ORIGINAL RESEARCH published: 11 February 2020 doi: 10.3389/fimmu.2020.00103 Cell-Mediated Immune Responses to in vivo-Expressed and Stage-Specific Mycobacterium tuberculosis Antigens in Latent and Active Tuberculosis Across Different Age Groups Edited by: Rajko Reljic, Mariateresa Coppola 1*† , Raquel Villar-Hernández 2† , Krista E. van Meijgaarden 1 , St George’s, University of London, Irene Latorre 2 , Beatriz Muriel Moreno 2 , Esther Garcia-Garcia 2 , Kees L. M. C. Franken 1 , United Kingdom Cristina Prat 2 , Zoran Stojanovic 3 , Maria Luiza De Souza Galvão 4 , Joan-Pau Millet 5 , Reviewed by: Josefina Sabriá 6 , Adrián Sánchez-Montalva 7,8 , Antoni Noguera-Julian 9 , Paulo Bettencourt, Annemieke Geluk 1 , Jose Domínguez 2† and Tom H. M. Ottenhoff 1*† University of Oxford, United Kingdom 1 Cecilia Lindestam Arlehamn, Department of Infectious Diseases, Leiden University Medical Center, Leiden, Netherlands, 2 Institut d’Investigació Germans La Jolla Institute for Immunology (LJI), Trias i Pujol, CIBER Enfermedades Respiratorias, Universitat Autònoma de Barcelona, Barcelona, Spain, 3 Servei de United States Neumología Hospital Universitari Germans Trias i Pujol, Institut d’Investigació Germans Trias i Pujol, CIBER Enfermedades Respiratorias, Universitat Autònoma de Barcelona, Barcelona, Spain, 4 Unitat de Tuberculosi de Drassanes, Hospital *Correspondence: Universitari Vall d’Hebron, Barcelona, Spain, 5 Serveis Clínics, Unitat Clínica de Tractament Directament Observat de la Mariateresa Coppola Tuberculosi, CIBER de Epidemiología y Salud Pública (CIBEREESP), Madrid, Spain, 6 Servei de Pneumologia, Hospital Sant m.coppola@lumc.nl Joan Despí Moises Broggi, Sant Joan Despí, Spain, 7 Infectious Diseases Department, Vall d’Hebron University Hospital, Tom H. M. Ottenhoff PROSICS Barcelona, Universitat Autònoma de Barcelona, Barcelona, Spain, 8 Grupo de Estudio de Micobacterias (GEIM), t.h.m.ottenhoff@lumc.nl Sociedad Española de Enfermedades Infecciosas y Microbiología Clínica (SEIMC), Madrid, Spain, 9 Malalties Infeccioses i † These Resposta Inflamatòria Sistèmica en Pediatria, Unitat d’Infeccions, Servei de Pediatria, Institut de Recerca Pediàtrica Hospital authors have contributed Sant Joan de Déu, CIBER de Epidemiología y Salud Pública (CIBERESP), Madrid, Spain equally to this work Specialty section: A quarter of the global human population is estimated to be latently infected by This article was submitted to Mycobacterium tuberculosis (Mtb), the causative agent of tuberculosis (TB). TB remains Vaccines and Molecular Therapeutics, a section of the journal the global leading cause of death by a single pathogen and ranks among the top-10 Frontiers in Immunology causes of overall global mortality. Current immunodiagnostic tests cannot discriminate Received: 14 October 2019 between latent, active and past TB, nor predict progression of latent infection to Accepted: 15 January 2020 Published: 11 February 2020 active disease. The only registered TB vaccine, Bacillus Calmette-Guérin (BCG), does Citation: not adequately prevent pulmonary TB in adolescents and adults, thus permitting Coppola M, Villar-Hernández R, van continued TB-transmission. Several Mtb proteins, mostly discovered through IFN-γ Meijgaarden KE, Latorre I, Muriel centered approaches, have been proposed as targets for new TB-diagnostic tests Moreno B, Garcia-Garcia E, Franken KLMC, Prat C, Stojanovic Z, or -vaccines. Recently, however, we identified novel Mtb antigens capable of eliciting De Souza Galvão ML, Millet J-P, multiple cytokines, including antigens that did not induce IFN-γ but several other Sabriá J, Sánchez-Montalva A, Noguera-Julian A, Geluk A, cytokines. These antigens had been selected based on high Mtb gene-expression in Domínguez J and Ottenhoff THM the lung in vivo, and have been termed in vivo expressed (IVE-TB) antigens. Here, (2020) Cell-Mediated Immune we extend and validate our previous findings in an independent Southern European Responses to in vivo-Expressed and Stage-Specific Mycobacterium cohort, consisting of adults and adolescents with either LTBI or TB. Our results confirm tuberculosis Antigens in Latent and that responses to IVE-TB antigens, and also DosR-regulon and Rpf stage-specific Mtb Active Tuberculosis Across Different Age Groups. Front. Immunol. 11:103. antigens are marked by multiple cytokines, including strong responses, such as for doi: 10.3389/fimmu.2020.00103 TNF-α, in the absence of detectable IFN-γ production. Except for TNF-α, the magnitude Frontiers in Immunology | www.frontiersin.org 1 February 2020 | Volume 11 | Article 103 Coppola et al. Cell-Mediated Responses to Mtb Antigens of those responses were significantly higher in LTBI subjects. Additional unbiased analyses of high dimensional flow-cytometry data revealed that TNF-α+ cells responding to Mtb antigens comprised 17 highly heterogeneous cell types. Among these 17 TNF-α+ cells clusters identified, those with CD8+TEMRA or CD8+CD4+ phenotypes, defined by the expression of multiple intracellular markers, were the most prominent in adult LTBI, while CD14+ TNF-α+ myeloid-like clusters were mostly abundant in adolescent LTBI. Our findings, although limited to a small cohort, stress the importance of assessing broader immune responses than IFN-γ alone in Mtb antigen discovery as well as the importance of screening individuals of different age groups. In addition, our results provide proof of concept showing how unbiased multidimensional multiparametric cell subset analysis can identify unanticipated blood cell subsets that could play a role in the immune response against Mtb. Keywords: IVE-TB antigens, Mycobacterium tuberculosis (Mtb), TB, LTBI, cytokines, cell responses INTRODUCTION contain a large number of epitopes predicted to bind to HLA-Ia and HLA-II alleles (coverage of 85% of the human population); Tuberculosis (TB) kills more than 4,000 persons per day and is (v) they are well-recognized by immune blood cells from Mtb the most life-threatening disease caused by a single infectious exposed subjects (as shown in 37 Mtb exposed individuals); agent, Mycobacterium tuberculosis (Mtb) (1). Although one and (vi) they elicit immune cells that are producing multiple fourth of the global population is estimated to be latently infected cytokines besides IFN-γ, a key cytokine known to be necessary (2), none of the current immunodiagnostics can discriminate but not sufficient in conferring protection against TB (12). between recent and past Mtb infection, or predict risk of possible IFN-γ has been used as the main readout to study classical TB progression in LTBI (3–5). Although neonatal vaccination Mtb antigens such as latency antigens (DosR regulon encoded with Bacillus Calmette-Guérin (BCG), the only licensed TB antigens, HBHA), resuscitation-promoting factors (Rpfs), and vaccine, protects infants against severe forms of TB, it is unable secreted antigens (such as ESAT-6 and CFP-10) (5, 13). Latency to prevent active pulmonary TB in adults and adolescents, such antigens are thought to be mostly expressed during latent that it has little impact on Mtb transmission (6). To boost or stages of Mtb infection (14), while Rpf proteins appear to be replace BCG, multiple TB vaccine candidates have been proposed functionally required in the transition from a dormant to an and are being evaluated in different higher end preclinical or actively replicating state of Mtb (15). In support of this, several clinical phase 1 and phase 2 studies (7). Recently one subunit Mtb stage specific antigens were recognized more strongly by vaccine candidate, M72/AS01E, induced significant protection IFN-γ producing cells from LTBI than from TB patients and from developing TB in a LTBI population (8, 9). Although the therefore have been proposed as novel Mtb antigen specific numbers of prevented cases were relatively low as yet (10 TB cases tools to differentiate latent Mtb infection from active TB disease in the M72/AS01E group vs. 22 TB cases in the placebo group, (16–20). Additionally, although to only a limited extent, the a vaccine efficacy of 54%), the outcome of this phase IIb trial recognition of some of these Mtb stage specific antigens has been supports the power of rational Mtb antigen-based subunit TB assessed by immune parameters other than IFN-γ, including vaccine development. IL-12, IP-10, IL-10, TNF-α (20–25). Ideally, Mtb proteins selected as targets for new TB vaccines To date, the most extensively characterized cellular subsets should be expressed during active Mtb lung infection and participating in the response to Mtb antigens have been mono- or efficiently trigger immune effector cells capable of controlling poly-functional CD4+ T cells producing IFN-γ, TNF-α, and/or or clearing the infection without inflicting major tissue damage. IL-2 (26). However, it remains understudied which other cell Recently, we identified a new class of Mtb antigens, named IVE- subsets recognizing Mtb antigens may be involved in the overall TB, encoded by Mtb genes that were highly and consistently response. Recent evidence for example has highlighted a role for expressed in the lung of susceptible (C3HeB/FeJ) as well as NK cells (27) as well as ILC3 (28) in protective immunity to resistant (C57BL/6J) mice following aerosol Mtb (Erdman) TB. In this study, we first validated the recognition of several challenge (10, 11). Besides their high expression in Mtb infected recently identified Mtb antigens by multicomponent cytokine lungs, these IVE-TB