Human MAIT cell cytolytic effector proteins synergize to overcome carbapenem resistance in Escherichia coli

Abstract

Mucosa-associated invariant T (MAIT) cells are abundant antimicrobial T cells in humans and recognize antigens derived from the microbial riboflavin biosynthetic pathway presented by the MHC-Ib-related protein (MR1). However, the mechanisms responsible for MAIT cell antimicrobial activity are not fully understood, and the efficacy of these mechanisms against antibiotic resistant bacteria has not been explored. Here, we show that MAIT cells mediate MR1-restricted antimicrobial activity against Escherichia coli clinical strains in a manner dependent on the activity of cytolytic proteins but independent of production of pro-inflammatory cytokines or induction of apoptosis in infected cells. The combined action of the pore-forming antimicrobial protein granulysin and the serine protease granzyme B released in response to T cell receptor (TCR)-mediated recognition of MR1-presented antigen is essential to mediate control against both cell-associated and free-living, extracellular forms of E. coli. Furthermore, MAIT cell-mediated bacterial control extends to multidrug-resistant E. coli primary clinical isolates additionally resistant to carbapenems, a class of last resort antibiotics. Notably, high levels of granulysin and granzyme B in the MAIT cell secretomes directly damage bacterial cells by increasing their permeability, rendering initially resistant E. coli susceptible to the bactericidal activity of carbapenems. These findings define the role of cytolytic effector proteins in MAIT cell-mediated antimicrobial activity and indicate that granulysin and granzyme B synergize to restore carbapenem bactericidal activity and overcome carbapenem resistance in E. coli.

Fig 1. MAIT cells kill bacteria-infected cells and suppress bacterial loads.

(A) Assessment of apoptosis of E. coli EC120S-infected HeLa cells by Casp3 activity and MAIT cell degranulation by CD107a, Gnly, and GrzB expression. (B) Apoptosis of infected HeLa cells at indicated time point (n = 9–11). (C) Measurement of early apoptosis (Casp3⁺DCM⁻) on uninfected or EC120S-infected HeLa cells, (D) degranulation by MAIT cells, and (E) bacterial counts following lysis of infected HeLa cells after 3 h of co-culture with and without MAIT cells in the presence of anti-MR1 or isotype control (n = 7–8 for EC120S-infected HeLa cells+anti-MR1, n = 24–25 for others). (F) Bacterial counts in infected HeLa cell lysates or in total lysates (cell lysates plus supernatants) after 3 h of co-culture with or without MAIT cells (n = 14). (G) MAIT cell degranulation, (H) apoptosis of EC120S-infected HeLa cells, and (I) relative bacterial loads following co-culture of EC120S-infected HeLa cells with MAIT cells derived from D0, 2, and 15 of culture (n = 4). (J) Percentage of cytolytic proteins expressed by MAIT cells from D0, 2, and 15 of culture assessed by flow cytometry (n = 3–10). (K, L) Flow cytometry analysis on frequency (K) and levels (MFI) (L) of cytolytic protein expression by PB (n = 4) and NP (n = 3) MAIT cells at various time points. (M) Representative FACS plot of apoptosis of EC120S-infected HeLa cells and MAIT cell degranulation following co-culture with PB or NP MAIT cells (n = 3). Data presented as heat map shows the mean, whereas data presented as line or bar graphs with error bars represent the mean and standard error. Box and whisker plots show median, the 10th to 90th percentile, and the interquartile range. Statistical significance was calculated using mixed-effects analysis followed by Tukey’s multiple comparison test (C–E), Wilcoxon’s signed-rank test or Friedman’s test with Dunn’s multiple comparisons test (F), or repeated-measure one-way ANOVA (G–I). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05. The underlying data of this figure can be found in S1 Data. Casp, caspase; CFU, colony-forming units; D, day; DCM, dead cell marker; FACS, fluorescence-activated cell sorting; Gnly, granulysin; GrzB, Granzyme B; MAIT, Mucosa-associated invariant T; MFI, mean fluorescence intensisty; MR1, MHC-Ib-related protein; NP, nasopharyngeal; ns, not significant; PB, peripheral blood.

