Fig 1: Cellular screening of SARS-CoV-2-specific response.a Memory CD8+ T cells (CD45RO+) from peripheral blood mononuclear cells (PBMC) of COVID-19 convalescents were seeded at 2000 cells/well in multiple wells containing irradiated allogenic PBMC, PHA, IL-2, IL-7, and IL-15 as described in Methods. The individual cultures were expanded to make up the libraries of polyclonally amplified CD8+ T cells. Then the libraries were washed and examined as to their antigen specificity by stimulating with a series of viral protein-expressing aAPC. Once being judged as positive by IFNγlevel (measured by ELISA), they were further expanded with relevant aAPC and cytokines. These libraries (cell lines) were subjected to several downstream analyses including epitope screening, cytokine profiling, and/or cytotoxicity assays, etc. b A representative result of the CD8+ T cell library assay. Libraries from a healthy donor (HC-010) or COVID-19 convalescent (CV-004) were divided into 7 groups, and each group was co-cultured with indicated aAPCs: S (n = 15 for HC-010; n = 47 for CV-004), M (n = 15 for HC-010; n = 47 for CV-004), N (n = 15 for HC-010; n = 47 for CV-004), ORF3a (n = 14 for HC-010; n = 41 for CV-004), or ORF1ab NSP6 (n = 14 for HC-010; n = 53 for CV-004) -expressing aAPCs, influenza virus and cytomegalovirus (Flu/CMV) derived peptides (n = 15 for HC-010; n = 47 for CV-004)-pulsed aAPC or aAPCs (None) (n = 15 for HC-010; n = 47 for CV-004). Each dot represents the IFNγ level of each library. The threshold of positivity was determined by mean + 3 SD of IFNγ level in the group co-cultures with aAPCs (indicated as “None”). c Heatmap representation of results from all participants (healthy controls (HC), convalescents from mild, moderate (mod), and severe (sev) disease) subjected to the library assay. The frequency of positive library in response to aAPCs (None), indicated viral protein-expressing, or influenza virus and cytomegalovirus (Flu/CMV) derived peptides-pulsed aAPC are shown. d Comparison of positive library frequency between healthy donors (HC) (open circles, n = 8) and COVID-19 convalescents (COVID-19) (filled circles, n = 20) in each group. p values were calculated by two-sided Mann–Whitney test. Data represent mean ± SD. **p = 0.0064. e Comparison of positive library frequency in response to SARS-CoV-2-M protein-expressing aAPC among indicated groups (healthy donors (HC) (n = 8), convalescents from mild severity COVID-19 (mild COVID-19) (n = 8), those from moderate severity COVID-19 (moderate COVID-19) (n = 8) and those from severe severity COVID-19 (severe COVID-19) (n = 4)). p values were calculated by two-sided Mann–Whitney test. Data represent mean ± SD. ***p = 0.0002; *p = 0.024.
Fig 2: Human CD8+ and γδ T cell proliferation and activation relies on three signals. Proliferation of CD8+ T cells in the presence of serial dilutions of peptide P26‐35, and (A) either CD86/4‐1BBL/HLA‐A*02‐expressing (red squares) or CD86/4‐1BBL‐expressing (blue circles) K562 cells and cytokines (IL‐7 + IL‐15 + IL‐21) (n = 5); (B) either CD86/4‐1BBL/HLA‐A*02‐expressing (red squares) or HLA‐A*02‐expressing (blue circles) K562 cells and cytokines (n = 5); and (C) CD86/4‐1BBL/HLA‐A*02‐expressing K562 cells, and the absence (open black squares) or presence of IL‐2 (red circles), IL‐7 (blue squares), IL‐15 (red triangles), IL‐21 (blue circles), or IL‐7 + IL‐15 + IL‐21 (red squares) (n = 5). Proliferation of γδ T cells in the presence of serial dilutions of an anti‐human TCR γδ antibody, and (D) either the absence (blue circles) or presence (red squares) of K562‐derived aAPC expressing CD86 and 4‐1BBL and cytokines (n = 6); (E) either parental (blue circles) or CD86/4‐1BBL‐expressing (red squares) K562 cells and cytokines (n = 6); and (F) K562‐derived aAPC expressing CD86 and 4‐1BBL, and either the absence (open black squares) or presence of IL‐2 (red circles), IL‐7 (blue squares), IL‐15 (red triangles), IL‐21 (blue circles), IL‐15 + IL‐21 (open red squares) or IL‐7 + IL‐15 + IL‐21 (solid red squares) (n = 6). Data is shown as the mean ± SEM. Normalization of the data and statistical analysis were performed as described in the Section Methods, 2.