proteins constitute an attractive group of signatures in an independent cohort of LTBI and TB patients. The novel candidate antigens for multiple other reasons (11): (i) they magnitude of those responses was higher in latently Mtb infected are conserved among 219 Mtb clinical isolates and thus cover a subjects. Additionally, the use of high dimensional single cell wide array of Mtb strains; (ii) they have high homology with BCG data analysis revealed numerous clusters of antigen specific TNF- and thus have potential as booster vaccines; (iii) they have high α+ cells, uncovering immunological heterogeneity in cellular homology with pathogenic mycobacteria, including M. leprae subsets responding to Mtb antigens in different age groups of and NTM and thus have potential as broader vaccines; (iv) they LTBI and TB. Frontiers in Immunology | www.frontiersin.org 2 February 2020 | Volume 11 | Article 103 Coppola et al. Cell-Mediated Responses to Mtb Antigens TABLE 1 | Demographic and clinical characteristics of all patients included in the study. Overall Adults Adolescents (n = 35) (n = 20) (n = 15) TB LTBI TB LTBI TB LTBI (n = 18) (n = 17) (n = 12) (n = 8) (n = 6) (n = 9) Age, average (years) ± SD 32.7 ± 14.1 26.8 ± 13.5 41.1 ± 8.6 40 ± 6.1 15.8 ± 1.2 15.0 ± 1.0 Gender (%) Female 5 (27.8) 10 (58.8) 2 (16.7) 4 (50.0) 3 (50.0) 6 (66.7) Male 13 (72.2) 7 (41.2) 10 (83.3) 4 (50.0) 3 (50.0) 3 (33.3) Country of birth (%) High TB burdena 2 (11.1) 0 (0.0) 0 (0.0) 0 (0.0) 2 (33.3) 0 (0.0) Low TB burden 16 (88.9) 17 (100.0) 12 (100.0) 8 (100.0) 4 (66.7) 9 (100.0) BCG vaccination (%) 11 (61.1) 5 (29.4) 7 (58.3) 5 (62.5) 4 (66.7) 0 (0.0) Known TB contact (%) 4 (22.2) 14 (82.3) 2 (16.7) 5 (62.5) 2 (33.3) 9 (100.0) TST >15 mm (%) ≥15 mm (%) 5 (27.8) 10 (58.8) 1 (8.3) 5 (62.5) 4 (66.7) 5 (55.6) <15mm (%) 5 (27.8) 5 (29.4) 3 (25.0) 1 (12.5) 2 (33.3) 4 (44.4) Unknown/not tested 8 (44.4) 2 (11.8) 8 (66.7) 2 (25.0) 0 (0.0) 0 (0.0) QFN-G-IT Positive 17 (94.4) 16 (94.1) 11 (91.7) 7 (87.5) 6 (100.0) 9 (100.0) Negative 1 (5.6) 0 (0.0) 1 (8.3) 0 (0.0) 0 (0.0) 0 (0.0) Unknown/not tested 0 (0.0) 1 (5.9) 0 (0.0) 1 (12.5) 0 (0.0) 0 (0.0) Prophylaxis (%) <30 days – 14 (82.3) – 6 (75.0) – 0 (0.0) >30 days – 1 (5.9) – 1 (12.5) – 8 (88.9) Average (days) ± SD – 22.9 ± 7.1 – 27.4 ± 8.2 – 19.0 ± 2.5 No prophylaxis – 2 (11.8) – 1 (12.5) – 1 (11.1) Anti-TB treatment (%) <30 days 6 (33.3) – 4 (33.3) – 2 (33.3) – >30 days 10 (55.6) – 6 (50.0) – 4 (66.7) – Average (days) ± SD 40.2 ± 17.0 – 37.6 ± 18.0 – 48.8 ± 33.2 – No treatment 2 (11.1) – 2 (16.7) – 0 (0.0) – a About 150 incident cases per 100,000 population. BCG, bacillus Calmette-Guérin; LTBI, latent tuberculosis infection; QFN-G-IT, QuantiFERON Gold In-Tube; SD, standard deviation; TB, tuberculosis; TST, tuberculin skin test. MATERIALS AND METHODS to TB treatment (the antidepressant Citalopram). The study was approved by the Ethics Committee of all participating centers Study Setting and Patients Recruitment (reference CEIC: PI-15-073) (http://www.ceicgermanstrias.cat/) In this study, whole blood samples were collected from 20 and performed following the guidelines and regulations. For each adults (age range = 27–51) and 15 adolescents (age range = participant a written informed consent was collected together 13–17) with pulmonary active TB (n = 18) or latent Mtb with a detailed questionnaire about clinical and demographic infection (LTBI) (n = 17) (Table 1). Active TB patients were data of the study participant. determined by a compatible X-ray, positive Mtb sputum culture and/or positive PCR. In adolescents the TB diagnosis was Recombinant Proteins also supported by a positive tuberculin skin test (TST) and A total of 59 Mtb recombinant proteins, previously identified a known TB contact. LTBI cases were defined by a positive by different Mtb antigen discovery approaches (13), were tested TST and/or QuantiFERON-TB Gold In tube (QFN-G-IT) test. in this study (Table 2). As described previously (11), Mtb Donors were recruited from six centers located in Barcelona, genes were amplified by PCR from genomic H37Rv DNA Spain: Germans Trias i Pujol University Hospital, Unitat de and cloned by Gateway technology (Invitrogen, Carlsbad, CA, Tuberculosi Vall d’Hebron-Drassanes, Serveis Clínics TB Directly USA) in a bacterial expression vector containing, overexpressed Observed Treatment Unit, Vall d’Hebron University Hospital, in Escherichia coli (E. coli) BL21 (DE3) and purified. Gel Sant Joan Despí Moises Broggi Hospital and Sant Joan de Déu electrophoresis and western blotting with an anti-His Ab Barcelona Children’s Hospital. Among all the subjects included in (Invitrogen) and an anti-E. coli polyclonal antibody (a gift of this study only one (TB21) had a record of drug intake unrelated Statens Serum Institute, Copenhagen, Denmark) were used to Frontiers in Immunology | www.frontiersin.org 3 February 2020 | Volume 11 | Article 103 Coppola et al. Cell-Mediated Responses to Mtb Antigens TABLE 2 | List of Mtb antigens included in this study. TABLE 2 | Continued List Rv number Function Category Ref List Rv number Function Category Ref 1 Rv0066 oxalosuccinate IVE-TB (11) 40 Rv2941 28 IVE-TB (11) decarboxylase 41 Rv3048c R1F protein IVE-TB (11) 2 Rv0287/ EsxG/EsxH IVE-TB (11) 42 Rv3052 FadB4 IVE-TB (11) Rv0288 43 Rv3407 VapB47 IVE-TB (11) 3 Rv0383c Possible conserved IVE-TB (11) 44 Rv3462 IF-1 infA IVE-TB (11) secreted protein 45 Rv3583c Possible IVE-TB (11) 4 Rv0423c ThiC IVE-TB (11) transcription factor 5 Rv0440 GroEL2 IVE-TB (11) 46 Rv3614/ /EspC IVE-TB (11) 6 Rv0467 icl1 IVE-TB (11) 3615 7 Rv0468 FadB2 IVE-TB (11) 47 Rv3615 EspC IVE-TB (11) 8 Rv0470c PcaA IVE-TB (11) 48 Rv3616* EspA IVE-TB (11) 9 Rv0501 GalE2 IVE-TB (11) 49 Rv3846 SodA IVE-TB (11) 10 Rv0640 RplK IVE-TB (11) 50 Rv3865* EspF IVE-TB (11) 11 Rv0642c MmaA4 IVE-TB (11) 51 Rv3874/ CFP10/ESAT6 IVE- (5, 11) 12 Rv0645 MmaA1 IVE-TB (11) Rv3875 TB/secreted antigens 13 Rv0824c Acyl-desaturase IVE-TB (11) DesA1 52 Rv0867c RpfA Rpf (15) 14 Rv0826 Conserved IVE-TB (11) 53 Rv1009 RpfB Rpf (15) hypothetical protein 54 Rv1733c Rv1733c latency (14) 15 Rv0991 Conserved serine IVE-TB (11) antigen rich protein 55 Rv2032 acg latency (14) 16 Rv1038c EsxJ IVE-TB (11) antigen 56 Rv2034 Rv2034 IVE-TB (11) 17 Rv1131 PrpC IVE-TB (11) 57 Rv2389c RpfD Rpf (15) 18 Rv1221 SigE IVE-TB (11) 58 Rv2450c RpfE Rpf (15) 19 Rv1284 Beta-carbonic IVE-TB (11) anhydrase 59 Rv3353c Rv3353c IVE-TB (11) 20 Rv1390 RpoZ IVE-TB (11) Rpf, resuscitation-promoting factors; Ref, reference(s). 21 Rv1479 MoxR1 IVE-TB (11) *Antigens included in a patented TB vaccine candidate. 22 Rv1738 Conserved IVE- (11, 14) hypothetical protein TB/latency antigen check the size and purity of the recombinant proteins. Rv0287- 23 Rv1791 PE19 IVE-TB (11) Rv0288, Rv2346c-Rv2347, and Rv3614-Rv3615 were produced 24 Rv1792 EsxM IVE-TB (11) as fusion proteins to mirror the pairwise dependent secretion 25 Rv1846 BlaI IVE-TB (11) pathway followed by T7S systems. All recombinant proteins were 26 Rv1872 IldD2 IVE-TB (11) tested to exclude protein-non-specific T cell stimulation and 27 Rv1980c Mpt64 IVE-TB (11) cellular toxicity (11). 28 Rv2007 FdxA IVE-TB (11) 29 Rv2031 HspX IVE-TB (11) Whole Blood Assay 30 Rv2215 DlaT IVE-TB (11) Within 3 h of collection, heparinized venous blood was diluted 31 Rv2245 KasA IVE-TB (11) 1:10 in AIM-V medium (Invitrogen, Breda, the Netherlands). 32 Rv2346c/ EsxO/EsxP IVE-TB (11) Samples were incubated (450 µl/well) in 48-well-plates at 37◦ C, Rv2347c 5% CO2 , with 50 µl antigen solution (final concentration 33 Rv2382 Polyketide IVE-TB (11) of 10 µg/ml). After 6 days, 200 µl of the supernatants was synthetase mbtC collected from each well and frozen in aliquots at −20◦ C until 34 Rv2431c PE25 IVE-TB (11) further analysis. 35 Rv2461 ClpP1 IVE-TB (11) 36 Rv2626 Hrp1 IVE- (11, 14) Multiple Cytokine Array and Analysis of TB/latency antigen Diluted Whole Blood Supernatant As it has been recommended by a cross-laboratory evaluation 37 Rv2657c PhiRv2 prophage IVE-TB (11) protein of multiplex bead assays, in this study we used one multiplex 38 Rv2710 SigB IVE-TB (11) kit from the same manufacturer (29). The R&D TM premixed 39 Rv2873 Mpt83 IVE-TB (11) Multi-analyte kit (Cat #: 1415903) was used to measure the concentrations of eight analytes (IL-13, IL-22, IL-17A, IFN- (Continued) γ, induced protein 10 [IP-10 (CXCL10)], IL-10, GM-CSF, and Frontiers in Immunology | www.frontiersin.org 4 February 2020 | Volume 11 | Article 103 Coppola et al. Cell-Mediated Responses to Mtb Antigens TNF-α) in diluted whole blood culture supernatants according for 15 min at room temperature in order to detach adherent to manufacturer’s instructions. These cytokines were selected cells. Erythrocytes were lysed and white blood cells fixed with to validate and extend our previously published findings (11). FACS lysing solution (BD Biosciences). The fixed cells were To rule out batch effects, for each multiplex run samples pelleted and cryopreserved in 1 ml of FCS with 10% dimethyl with different disease status and age were always mixed. Data sulfoxide (DMSO). were acquired using Luminex 200 (Luminex Corp., Austin, TX) and analyzed using Bio-Plex Manager software 6.0 (Bio- Flow Cytometry Reagents Rad Laboratories), as previously described (11). The median Fixed whole-blood samples were thawed in a water bath at background values of unstimulated samples were: 14 pg/ml (GM- 37◦ C for 2 min and stained with a 14-color FACS panel in CSF, LTBI = 14 pg/ml and TB = 14 pg/ml), 47 pg/ml (IFN-γ, permeabilization solution (Fix&Perm cell permeabilization kit, LTBI = 47 pg/ml and TB = 47 pg/ml), 3 pg/ml (IL-10, LTBI = An Der Grub BioResearch GMBH, Susteren, the Netherlands). 