homeostasis. However, there is considerable controversy regarding whether immunosuppressive cells promote or constrain the formation of fibrosis induced by pathogenic T cells (21) (22) (23) (24) . We next explored the distribution of immunosuppressive cells in recovered patients. After lysis of the erythrocytes, we directly used cell surface markers to stain monocytic myeloid-derived suppressor cells (M-MDSCs; HLA-DR 2 /low CD33 1 CD11b 1 CD14 1 ), granulocytic MDSCs (G-MDSCs; HLA-DR 2 /low CD33 1 CD11b 1 CD14 2 CD15 1 ), regulatory T cells (Tregs; CD4 1 CD127 2 /low CD25 1 ), and regulatory B cells (Bregs; CD19 1 CD24 1 CD38 1 ) (25, 26) . Certainly, it is not stringent to define G-MDSCs by using this method. Patients with acute COVID-19 produce emergency myelopoiesis-generating immunosuppressive myeloid cells (27) . But M-MDSC and G-MDSC frequencies did not differ between PPSs and NPSs ( Fig. 6A and B) . A decrease in HLA-DR on monocytes in acute COVID-19 is associated with severe respiratory failure (28). We did not observe any significant difference in HLA-DR expression on monocyte between the two groups 1 year after discharge (data not shown). Additionally, Tregs and Bregs showed no significant differences among the groups (Fig. 6C and D) . Together, our data indicate that immunosuppressive cells may be insufficient to constrain the robust activation of a variety of immune cells in PPSs.

The interrelationship of immune cells and its association with clinical features in COVID-19. Having established that senescent signatures coexist within cytotoxic molecule-secreting cells, we further investigated the interrelationship of immune cells in convalescent patients with COVID-19 (Fig. 7A) . Correlation analysis verified that T cells secreting cytotoxic molecules correlated negatively with CD27 1 CD62L 1 T cells and CD28 1 T cells but positively with short-lived effector-like CD27 2 CD62L 2 T cells, CD57 1 T cells, and KLRG1 1 T cells (Fig. 7A ). CD57 and KLRG-1 are thought to be associated with T cell senescence and also serve as a marker of cytotoxic function (29) . Importantly, IL-2 expression within CD8 1 T cells correlated negatively with CD8 1 CD27 2 CD62L 2 T cells, CD8 1 CD57 1 T cells, CD8 1 Perforin 1 T cells, and CD8 1 GZBM 1 T cells (Fig. 7A) . The downregulated IL-2 expression in CD8 1 T cells from PPSs further verifies that these patients retained large amounts of short-lived like CD8 1 T cells that abundantly secrete cytotoxic molecules. Moreover, the expression of perforin and GZMB within NK cells was also positively associated with effector molecule CD57 but inversely with PD-1, NKP30, and NKP46 expression (Fig. 7A) . Additionally, GZMB and perforin expression exhibited intrinsic positive correlations among CD4 1 T, CD8 1 T, NK, NKT, and gd T cells, suggesting that SARS-CoV-2 infection may simultaneously trigger robust activation of a variety of immune cells. More detailed information is displayed in the correlation heatmap depicted in Fig. 7A .

We next sought to determine whether the phenotypic and functional features of immune cells correlate with age and disease type in COVID-19. Correlation analysis suggested that age significantly influenced the phenotypic and functional features of immune cells (Fig. 7B ). Senescent and short-lived, such as CD8 1 T cells, CD8 1 T cells secreting perforin/granzyme B/IFN-g, NK cells secreting perforin/granzyme B, NKP30 1 NKT, TIM-3 1 NKT, NKP30 1 gd T, KLRG1 1 gd T, and CD57 1 gd T cell percentages correlated positively with age, whereas B cells, CD3 1 T cells, NKT cells, CD8 1 CD27 1 CD62L 1 T cells, NKB1 1 gd T cells, NKG2D 1 gd T cells, CD27 1 NK cells, NKG2D 1 NK cells, and NKG2D 1 NKT cell frequencies correlated negatively with age (Fig. 7B) . Furthermore, correlation analysis of immunological parameters in PPSs and NPSs revealed that senescent, exhausted GZMB-and perforin-secreting immune cells correlated positively with PS (Fig. 7B) . Additionally, CD8 1 CD27 2 CD62L 2 T cell, NK cells, M-MDSCs, NKP30 1 gd T cell, NKP30 1 NKT cell, CD4 1 IFNg 1 , and CD8 1 IFN-g 1 T cell percentages correlated positively with PS (Fig. 7B) .

