phosphorylation of mitogen-activated protein kinases contributes to interferon   production in...

M A J O R A R T I C L E

Phosphorylation of Mitogen-Activated ProteinKinases Contributes to Interferon γ Productionin Response to Mycobacterium tuberculosis

Virginia Pasquinelli,1 Ana I. Rovetta,1 Ivana B. Alvarez,1 Javier O. Jurado,1 Rosa M. Musella,2 Domingo J. Palmero,2

Alejandro Malbrán,3 Buka Samten,4 Peter F. Barnes,4 and Verónica E. García1

1Departamento de Química Biológica, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Instituto de Química Biológica, CienciasExactas y Naturales, Consejo Nacional de Investigaciones Científicas y Tecnológicas, 2División de Neumotisiología, Hospital F. J. Muñiz, and 3Unidadde Alergia, Asma e Inmunología Clínica, Hospital Británico de Buenos Aires, Argentina; and 4Center for Pulmonary and Infectious Disease Control, TheUniversity of Texas Health Science Center, Tyler

Immune control of Mycobacterium tuberculosis depends on interferon γ (IFN-γ)–producing CD4+ lympho-cytes. Previous studies have shown that T cells from patients with tuberculosis produce less IFN-γ, comparedwith healthy donors, in response to mycobacterial antigens, although IFN-γ responses to mitogens are pre-served. In this work, we found that M. tuberculosis–induced IFN-γ production by human T cells correlatedwith phosphorylation of the mitogen-activated protein kinases (MAPKs), extracellular signal-regulated kinase(ERK), and p38. Moreover, the majority of IFN-γ–producing T cells expressed signaling lymphocyte activa-tion molecule (SLAM), and SLAM activation further increased ERK phosphorylation. Interestingly, patientswith tuberculosis had delayed activation of ERK and p38, and this was most marked in patients with thepoorest IFN-γ responses (ie, low responders). Besides, SLAM signaling failed to phosphorylate ERK in lowresponders. Our findings suggest that activation of p38 and ERK, in part through SLAM, mediates T-cellIFN-γ production in response to M. tuberculosis, a pathway that is defective in patients with tuberculosis.

Keywords. Tuberculosis; signaling; CREB; MAPK; IFN-γ.

Tuberculosis remains a substantial global healthproblem, its resurgence fueled by a large population ofimmunocompromised human immunodeficiency virus(HIV)–infected persons and by an alarming increasein the prevalence of drug-resistant organisms. BCGvaccine, the only available vaccine against tuberculosis,is of variable efficacy, and development of a more ef-fective vaccine hinges on a better understanding of thehuman immune response to this pathogen. During

infection with Mycobacterium tuberculosis, innatemechanisms help to control bacillary spread, butT-lymphocyte recruitment to the lung is required tocontain the infection in granulomas [1]. Protectiveimmunity requires the generation of T-helper 1 (Th1)cytokine responses [2] and interferon γ (IFN-γ), whichactivates macrophages to inhibit mycobacterial growth[1]. Persons with mutations linked to IFN-γ signalinghave increased susceptibility to mycobacterial infectionand disseminated infection after BCG vaccination[3, 4]. Besides, T cells from patients with tuberculosisproduce less IFN-γ than those from persons with latentM. tuberculosis infection [5], and IFN-γ production islowest in patients with the most severe manifestation oftuberculosis [6].

Thus, elucidation of the mechanisms for reducedIFN-γ production in individuals who developtuberculosis will enhance our knowledge of diseasepathogenesis, contributing to a better understandingof the immune response to this intracellular pathogen.

Received 6 June 2012; accepted 21 August 2012; electronically published 2November 2012.

aV. P. and A. I. R. contributed equally to the study.bP. F. B. and V. E. G. share senior authorship for this article.Correspondence: Verónica E. García, PhD, Departamento de Química Biológica,

Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Inten-dente Guiraldes 2160, Pabellón II, piso 4, Capital Federal, 1428, Buenos Aires,Argentina (vgarcia@qb.fcen.uba.ar).

The Journal of Infectious Diseases 2013;207:340–350© The Author 2012. Published by Oxford University Press on behalf of the InfectiousDiseases Society of America. All rights reserved. For Permissions, please e-mail:journals.permissions@oup.com.DOI: 10.1093/infdis/jis672

340 • JID 2013:207 (15 January) • Pasquinelli et al

at Serials Section Norris M

edical Library on A

pril 3, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

To understand these mechanisms, it is important to delineatehow T-helper precursor cells activate the gene encoding IFN-γand become committed to the Th1 phenotype [7].

Mitogen-activated protein kinases (MAPKs) are involved inmany aspects of immune responses, including initiation ofinnate immunity, activation of adaptive immunity, and termi-nation of immune responses through cell death and regulatoryT cells [8, 9]. MAPKs are essential for macrophage activation,and the extent of MAPK phosphorylation in human monocyte-derived macrophages controls growth of intracellular Myco-bacterium avium [10]. Moreover, MAPKs phosphorylateand activate downstream molecules, resulting in T-cell activa-tion, proliferation, and differentiation into T-helper pheno-types [8, 11, 12]. Mycobacterium-induced production ofproinflammatory cytokines, such as TNF-α and interleukin 1,depends on MAPK activation [10, 13, 14]. However, althoughthere is compelling evidence for the role of extracellularsignal-regulated kinase (ERK) and p38 protein kinases in theantimycobacterial activity of antigen-presenting cells, little isknown about the contribution of these kinases to humanT-lymphocyte responses to mycobacteria.