3 pg/ml and TB = 4 pg/ml), 5 pg/ml (IL-17A, LTBI = 4 pg/ml The 14-color FACS panel included: CD3-PE-TexasRed (clone and TB = 5 pg/ml), 10 pg/ml (IL-22, LTBI = 12 pg/ml and S4.1) (Thermo Fischer), CD4-Pacific Blue (clone RPA-T4), CD28 TB = 9 pg/ml), 17 pg/ml (IP-10, LTBI = 10 pg/ml and TB = PerCP-Cy5.5 (clone L293), CD8-PECy5 (clone RPA-T8), CD14- 20 pg/ml), and 3 pg/ml (TNF-α, LTBI = 3 pg/ml and TB = 4 HorizonV500 (clone RPA-T8), IFN-γ-AlexaFluor700 (clone B27) pg/ml). Values outside the upper (ULOQ) or lower (LLOQ) limits (all BD Biosciences), CD45RA-Brilliant Violet 650 (clone HI100), of quantification were set as the values of the analyte detection CD27-Brilliant Violet 605 (clone O323), TNF-α-APC-Cy7 (clone limits. Due to the high LLOQ for IL-13 (median = 177 pg/ml), MAb11), IL-10-PE-Cy7 (clone JES3-9D7), IL-13-PE (clone the values detected for this cytokine did not fall in the linear part JES10-5A2) (all BioLegend), IL-17-A-FITC (clone eBio17B7), IL- of the standard curve and therefore were not further analyzed. 22-PerCP-eFluor710 (clone 22 URTI) (all eBioscience), IL-32- allo-phycocyanin (APC) (clone 373821) (R&D). Samples were IL-32 ELISA acquired on a BD LSRFortessa using FACSDiva software (version Based on new reports describing the possible role of IL-32 6.2, BD Biosciences) with compensated parameters. in protection against TB (30, 31), we decided to assess the IL-32 alpha concentration in supernatants, using the DuoSet Data Analysis ELISA (R&D Systems, Catalog #: DY3040-05) according to Significant differences (p < 0.05) between cytokine levels the manufacturer’s instructions. Samples were diluted 2-fold in stimulated and unstimulated samples were evaluated by in reagent diluent (R&D Systems, Catalog #: DY995). The Mann–Whitney U-test corrected for multiple comparison optical density (O.D.) was acquired by Varioskan Flash (Thermo (FDR, Benjamini-Hochberg test correction). The R package Electron Corporation) using the SkanIt software version 2.4.1. “phenotypicForest” was used to construct the polar histogram Data were linearized by plotting the log of the human IL- (35). After dividing by the background-values, i.e., AIM- 32 concentration (pg/ml) vs. the log of the O.D. and the fit V medium, for each response in each donor, differences line was determined by regression analysis using GraphPad between cytokine levels in LTBI donors and TB patients were Prism (version 7.0). The concentration read from the standard assessed by Mann-Whitney U-test (p < 0.05) and median curve was multiplied by the dilution factor of 2. The median log2 fold-changes [log2FC(LTBI/TB)> 1 or <−1]. Results are background value of unstimulated samples was 2 pg/ml for both reported only for Mtb antigens that induced increased cytokine TB ad LTBI groups. levels compared to the background-values. Multiple R-squared derived by a linear model was used to compare responses Blood Processing for Whole Blood between different age groups (https://www.rdocumentation.org/ Intracellular Staining (WB-ICs) Assay packages/stats/versions/3.6.1/topics/lm). As previously described (32, 33), venous blood was collected from Flow cytometry data were analyzed using FlowJo v10. Results 15 study participants (described in Figure 3A) in sodium heparin from single-stained and unstained mouse/rat κ beads were used tubes and, within 60 min (34), 1 ml of blood was transferred to calculate compensations. Cell doublets were excluded using into Sarstedt tubes containing either AIM-V medium alone or forward scatter-area vs. forward scatter-height parameters. After combined with a pool of Mtb antigens (Rv1131, Rv2461, and debris exclusion, TNF-α+ cells were gated and FCS files of Rv3616c, selected based on the multiplex results), or PPD, in the stimulated samples (antigen pool and PPD) were exported. FACS presence of co-stimulants (anti-CD28 and anti-CD49d, each at data were transformed using hyperbolic arcsin with a cofactor 1 µg/ml, BD Biosciences, Erembodegem, Belgium). The IVE-TB of 150 directly within Cytosplore+HSNE (36). Then, for each proteins were tested at a final concentration of 10 µg/ml while donor and condition the gated events were randomly down- PPD (Statens Serum Institut, Copenhagen, Denmark) was used sampled to the same number of TNF-α+ events as the sample at a final concentration of 5 µg/ml. with the lowest number of TNF-α+ cells (n = 153), to ensure After 3 h of incubation in a water bath set at 37◦ C, Brefeldin equal representation of all samples. Next, a HSNE analysis was A (3 µg/ml; Sigma-Aldrich, Zwijndrecht, the Netherlands) and performed on a total of 3,366 TNF-α+ cells after defining the Monensin (1:1,000; BD Biosciences) were added. Samples were markers used for the similarity computation (CD45, CD3, CD4, then transferred back to the water bath, programmed to switch CD28, CD8, CD14, IFN-γ, CD27, IL-10, IL-13, IL-17A, IL-22, off after 12 h. Samples remained in the water bath and EDTA and IL-32). We used the standard parameters for the hierarchy was added (2 mM final concentration) 9 h later and incubated construction; number of random walks for landmark selection: Frontiers in Immunology | www.frontiersin.org 5 February 2020 | Volume 11 | Article 103 Coppola et al. Cell-Mediated Responses to Mtb Antigens N = 100, random walk length: L = 15, number of random walks the same type of cytokines response (Figure 1A). However, for influence computation: N = 15, number of scale = 2. For most of the antigens studied (n = 54) induced strong TNF- any clustering that occurred the GMS grid size was set to S α production. Interestingly, out of 59 Mtb proteins, 30 were = 256 (36). The iterations chosen of the HSNE analysis were significantly recognized by cells producing two to five cytokines, 1,000. We clustered the data with a kernel size sigma of 30 on which not always included canonical IFN-γ (Figure 1B). the overview level without manual modifications. Cell clusters Together these results extend, and most importantly, validate were inspected using the integrated heatmap visualization. The our previous findings in a completely independent cohort of median abundance (i.e., median cell frequency) of the cells donors, by demonstrating multi- rather than single- component composing each cluster was calculated per each group of cytokine signatures in the response to IVE-TB antigens, and for donors and conditions (heatmap generated in Morpheus: https:// the first time also to DosR and Rpf stage specific Mtb antigens. software.broadinstitute.org/morpheus). Based on the median Importantly, we confirm immune recognition of Mtb antigens fluorescence intensity reported in the Supplementary Table 3, in the absence of detectable IFN-γ production in a significant manual gating using OMIQ (www.omiq.ai) was performed to number of cases (Figure 1B). investigate seven populations, each of those representing a best approximation of the main clusters discussed in the paper LTBI Individuals’ and TB Patients’ Blood (Supplementary Figure 1A). GraphPad Prism 8.2.1 was used to depict the frequencies of these cell populations for each donor Cells Secrete Different Amounts of (Supplementary Figure 1B). Cytokines in Response to IVE-TB and Stage Specific Mtb Antigens Next, we assessed whether the in vitro cytokine profiles in RESULTS response to the selected 59 Mtb antigens (Table 2) were similar IVE-TB Mtb Antigens, Mtb DosR-regulon, for stimulated blood cells from LTBI subjects (n = 17) or active (n = 18) TB patients (Table 1). This comparison was and Rpf Antigens Activate performed for five cytokines, TNF-α, GM-CSF, IFN-γ, IP-10, and Multiple-Cytokine-Producing Blood Cells IL-17A, that had been found to be significantly increased in Mtb of Mtb Exposed Subjects antigen stimulated samples above (Figure 1). Prior to analysis, We recently described a new class of Mtb antigens, termed IVE- the concentrations detected in response to every antigen within TB antigens, that were recognized by blood cells of Mtb exposed each donor were normalized to background-values, i.e., divided subjects from two small Northern European LTBI cohorts (n = by the medium concentration, and log2 transformed. Cytokine 37). IVE-TB antigens not only elicited canonical IFN-γ but also differences were assessed by Mann–Whitney U-test (significant alternative cytokines including GM-CSF, IP-10, IL-13, TNF-α, for p < 0.05) and by median log2 fold-changes [log2 FC(LTBI/TB) and IL-17 (11). To formally and independently validate these > 1 or < −1]. Results described below focus only on those findings, we here tested these antigens in a Southern European Mtb antigens that induced significantly different responses as cohort of 35 Mtb exposed individuals, which included LTBI as confirmed by both p-values and median log2 fold-changes well as TB patients (Table 1). In addition to IVE-TB proteins we (Figure 2, Supplementary Table 2). Of note, although the TB also included two latency antigens (Rv1733c and Rv2032) and group consisted of untreated patients and patients treated for less four resuscitation-promoting factors (Rv0867c, Rv1009, Rv2389c, or more than 30 days, we did not find significant differences in and Rv2450c), accumulating to a total of 59 Mtb proteins the level of cytokine secreted among those groups possibly in (Table 2). These highly immunogenic DosR- and Rpfs- antigens, part