Next, we examined the possibility of using the above-mentioned parameters as prognostic factors for identifying determinants for residual lesions in COVID-19 patients. Multivariate logistic regression analyses identified CD8 1 CD27 1 CD62L 1 T cells (odds ratio [OR], 0.738; 95% confidence interval [CI], 0.590 to 0.924; P = 0.008), CD57 1 NK cells (OR, 1.181; 95% CI, 1.038 to 1.343; P = 0.012), and CD4 1 perforin-positive T cells (OR, 1.153; 95% CI, 0.953 to 1.396; P = 0.143) as independent predictors for residual lung lesions (Table 2) . Additionally, receiver operating characteristic (ROC) curve was carried out to assess the capacity of the three cell populations to differentiate disease type (PPSs and NPSs). The cutoff values (sensitivity and specificity) are as follows: CD8 1 CD27 1 CD62L 1 T cells, 26.045% (0.885% and 0.708%); CD57 1 NK cells, 74.095% (0.708% and 0.885%); and CD4 1 Perforin 1 T cells, 3.245% (0.875% and 0.423%). The area under the concentration-time curve (AUC) value of the combination of the three cells (CD57 1 NK cells, CD4 1 Perforin 1 T cells, and CD8 1 CD27 1 CD62L 1 T cells) was highest (0.942) (Fig. 7C to F) .

Although a tremendous global effort by the scientific community has greatly improved our understanding of COVID-19, the immunopathogenesis of PS remains to be elucidated. Here, we describe the circulating immune landscape of COVID-19 patients with PS compared with those without PS. Residual lesions in PPSs were mainly GGO and fiber streak shadow. Our study demonstrates that there existed significantly divergent immunological characteristics between PPSs and NPSs 1 year after discharge. Immunological signatures in PPSs were predominated by the accumulation of CD4 1 T cells secreting IL-17A and short-lived effector-like CD8 1 T cells with excessive secretion (B) Correlation heatmap exhibiting the correlation of immunological parameters with age and disease type (two types, patients with pulmonary sequelae and patients without pulmonary sequelae). Spearman's rank coefficient was used to determine correlations between two variables. The three most contributing variables (CD8 1 CD27 1 CD62 1 T cells, CD57 1 NK cells, and CD4 1 Perforin 1 T cells) were identified by multivariate logistic regression analyses. (C to F) ROC curves were calculated for these selected parameters by using SPSS.

Immunological Profiling of COVID-19 Patients of IFN-g/granzyme B/perforin. NK cells were characterized by excessive secretion of granzyme B and perforin and downregulation of NKP30 and NKP46; NKT and gd T cells demonstrated highly activated and exhausted states in PPSs. Overall, we observed robust activation of a variety of immune cells in response to SARS-CoV-2 infection and specific features unique to COVID-19 with PS and hyperinflammation, providing a base for understanding the role of cellular immunity in patients with COVID-19.