Regulation of the IFN-γ gene is complex, involving multipletranscription factors that bind to the proximal and distal pro-moter elements [15–17], including members of the cyclicadenosine monophosphate response element binding protein(CREB)/activating transcription factor family. We previouslydemonstrated that CREB increased IFN-γ secretion to M.tuberculosis by binding to the IFN-γ proximal promoter [5].We also showed that activation of signaling lymphocyte activa-tion molecule (SLAM) in T cells from patients with tuberculosisincreased CREB phosphorylation and IFN-γ production [18].CREB is regulated by several MAPKs [19], but it is unknownwhether these MAPKs contribute to the signaling pathwaysthat enhance transcription of IFN-γ by human T cells inresponse to M. tuberculosis. In this report, we found thatM. tuberculosis–induced IFN-γ production correlated withphosphorylation of the MAPKs ERK and p38. The majority ofIFN-γ–producing T cells expressed SLAM, and SLAM activa-tion further increased ERK phosphorylation. Patients with tu-berculosis had delayed activation of ERK and p38, and SLAMsignaling failed to phosphorylate ERK in patients whoseT cells produced little M. tuberculosis–induced IFN-γ. There-fore, our findings suggest that activation of p38 and ERK me-diates T-cell IFN-γ production in response to M. tuberculosisand that this pathway is defective in patients with tuberculosis.

METHODS

PatientsWe studied HIV-seronegative patients with tuberculosis diag-nosed at the Hospital Muñiz (Buenos Aires, Argentina) on thebasis of clinical and radiological data, identification of acid-

fast bacilli in sputum, and isolation of M. tuberculosis inculture. All patients had received antituberculosis therapy for<1 week and were classified on the basis of lymphocyte re-sponses to M. tuberculosis antigen, as previously reported[20]. High responders were defined as patients who showedsignificant lymphocyte proliferation, a high level of IFN-γ pro-duction, and a high level of SLAM expression in response tothe antigen. Low responders were defined as patients who ex-hibited low proliferative responses, a low level of IFN-γ secre-tion, and a reduced level of SLAM expression [20]. We alsostudied healthy adults with no history of tuberculosis who re-ceived BCG vaccine at birth and had negative results of Quan-tiferon tuberculosis Gold tests, indicating the absence of latentM. tuberculosis infection. Two subjects with X-linked lympho-proliferative disease (XLP), confirmed at the InternationalXLP Registry Headquarters, University of Nebraska MedicalCenter (Omaha, NE) [20, 21], were also studied. Both patientsdisplay defective expression of the SLAM-associated protein.All participants provided written informed consent for samplecollection and subsequent analysis.

AntigenCells were stimulated in vitro with a cell lysate from the viru-lent M. tuberculosis H37Rv strain (NR-14822), prepared byprobe sonication (M. tuberculosis-Ag), and obtained throughthe Biodefense and Emerging Infections Research ResourcesRepository, housed within the National Institute of Allergyand Infectious Diseases of the National Institute of Health.

Cell Preparation and ReagentsPeripheral blood mononuclear cells (PBMCs) were isolated bycentrifugation over Ficoll-Hypaque (Amersham Biosciences)and cultured (1 × 106 cells/mL), with or without M. tuberculosis-Ag (10 μg/mL) with Roswell Park Memorial Institute 1640medium (Gibco) supplemented with L-glutamine, gentamicin,and 10% human serum (Sigma-Aldrich). CD3+ T cells were pu-rified by immunomagnetic negative selection, with a purity of95%–98%, as confirmed by flow cytometry.

In some experiments, cells were cultured with or withoutERK (PD98059, 50 μM, Calbiochem) or p38 (SB220025, 10μM, Calbiochem) inhibitors for 1 hour and incubated withM. tuberculosis-Ag for 48 hours, after which IFN-γ productionwas determined by an enzyme-linked immunosorbent assay(eBioscience). PD98059 specifically inhibits activation of ERKbut does not affect activation of other related dual-specificityprotein kinases or that of 18 other serine/threonine proteinkinases. SB220025 is a specific inhibitor of human p38MAPK, which binds to an extended pocket in the active site ofthe enzyme. It has a selectivity for p38 MAPK that is 2000-fold greater than that for ERK and 500-fold greater than thatfor protein kinase A. In some experiments, agonistic anti-SLAM monoclonal antibody (10 μg/mL, A12, eBioscience)

Role of MAPKs in Patients With Tuberculosis • JID 2013:207 (15 January) • 341

at Serials Section Norris M

edical Library on A

pril 3, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

was used as described below (ie, in the Western blotsubsection).

Flow CytometryPBMCs were incubated with or without ERK or p38 inhibitorsfor 1 hour. They were then stimulated with M. tuberculosis-Agfor 5 days, with Golgi Stop reagent containing monensin(1 μL/mL, BD Biosciences) added for the final 5 hours ofculture. Cells were subsequently stained with antibodies toCD3 and SLAM (both from BD Biosciences), and intracellularstaining was performed to detect IFN-γ and phospho-CREB(p-CREB, BD Biosciences), as previously described [18]. Neg-ative control samples were incubated with irrelevant isotype-matched monoclonal antibodies, and all samples were ana-lyzed on a FACS ARIA II flow cytometer (BD Biosciences).

Western BlotTotal cell protein extracts were prepared from PBMCs stimu-lated with M. tuberculosis-Ag for 24 or 48 hours or from puri-fied CD3+ lymphocytes obtained from M. tuberculosis-Ag–stimulated PBMCs after 48 hours.

In different experiments, PBMCs were stimulated withM. tuberculosis-Ag for 5 days. Afterwards, the cells were har-vested, rested for 5 hours in low (1%) serum medium, andrestimulated with anti-SLAM agonistic antibody for differentperiods. Western blotting was performed by standardmethods. Each nitrocellulose membrane was blotted, stripped,and reblotted with rabbit monoclonal antibodies to phospho-ERK (p-ERK, Thr-202/Tyr-204), total ERK, phospho-p38 (p-p38, Thr-180/Tyr-182), and total p38 (all from Cell Signaling).Bound antibodies were revealed with horseradish peroxidase–conjugated anti-rabbit antibody (1:3000; Bio-Rad, Hercules,CA), using ECL PLUS (Amersham Biosciences). Images wereobtained with an Intelligent Dark Box (Fujifilm LAS1000) andanalyzed with ImageJ Analysis software. The intensity of eachband was expressed in arbitrary units.