due to the small samples size (p > 0.05 for all conditions known to distinguish latent Mtb infection from active TB disease according to a Kruskal–Wallis H-test, Supplementary Table 2). based on IFN-γ production levels as previously reported (13, Of all 59 Mtb antigens, 16 were preferentially recognized by 16–20), were here included to investigate whether they could blood cells from LTBI donors (Rv0066, Rv0287/0288, Rv0470, also induce additional cytokines. To allow a comparison with Rv0501, Rv0824, Rv1131, Rv1390, Rv1733, Rv1980, Rv2034, our previously published study (11), we selected the same Rv2245, Rv2873, Rv3462, Rv3615, Rv3616, and Rv3614/3615) cytokines (IL-13, IL-22, IL-17A, IFN-γ, induced protein 10 [IP- (Figure 2A). Except for TNF-α, the cytokines detected in 10 (CXCL10)], IL-10, GM-CSF, and TNF-α) and assay (multiplex response to Mtb antigen stimulation were consistently higher in on the supernatants of diluted whole blood incubated for 6 days blood samples from LTBI individuals compared to TB patients with single or fusion Mtb antigens) as read-out, with only the new (Figure 2A). Of note, the recognition of these 16 antigens by addition of IL-32 (rationale explained in the methods section). samples from LTBI was primarily mediated by IP-10 production When measuring these multicomponent cytokine signatures in (Figure 2A). Interestingly, this changed upon age-subgroup the supernatants of diluted whole blood (incubated for 6 days analyses comparing adults (TB n = 8; LTBI n = 12) and with single or fusion Mtb antigens), secretion of TNF-α, GM- adolescents (TB n = 9; LTBI n = 6) (Figure 2B). Although IFN- CSF, IFN-γ, IP-10, and IL-17A (but not of IL-13, IL-22, IL-10, γ and IL-17A were abundantly secreted by blood cells of LTBI and IL-32) was significantly increased in stimulated compared donors independent of age (multiple R-squared derived by linear to unstimulated samples (Mann–Whitney U-test with FDR models: 0.1037 and 0.4192, for IFN-γ and IL-17, respectively), multiple test correction) (Figure 1A, Supplementary Table 1). In differences in IP-10 responses between LTBI and TB were less agreement with our previous findings, not all antigens induced marked in adolescents (multiple R-squared derived by linear Frontiers in Immunology | www.frontiersin.org 6 February 2020 | Volume 11 | Article 103 Frontiers in Immunology | www.frontiersin.org Coppola et al. 7 Cell-Mediated Responses to Mtb Antigens February 2020 | Volume 11 | Article 103 FIGURE 1 | IVE-TB Mtb antigens, Mtb DosR-regulon, and Rpf antigens activate multiple-cytokine-producing blood cells of Mtb exposed subjects. The levels of eight cytokines were measured in diluted whole blood supernatants from a cohort of Mtb exposed individuals (n = 35) after 6 days stimulation with either IVE-TB, DosR, Rpf (10 µg/ml), control antigens ESAT6/CFP10 (E/C) (10 µg/ml), PPD (5 µg/ml), or the positive control, mitogen PHA (2 µg/ml). The statistical significance of the differences between cytokine levels in stimulated and unstimulated samples was evaluated by Mann-Whitney U-test with FDR multiple test correction for IL-13, IL-22, IL-17A, IFN-γ, IP-10, IL-10, GM-CSF, IL-32, and TNF-α. (A) Results are shown only for TNF-α, GM-CSF, IP-10, IFN-γ, and IL-17A for which significant differences were found among single Mtb antigens and unstimulated samples. Each dot represents a donor. Bars indicate medians. Purple bars indicate significantly increased responses compared to the unstimulated samples with a p < 0.05 (Mann-Whitney U-test with FDR multiple test correction). (B) The polar histogram displays the different p-values found among single Mtb antigens (or control) and unstimulated samples. Stimuli are ordered clockwise as follows: controls and Mtb antigens grouped according to how many cytokines were induced. These groups of antigens are separated by interruptions of the histogram. In red are indicated Mtb antigens that induced cytokines other than IFN-γ. Coppola et al. Cell-Mediated Responses to Mtb Antigens FIGURE 2 | LTBI individuals’ and TB patients’ blood cells secrete different amounts of cytokines in response to IVE-TB and stage specific Mtb antigens. The levels of nine cytokines were measured in diluted whole blood supernatants after 6 days stimulation with either IVE-TB, DosR, Rpf antigens (10 µg/ml), ESAT6/CFP10 (E/C) (Continued) Frontiers in Immunology | www.frontiersin.org 8 February 2020 | Volume 11 | Article 103 Coppola et al. Cell-Mediated Responses to Mtb Antigens FIGURE 2 | (10 µg/ml), PPD (5 µg/ml), or PHA (2 µg/ml). The cytokine responses were compared between LTBI donors (n = 17) and TB patients (n = 18). This comparison is displayed for five cytokines, TNF-α, GM-CSF, IFN-γ, IP-10, and IL-17A, which were found to be significantly increased in stimulated samples compared to unstimulated ones (Figure 1). Within each donor, prior to further analysis, the pg/ml values detected in response to every antigen were normalized to the background-values, i.e., divided by the medium, and then log2 transformed. Results are reported only for Mtb antigens that induced increased cytokine levels compared to the background-values. (A) Cytokine differences were assessed by Mann–Whitney U-test (significant for p-values < 0.05) and by median log2 fold-changes [log2 FC(LTBI/TB) > 1 or <-1). P-values and median log2 fold-changes are reported on the y and x axes, respectively. Thresholds for p-values and median log2 fold-changes are indicated by dotted lines intersecting the axes (x:−1 and 1; y: 1.30103). Closed circles define when the log2 FC(LTBI/TB) is >1 (red) or <−1 (blue), such that closed red circles define antigens inducing higher responses in LTBI donors than TB patients, while blue circles depict antigens inducing higher responses in TB patients than LTBI subjects. Open circles define when the log2 FC(LTBI/TB) is >−1 and <1 (gray). (B) Differences between LTBI and TB patients were analyzed separately among adults (TB n = 8; LTBI n = 12) and adolescents (TB n = 9; LTBI n = 6). X and y-axes indicate the median log2 fold-changes in adolescents and adults, respectively. The color-coded dots indicate p-values: ≥0.05 in adults and adolescents (gray), <0.05 in adults (orange), <0.05 in adolescents (blue), and <0.05 in adults and adolescents (green). The gray area around the regression line (dotted line) represents the range in which the true regression line lies at a 95% level of confidence. Both in (A,B), the Rv numbers are coupled to antigens recognized differently by LTBI and TB according to both p-values and median log2 fold-changes. models: 0.009575) (Figure 2B, upper plots). While the highest the Mtb antigen pool but only in the adult group, thus partially TNF-α responses were found in adult TB patients (multiple R- mirroring the amount of TNF-α protein secreted in supernatants squared derived by linear models: 0.04714), GM-CSF responses of cells stimulated 6 days with single antigens (Figure 2B). were higher in adolescent LTBI donors than in adolescent TB We next imported the TNF-α+ cells from all stimulated patients; they were also higher in adult TB patients than adult conditions (both with the antigen pool or with PPD) LTBI subjects (multiple R-squared derived by linear models: into Cytosplore, and analyzed all data using hierarchical 0.05153) (Figure 2B, lower plots). Interestingly, these contrasting stochastic neighbor embedding (HSNE) (36). HSNE responses were consistent following stimulation with specific landmarks revealed a clear heterogeneity in the memory antigens such as Rv2461 or Rv1872 (Figure 2B), lending further cell compartment, distinguishing central memory phenotype validity to these observations. like subsets (CD45RA–CD27+CD28+) from effector Collectively, the data show that blood cells from LTBI subjects (CD45RA+CD27+/–CD28+/–) and terminally differentiated and TB patients produce qualitatively and quantitatively different effector (CD45RA+CD27–CD28–) memory like subsets cytokines in response to stimulation with Mtb antigens. The data, (Figure 3C). By applying a Gaussian mean-shift clustering although based on a limited number of donors, also suggest that to the flow cytometry data, 17 distinct cell clusters were certain cytokine responses such as IP-10, GM-CSF, and TNF- α defined by unique combinations of expressed markers, ranging can differ between age groups. from 22 to 454 cells per cluster (Figure 3D, upper panel). Hierarchical clustering of the heat map revealed two major Quantitative and Qualitative Differences in groups, corresponding either to the absence or presence of Specific TNF-α+ Responding Cellular CD45RA (Figure 3D, middle panel). The first major group, Subsets Between LTBI and TB Patients composed of cells lacking CD45RA but expressing CD28, To further characterize the functional blood cells involved in the clustered in 3 main subpopulations: co-expressing either CD14 recognition of Mtb antigens in a manner unbiased with respect but not CD3 [cluster 3, number of cells in the cluster (N) = 310], to phenotype, we focused on TNF-α+ cells, since TNF-α was co-expressing CD3 and CD4 (cluster 4, N = 352), or expressing the most abundant cytokine found in the supernatant of diluted IL-17 (cluster 12, N = 235) (Figure 3D, middle panel). A fourth whole blood in response to Mtb antigen stimulation (Figure 1). CD45RA negative cluster, mostly characterized by increased To address this question in the same samples as used in the IL-32 and IFN-γ expression was also identified but since