Fig 3: Functional effects of IL-7 and IL-7-receptor internalization in HPB-ALL cells and human CD8+ PBMCs remain intact upon cleavage by MMP-9. (A) Analysis of cell-surface IL-7R, 5 minutes after stimulation of HPB-ALL cells with 50 ng/ml IL-7, cIL-7 (1/100 molar ratio MMP-9/IL-7, 4h at 37°C) or the equivalent of MMP-9. Data represent four independent experiments, with experimental replicates shown in the same color. Histograms represent median values and error bars represent the IQR. Data were normalized to the unstimulated condition, representing steady-state quantities of cell surface IL-7R (100% IL-7Rα). *p < 0.05, as determined by Kruskal-Wallis test with Dunn’s correction for multiple comparisons and calculated on mean values of each experiment. (B) Analysis of pSTAT3 (Tyr705) in HPB-ALL cells, 15 minutes after stimulation with 50 ng/ml IL-7, cIL-7 (1/100 molar ratio MMP-9/IL-7, 4h at 37°C) or the equivalent of MMP-9. Each data point indicates an independent experiment. Histograms represent median values and error bars represent the IQR. Data were normalized to β-actin. (C) Growth of HBP-ALL cells, 4 days upon stimulation with 50 ng/ml IL-7, cIL-7 (1/100 molar ratio MMP-9/IL-7, 4h at 37°C) or the equivalent of MMP-9 (right panel). Data represent six independent experiments, with all experimental replicates shown in the same color. Histograms represent medians and error bars represent the IQR. Data were normalized to cells stimulated with IL-7 alone, representing 100% growth. ***p ≤ 0.001, as determined by Kruskal-Wallis test with Dunn’s correction for multiple comparisons and calculated on the mean values of each experiment. (Left panel) representative images of HPB-ALL cells with the indicated stimulations. Scale bar = 200 µm. (D) Analysis of cell-surface IL-7R, 24h after stimulation of human CD8+ PBMCs with 15 ng/ml IL-7, cIL-7 (1/100 molar ratio MMP-9/IL-7, 4h at 37°C) or the equivalent of MMP-9. Histograms represent median values and error bars represent the IQR. *p < 0.05, as determined by Kruskal-Wallis test with Dunn’s correction for multiple comparisons. ns, not significant.
Fig 4: Characterization of IL-7 cleavage by MMP-9. (A) Recombinant IL-7 produced in E. coli (left), insect cells (middle) or human HEK-293 cells (right) were incubated with different concentrations of active MMP-9 at the indicated molar ratios (MMP-9/IL-7), for a period of 2h. Fragments were separated by SDS-PAGE under reducing conditions followed by silver staining. MMP-9, intact IL-7 (IL-7i) and two cleavage products (IL-7* & IL-7**) are indicated with arrowheads, respectively, in black and green color. inh; negative control in the presence of 500 µM of the MMP inhibitor SB-3CT (inh). (B) Recombinant IL-7 produced in E. coli (top, non-glycosylated), insect cells (middle, partially glycosylated) or human HEK-293 cells (bottom, fully glycosylated) were incubated with active MMP-9 at a molar ratio of 1:100 (MMP-9/IL-7) and samples were taken at the indicated time-points. Fragments were resolved by SDS-PAGE under reducing conditions. Intact IL-7 (IL-7i) and two cleavage products (IL-7* & IL-7**) are indicated with arrowheads, respectively, in black and green color. inh; negative control in the presence of 500 µM of the MMP inhibitor SB-3CT. (C), Edman sequencing data. Identification of the neo-N-termini generated upon digestion of IL-7 with MMP-9 (1:100, MMP-9:IL-7 for 6h). Identification of the LGEAQ N-terminus from fragment IL-7**, indicates an MMP-9 cleavage site between the amino acid residues A128 and L129. Variations in N-terminal amino acids of fragments IL-7* are due to the different IL-7 expression systems ( Supplementary Table 1 ). (D) Amino acid sequence of human IL-7 with indication of glycosylation sites (blue), disulfide bridges (yellow), MMP-9 cleavage site (green triangle) and IL-7 fragments (green arrows). (E) 3D model of IL-7 [PDB structure 3DI3 (29)] with indication of the location of the MMP-9 cleavage site (green triangle), glycosylation sites (blue hexagons) and disulfide bridges (yellow).
Fig 5: IL-7 interaction with the IL-7 receptor remains intact upon cleavage by MMP-9. (A) His-tagged IL-7 (produced in insect cells) was subjected to histidine tag pull-down. Bead-bound proteins and non-bound proteins were resolved by reducing SDS-PAGE. (B) IL-7 in the presence (+) or absence (-) of MMP-9, subjected to non-reducing or reducing electrophoretic separation. (C) 3D model of the interaction between IL-7 and the IL-7Rα (blue) [based on PDB structure 3DI3 (29)] with indication of the IL-7 fragments generated by MMP-9 [IL-7* (orange) and IL-7** (red)]. (D) Binding of IL-7 [from E. coli (top) or mammalian cells (bottom)] and MMP-9-cleaved IL-7 (cIL-7, 1/100 molar ratio MMP-9/IL-7, 4h at 37°C) to rhIL-7Rα, as analyzed by SPR (results representative for 2 experiments, see Supplementary Figure 3 ). Colors represent different concentrations of IL-7 as indicated and black lines represent curve fits with a two-state reaction model (IL-7-IL-7Rα) according to McElroy et al. (29).
Supplier Page from BioLegend for Recombinant Human IL-7 (carrier-free)