Transfection of SLAM Small-Interfering RNA (siRNA)SLAM siRNA was synthesized as previously described [18].Briefly, the DNA template for transcription was preparedby published methods [22]: forward T7-tagged primer, 5′-GCGTAATACGACTCACTATAGGGAGAGAGCATGAACAAAAGCATCCAC-3′; reverse T7-tagged primer, 5′-GCGTAATACGACTCACTATAGGGAGACTTCTGAAGAGGACCTGGT

Figure 1. A, Peripheral blood mononuclear cells (PBMCs) from 15 pa-tients with tuberculosis (TB) and 16 healthy donors (HDs) were incubatedwith or without inhibitors of extracellular signal-regulated kinase(PD98059 [PD]) or p38 mitogen-activated protein kinase (SB220025 [SB])for 1 hour and then stimulated with Mycobacterium tuberculosis antigen(Mtb) for 48 hours. Interferon γ (IFN-γ) levels were then measured byenzyme-linked immunosorbent assay. Values are expressed as mean ±standard error of the mean (SEM). B and C, PBMCs from 8 tuberculosisand 8 HDs were incubated with or without PD or SB for 1 hour andstimulated with Mtb for 5 days, after which IFN-γ production by CD3+

Figure 1 Continued. lymphocytes was measured by flow cytometry.B, A representative histogram is shown. C, Bars represent the mean ±SEM of the percentages of IFN-γ–producing CD3+ cells. *P < .05,**P < .01, by the Wilcoxon rank sum test; #P < .05, by the Mann–WhitneyU test.

342 • JID 2013:207 (15 January) • Pasquinelli et al

at Serials Section Norris M

edical Library on A

pril 3, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

TTC-3′ (Integrated DNA Technologies). The polymerasechain reaction product was sequenced (Northwood DNA),and DNA purification was performed using the QIAquick GelExtraction Kit Protocol (Qiagen). siRNA was generated usingthe Dicer siRNA generation kit (Genlantis). PBMCs weretransfected (GeneSilencer siRNA Transfection Reagent, Gen-lantis) with 500 ng of SLAM or control (GFP) siRNA for 48hours. Then cells were washed and incubated for 72 hours,with or without M. tuberculosis-Ag. Protein kinase expressionwas determined by Western blot, as described above.

Statistical AnalysisWe used GraphPad Prism v5.0 for statistical analysis. TheMann–Whitney U and the Wilcoxon rank sum tests wereused to analyze differences between unpaired samples andpaired samples, respectively. P values of < .05 were consideredstatistically significant.

RESULTS

ERK and p38 MAPKs Regulate IFN-γ Production AgainstM. tuberculosis Through CREBM. tuberculosis–stimulated PBMCs from patients with tuber-culosis produced less IFN-γ than those from healthy donors(P = .04; Figure 1A). Pretreatment of PBMCs with ERK(PD98059) and p38 (SB220025) inhibitors significantlyreduced M. tuberculosis–induced IFN-γ levels by both groups(Figure 1A), and the effect of PD98059 was greater in healthydonors than in patients with tuberculosis (P = .02). Pretreat-ment of PBMCs with the ERK and p38 MAPK inhibitors didnot affect cell viability, as determined by an MTT assay (datanot shown). Both kinase inhibitors also reduced the percent-ages of IFN-γ–producing T lymphocytes expanded byM. tuberculosis-Ag (Figure 1B and 1C)

On the basis of our current results (Figure 1) and our previ-ous demonstration that CREB phosphorylation contributes toM. tuberculosis–stimulated production of IFN-γ [23], we nextinvestigated whether ERK and p38 controlled CREB phos-phorylation. M. tuberculosis-Ag induced expression of bothIFN-γ and p-CREB in T lymphocytes from patients with tu-berculosis and those from healthy donors, with the latterhaving higher percentages of IFN-γ–producing p-CREB+

T cells (Figure 2). Both MAPK inhibitors significantly de-creased the number of these lymphocytes (Figure 2), suggest-ing that ERK and p38 MAPKs regulate IFN-γ productionagainst M. tuberculosis by a mechanism that involves CREB.

Mtb-Ag Induces ERK and p38 Phosphorylation in T Cells FromPatients With TuberculosisTo determine whether M. tuberculosis-Ag induces MAPK acti-vation, we measured phosphorylation of ERK and p38 byWestern blot. Although there was notable individual variation,

M. tuberculosis stimulation markedly induced ERK and p38phosphorylation, with maximal induction after 48 hours(Figure 3). Phosphorylation of ERK and p38 occur veryrapidly in response to stimulation through the T-cell receptor,but phosphorylation of these enzymes was not significantly in-creased until 24 hours after exposure to M. tuberculosis (datanot shown), perhaps because of delays related to antigen pro-cessing and presentation.

Patients with XLP have uncontrolled Th1 responses and highbasal levels of IFN-γ [20]. Therefore, to determine whether ERKand p38 MAPKs regulate IFN-γ production in response toM. tuberculosis specifically or to other stimuli, we evaluatedMAPK activation in PBMCs from 2 BCG-vaccinated patientswith XLP that were stimulated with M. tuberculosis-Ag. PBMCsfrom patients with XLP produced high concentrations of IFN-γwithout stimulation, but there was no significant MAPK phos-phorylation (Figure 4A–C). However, ERK and p38 phosphory-lation were greatly increased in Ag-stimulated cells from patientswith XLP. In addition, in patients with tuberculosis, ERK acti-vation was minimal at 24 hours and modest at 48 hours,whereas ERK phosphorylation was maximal at 24 hours inhealthy donors and individuals with XLP (Figure 4D). Similarfindings were noted for phosphorylation of p38 (data notshown). Thus, patients with tuberculosis who have diminishedIFN-γ production in response to M. tuberculosis and ineffectiveimmunity [23] have delayed kinetics of MAPK activation.

Because the experiments described above were performedwith PBMCs that included different cell types, we wished toconfirm that M. tuberculosis activates ERK and p38 in T cells.After stimulation of PBMCs with M. tuberculosis-Ag, we puri-fied CD3+ cells by negative selection and found that M. tuber-culosis stimulation augmented expression of p-ERK and p-p38(Figure 5A).