Over 80% of patients with COVID-19 experience lymphopenia and exhibit drastically reduced numbers of various lymphocyte subsets, including CD4 1 T, CD8 1 T, B, gd T, and NK cells, especially in the peripheral blood of those with severe COVID-19 during the acute phase (16, (30) (31) (32) (33) (34) . Early-stage lymphopenia is thought to correlate with lymphocyte chemotaxis in COVID-19 (35, 36) . To our surprise, lymphocyte counts were higher in PPSs than in NPSs 1 year after discharge, suggesting that lymphocytes may exit inflamed tissues and undergo clonal expansion after the resolution of SARS-CoV-2 infection. In general, clonal expansion of both innate and adaptive lymphocytes is a critical process for host defense via amplification of lymphocytes specific to the invading pathogen. The NK cell percentage was significantly higher in PPSs than in NPSs, whereas the B cell frequency was lower in the former, suggesting that NK cells underwent more significant clonal expansion. A previous study found that decreased B cells were associated with prolonged viral RNA shedding from the respiratory tract in COVID-19 (37) . Nevertheless, the reason for the obvious expansion of NK cells following SARS-CoV-2 clearance in PPSs remains to be elucidated. Prolonged viral RNA shedding may induce clonal expansion of NK cells in an antigen-specific manner, similar to the response to cytomegalovirus infection of acquiring memory features (38) . Since reverse-transcribed SARS-CoV-2 RNA can integrate into the genome of cultured human cells and can be expressed in patient-derived tissues (39) , there might also be latent SARSCoV-2 virus in the reservoir cells in PPSs that could lead to the prolonged expansion of NK cells. Excessive activation of proinflammatory immune cells can lead to enhanced inflammation and injury during pulmonary viral infection (40) . COVID-19 promotes cell polarization of naive and memory cells to effector, cytotoxic, exhausted, and regulatory cells, along with increased late NK cells, and induces gene expression related to inflammation and cellular senescence (41) . By examining the phenotypic characteristics of T cells, we observed a low percentage of CD8 1 CD27 1 CD62L 1 T cells but a high level of CD8 1 CD27 2 CD62L 2 T cells in PSSs. T cells in PSSs displayed an overall exhausted and senescent phenotype, with overexpression of CD57, KLRG-1, and TIM-3. Functional analysis further revealed that upregulation of degranulation capacity and cytotoxic molecules in CD4 1 and CD8 1 T cells in PPSs compared with NPSs. Our findings are in line with published papers (42) (43) (44) . SARS-CoV-2 infection may induce a cytotoxic response, characterized by simultaneous production of GZMB and perforin in T cells, NK cells, and gd T cells in PPSs. Excessive activation of cytotoxic T, NK, and gd T cells is not protective but, rather, drives pulmonary damage after SARS-CoV-2 infection.

Emerging evidence suggests that COVID-19 survivors have impaired lung function, with the development of pulmonary fibrosis (2, 14) . Tissue-resident CD8 1 T cells drive age-associated chronic lung sequelae after viral pneumonia (45) . Also, clonally expanded tissue-resident memory-like Th17 cells may interact with profibrotic macrophages and cytotoxic CD8 1 T cells, leading to the formation of pulmonary fibrosis (4). We observed that peripheral blood CD4 1 T cells overexpressed IL-17A/IFN-g and that CD8 1 T cells overexpressed IFN-g/GZMB/perforin in PSSs. CD103 high Tregs can constrain lung fibrosis induced by CD103 low tissue-resident pathogenic CD4 1 T cells with higher production of effector cytokines, such as IL-4, IL-5, IL-13, IL-17A, and IFN-g (23) . However, immunosuppressive cells were comparable between PPSs and NPSs in our study, indicating that these cells may be insufficient to constrain the robust activation of a variety of immune cells in PSSs. Considering that Tregs are equally important to prevent inflammation-induced tissue damage during acute infections and to promote tissue repair, the scholars suggest that Tregs-based strategies could be considered for COVID-19 patient management (46) . Hence, immune intervention may be one of the effective treatment measures to reduce the occurrence of PS after the resolution of SARS-CoV-2 infection.