it was cytokine analyses, we performed intracellular cytokine staining defined based on a very low number of cells (cluster 17, N = on whole blood (ICS-WB) as previously described by others 22) (Figure 3D, middle panel), it is not further described here. (32, 34, 37). Due to restriction in blood collection, the ICS- The second major group was characterized by a progressively WB could be performed only for a limited number of donors higher expression of CD45RA, combined with differential (n = 11). Samples were left unstimulated or stimulated over expression of CD27 molecules, co-expressing either CD4 (cluster night with a pool of three Mtb antigens (Rv1131, Rv2461, and 6, N = 219; cluster 15, N = 125 and cluster 4, N = 352), CD8 Rv3616c), that had been selected prior to commencing the study (cluster 8, N = 92; cluster 9, N = 93; and cluster 14, N = 87) or on the basis of previous results (11), or PPD. Interestingly, two both (cluster 2, N = 196 and cluster 10, N = 335) and several antigens of the pool (Rv1131 and Rv2461) were among those intracellular markers including IL-32, IL-17A, IFN-γ, and IL-22 recognized differently by LTBI and TB patients of different age but not IL-10 and IL-13 (Figure 3D). The phenotype of three cell groups in the cytokine screening (Figure 2B). After staining clusters expressing IL-17 could not be fully captured and better cryopreserved fixed blood cells with a 14-color FACS panel, all defined by the markers included in our analysis (cluster 11, N = TNF-α+ cells were enumerated in each sample (Figure 3A). 454; cluster 12, N = 235; and cluster 13, N = 88) (Figure 3D, Already after 12 h of stimulation, TB patients had larger numbers middle panel). Two cell clusters were only defined by different of TNF-α+ cells [quantified by log2 FC(LTBI/TB)] in response to expression of CD45RA, CD27, and CD28 but not in relation to PPD compared to LTBI subjects, after correction for background any of the cytokines measured (cluster 16, N = 87; cluster 7, N = values (Figure 3B). The same trend was observed in response to 302) (Figure 3D, middle panel). Frontiers in Immunology | www.frontiersin.org 9 February 2020 | Volume 11 | Article 103 Coppola et al. Cell-Mediated Responses to Mtb Antigens FIGURE 3 | Quantitative and qualitative differences are present in specific TNF-α+ cell subsets between LTBI and TB patients. Intracellular cytokine staining was performed on whole blood (ICS-WB) of a limited number of Mtb exposed individuals recruited in this study (n = 11). Samples were left unstimulated or stimulated for (Continued) Frontiers in Immunology | www.frontiersin.org 10 February 2020 | Volume 11 | Article 103 Coppola et al. Cell-Mediated Responses to Mtb Antigens FIGURE 3 | 12 h with a pool of three Mtb antigens (Rv1131, Rv2461, and Rv3616c) or PPD, in the presence of (T cell) co-stimulants (anti-CD28 and anti-CD49d). After staining cryopreserved fixed blood cells with an extensive 14-color FACS panel, TNF-α+ cells were quantified in each sample. (A) The gating strategy is shown for one sample (TB02) to illustrate the selection of single debris-free TNF-α+ cells in unstimulated (AIM-V i.e., medium) and stimulated (antigen pool or PPD) sample. In the left upper table, the number of TNF-α+ cells is indicated for each donor. (B) The difference in the absolute number of TNF-α+ cells between LTBI and TB was defined by median log2 fold change [log2 FC(LTBI/TB)] and calculated separately in adolescents and adults. Blue or red bars represent log2 FC(LTBI/TB) >0.5 or <-0.5, respectively. (C) For each of the 11 donors, TNF-α+ cells responding to the antigen pool and PPD were used as input in Cytosplore. Input cells were randomly downsampled to 153 events leading to a total of 3,366 TNF-α+ cells to explore via hierarchical stochastic neighbor embedding (HSNE). In the upper left HSNE plot, blue, and red dots indicate cells isolated from LTBI or TB patients (color tonality distinguishes adolscents from adults). In the other 11 HSNE plots the colored dots indicate the expression range of 11 singular cellular markers on the cells analyzed (high expression, red; medium expression, yellow; low expression, blue). (D) By applying a Gaussian mean-shift clustering to the FACS data, 17 distinct TNF-α+ cell clusters (displayed in the HSNE plot, upper panel) were defined by a unique combination of expressed markers as shown by the heatmap (middle panel). Differences in the size of colored squares indicate the contribution of each marker in defining the specific cell subset, while colors indicate the relative expression of each cellular marker compared to all the other markers (red indicates high; blue indicates low). Lower heatmap (lower panel) shows the proportion of cells composing each cluster per age group of donors (LTBI and TB) and stimuli (PPD and antigen pool). The numbers in the last row refer to the cell clusters. The proportion of cells (median relative abundance), derived infection, irrespective of age. Additionally, through unbiased from LTBI subjects and TB patients, which contributed to analysis of high dimensional single cell data, 17 clusters of Mtb the above cell clusters was different between adolescents and antigen-specific TNF-α+ cells of both myeloid and lymphoid adults (Figure 3D, lower panel). More specifically, in our origin were defined. This suggests the presence of as yet analysis TNF-α+ CD8+ (CD3+CD4–) cells with a poly- poorly defined but phenotypically and functionally different Mtb functional (IL-22, IFN-γ, and IL-32) terminal effector memory- antigen-responsive TNF-α+ cell subsets in LTBI subjects and TB like phenotype were mainly found in LTBI adults although patients of different age groups. their presence did not seem to require antigen specific stimulation (clusters 8, 9, and 14 Figure 3D, lower panel; DISCUSSION Supplementary Figure 1, Supplementary Table 3). By contrast, TNF-α+ cells co-expressing CD14 (cluster 1, N = 301; cluster Mtb antigens selected as targets for TB diagnostic tests or TB 3, N = 310; and cluster 5, N = 68; Figure 3D, lower vaccines need to be able to activate immune cells in Mtb exposed panel; Supplementary Table 3) mostly originated from antigen individuals (13). This has classically been studied, at least in stimulated samples of LTBI adolescents and TB adults while humans, by measuring Mtb antigen induced IFN-γ production they were almost absent in samples from antigen unstimulated in in vitro assays. While useful, sensitive and robust, clearly and stimulated TB adolescents and LTBI adults (Supplementary many other molecules are secreted by immune cells, often in Figure 1). Two TNF-α+ CD4 central memory/effector like the absence of IFN-γ. Our previous work showed that these clusters moderately co-expressing IFN-γ (clusters 4 and 6) were molecules include GM-CSF, IP-10, IL-13, TNF-α, and IL-17 (11). observed only in antigen stimulated samples of adolescent LTBI This concept was recently also confirmed by others (12). In and TB patients as well as of adult TB patients two (Figure 3D, the current study, we validate and confirm these findings in an lower panel; Supplementary Figure 1, Supplementary Table 3). independent cohort across different age groups and phases of One cluster of TNF-α+ T cells co-expressing CD4 and CD8 and Mtb infection (TB vs. LTBI). These molecules all participate in IFN-γ (cluster 10) was more abundantly present in adults LTBI the immune response against Mtb antigens, including several than in TB patients in response to PPD, but this was not found IVE-TB and stage specific antigens. Additionally, quantitative for adolescents (Supplementary Figure 1). differences in cytokine production were found between LTBI and Despite the limited number of donors these results support TB patients after Mtb antigen stimulation. Interestingly, high the presence of higher numbers of TNF-α+ cell subsets in the dimensional unbiased single cell data analysis defined multiple blood of adult TB patients compared to adult LTBI subjects clusters of TNF-α+ cells and suggested distinct abundancies in response to specific Mtb antigens (IVE-TB antigen pool or in these subsets in LTBI and TB patients across different PPD). In addition to these quantitative differences, our multi- age groups in response to Mtb antigen stimulation. These dimensional analysis, looking beyond the key cytokine IFN-γ, heterogeneous TNF-α+ cell clusters have potential implications defined phenotypically and functionally different TNF-α+ cell for the identification of correlates of infection and of protection, subsets in the blood of LTBI and TB patients. Future studies which are urgently needed in TB. Due to the limited number should confirm and assess the immunological relevance of these of subjects studied here, no definitive conclusions can be drawn cell subsets in larger and different cohorts. with regard to the immunological relevance of these cell subsets Collectively, our findings confirm and validate the recognition in relation to clinical outcome of infection. This will need of recently described Mtb antigens in an independent South validation and further analyses in future studies, including European cohort of LTBI and extend this also to active prospective cohorts. TB patients. We further confirmed that these responses are Although IL-12/IFN-γ axis deficiency results in susceptibility defined by multiple cytokines besides IFN-γ in both adults and to unusual mycobacterial infections (38, 39), the role of IFN-γ adolescents. Of note is