Reduced Production of IFN-γ in Response to M. tuberculosisin Patients With Active Disease is Associated With ImpairedMAPK ActivationDuring tuberculosis, low responders have reduced IFN-γproduction and more severe disease, compared with high re-sponders [20]. Mean M. tuberculosis–induced IFN-γ produc-tion (±SD) was 0.49 ± 0.17 ng/mL among low responders and11.95 ± 3.89 ng/mL among high responders (P = .0007). IFN-γresponse status (ie, low or high) correlated with M. tuberculo-sis–induced phosphorylation of ERK and p38, as low respond-ers had impaired activation of ERK and p38, both in PBMCsand purified T cells, compared with high responders (Figure 5Band 5C).

Costimulation Through SLAM Increases Phosphorylation ofERK but Not p38 in Patients With TuberculosisWe previously demonstrated that SLAM, a protein expressedon activated T cells, induces Th1 responses in patients with

Role of MAPKs in Patients With Tuberculosis • JID 2013:207 (15 January) • 343

at Serials Section Norris M

edical Library on A

pril 3, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

tuberculosis through a mechanism that involves CREB phos-phorylation [18, 20]. We extended these studies to show thatnearly 60% of IFN-γ–producing T cells coexpressed SLAMand p-CREB after M. tuberculosis stimulation (Figure 6A).

Unlike high responders, low responders do not upregulateSLAM expression after M. tuberculosis stimulation (Supple-mentary Figure 1), a defect that has been found to contributeto reduced IFN-γ production [20]. Engagement throughSLAM significantly increased M. tuberculosis–induced IFN-γproduction from patients with tuberculosis, and this increasewas much higher in high responders than in low responders(Supplementary Figure 1) [20]. Because low respondersdisplay impaired activation of ERK and p38 in response to M.tuberculosis (Figure 5), we investigated whether signalingthrough SLAM participated in MAPK activation during tuber-culosis. PBMCs from high responders and low responderswere stimulated with M. tuberculosis-Ag, with or without an

agonistic anti-SLAM monoclonal antibody, and ERK and p38activation were measured. Costimulation through SLAM in-creased ERK phosphorylation in high responders but not inlow responders (Figure 6B). Furthermore, SLAM siRNAreduced ERK phosphorylation (Figure 6C). In contrast, activa-tion of SLAM did not increase p38 expression, either in highresponders or low responders (Figure 6B), and SLAM siRNAdid not affect p38 phosphorylation (Figure 6C).

DISCUSSION

Because MAPK pathways regulate Th1 development [24], wehypothesized that certain MAPKs could mediate M. tuberculo-sis–induced IFN-γ production. We found that stimulation ofT cells from healthy donors and patients with tuberculosis withM. tuberculosis activates ERK and p38 MAPK and that inhibi-tion of these 2 signaling molecules reduces the number of

Figure 2. Peripheral blood mononuclear cells (PBMCs) from 9 healthy donors (HDs) and 7 high responders (tuberculosis; defined in Methods) wereincubated with or without extracellular signal-regulated kinase (PD98059 [PD]) or p38 mitogen-activated protein kinase (SB220025 [SB]) inhibitors for 1hour. Cells were then stimulated with Mycobacterium tuberculosis antigen for 5 days, and cyclic adenosine monophosphate response element bindingprotein phosphorylation (p-CREB) and interferon γ (IFN-γ) production were analyzed in CD3+ cells by flow cytometry. A, Results for a representative HDare shown. IgG1, immunoglobulin G1. B, The bars show the mean ± standard error of the mean. The percentages were calculated by first gating onlymphocytes by light scatter and then gating on CD3+ T cells. *P < .05, **P < .01, by the Wilcoxon rank sum test.

344 • JID 2013:207 (15 January) • Pasquinelli et al

at Serials Section Norris M

edical Library on A

pril 3, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

IFN-γ–producing T cells, as well as total T-cell production ofIFN-γ. On exposure to mycobacterial antigens, T cells from pa-tients with tuberculosis showed delayed activation of ERK andp38, and patients whose T cells produced the lowest IFN-γ levelshad the greatest reductions in phosphorylation of these signalingmolecules. Inhibition of ERK and p38 MAPKs reduced thenumber of IFN-γ–producing p-CREB+ T cells. Costimulationthrough SLAM increased IFN-γ production and ERK phosphor-ylation by high responders, and inhibition with SLAM siRNAabrogated these effects. The sum of these data demonstrates thatERK and p38 MAPK pathways contribute to M. tuberculosis–induced IFN-γ production by primary human T cells throughmechanisms that increase phosphorylation of CREB. In addi-tion, SLAM contributes to ERK-mediated IFN-γ production.

We found that human T cells phosphorylate p38 MAPK andERK when stimulated with M. tuberculosis antigens and that in-hibition of these kinases reduces M. tuberculosis–induced IFN-γproduction (Figure 1). Patients with tuberculosis, whose T cellsproduce less IFN-γ than those of healthy donors, displayeddelayed activation of ERK and p38 (Figure 4D), with the mostmarked changes in low responders (Figure 5B and 5C), withpoor immunity and the most severe disease [20, 25]. These find-ings suggest that ERK and p38 contribute to IFN-γ productionin individuals with M. tuberculosis infection and that the activityof these pathways correlates with the clinical manifestations of

infection. Chemical inhibition of both ERK and p38 reducedM. tuberculosis–stimulated IFN-γ production to a greater extentthan either inhibitor alone but did not completely abolish IFN-γproduction (V. Pasquinelli, unpublished data), indicating thatother signaling molecules also control IFN-γ production in re-sponse to M. tuberculosis. Phosphorylation of p38 and ERKwere not observed in patients with XLP whose PBMCs produceIFN-γ without antigen stimulation (Figure 4C), indicating thatthese kinases do not mediate all mechanisms of IFN-γ produc-tion. ERK and p38 also contribute to IFN-γ production in re-sponse to polyclonal activation of T cells with antibodies toCD3 and CD28 (V. Pasquinelli, unpublished data), showing thatstimuli other than M. tuberculosis also act through these kinases.