Several other important findings emerged from our data. First, IL-2 1 T cells, IL-4 1 T cells, and IL-17A 1 T cells hardly expressed CD57 but were enriched among CD28 1 cells, indicating that autocrine cytokines may provide a tonic signal that inhibits senescence. Furthermore, CD4 1 T cells, CD8 1 T cells, NK cells, NKT cells, and gd T cells secreting GZMB and perforin were enriched among CD57 1 cells, CD28 2 cells, and CD27 2 cells. Although peripheral blood CD57 1 T cells exhibit phenotypic and functional features of terminally differentiated effector cells, these cells display enhanced cytotoxic function (47) (48) (49) . Moreover, GZMB and perforin expression exhibited intrinsic positive correlations among CD4 1 T cells, CD8 1 T cells, NK cells, NKT cells, and gd T cells, suggesting that SARS-CoV-2 infection may simultaneously activate a variety of immune cells. In addition, we also noted downregulated expression of NKP30 and NKP46 in NK cells despite excessive secretion of perforin and GZMB in PPSs. Combined with previous research showing reduced surface expression of NKP30 and NKP46 on adaptive memory NK cells (50) , downregulated expression of NKP30 and NKP46 may act as a protective mechanism against tissue damage induced by excessive secretion of perforin and GZMB. Certainly, SARS-CoV-2 may escape the killing of NK cells and damage lung tissue due to downregulated expression of NKP30 and NKP46.

Although our study confirms some findings and provides new data on the innate and adaptive immune landscape of patients with PS who have recovered from COVID-19 1 year after discharge, we recognize limitations that might be overcome with larger sample sizes and matched control populations. Furthermore, there is a lack of understanding of the phenotype and function of immune cells from the lungs, which may directly participate in the formation of PS. Hence, the hierarchy of immunodominant circulating blood immune cells may not exactly reflect immunophenotypic features in the lungs. In summary, our study first shows significant differences in immunological characteristics between PPSs and NPSs 1 year after discharge. Although the detailed mechanisms by which cellular immunity participates in the development of PS remain to be investigated, our in-depth analysis of immunological profiling contributes to our understanding of the immunopathogenesis of COVID-19, facilitating the tailoring of more effective and proactive therapies for these patients.

Study design and participants. In order to determine the immunopathogenesis of residual lung lesions in COVID-19 survivors 1 year after discharge, a total of 50 convalescent patients were recruited at union hospitals. At the visit, routine blood test, chest computed tomography (CT) scans, and nucleic acid test and antibody detection for SARS-CoV-2 were performed for each participant. The patients were classified as 24 PPSs and 26 NPSs according to radiological characteristics. Patients who reached complete radiological resolution were regarded as NPSs. Complete radiological resolution was defined as the absence of any chest radiographic abnormality potentially related to infection (51) . PS, including residual GGO, fibrous stripe shadow, tractive bronchiectasis, reticulation, and bronchovascular bundle distortion, were evaluated by two radiologists. During our recruitment process, we excluded participants with underlying chronic lung diseases and cancers. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology (no. 2020/0004), and written informed consents were obtained from all participants.

Detection of SARS-CoV-2 mRNA and serum SARS-CoV-2 IgG and IgM. SARS-CoV-2 RNA was detected by reverse transcription-PCR (RT-PCR). Total nucleic acid extraction from nasopharyngeal specimens was performed using the QIAamp RNA viral kit (Qiagen), and two sets of primers were taken for two target genes (i.e., open reading frame 1ab [ORF1ab] and nucleocapsid protein [N]). Detection of serum SARS-CoV-2 IgG and IgM antibodies was evaluated by IgM/IgG antibody detection kit (Abbott Laboratories, Inc.).

Multiparametric flow cytometric analysis. Peripheral blood (200 ml) from convalescent patients with COVID-19 was added into a tube, and 2 ml of red cell lysis buffer was added and incubated for 10 min. After lysis, the sample was washed twice with phosphate-buffered saline (PBS) containing 1% fetal bovine serum (FBS). Immune cells were surface stained with fluorochrome-conjugated antibodies. The samples were incubated with antibodies for 15 min at 4°C. Cells were resuspended in PBS and washed at 400 Â g for 6 min. The specimens were immediately evaluated by flow cytometry.