Fig 2. MAIT cell antimicrobial activity is associated with cytolytic protein expression.

(A) Detection of cytolytic protein contents in effector MAIT cells and target E. coli EC120S-infected or uninfected HeLa cells following 3 h co-culture with MAIT cells obtained at different days following MAIT cell activation. Representative histograms from 4 independent donors are shown. (B) Correlation between live (n = 28) and (C) apoptotic (n = 28) EC120S-infected HeLa cells, (D) GrzB intensity (n = 19), (E) GrzB expression (n = 28), (F) Gnly expression (n = 27), and (G) Gnly-GrzB co-expression (n = 27) in MAIT cells with the inhibition of bacterial growth. (H) Levels of the GrzB substrate GranToxiLux activity in uninfected or EC120S-infected HeLa cells with or without MAIT cells (n = 4). (I) Flow cytometry plots of (J) degranulation by MAIT cells, (K) apoptosis in EC120S-infected HeLa cells, and (L) the relative bacterial loads in the presence of the indicated inhibitors or mAbs. Bar graphs with error bars represent the mean and standard error. Box and whisker plots show median, the 10th to 90th percentile, and the interquartile range. Statistical significance was determined using the ordinary one-way ANOVA (H), or mixed-effects analysis (J–L) followed by Dunnett’s or Tukey’s post hoc test as appropriate. Correlations were calculated using the Pearson (B, C) or the Spearman test (D–G). ****p < 0.0001, ***p < 0.001, **p < 0.01,*p < 0.05. The underlying data of this figure can be found in S1 Data. α, anti; Casp, caspase; d, day; ctrl, control; EGTA, ethylene glycol tetraacetic acid; Gnly, granulysin; Grz, Granzyme; IETD-CHO, N-acetyl-L-isoleucyl-L-α-glutamyl-N-[(1S)-2-carboxy-1-formylethyl]-L-threoninamide trifluoroacetate; mAbs, monoclonal antibodies; MAIT, Mucosa-associated invariant T; Nil, untreared; ns, not significant; Prf, perforin; Pro-VAD, valyl-alanyl-aspartyl-[O-methyl]-fluoromethylketone.

Fig 3. MAIT cells respond to and control CREC.

(A) Expression of CD107a, GrzB, IFNγ, TNF, and IL-17A in MAIT cells stimulated for 24 h with E. coli strains DH5α (n = 16), EC120S (n = 7), and the carbapenem-resistant strains EC234 (n = 16), EC241 (n = 5), EC362 (n = 16), and EC385 (n = 5). Unstimulated, n = 16. (B) MR1-dependency of effector protein and cytokine production by MAIT cells stimulated with indicated strains. MR1-dependency was calculated as previously described [2] (n = 7). (C) Representative flow cytometry plots of Casp3 activation and apoptosis in HeLa cells alone, HeLa cells infected with EC241, or co-cultured with MAIT cells with or without EC241 for 3 h. (D) Apoptosis of HeLa cells alone or co-cultured with MAIT cells for 3 h in the presence of 5-OP-RU, EC120S, EC234, EC241, or EC362 (n = 5). (E, F) Casp3 activation (E) and bacterial loads (F) in HeLa cells infected with EC241 co-cultured with MAIT cells in the presence of anti-MR1 antibody or isotype control (n = 5–8 for EC241-infected HeLa cells+anti-MR1 mAb, n = 13–25 others). Data presented as bar graphs with error bars represent the mean and standard error. Box and whisker plots show median, the 10th to 90th percentile, and the interquartile range. Statistical significance was determined using the Kruskal-Wallis ANOVA (A) or the Friedman test (B, D) followed by Dunn’s multiple comparison test, or mixed-effects analysis followed by Dunnett’s multiple comparison test (E, F). ****p < 0.0001, ***p < 0.001,**p < 0.01,*p < 0.05, [*]p < 0.1. The underlying data of this figure can be found in S1 Data. Casp, caspase; CFU, colony-forming units; CREC, carbapenem-resistant E. coli; Ctrl, control; DCM, dead cell marker; FCS-A, forward scatter area; Grz, Granzyme; IFNγ, interferon γ; IL-17A, interleukin-17A; mAb, monoclonal antibody; MAIT, Mucosa-associated invariant T; MR1, MHC-Ib-related protein; TNF, tumor necrosis factor; 5-OP-RU, 5-(2-oxopropylideneamino)-6-D-ribitylaminouracil.