The role of p38 MAPK in T-cell production of IFN-γdepends on the mode of T-cell stimulation and may differ inmice and humans. Elegant studies in mice with a p38 MAPKgene deletion in T cells showed that p38 is not required forIFN-γ production in response to antigen but is central for elic-iting IFN-γ secretion in response to interleukin 12 (IL-12) andinterleukin 18 [26]. p38 MAPK phosphorylates STAT4, a tran-scription factor through which IL-12 upregulates IFN-γ tran-scription in T cells [27, 28]. In humans, some investigatorsfound that T-cell receptor (TCR) stimulation induces p38MAPK phosphorylation and IFN-γ production by antigen-experienced T cells [29], whereas others reported that p38MAPK does not affect T cells activated through the TCR [30].In the current report, we found that primary human T cellsphosphorylate p38 MAPK when stimulated with M. tuber-culosis antigens and that inhibition of p38 MAPK reducesM. tuberculosis–induced IFN-γ production. Mycobacteriaelicit IL-12 production by mononuclear phagocytes, andT-cell IFN-γ production in response to M. tuberculosis antigensis inhibited by anti–IL-12 [31]. In combination with prior pub-lications, our current results suggest that M. tuberculosis stimu-lates IFN-γ production by stimulation of antigen-experiencedT cells through the TCR and through IL-12, with at least thelatter process dependent on p38 MAPK. The signaling path-ways that control IFN-γ production are likely to depend on thepathogen or other T-cell stimulus. For example, ERK but notp38 MAPK plays a dominant role in IL-12–mediated restora-tion of T-cell IFN-γ production following burn injury [32].

Most studies of p38 MAPK signaling in the setting of my-cobacterial infection have evaluated the effects on mononucle-ar phagocytes. Accordingly, MAPKs contribute to productionof TNF-α and interleukin 10 by macrophages in response toM. tuberculosis [33–37], and M. tuberculosis delays phagoso-mal maturation by reducing recruitment of early endosomeautoantigen 1 to the phagosomal membrane through p38MAPK [38] and inhibits release of TNF-α by macrophagesduring murine infection with M. tuberculosis [39]. Fur-thermore, pathogenic M. avium complex strains reduce p38MAPK activity in macrophages, whereas less virulent

Figure 3. Peripheral blood mononuclear cells (PBMCs) from 8 high re-sponders (defined in Methods) were cultured, with or without Mycobac-terium tuberculosis antigen for 1–5 days. Phosphorylated and totalextracellular signal-regulated kinase (p-ERK and tERK, respectively) andp38 (p-p38 and tp38, respectively) expression were then measuredby Western blot. A, Results from a representative patient are shown.B, Densitometry of the images was performed, and the ratios ofphosphorylated to total protein expression were expressed as arbitraryunits. *P < .05, **P < .01, by the Wilcoxon rank sum test.

Role of MAPKs in Patients With Tuberculosis • JID 2013:207 (15 January) • 345

at Serials Section Norris M

edical Library on A

pril 3, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

mycobacteria do not [13, 35], suggesting that inhibition of p38MAPK activity constitutes a mycobacterial virulence mecha-nism. Early secreted antigenic target of 6 kDa (ESAT-6) is asecreted protein of M. tuberculosis that confers virulence inanimal models [40, 41]. We recently found that high concen-trations of ESAT-6 inhibit T-cell IFN-γ production by activat-ing p38 MAPK [42], contrasting with our current findings, inwhich ERK and p38 pathways enhanced M. tuberculosis–

induced IFN-γ production by T cells from healthy donors.M. tuberculosis produces many proteins, lipids, and lipogly-cans, which have differential effects on cell signaling pathwaysand on the human immune response. We speculate that myco-bacterial components other than ESAT-6 induce IFN-γ pro-duction through p38 MAPK pathways and that these factorsdominate when T cells are exposed to M. tuberculosis celllysates used in the current study, which contain limited

Figure 4. Peripheral blood mononuclear cells (PBMCs) from 2 patients with X-linked lymphoproliferative disease (XLPs; from 3 independent expe-riments) and 7 healthy donors (HDs) were cultured, with or without M. tuberculosis Ag, for 48 hours. Phosphorylated and total extracellular signal-regulated kinase (p-ERK and tERK, respectively) and p38 (p-p38 and tp38, respectively) expression were then measured by Western blot.A, Representative blots are shown. B, Densitometry was performed, and the ratios of phosphorylated to total protein were expressed as arbitrary units.C, Interferon γ (IFN-γ) levels were determined by enzyme-linked immunosorbent assay. *P < .05, by the Wilcoxon rank sum test, comparing values formedia and M. tuberculosis. D, PBMCs from 8 high responder patients with tuberculosis (defined in Methods), 7 HDs, and 2 XLPs (3 independentexperiments) were cultured with or without M. tuberculosis-Ag for 24 or 48 hours. p-ERK and tERK expression was measured by Western blot. Densi-tometry was performed, and the ratios of p-ERK to tERK expression were expressed as arbitrary units. The fold increase (relative to media) is shown.#P < .05, ##P < .01, by the Mann–Whitney U test.

346 • JID 2013:207 (15 January) • Pasquinelli et al

at Serials Section Norris M

edical Library on A

pril 3, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

amounts of ESAT-6, as this protein is secreted and is presentprimarily in the culture filtrate. Further studies are needed todetermine whether p38 MAPK–dependent IFN-γ productionoccurs when T cells encounter live M. tuberculosis.

CREB is a transcription factor that controls genes thatmediate immune responses, including inhibition of nuclearfactor κB activation, induction of macrophage survival, andpromotion of proliferation, survival, and regulation of lympho-cytes [43]. We previously demonstrated that CREB binds to theIFN-γ promoter and upregulates the production of IFN-γinduced by M. tuberculosis [23]. Protein kinase A is the classicactivator of CREB, but other signaling pathways can also phos-phorylate CREB in T cells, including protein kinase C, Ras,ERK1/2, and MAPKs [44–46]. Two of the substrates of p38

MAPK are mitogen- and stress-activated kinases 1 and 2, bothof which can phosphorylate CREB in T cells [47]. In line withthese reports, we found that inhibition of the ERK and p38MAPK pathways reduced the number of T cells expressing p-CREB and IFN-γ in response to M. tuberculosis (Figure 2).These findings suggest that ERK and p38 contribute to CREBphosphorylation and IFN-γ production, but it is uncertainwhether this is mediated directly through these kinases or indi-rectly through other linked cell-signaling pathways.