For cytotoxic molecule detection, we isolated peripheral blood mononuclear cells (PBMCs) from heparinized blood by Ficoll-Hypaque gradient centrifugation (Pharmacia, Uppsala, Sweden). For surface staining, PBMCs were washed twice with PBS containing 1% FBS and stained with fluorochrome-conjugated antibodies. Intracellular staining for granzyme B (GZMB) and perforin was performed after cell fixation and permeabilization (eBioscience), and then intracellular proteins were labeled with the corresponding antibodies conjugated with fluorescent molecules according to the manufacturer's instructions.

A list of the antibodies used is provided in Table S1 in the supplemental material, and the gating strategy is presented in Fig. S1 . Flow cytometry was performed using a BD LSRFortessa X-20 (BD Biosciences), and data were analyzed with FlowJo v10 software.

Cytokine production assays. PBMCs were cultured in RPMI 1640 supplemented with 10% FBS, and cytokine production assays were performed after lymphocytes were stimulated with polymethyl acrylate (PMA; 50 ng/ml) and ionomycin (1 mM) in the presence of GolgiStop. After 5 h at 37°C, the cells were stained with fluorochrome-associated antibodies specific for surface molecules; next, the cells underwent fixation and permeabilization for intracellular staining with antibodies specific for the following intracellular proteins: IL-2, IL-4, IL-17A, IFN-g, and TNF-a.

Statistical analysis. The Shapiro-Wilk test was used to evaluate the distribution of variance. Continuous variables with normally and nonnormally distributed data were assessed using unpaired two-tailed Student's t tests or Mann-Whitney U test, respectively. Fisher's exact test was applied to examine categorical variables. Spearman's rank coefficient was used to determine correlations between two variables. Multivariate logistic regression analyses were performed to identify the independent predictive factors of residual lesions. The final model was determined using stepwise logistic regression, with significance level for selection set at a P value of 0.05. The optimum cutoff values were defined based on their maximum Youden index (sensitivity, 11; specificity, 21). All tests were 2-sided, and significance levels were set to *, P , 0.05; **, P , 0.01; ***, P , 0.001; ****, P , 0.0001; and ns, not significant. All statistical data were analyzed using SPSS version 25.0 statistical software (Chicago, IL, USA), GraphPad Prism 8 software (GraphPad Software, La Jolla, CA), or R software version 4.0.2 (Institute for Statistics and Mathematics, Vienna, Austria).

Supplemental material is available online only. FIG S1, PDF file, 0.7 MB. TABLE S1, PDF file, 0.1 MB.

Characteristics of and important lessons from the coronavirus disease 2019 (COVID-19) outbreak in China: summary of a report of 72314 cases from the Chinese Center for Disease Control and Prevention

Plasma metabolomic profiling of patients recovered from COVID-19 with pulmonary sequelae 3 months after discharge

Immune response to SARS-CoV-2 and mechanisms of immunopathological changes in COVID-19

Clonal expansion and activation of tissue-resident memory-like Th17 cells expressing GM-CSF in the lungs of patients with severe COVID-19

Functional SARS-CoV-2-specific immune memory persists after mild COVID-19

Lasting antibody and T cell responses to SARS-CoV-2 in COVID-19 patients three months after infection

Robust T cell immunity in convalescent individuals with asymptomatic or mild COVID-19

Targets of T cell responses to SARS-CoV-2 coronavirus in humans with COVID-19 disease and unexposed individuals

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

Detection of SARS-CoV-2-specific humoral and cellular immunity in COVID-19 convalescent individuals

Broad and strong memory CD4(1) and CD8(1) T cells induced by SARS-CoV-2 in UK convalescent individuals following COVID-19