that, except for TNF-α+, the magnitude in TB resistance and immunopathology, especially in the lung, of those responses was more pronounced during latent Mtb is not undisputed (4, 40, 41). In spite of this, most Mtb proteins Frontiers in Immunology | www.frontiersin.org 11 February 2020 | Volume 11 | Article 103 Coppola et al. Cell-Mediated Responses to Mtb Antigens have been proposed as candidate antigens almost exclusively on producing cells were not compromised by active TB disease or the basis of their recognition by IFN-γ producing cells obtained age (or time of Mtb exposure), since blood cells of all groups from latently infected individuals (LTBI) and TB patients (13, 20). were similarly effective in secreting IFN-γ in response to antigens Despite the fact that several Mtb proteins triggered different type present in the QuantiFERON-TB Gold test [ESAT-6/CFP-10/TB- of cytokines and chemokines, including IL-12, IP-10, IL-10, and 7.7(p4)] (Table 1) and to other specific Mtb antigens (Figure 1). TNF-α, these broader immune responses have been examined However, such impairment could not be excluded for IFN-γ- almost only for candidates that had already been selected by prior independent immune mechanisms, such as for GM-CSF and IFN-γ screening approaches (20–25, 42). Here, in a new cohort of TNF-α responses. Mtb exposed individuals, we corroborated our previous finding Finding age-related differences among TB and LTBI might on the added value of TNF-α, GM-CSF, IP-10, and IL-17A in suggest that either the abundance, the functionality or the the screening of putative Mtb antigens using a non-IFN-γ centric phenotype of GM-CSF and TNF-α producing cells might be approach (Figure 1). heterogeneous in relation to age. To start addressing this issue As for IFN-γ, there is no unequivocal evidence defining IL- we focused first on TNF-α+ cells, since the levels of TNF-α 17A, GM-CSF, and IP-10 as essential to controlling latent Mtb were the highest detectable ones in the supernatants of diluted infection (43–47) or active TB disease (48–51). TNF-α, generally whole blood in response to antigen stimulation. Although the associated with tissue damage (42, 52), is also considered a number of donors was limited (n = 11), we found that TNF-α+ key factor in TB granuloma integrity and protective immunity cells stimulated with three Mtb antigens (Rv1131, Rv2461, and (53). In our cohort, the concentrations of GM-CSF, IL-17A, Rv3616c) or PPD were more abundant in TB patients than adult IFN-γ, and IP-10 were higher in the stimulated blood cell LTBIs (Figure 3A). An opposite trend was seen in adolescent supernatants from LTBI compared to TB patients (Figure 2A). LTBI samples compared to TB patients when stimulated with the If LTBI control Mtb infection better than TB patients (54), pool of Mtb antigens (Figure 3B). Of note, this pool contained those Mtb antigens more strongly recognized by blood cells two antigens (Rv1131 and Rv2461) recognized differently by of LTBI correlate with, and potentially could contribute to the LTBI and TB patients of different age groups in the cytokine host immune response containing Mtb. This rational has led screening (Figure 2B). Therefore, one could speculate that part to the identification of several TB vaccine candidates whose of the differences found in the secreted cytokine profiles reflected efficacy has been proven in animal models (55–57), but not the number of TNF-α+ cells recognizing the pool of Mtb antigens yet tested in humans. It therefore remains to be seen whether (Figure 3A). there is a difference in efficacy between TB vaccine candidates More interestingly, by analyzing high dimensional single cell based on antigens preferentially recognized by LTBI vs. TB data (36), we observed additional differences in the phenotype patients (7, 8, 58–60). and functionality of TNF-α+ cells between age groups. TNF-α+ Despite the fact that most of the cytokines in response to cells co-expressing CD14 (Cluster 1, 5, and 3; Figure 3D) were Mtb antigens were found more abundantly in LTBI, as expected mainly found in adolescent LTBIs after stimulation with the Mtb by previous reports (42, 52) we found TNF-α responses to be antigen pool or PPD while almost absent in adult LTBIs. This is higher in TB patients (Figure 2A) and as such correlated with interesting since the major source of TNF in human blood seems pathology. The role of TNF-α is not unequivocal: on the one to be a subset of pro-inflammatory, non-classical monocytes, hand it has a pivotal role in the host defense against tuberculosis which have previously been associated with recent Mtb exposure (as proven by an increased risk of active disease following anti- (46, 65). TNF-α therapies), on the other hand its excess is also associated In contrast, among the cells from adult LTBIs, the TNF-α+ with TB pathogenesis both in animal models and humans cellular subsets were found to be CD8+ T cells with terminal (61, 62). Interestingly, when analyzing adults and adolescents effector memory (TEMRA) like phenotype co-expressing IL-22, separately, this difference was more pronounced in the adult IL-32, and IFN-γ although that did not require Mtb antigen group (Figure 2B). This trend in cytokine secretion between stimulation (Clusters 8, 9, and 14; Figure 3D, Supplementary the two age groups was also found for GM-CSF (Figure 2B). Figure 1). TNF-α+ CD8+ TEMRA cells have been already Adolescents and adults represent the primary target population described in adult LTBI and suggested to play a role in for subunit TB vaccination (63). However, the immune response antimicrobial activity against TB reactivation (66, 67). In one of adolescents is not often analyzed as a separate entity but of these studies, it was elegantly demonstrated that anti-TNF merged with that found in either children (48, 64) or adults therapy correlated with lower numbers of TNF-α+ CD8+ (37). The need for carefully considering different age groups TEMRA cells and decreased anti-mycobacterial activity that was is supported by the M72/AS01E efficacy trial results, which reverted by the addition of TNF-α+ CD8+ TEMRA cells (66). showed that the highest vaccine efficacy in preventing active TB Thus, it would be interesting to assess whether there is a link among LTBI was in individuals of 25 years of age or younger, between the “protection associated” TNF-α+ CD8+ TEMRA after 2.5 years of follow up (8, 9). This observed age-effect subset previously described and the ones found in our cohort. has been explained by possibly different timing of primary One subset co-expressing CD4 and CD8 with IFN-γ (cluster Mtb infection occurrence. Based on this reasoning, younger 10) was also mainly formed by cells originating from LTBI LTBI due to a more-recent primary Mtb infection would be adults. Recently (68), a new subset of CD4 CD8 double positive less likely to have the infection under immune control, and T cells able to produce cytokines and cytolytic markers has therefore could benefit more from boosting immune responses been identified in the blood, airways and lung granulomas of by the administration of M72/AS01E (8, 9). In our cohort, IFN-γ Mtb infected cynomolgus macaques. Additional studies would Frontiers in Immunology | www.frontiersin.org 12 February 2020 | Volume 11 | Article 103 Coppola et al. Cell-Mediated Responses to Mtb Antigens be required to clarify the role of those cells in the context AUTHOR CONTRIBUTIONS of active and latent Mtb infection. Larger cohorts studied by advanced multiparametric technologies might help to resolve the MC designed and performed experiments and analysis and cell heterogeneity, especially present in LTBI, found in our as wrote the manuscript. RV-H performed experiments and revised well as in previous studies (69). High-dimensional cytometry the manuscript. KM designed and performed experiments and analyses have been already harnessed to explore cells from revised the manuscript. BM and EG-G performed experiments. differently exposed Mtb adolescents uncovering new cell subsets, KF developed recombinant proteins. CP, ZS, MD, J-PM, JS, including those expressing CD16+ and other populations AS-M, and AN-J were responsible for the recruitment of defined as NK cells, CD27–CD8+ αβ T cells, B cells (27), and donors. RV-H, BM, and EG-G organized sample acquisition and ILC3 (28). database. AG and IL supervised experiments and revised the Despite the relatively limited number of subjects that could be manuscript. JD and TO supervised, designed experiments, and included, our cell subset-unbiased analysis reveals the presence revised the manuscript. All authors have read and approved the of heterogeneous TNF-α+ cell subsets associated with responses final manuscript. to Mtb antigens as well as PPD. These responses could reflect the spectrum of Mtb infection and Mtb exposure, but their identity FUNDING and exact function needs further elucidation (70, 71). In conclusion, our data validate our previous antigen This study was supported by the European Commission [FP7 discovery approach, support the value of assessing broader NEWTBVAC project contract no. LSHP-CT-2003-503367, EC immune responses than IFN-γ alone at an early stage of Mtb ITN FP7 VACTRAIN project; EC HORIZON2020 TBVAC2020 antigen discovery, and suggest how in depth unbiased profiling (Grant Agreement No. 643381)]. The text represents the of functional cells recognizing Mtb antigens can identify a yet authors’ views and does not necessarily represent a position ill-defined