Costimulation lowers the threshold for TCR signaling butalso increases the probability that activated cells will undergoproliferation and subsequent differentiation [48]. We have pre-viously shown that T cells responding to M. tuberculosis anti-gens rapidly upregulate the costimulator SLAM and that these

Figure 5. A, Peripheral blood mononuclear cells (PBMCs) from 3 high responders (defined in Methods) and 3 healthy donors (HDs) were stimulatedwith Mycobacterium tuberculosis antigen for 48 hours. Cells were then harvested, and CD3+ T cells were purified by negative selection. Phosphorylatedand total extracellular signal-regulated kinase (p-ERK and tERK, respectively) and p38 (p-p38 and tp38, respectively) expression were determined byWestern blot. A representative example is shown for each kinase. B, PBMCs from 6 high responders (HRs) and 6 low responders (LRs; defined inMethods) were stimulated with M. tuberculosis for 48 hours. p-ERK, tERK, p-p38, and tp38 expression were then measured by Western blot. Densitom-etry was performed, and the ratios of phosphorylated to total protein expression were expressed as arbitrary units. *P < .05, by the Wilcoxon rank sumtest; #P < .05, by the Mann–Whitney U test. C, PBMCs from 3 HRs and 3 LRs; were stimulated with M. tuberculosis for 48 hours. Cells were thenharvested, and CD3+ T cells were purified by negative selection. p-ERK, tERK, p-p38, and tp38 expression was determined by Western blot. Onerepresentative result is shown.

Role of MAPKs in Patients With Tuberculosis • JID 2013:207 (15 January) • 347

at Serials Section Norris M

edical Library on A

pril 3, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

signals induce IFN-γ production [20] and lead to CREB phos-phorylation [18]. In this work, we extended these results,showing that most cells producing IFN-γ in response toM. tuberculosis expressed SLAM and p-CREB, suggesting thatSLAM activation increased IFN-γ production by a signalingpathway that involves CREB. Moreover, costimulation throughSLAM increased ERK phosphorylation and IFN-γ productionin M. tuberculosis–activated cells, and neutralization of SLAMwith siRNA prevented ERK phosphorylation, suggesting thatSLAM contributes to antigen-induced IFN-γ productionthrough the ERK pathway. The unresponsiveness of T cells oflow responders is due in part to failure to upregulate SLAM[20], and our current findings suggest that this results in im-paired ERK activation and IFN-γ production.

In summary, we found that IFN-γ production in patientswith tuberculosis is regulated by ERK and p38 MAPK signal-ing pathways that involve CREB activation. Costimulationthrough SLAM augmented phosphorylation of ERK and IFN-γproduction, adding another layer of complexity to MAPK acti-vation in T cells and cytokine production. We also showedthat patients with tuberculosis whose T cells secrete little IFN-γin response to M. tuberculosis have impaired activation ofERK and p38 MAPK. Further study of the dysfunctional acti-vation of MAPKs during tuberculosis could identify newtargets to increase IFN-γ production and improve patientoutcome, particularly in those with drug-resistant disease.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseasesonline (http://jid.oxfordjournals.org/). Supplementary materials consist ofdata provided by the author that are published to benefit the reader. Theposted materials are not copyedited. The contents of all supplementarydata are the sole responsibility of the authors. Questions or messagesregarding errors should be addressed to the author.

Notes

Acknowledgments. We thank Rodrigo E. Hernández Del Pino, DelfinaPeña, and Guadalupe I. Álvarez for helpful discussions.Financial support. This work was supported by Agencia Nacional de

Promoción Científica y Tecnológica (PAE-PID-2007-00127 and PICT 01384to V. E. G.; PICT 01769 to V. P.); University of Buenos Aires (UBACYTX087 and UBACYT 20020100100221 to V. E. G.), and the National Insti-tutes of Health, National Institutes of Allergy and Infectious Diseases (R01AI079007 to V. E. G.). A. I. R and I. B. A. are fellows of Consejo Nacionalde Producción Científica y Tecnológica (CONICET), Argentina. V. P. andV. E. G. are members of the Researcher Career of CONICET, Argentina.Potential conflicts of interest. All authors: No reported conflicts.All authors have submitted the ICMJE Form for Disclosure of Potential

Conflicts of Interest. Conflicts that the editors consider relevant to thecontent of the manuscript have been disclosed.

References

1. Giacomini E, Iona E, Ferroni L, et al. Infection of human macrophag-es and dendritic cells with Mycobacterium tuberculosis induces a dif-ferential cytokine gene expression that modulates T cell response. JImmunol 2001; 166:7033–41.

Figure 6. A, Peripheral blood mononuclear cells (PBMCs) from 8 highresponders (HRs; defined in Methods) were stimulated with Mycobacteri-um tuberculosis antigen for 5 days, and expression of signaling lympho-cyte activation molecule (SLAM), phosphorylated cyclic adenosinemonophosphate response element binding protein phosphorylation (p-CREB), and interferon γ (IFN-γ) was measured by flow cytometry. Thepercentages of p-CREB+ and SLAM+ T cells were determined by gatingon IFN-γ–producing CD3+ cells. B, PBMCs from 6 HRs and 5 lowresponders (HRs and LRs, respectively; defined in Methods) werestimulated with M. tuberculosis. After 5 days, cells were harvested,rested for 5 hours in low (1%) serum media, and restimulated with anagonistic anti-SLAM monoclonal antibody (α-SLAM) for 30 minutes.Phosphorylated and total extracellular signal-regulated kinase (p-ERK andtERK, respectively) and p38 (p-p38 and tp38, respectively) expressionwas then measured by Western blot. Densitometry was performed, andthe results were expressed as arbitrary units. The levels of phosphorylat-ed protein were normalized to those of total protein, and the fold-in-crease (relative to time 0) is shown. **P < .01, by the Mann–Whitney Utest. C, SLAM or GFP small-interfering RNA (siRNA) were incubated withPBMCs from 5 HRs for 48 hours. Cells were collected, washed, and stim-ulated with M. tuberculosis for another 72 hours. p-ERK, tERK, p-p38,tp38 expression was then measured by Western blot. Densitometry wasperformed, and the ratios of phosphorylated to total protein levels wereexpressed as arbitrary units. *P < .05, by the Wilcoxon rank sum test.