Influenza virus-induced lung injury: pathogenesis and implications for treatment

A molecular single-cell lung atlas of lethal COVID-19

Six-month follow-up chest CT findings after severe COVID-19 pneumonia

The pulmonary sequalae in discharged patients with COVID-19: a short-term observational study

China Medical Treatment Expert Group for Covid-19. 2020. Clinical characteristics of coronavirus disease 2019 in China

Stochastic inheritance of division and death times determines the size and phenotype of CD8(1) T cell families

Natural killer cell specificity for viral infections

Raising the NKT cell family

Making a case for using gammadelta T cells against SARS-CoV-2

Peripheral blood myeloid-derived suppressor cells reflect disease status in idiopathic pulmonary fibrosis

CCR2(1) monocytic myeloidderived suppressor cells (M-MDSCs) inhibit collagen degradation and promote lung fibrosis by producing transforming growth factor-beta1

CD103(hi) Treg cells constrain lung fibrosis induced by CD103(lo) tissue-resident pathogenic CD4 T cells

T cells in fibrosis and fibrotic diseases

Deciphering myeloid-derived suppressor cells: isolation and markers in humans, mice and non-human primates

Daratumumab depletes CD381 immune regulatory cells, promotes T-cell expansion, and skews T-cell repertoire in multiple myeloma

Elevated calprotectin and abnormal myeloid cell subsets discriminate severe from mild COVID-19

A dynamic immune response shapes COVID-19 progression

Exhausted-like CD81 T cell phenotypes linked to C-peptide preservation in alefacept-treated T1D subjects

Longitudinal characteristics of lymphocyte responses and cytokine profiles in the peripheral blood of SARS-CoV-2 infected patients

Clinical and pathological investigation of patients with severe COVID-19

The laboratory tests and host immunity of COVID-19 patients with different severity of illness

Reappearance of effector T cells is associated with recovery from COVID-19

Functional exhaustion of antiviral lymphocytes in COVID-19 patients

Ten Brinke A, Hombrink P. 2020. Divergent SARS-CoV-2-specific T-and B-cell responses in severe but not mild COVID-19 patients

Pathophysiology, transmission, diagnosis, and treatment of coronavirus disease 2019 (COVID-19): a review

Decreased B cells on admission associated with prolonged viral RNA shedding from the respiratory tract in coronavirus disease 2019: a case-control study

Adaptive immune features of natural killer cells

Reverse-transcribed SARS-CoV-2 RNA can integrate into the genome of cultured human cells and can be expressed in patient

Viral infection of the lung: host response and sequelae

A human circulating immune cell landscape in aging and COVID-19

Impaired cytotoxic CD8(1) T cell response in elderly COVID-19 patients

Imbalance of regulatory and cytotoxic SARS-CoV-2-reactive CD4(1) T cells in COVID-19

Natural killer cell immunotypes related to COVID-19 disease severity

Tissue-resident CD8(1) T cells drive age-associated chronic lung sequelae after viral pneumonia

Potential of regulatory T-cell-based therapies in the management of severe COVID-19

CD8(1)CD57(1) T cells exhibit distinct features in human non-small cell lung cancer

Comparison of primary human cytotoxic T-cell and natural killer cell responses reveal similar molecular requirements for lytic granule exocytosis but differences in cytokine production

The cytolytic enzymes granyzme A, granzyme B, and perforin: expression patterns, cell distribution, and their relationship to cell maturity and bright CD57 expression

The natural cytotoxicity receptors in health and disease

Pneumonia recovery: discrepancies in perspectives of the radiologist, physician and patient

We thank all participating patients for their support and cooperation as well as our colleagues from Imaging Department of Union Hospital. This work was supported by funds from the Key Special Project of Ministry of Science and Technology, China (no. 2020YFC0845700), Scientific Research Projects of Chinese Academy of Engineering (no. 2020-KYGG-01-07), and Fundamental Research Funds for the Central Universities (no. 2020kfyXGYJ029).