Fig 4. MAIT cell secretomes mediate antimicrobial activity against CREC.

(A) Concentration of cytolytic proteins in the supernatant of MAIT cells following 3 h co-culture with E. coli EC120S-infected HeLa cells (n = 5–6). (B) Gating strategy for E. coli identification and quantification of SYTOX Green by flow cytometry. (C) SYTOX Green levels (n = 10) and (D) bacterial counts of various E. coli strains after incubation with MAIT cell or control supernatants for 2 h (n = 6 [EC120S], 12 [EC234, EC362]). (E, F) Levels of SYTOX Green^bright in E. coli EC234 and EC362 in the presence of MAIT cell or control supernatants supplemented with imipenem for 2 h (n = 10) or (G, H) the live bacterial counts over 24 h (n = 11–16 [EC234], 11–13 [EC362]). Data presented as line or bar graphs with error bars represent the mean and standard error. Statistical significance was determined using paired t test (C), Wilcoxon’s signed-rank test (D), RM one-way ANOVA with Dunnett’s post hoc test (F), and mixed-effects analysis with Sidak’s post hoc test (G, H). Significant differences in bacterial counts cultured in control versus MAIT cell supernatants at indicated time points and imipenem concentrations in (G) and (H) were indicated by asterisks. ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05. The underlying data of this figure can be found in S1 Data. CFU, colony-forming units; CREC, carbapenem-resistant E. coli; Ctrl, control; Gnly, granulysin; Grz, Granzyme; MAIT, Mucosa-associated invariant T; Prf, perfori; s/n, supernatant.

Fig 5. Cytolytic proteins contribute to the antimicrobial activity of MAIT cell secretomes.

(A, B) Concentration of cytolytic proteins in the MAIT cell supernatants spiked with imipenem in the E. coli EC234 (n = 12) and EC362 (n = 28) cultures in which growth was detected or not at the indicated time points. (C, D) Correlation between Gnly and (E, F) GrzB levels in the MAIT cell supernatants and E. coli EC234 and EC362 bacterial loads after 24 h incubation in the MAIT cell supernatants spiked with imipenem. (G) Gating strategy of Gnly flow cytometry staining and (H) proportion of Gnly in E. coli EC234 (n = 4) and EC362 (n = 6) following 30 min incubation with MAIT cell or control supernatants. (I) Representative histograms of GrzB, and (J) SYTOX Green staining and (K) intensity in E. coli EC234 (n = 4) and EC362 (n = 6) following 30 min incubation in MAIT cell or control supernatants. Data presented as bar graphs with error bars represent the mean and standard error. Box and whisker plots show median, the 10th to 90th percentile, and the interquartile range. Statistical significance was determined using the Mann-Whitney test (A, B), paired t test for intrastrain or unpaired t test for interstrain analyses (H), and Friedman’s test with Dunn’s post hoc test (K). Correlations were calculated using the Pearson test (C, E) and the Spearman test (D, F). ***p < 0.001, **p < 0.01,*p < 0.05, [*] p < 0.1. The underlying data of this figure can be found in S1 Data. CFU, colony-forming units; Gnly, granulysin; Grz, Granzyme; MAIT, Mucosa-associated invariant T; MFI, mean fluorescence intensity; ns, not significant; Prf, perforin; s/n, supernatant.

Fig 6. A model of MAIT cell antimicrobial activity against cell-associated (A) and extracellular (B) drug-sensitive E. coli and CREC, and (C) MAIT cell secretome potentiation of carbapenem killing activity against carbapenem-resistant E. coli strains.

CREC, carbapenem-resistant E. coli; IFNγ, interferon-γ; IL-17A; interleukin-17A MAIT, Mucosa-associated invariant T; MR1, MHC-Ib-related protein; TCR, T cell receptor; TNF, tumor necrosis factor.