348 • JID 2013:207 (15 January) • Pasquinelli et al

at Serials Section Norris M

edical Library on A

pril 3, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

2. Cooper AM, Dalton DK, Stewart TA, Griffin JP, Russell DG, OrmeIM. Disseminated tuberculosis in interferon gamma gene-disruptedmice. J Exp Med 1993; 178:2243–7.

3. Cunningham JA, Kellner JD, Bridge PJ, Trevenen CL, McLeod DR,Davies HD. Disseminated bacille Calmette-Guerin infection in aninfant with a novel deletion in the interferon-gamma receptor gene.Int J Tuberc Lung Dis 2000; 4:791–4.

4. Newport MJ, Huxley CM, Huston S, et al. A mutation in the interferon-gamma-receptor gene and susceptibility to mycobacterial infection.N Engl J Med 1996; 335:1941–9.

5. Samten B, Ghosh P, Yi AK, et al. Reduced expression of nuclear cyclicadenosine 5’-monophosphate response element-binding proteins andIFN-gamma promoter function in disease due to an intracellular path-ogen. J Immunol 2002; 168:3520–6.

6. Sodhi A, Gong J, Silva C, Qian D, Barnes PF. Clinical correlates ofinterferon gamma production in patients with tuberculosis. ClinInfect Dis 1997; 25:617–20.

7. Ho IC, Glimcher LH. Transcription: tantalizing times for T cells. Cell2002; 109 (Suppl):S109–20.

8. Dong C, Davis RJ, Flavell RA. MAP kinases in the immune response.Annu Rev Immunol 2002; 20:55–72.

9. Adler HS, Kubsch S, Graulich E, Ludwig S, Knop J, Steinbrink K. Ac-tivation of MAP kinase p38 is critical for the cell-cycle-controlled sup-pressor function of regulatory T cells. Blood 2007; 109:4351–9.

10. Blumenthal A, Ehlers S, Ernst M, Flad H-D, Reiling N. Control ofMycobacterial Replication in Human Macrophages: Roles of Extracel-lular Signal-Regulated Kinases 1 and 2 and p38 Mitogen-ActivatedProtein Kinase Pathways. Infection and Immunity 2002; 70:4961–7.

11. Cantrell DA. T-cell antigen receptor signal transduction. Immunology2002; 105:369–74.

12. Huang Y, Wange RL. T cell receptor signaling: beyond complex com-plexes. J Biol Chem 2004; 279:28827–30.

13. Schorey JS, Cooper AM. Macrophage signalling upon mycobacterialinfection: the MAP kinases lead the way. Cellular Microbiology 2003;5:133–42.

14. Fietta AM, Morosini M, Meloni F, Bianco AM, Pozzi E. Pharmacolog-ical analysis of signal transduction pathways required for Mycobacteri-um tuberculosis-induced IL-8 and MCP-1 production in humanperipheral monocytes. Cytokine 2002; 19:242–9.

15. Shnyreva M, Weaver WM, Blanchette M, et al. Evolutionarily con-served sequence elements that positively regulate IFN-gamma expres-sion in T cells. Proc Natl Acad Sci U S A 2004; 101:12622–7.

16. Cippitelli M, Sica A, Viggiano V, et al. Negative transcriptionalregulation of the interferon-gamma promoter by glucocorticoidsand dominant negative mutants of c-Jun. J Biol Chem 1995;270:12548–56.

17. Penix LA, Sweetser MT, Weaver WM, Hoeffler JP, Kerppola TK,Wilson CB. The proximal regulatory element of the interferon-gammapromoter mediates selective expression in T cells. J Biol Chem 1996;271:31964–72.

18. Pasquinelli V, Townsend JC, Jurado JO, et al. IFN-gamma productionduring active tuberculosis is regulated by mechanisms that involve IL-17, SLAM, and CREB. J Infect Dis 2009; 199:661–5.

19. Roux PP, Blenis J. ERK and p38 MAPK-Activated Protein Kinases: aFamily of Protein Kinases with Diverse Biological Functions. Microbi-ology and Molecular Biology Reviews 2004; 68:320–344.

20. Pasquinelli V, Quiroga MF, Martinez GJ, et al. Expression of signalinglymphocytic activation molecule-associated protein interrupts IFN-gammaproduction in human tuberculosis. J Immunol 2004; 172:1177–85.

21. Malbran A, Belmonte L, Ruibal-Ares B, et al. Loss of circulating CD27+ memory B cells and CCR4+ T cells occurring in association withelevated EBV loads in XLP patients surviving primary EBV infection.Blood 2004; 103:1625–31.

22. Elbashir SM, Harborth J, Weber K, Tuschl T. Analysis of gene func-tion in somatic mammalian cells using small interfering RNAs.Methods 2002; 26:199–213.

23. Samten B, Howard ST, Weis SE, et al. Cyclic AMP response element-binding protein positively regulates production of IFN-gamma by T cellsin response to a microbial pathogen. J Immunol 2005; 174:6357–63.

24. Smith-Garvin JE, Koretzky GA, Jordan MS. T cell activation. AnnuRev Immunol 2009; 27:591–619.

25. Jurado JO, Pasquinelli V, Alvarez IB, et al. IL-17 and IFN-gamma ex-pression in lymphocytes from patients with active tuberculosis corre-lates with the severity of the disease. J Leukoc Biol 2012; 91:991–1002.

