An Unbiased Genome-Wide Mycobacterium tuberculosis … · An Unbiased Genome-Wide Mycobacterium tuberculosis Gene Expression Approach To Discover Antigens Targeted by ... The Journal

Page 1

of July 7, 2018.This information is current as

Cells Expressed during Pulmonary InfectionDiscover Antigens Targeted by Human T

Gene Expression Approach TotuberculosisMycobacteriumAn Unbiased Genome-Wide

Geluk and Tom H. M. OttenhoffSchoolnik, Fredrik Oftung, Gro Ellen Korsvold, AnnemiekeFranken, Gregory Dolganov, Igor Kramnik, Gary K. van den Eeden, Annemieke H. Friggen, Kees L. M. C.Prins, Alexander V. Pichugin, Karin Dijkman, Susan J. F. Susanna Commandeur, Krista E. van Meijgaarden, Corine

http://www.jimmunol.org/content/190/4/1659doi: 10.4049/jimmunol.1201593January 2013;

2013; 190:1659-1671; Prepublished online 14J Immunol 

MaterialSupplementary

3.DC1http://www.jimmunol.org/content/suppl/2013/01/15/jimmunol.120159

Referenceshttp://www.jimmunol.org/content/190/4/1659.full#ref-list-1

, 26 of which you can access for free at: cites 69 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2013 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on July 7, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Page 2

The Journal of Immunology

An Unbiased Genome-Wide Mycobacterium tuberculosis GeneExpression Approach To Discover Antigens Targeted byHuman T Cells Expressed during Pulmonary Infection

Susanna Commandeur,* Krista E. van Meijgaarden,* Corine Prins,*

Alexander V. Pichugin,†,1 Karin Dijkman,* Susan J. F. van den Eeden,*

Annemieke H. Friggen,* Kees L. M. C. Franken,* Gregory Dolganov,‡ Igor Kramnik,†,2

Gary K. Schoolnik,‡ Fredrik Oftung,x Gro Ellen Korsvold,x Annemieke Geluk,*

and Tom H. M. Ottenhoff*

Mycobacterium tuberculosis is responsible for almost 2 million deaths annually.Mycobacterium bovis bacillus Calmette-Guerin, the

only vaccine available against tuberculosis (TB), induces highly variable protection against TB, and better TB vaccines are

urgently needed. A prerequisite for candidate vaccine Ags is that they are immunogenic and expressed by M. tuberculosis during

infection of the primary target organ, that is, the lungs of susceptible individuals. In search of new TB vaccine candidate Ags, we

have used a genome-wide, unbiased Ag discovery approach to investigate the in vivo expression of 2170 M. tuberculosis genes

duringM. tuberculosis infection in the lungs of mice. Four genetically related but distinct mouse strains were studied, representing

a spectrum of TB susceptibility controlled by the supersusceptibility to TB 1 locus. We used stringent selection approaches to

select in vivo–expressedM. tuberculosis (IVE-TB) genes and analyzed their expression patterns in distinct disease phenotypes such

as necrosis and granuloma formation. To study the vaccine potential of these proteins, we analyzed their immunogenicity. Several

M. tuberculosis proteins were recognized by immune cells from tuberculin skin test-positive, ESAT6/CFP10-responsive individ-

uals, indicating that these Ags are presented during natural M. tuberculosis infection. Furthermore, TB patients also showed

responses toward IVE-TB Ags, albeit lower than tuberculin skin test-positive, ESAT6/CFP10-responsive individuals. Finally, IVE-

TB Ags induced strong IFN-g+/TNF-a+ CD8+ and TNF-a+/IL-2+ CD154+/CD4+ T cell responses in PBMC from long-term latently

M. tuberculosis–infected individuals. In conclusion, these IVE-TB Ags are expressed during pulmonary infection in vivo, are

immunogenic, induce strong T cell responses in long-term latently M. tuberculosis–infected individuals, and may therefore

represent attractive Ags for new TB vaccines. The Journal of Immunology, 2013, 190: 1659–1671.

Tuberculosis (TB) remains a leading cause of death, par-ticularly in low and middle income countries (1). One thirdof the world population is estimated to be latently infected

with Mycobacterium tuberculosis, and 3–10% of these will developactive TB during their lifetime. In HIV-infected individuals thisproportion increases to 7–10% per life year. The emergence ofmultidrug-resistant, extensively drug-resistant, and more recentlyalso totally drug-resistant M. tuberculosis strains is further ag-gravating the TB epidemic. Currently, Mycobacterium bovis ba-cillus Calmette-Guerin (BCG) is the only available vaccine against

TB. Although BCG vaccination can prevent severe childhood TB(2), it induces highly variable and inconsistent protection againstpulmonary TB, the contagious form of TB in adults (3). A morerecently identified drawback of live BCG vaccination is the oc-currence of disseminating BCG infections in HIV-infected chil-dren (4), similar to severe BCG infections in individuals withgenetic defects in the IL-12/IL-23/IFN-g axis (5, 6). Thus, new TBvaccines are needed that are more effective and safer than BCG.Understanding the intracellular behavior of M. tuberculosis

during in vivo infection is important not only for understanding its

*Department of Infectious Diseases, Leiden University Medical Center, 2300 RCLeiden, The Netherlands; †Department of Immunology and Infectious Diseases,Harvard School of Public Health, Boston, MA 02115; ‡Department of Microbiologyand Immunology, Stanford University School of Medicine, Stanford, CA 94305; andxDivision of Infectious Disease Control, Department of Bacteriology and Immunol-ogy, Norwegian Institute of Public Health, NO-0403 Oslo, Norway

1Current address: Malaria Vaccine Branch, Military Malaria Vaccine Program, WalterReed Army Institute of Research, Silver Spring, MD.