cell subsets participating in the immune response of the Commission who will not be liable for the use made of against Mtb. such information. This research was also supported by: (i) a grant from the Instituto de Salud Carlos III (PI16/01912 and DATA AVAILABILITY STATEMENT PI18/00411), integrated in the Plan Nacional de I+D+I and cofunded by the ISCIII Subdirección General de Evaluación All datasets generated for this study are included in the and the Fondo Europeo de Desarrollo Regional (FEDER); article/Supplementary Material. and (ii) a grant from the Sociedad Española de Neumología y Cirugía Torácica (project 25/2016; SEPAR; Barcelona, ETHICS STATEMENT Spain). AS-M was supported by a Juan Rodés (JR18/00022) postdoctoral fellowship from ISCIII. We acknowledge the This study was approved by the Ethics Committee of all support of the European Respiratory Society, Fellowship participating centers (reference CEIC: PI-15-073) (http://www. STRF 2016. ceicgermanstrias.cat/) and performed following the guidelines and regulations. For each participant a written informed consent SUPPLEMENTARY MATERIAL was collected together with a detailed questionnaire about clinical and demographic data of the study participant. Written The Supplementary Material for this article can be found informed consent to participate in this study was provided by the online at: https://www.frontiersin.org/articles/10.3389/fimmu. participants’ legal guardian/next of kin. 2020.00103/full#supplementary-material REFERENCES with Mycobacterium tuberculosis. Infect Immun. (2002) 70:2544–8. doi: 10.1128/IAI.70.5.2544-2548.2002 1. World Health Organization. Global Tuberculosis Report 2019 (2019). 6. Mathema B, Andrews JR, Cohen T, Borgdorff MW, Behr M, Glynn JR, 2. Houben RM, Dodd PJ. The global burden of latent tuberculosis infection: a et al. Drivers of tuberculosis transmission. J Infect Dis. (2017) 216(Suppl. re-estimation using mathematical modelling. PLoS Med. (2016) 13:e1002152. 6):S644–53. doi: 10.1093/infdis/jix354 doi: 10.1371/journal.pmed.1002152 7. TBVI. Pipeline of vaccines 2019. Available online at: https://www.tbvi.eu/ 3. Goletti D, Lee MR, Wang JY, Walter N, Ottenhoff THM. Update what-we-do/pipeline-of-vaccines/ (accessed October 12, 2019). on tuberculosis biomarkers: from correlates of risk, to correlates of 8. Van Der Meeren O, Hatherill M, Nduba V, Wilkinson RJ, Muyoyeta active disease and of cure from disease. Respirology. (2018) 23:455–66. M, Van Brakel E, et al. Phase 2b controlled trial of M72/AS01E doi: 10.1111/resp.13272 vaccine to prevent tuberculosis. N Engl J Med. (2018) 379:1621–34. 4. Simmons JD, Stein CM, Seshadri C, Campo M, Alter G, Fortune S, doi: 10.1056/NEJMoa1803484 et al. Immunological mechanisms of human resistance to persistent 9. Tait DR, Hatherill M, Van Der Meeren O, Ginsberg AM, Van Brakel E, Salaun Mycobacterium tuberculosis infection. Nat Rev Immunol. (2018) 18:575–89. B, et al. Final analysis of a trial of M72/AS01E vaccine to prevent tuberculosis. doi: 10.1038/s41577-018-0025-3 N Engl J Med. (2019) 381:2429–39. doi: 10.1056/NEJMoa1909953 5. Geluk A, van Meijgaarden KE, Franken KL, Subronto YW, Wieles 10. Commandeur S, van Meijgaarden KE, Prins C, Pichugin AV, Dijkman K, van B, Arend SM, et al. Identification and characterization of the ESAT- den Eeden SJ, et al. An unbiased genome-wide Mycobacterium tuberculosis 6 homologue of Mycobacterium leprae and T-cell cross-reactivity gene expression approach to discover antigens targeted by human T cells Frontiers in Immunology | www.frontiersin.org 13 February 2020 | Volume 11 | Article 103 Coppola et al. Cell-Mediated Responses to Mtb Antigens expressed during pulmonary infection. J Immunol. (2013) 190:1659–71. associated with M. tuberculosis infection outcomes. Nature. (2018) 560:644–8. doi: 10.4049/jimmunol.1201593 doi: 10.1038/s41586-018-0439-x 11. Coppola M, van Meijgaarden KE, Franken KL, Commandeur S, Dolganov 28. Ardain A, Domingo-Gonzalez R, Das S, Kazer SW, Howard NC, Singh A, G, Kramnik I, et al. New genome-wide algorithm identifies novel in- et al. Group 3 innate lymphoid cells mediate early protective immunity against vivo expressed Mycobacterium tuberculosis antigens inducing human T-cell tuberculosis. Nature. (2019) 570:528–32. doi: 10.1038/s41586-019-1276-2 responses with classical and unconventional cytokine profiles. Sci Rep. (2016) 29. van Meijgaarden KE, Khatri B, Smith SG, Drittij A, de Paus RA, Goeman 6:37793. doi: 10.1038/srep37793 JJ, et al. Cross-laboratory evaluation of multiplex bead assays including 12. Lu LL, Smith MT, Yu KKQ, Luedemann C, Suscovich TJ, Grace PS, et al. independent common reference standards for immunological monitoring IFN-gamma-independent immune markers of Mycobacterium tuberculosis of observational and interventional human studies. PLoS ONE. (2018) exposure. Nat Med. (2019) 25:977–87. doi: 10.1038/s41591-019-0441-3 13:e0201205. doi: 10.1371/journal.pone.0201205 13. Coppola M, Ottenhoff TH. Genome wide approaches discover novel 30. Bai X, Dinarello CA, Chan ED. The role of interleukin-32 against tuberculosis. Mycobacterium tuberculosis antigens as correlates of infection, disease, Cytokine. (2015) 76:585–7. doi: 10.1016/j.cyto.2015.06.013 immunity and targets for vaccination. Semin Immunol. (2018) 39:88–101. 31. Montoya D, Inkeles MS, Liu PT, Realegeno S, Teles RM, Vaidya P, et al. IL-32 doi: 10.1016/j.smim.2018.07.001 is a molecular marker of a host defense network in human tuberculosis. Sci 14. Voskuil MI, Schnappinger D, Visconti KC, Harrell MI, Dolganov GM, Transl Med. (2014) 6:250ra114. doi: 10.1126/scitranslmed.3009546 Sherman DR, et al. Inhibition of respiration by nitric oxide induces a 32. Kagina BM, Mansoor N, Kpamegan EP, Penn-Nicholson A, Nemes E, Mycobacterium tuberculosis dormancy program. J Exp Med. (2003) 198:705– Smit E, et al. Qualification of a whole blood intracellular cytokine 13. doi: 10.1084/jem.20030205 staining assay to measure mycobacteria-specific CD4 and CD8 T cell 15. Mukamolova GV, Kaprelyants AS, Young DI, Young M, Kell DB. immunity by flow cytometry. J Immunol Methods. (2015) 417:22–33. A bacterial cytokine. Proc Natl Acad Sci USA. (1998) 95:8916–21. doi: 10.1016/j.jim.2014.12.003 doi: 10.1073/pnas.95.15.8916 33. Hanekom WA, Hughes J, Mavinkurve M, Mendillo M, Watkins M, 16. Leyten EM, Lin MY, Franken KL, Friggen AH, Prins C, van Meijgaarden Gamieldien H, et al. Novel application of a whole blood intracellular cytokine KE, et al. Human T-cell responses to 25 novel antigens encoded by genes of detection assay to quantitate specific T-cell frequency in field studies. J the dormancy regulon of Mycobacterium tuberculosis. Microbes Infect. (2006) Immunol Methods. (2004) 291:185–95. doi: 10.1016/j.jim.2004.06.010 8:2052–60. doi: 10.1016/j.micinf.2006.03.018 34. Shey MS, Hughes EJ, de Kock M, Barnard C, Stone L, Kollmann TR, et al. 17. Masungi C, Temmerman S, Van Vooren JP, Drowart A, Pethe K, Menozzi Optimization of a whole blood intracellular cytokine assay for measuring FD, et al. Differential T and B cell responses against Mycobacterium innate cell responses to mycobacteria. J Immunol Methods. (2012) 376:79–88. tuberculosis heparin-binding hemagglutinin adhesin in infected healthy doi: 10.1016/j.jim.2011.11.011 individuals and patients with tuberculosis. J Infect Dis. (2002) 185:513–20. 35. C. L. phenotypicForest 0.2. Available online at: http://chrisladroue.com/ doi: 10.1086/338833 phorest/ (accessed October 13, 2014). 18. Hougardy JM, Schepers K, Place S, Drowart A, Lechevin V, Verscheure 36. van Unen V, Hollt T, Pezzotti N, Li N, Reinders MJT, Eisemann E, V, et al. Heparin-binding-hemagglutinin-induced IFN-gamma release as et al. Visual analysis of mass cytometry data by hierarchical stochastic a diagnostic tool for latent tuberculosis. PLoS ONE. (2007) 2:e926. neighbour embedding reveals rare cell types. Nat Commun. (2017) 8:1740. doi: 10.1371/journal.pone.0000926 doi: 10.1038/s41467-017-01689-9 19. Serra-Vidal MM, Latorre I, Franken KL, Diaz J, de Souza-Galvao ML, Casas I, 37. Rodo MJ, Rozot V, Nemes E, Dintwe O, Hatherill M, Little F, et al. et al. Immunogenicity of 60 novel latency-related antigens of Mycobacterium A comparison of antigen-specific T cell responses induced by six tuberculosis. Front Microbiol. (2014) 5:517. doi: 10.3389/fmicb.2014.00517 novel tuberculosis vaccine candidates. PLoS Pathog. (2019) 15:e1007643. 20. Meier NR, Jacobsen M, Ottenhoff THM, Ritz N. A systematic review on novel doi: 10.1371/journal.ppat.1007643 Mycobacterium tuberculosis antigens and their discriminatory potential for 38. de Jong R, Altare F, Haagen IA, Elferink DG, Boer T, van Breda the diagnosis of latent and active Tuberculosis. Front Immunol. (2018) 9:2476. Vriesman PJ, et al. Severe mycobacterial and Salmonella infections doi: 10.3389/fimmu.2018.02476 in interleukin-12 receptor-deficient patients. Science. (1998) 280:1435–8. 