26. Berenson LS, Yang J, Sleckman BP, Murphy TL, Murphy KM. Selec-tive Requirement of p38a MAPK in Cytokine-Dependent, but NotAntigen Receptor-Dependent, Th1 Responses. The Journal of Immu-nology 2006; 176:4616–21.

27. Morinobu A, Gadina M, Strober W, et al. STAT4 serine phosphoryla-tion is critical for IL-12-induced IFN-γ production but not for cellproliferation. Proceedings of the National Academy of Sciences 2002;99:12281–6.

28. Visconti R, Gadina M, Chiariello M, et al. Importance of the MKK6/p38 pathway for interleukin-12–induced STAT4 serine phosphoryla-tion and transcriptional activity. Blood 2000; 96:1844–52.

29. Adachi K, Davis MM. T-cell receptor ligation induces distinct signal-ing pathways in naive vs. antigen-experienced T cells. Proceedings ofthe National Academy of Sciences 2010; 108:1549–54.

30. Li C, Beavis P, Palfreeman AC, Amjadi P, Kennedy A, Brennan FM.Activation of p38 mitogen-activated protein kinase is critical step foracquisition of effector function in cytokine-activated T cells, but actsas a negative regulator in T cells activated through the T-cell receptor.Immunology 2011; 132:104–10.

31. Zhang M, Gately MK, Wang E, et al. Interleukin 12 at the site ofdisease in tuberculosis. J Clin Invest 1994; 93:1733–9.

32. Li X, Chaudry IH, Choudhry MA. ERK and not p38 pathway is re-quired for IL-12 restoration of T cell IL-2 and IFN-gamma in a rodentmodel of alcohol intoxication and burn injury. J Immunol 2009;183:3955–62.

33. Rajaram MV, Ni B, Morris JD, et al. Mycobacterium tuberculosis lip-omannan blocks TNF biosynthesis by regulating macrophage MAPK-activated protein kinase 2 (MK2) and microRNA miR-125b. Proc NatlAcad Sci U S A 2011; 108:17408–13.

34. Lee HM, Shin DM, Kim KK, Lee JS, Paik TH, Jo EK. Roles of reactiveoxygen species in CXCL8 and CCL2 expression in response to the 30-kDaantigen ofMycobacterium tuberculosis. J Clin Immunol 2009; 29:46–56.

35. Yang CS, Lee JS, Song CH, et al. Protein kinase C zeta plays an essen-tial role for Mycobacterium tuberculosis-induced extracellular signal-regulated kinase 1/2 activation in monocytes/macrophages via Toll-like receptor 2. Cell Microbiol 2007; 9:382–96.

36. Mendez-Samperio P, Trejo A, Miranda E. Activation of ERK1/2 andTNF-alpha production are mediated by calcium/calmodulin, and PKAsignaling pathways during Mycobacterium bovis infection. J Infect2006; 52:147–53.

37. Nair S, Ramaswamy PA, Ghosh S, et al. The PPE18 of Mycobacteriumtuberculosis interacts with TLR2 and activates IL-10 induction in mac-rophage. J Immunol 2009; 183:6269–81.

38. Fratti RA, Chua J, Deretic V. Induction of p38 Mitogen-activatedProtein Kinase Reduces Early Endosome Autoantigen 1 (EEA1) Re-cruitment to Phagosomal Membranes. Journal of Biological Chemistry2003; 278:46961–7.

39. van den Blink B, Juffermans NP, ten Hove T, et al. p38 Mitogen-Activated Protein Kinase Inhibition Increases Cytokine Release byMacrophages In Vitro and During Infection In Vivo. The Journal ofImmunology 2001; 166:582–7.

40. Pym AS, Brodin P, Brosch R, Huerre M, Cole ST. Loss of RD1 con-tributed to the attenuation of the live tuberculosis vaccines Mycobac-terium bovis BCG and Mycobacterium microti. MolecularMicrobiology 2002; 46:709–17.

41. Lewis KN, Liao R, Guinn KM, et al. Deletion of RD1 from Mycobacte-rium tuberculosis Mimics Bacille Calmette-Guérin Attenuation.Journal of Infectious Diseases 2003; 187:117–23.

Role of MAPKs in Patients With Tuberculosis • JID 2013:207 (15 January) • 349

at Serials Section Norris M

edical Library on A

pril 3, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

42. Peng H, Wang X, Barnes PF, Tang H, Townsend JC, Samten B. TheMycobacterium tuberculosis Early Secreted Antigenic Target of 6 kDaInhibits T Cell Interferon-γ Production through the p38 Mitogen-activated Protein Kinase Pathway. Journal of Biological Chemistry2011; 286:24508–18.

43. Wen AY, Sakamoto KM, Miller LS. The role of the transcrip-tion factor CREB in immune function. J Immunol 2010; 185:6413–9.

44. Grady GC, Mason SM, Stephen J, Zuniga-Pflucker JC, Michie AM.Cyclic adenosine 5’-monophosphate response element binding proteinplays a central role in mediating proliferation and differentiationdownstream of the pre-TCR complex in developing thymocytes.J Immunol 2004; 173:1802–10.

45. Muthusamy N, Leiden JM. A protein kinase C-, Ras-, and RSK2-dependent signal transduction pathway activates the cAMP-responsiveelement-binding protein transcription factor following T cell receptorengagement. J Biol Chem 1998; 273:22841–7.

46. Hughes-Fulford M, Sugano E, Schopper T, Li CF, Boonyaratanakorn-kit JB, Cogoli A. Early immune response and regulation of IL-2 recep-tor subunits. Cell Signal 2005; 17:1111–24.

47. Kaiser M, Wiggin GR, Lightfoot K, Arthur JS, Macdonald A. MSKregulate TCR-induced CREB phosphorylation but not immediateearly gene transcription. Eur J Immunol 2007; 37:2583–95.

48. Iezzi G, Karjalainen K, Lanzavecchia A. The duration of antigenicstimulation determines the fate of naive and effector T cells. Immunity1998; 8:89–95.

350 • JID 2013:207 (15 January) • Pasquinelli et al

at Serials Section Norris M

edical Library on A

pril 3, 2014http://jid.oxfordjournals.org/

Dow

nloaded from