21. Commandeur S, Lin MY, van Meijgaarden KE, Friggen AH, Franken KL, doi: 10.1126/science.280.5368.1435 Drijfhout JW, et al. Double- and monofunctional CD4(+) and CD8(+) T- 39. Smith NL, Denning DW. Clinical implications of interferon-gamma cell responses to Mycobacterium tuberculosis DosR antigens and peptides in genetic and epigenetic variants. Immunology. (2014) 143:499–511. long-term latently infected individuals. Eur J Immunol. (2011) 41:2925–36. doi: 10.1111/imm.12362 doi: 10.1002/eji.201141602 40. Bhatt K, Verma S, Ellner JJ, Salgame P. Quest for correlates of 22. Kassa D, Ran L, Geberemeskel W, Tebeje M, Alemu A, Selase A, et al. protection against tuberculosis. Clin Vaccine Immunol. (2015) 22:258–66. Analysis of immune responses against a wide range of Mycobacterium doi: 10.1128/CVI.00721-14 tuberculosis antigens in patients with active pulmonary tuberculosis. Clin 41. Sakai S, Kauffman KD, Sallin MA, Sharpe AH, Young HA, Ganusov VV, Vaccine Immunol. (2012) 19:1907–15. doi: 10.1128/CVI.00482-12 et al. CD4 T Cell-derived IFN-gamma plays a minimal role in control 23. Pathakumari B, Prabhavathi M, Raja A. Evaluation of cytokine and of pulmonary Mycobacterium tuberculosis infection and must be actively chemokine response elicited by Rv2204c and Rv0753c to detect latent repressed by PD-1 to prevent lethal disease. PLoS Pathog. (2016) 12:e1005667. tuberculosis infection. Cytokine. (2015) 76:496–504. doi: 10.1016/j.cyto.2015. doi: 10.1371/journal.ppat.1005667 07.028 42. Clifford V, Tebruegge M, Zufferey C, Germano S, Forbes B, Cosentino L, 24. Belay M, Legesse M, Mihret A, Bekele Y, Ottenhoff TH, Franken KL, et al. et al. Cytokine biomarkers for the diagnosis of tuberculosis infection and Pro- and anti-inflammatory cytokines against Rv2031 are elevated during disease in adults in a low prevalence setting. Tuberculosis. (2019) 114:91–102. latent tuberculosis: a study in cohorts of tuberculosis patients, household doi: 10.1016/j.tube.2018.08.011 contacts and community controls in an endemic setting. PLoS ONE. (2015) 43. Coulter F, Parrish A, Manning D, Kampmann B, Mendy J, Garand M, et al. 10:e0124134. doi: 10.1371/journal.pone.0124134 IL-17 production from T helper 17, mucosal-associated invariant T, and 25. Arroyo L, Rojas M, Franken KL, Ottenhoff TH, Barrera LF. Multifunctional gammadelta cells in tuberculosis infection and disease. Front Immunol. (2017) T cell response to DosR and Rpf antigens is associated with protection in 8:1252. doi: 10.3389/fimmu.2017.01252 long-term Mycobacterium tuberculosis-infected individuals in colombia. Clin 44. Shen H, Chen ZW. The crucial roles of Th17-related cytokines/signal Vaccine Immunol. (2016) 23:813–24. doi: 10.1128/CVI.00217-16 pathways in M. tuberculosis infection. Cell Mol Immunol. (2018) 15:216–25. 26. Lewinsohn DA, Lewinsohn DM, Scriba TJ. Polyfunctional CD4(+) T cells doi: 10.1038/cmi.2017.128 as targets for tuberculosis vaccination. Front Immunol. (2017) 8:1262. 45. Villar-Hernandez R, Latorre I, De Souza-Galvao ML, Jimenez MA, Ruiz- doi: 10.3389/fimmu.2017.01262 Manzano J, Pilarte J, et al. Use of IP-10 detection in dried plasma spots for 27. Roy Chowdhury R, Vallania F, Yang Q, Lopez Angel CJ, Darboe F, latent tuberculosis infection diagnosis in contacts via mail. Sci Rep. (2019) Penn-Nicholson A, et al. A multi-cohort study of the immune factors 9:3943. doi: 10.1038/s41598-019-40778-1 Frontiers in Immunology | www.frontiersin.org 14 February 2020 | Volume 11 | Article 103 Coppola et al. Cell-Mediated Responses to Mtb Antigens 46. Joosten SA, van Meijgaarden KE, Arend SM, Prins C, Oftung F, Korsvold 60. Al-Attiyah R, Mustafa AS, Abal AT, El-Shamy AS, Dalemans W, Skeiky YA. In GE, et al. Mycobacterial growth inhibition is associated with trained innate vitro cellular immune responses to complex and newly defined recombinant immunity. J Clin Invest. (2018) 128:1837–51. doi: 10.1172/JCI97508 antigens of Mycobacterium tuberculosis. Clin Exp Immunol. (2004) 138:139– 47. Robinson RT. T cell production of GM-CSF protects the host 44. doi: 10.1111/j.1365-2249.2004.02609.x during experimental tuberculosis. mBio. (2017) 8:e02087-17. 61. Mootoo A, Stylianou E, Arias MA, Reljic R. TNF-alpha in tuberculosis: doi: 10.1128/mBio.02087-17 a cytokine with a split personality. Inflamm Allergy Drug Targets. (2009) 48. Tebruegge M, Ritz N, Donath S, Dutta B, Forbes B, Clifford V, et al. 8:53–62. doi: 10.2174/187152809787582543 Mycobacteria-specific mono- and polyfunctional CD4+ T cell profiles in 62. Roca FJ, Ramakrishnan L. TNF dually mediates resistance and susceptibility to children with latent and active tuberculosis: a prospective proof-of-concept mycobacteria via mitochondrial reactive oxygen species. Cell. (2013) 153:521– study. Front Immunol. (2019) 10:431. doi: 10.3389/fimmu.2019.00431 34. doi: 10.1016/j.cell.2013.03.022 49. Wang S, Li Y, Shen Y, Wu J, Gao Y, Zhang S, et al. Screening and 63. World Health Organization. WHO Preferred Product Characteristics for New identification of a six-cytokine biosignature for detecting TB infection Tuberculosis Vaccine. World Health Organization (2018). and discriminating active from latent TB. J Transl Med. (2018) 16:206. 64. Dreesman A, Corbiere V, Dirix V, Smits K, Debulpaep S, De Schutter I, doi: 10.1186/s12967-018-1572-x et al. Age-stratified T cell responses in children infected with Mycobacterium 50. Balcells ME, Ruiz-Tagle C, Tiznado C, Garcia P, Naves R. Diagnostic tuberculosis. Front Immunol. (2017) 8:1059. doi: 10.3389/fimmu.2017.01059 performance of GM-CSF and IL-2 in response to long-term specific- 65. Belge KU, Dayyani F, Horelt A, Siedlar M, Frankenberger M, antigen cell stimulation in patients with active and latent tuberculosis Frankenberger B, et al. The proinflammatory CD14+CD16+DR++ infection. Tuberculosis. (2018) 112:110–9. doi: 10.1016/j.tube.2018. monocytes are a major source of TNF. J Immunol. (2002) 168:3536–42. 08.006 doi: 10.4049/jimmunol.168.7.3536 51. Latorre I, Diaz J, Mialdea I, Serra-Vidal M, Altet N, Prat C, et al. IP-10 is an 66. Bruns H, Meinken C, Schauenberg P, Harter G, Kern P, Modlin RL, et al. Anti- accurate biomarker for the diagnosis of tuberculosis in children. J Infection. TNF immunotherapy reduces CD8+ T cell-mediated antimicrobial activity (2014) 69:590–9. doi: 10.1016/j.jinf.2014.06.013 against Mycobacterium tuberculosis in humans. J Clin Invest. (2009) 119:1167– 52. Harari A, Rozot V, Bellutti Enders F, Perreau M, Stalder JM, Nicod LP, et al. 77. doi: 10.1172/JCI38482 Dominant TNF-alpha+ Mycobacterium tuberculosis-specific CD4+ T cell 67. Rozot V, Vigano S, Mazza-Stalder J, Idrizi E, Day CL, Perreau M, et al. responses discriminate between latent infection and active disease. Nat Med. Mycobacterium tuberculosis-specific CD8+ T cells are functionally and (2011) 17:372–6. doi: 10.1038/nm.2299 phenotypically different between latent infection and active disease. Eur J 53. O’Garra A, Redford PS, McNab FW, Bloom CI, Wilkinson RJ, Berry MP. The Immunol. (2013) 43:1568–77. doi: 10.1002/eji.201243262 immune response in tuberculosis. Annu Rev Immunol. (2013) 31:475–527. 68. Diedrich CR, Gideon HP, Rutledge T, Baranowski TM, Maiello P, Myers doi: 10.1146/annurev-immunol-032712-095939 AJ, et al. CD4CD8 double positive T cell responses during Mycobacterium 54. Weiner J III, Kaufmann SH. Recent advances towards tuberculosis tuberculosis infection in cynomolgus macaques. J Med Primatol. (2019) 48:82– control: vaccines and biomarkers. J Internal Med. (2014) 275:467–80. 9. doi: 10.1111/jmp.12399 doi: 10.1111/joim.12212 69. Spitzer MH, Nolan GP. Mass cytometry: single cells, many features. Cell. 55. Aagaard C, Hoang T, Dietrich J, Cardona PJ, Izzo A, Dolganov G, et al. A (2016) 65:780–91. doi: 10.1016/j.cell.2016.04.019 multistage tuberculosis vaccine that confers efficient protection before and 70. Lin PL, Flynn JL. Understanding latent tuberculosis: a moving target. J after exposure. Nat Med. (2011) 17:189–94. doi: 10.1038/nm.2285 Immunol. (2010) 185:15–22. doi: 10.4049/jimmunol.0903856 56. Geluk A, van den Eeden SJ, van Meijgaarden KE, Dijkman K, Franken 71. Cadena AM, Fortune SM, Flynn JL. Heterogeneity in tuberculosis. Nat Rev KL, Ottenhoff TH. A multistage-polyepitope vaccine protects against Immunol. (2017) 17:691–702. doi: 10.1038/nri.2017.69 Mycobacterium tuberculosis infection in HLA-DR3 transgenic mice. Vaccine. (2012) 30:7513–21. doi: 10.1016/j.vaccine.2012.10.045 Conflict of Interest: The authors declare that the research was conducted in the 57. Hansen SG, Zak DE, Xu G, Ford JC, Marshall EE, Malouli D, et al. Prevention absence of any commercial or financial relationships that could be construed as a of tuberculosis in rhesus macaques by a cytomegalovirus-based vaccine. Nat potential conflict of interest. Med. (2018) 24:130–43. doi: 10.1038/nm.4473 58. Dillon DC, Alderson MR, Day CH, Lewinsohn DM, Coler R, Bement T, et al. Copyright © 2020 Coppola, Villar-Hernández, van Meijgaarden, Latorre, Muriel Molecular characterization and human T-cell responses to a member of a Moreno, Garcia-Garcia, Franken, Prat, Stojanovic, De Souza Galvão, Millet, Sabriá, novel Mycobacterium tuberculosis mtb39 gene family. Infection Immunity. Sánchez-Montalva, Noguera-Julian, Geluk, Domínguez and Ottenhoff. This is an (1999) 67:2941–50. doi: 10.1128/IAI.67.6.2941-2950.1999 open-access article distributed under the terms of the Creative Commons Attribution 59. Skeiky YA, Lodes MJ, Guderian JA, Mohamath R, Bement T, Alderson License (CC BY). The use, distribution or reproduction in other forums is permitted, MR, et al. Cloning, expression, and immunological evaluation of two provided the original author(s) and the copyright owner(s) are credited and that the putative secreted serine protease antigens of Mycobacterium tuberculosis. original publication in this journal is cited, in accordance with accepted academic Infection Immunity. (1999) 67:3998–4007. doi: 10.1128/IAI.67.8.3998-4007. practice. No use, distribution or reproduction is permitted which does not comply 1999 with these terms. Frontiers in Immunology | www.frontiersin.org 15 February 2020 | Volume 11 | Article 103