2Current address: National Emerging Infectious Diseases Laboratories Institute, Bos-ton University School of Medicine, Boston, MA.

Received for publication June 14, 2012. Accepted for publication December 10,2012.

This work was supported by European Commission Projects FP7 NEWTBVAC andFP7 VACTRAIN (the text represents the authors’ views and does not necessarilyrepresent a position of the Commission, which will not be liable for the use made ofsuch information), Top Institute Pharma Project D-101-1, and European and Devel-oping Countries Clinical Trials Partnership Project AE-TBC. The funding agencieshad no role in study design, data collection and analysis, decision to publish, orpreparation of the manuscript.

Address correspondence and reprint requests to Prof. Tom H. M. Ottenhoff, Depart-ment of Infectious Diseases, Leiden University Medical Center, Building 1, C5-43,Albinusdreef 2, P.O. Box 9600, 2300 RC Leiden, The Netherlands. E-mail address:[email protected]

The online version of this article contains supplemental material.

Abbreviations used in this article: B6, C57BL/6J; BCG, Mycobacterium bovis bacil-lus Calmette-Guerin; C3H, C3HeB/FeJ; E/C, ESAT6/CFP10 hybrid; EHR, enduringhypoxic response; HC, healthy control (donor not exposed to Mycobacterium tuber-culosis); iMFI, integrated median fluorescence intensity; INH, isoniazid; Ipr1, intra-cellular pathogen resistance 1; IVE-TB, in vivo–expressed Mycobacteriumtuberculosis; LTBI, latent tuberculosis infection; ltLTBI, long-term latent tuberculo-sis infection; QFT-GIT, QuantiFERON-TB Gold In-Tube test; RGCN, relative genecopy number; RT, reverse transcriptase; sst1, supersusceptibility to tuberculosis 1;TB, tuberculosis; TST, tuberculin skin test; WBA, whole blood assay.

Copyright� 2013 by TheAmerican Association of Immunologists, Inc. 0022-1767/13/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1201593

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Page 3

infection biology, but it is also essential for the identification ofpossible novel TB vaccine candidate Ags. Infection stage and site-related differences in in vivo M. tuberculosis gene expressionpatterns can clearly affect the repertoire of potential M. tubercu-losis Ags that is available for immune recognition in the primaryinfected organ, the lung. Ags expressed in the lungs of M. tu-berculosis–infected, susceptible individuals could represent in-teresting new candidate Ags for TB vaccination, because theywould induce responses capable of recognizing in situ M. tuber-culosis–infected cells.M. tuberculosis has a remarkable ability to adapt to environ-

mental changes by altering its metabolic state. A major environ-mental stress factor that M. tuberculosis is thought to encounterduring host infection is the deprivation of oxygen and nutrients.In vitro hypoxia induces the expression of the M. tuberculosisdormancy regulon (7), which is controlled by the master regulatorDosR (Rv3133c). The expression of the M. tuberculosis DosRregulon is also induced by low-dose NO, carbon monoxide ex-posure, and during infection in IFN-g–activated macrophages (7,8). Previously, we have reported broad human T cell responses toM. tuberculosis DosR regulon–encoded Ags and showed that re-sponses to these Ags were prominent and associated with latentTB infection (LTBI) in ethnically and geographically distinct po-pulations (9–13). Other work has shown that nutrient limitationcan induce the expression of specific M. tuberculosis genes suchas Rv2660c (14). This gene was found to encode a “starvation” Agwith promising long-term vaccine efficacy in preclinical TB in-fection models, both in mice (15) and in nonhuman primates (16).The more recently described enduring hypoxic response (EHR)genes represent an alternative hypoxia-induced response model,which includes most of the DosR regulon–encoded genes com-plemented with an additional number (.200) of M. tuberculosisstress response genes (17). This model has also helped to identifynew M. tuberculosis Ags (18).A limitation of the models discussed above is that the identi-

fication of differentially regulated M. tuberculosis genes relies onin vitro models supposed to recapitulate relevant environmentalstress conditions that M. tuberculosis encounters upon host in-fection. First, however, many of these environmental stress factorsmay not be known as yet, limiting the value of such hypothesis-driven studies. Second, there may be additive or synergistic effectsbetween multiple stress factors in vivo that may easily be missedwhen studied in isolation in vitro. Third, and perhaps more im-portantly, certain key features of host response–induced stresscannot readily be recapitulated in vitro, including granuloma for-mation and TB necrosis, both being cardinal features of TB. Toovercome these limitations several laboratories have started toanalyze the gene expression profiles of intracellular M. tubercu-losis, either in infected human or murine macrophages (8, 19), inthe infected tissue of different mouse strains (BALB/c, SCID)(20), or in artificial granuloma mouse models (21). However, noneof these mouse models developed granulomatous necrotic TBlesions (22). We therefore have studied M. tuberculosis genome-wide gene expression patterns in mice strains carrying differentgenotypes of the supersusceptibility to TB 1 (sst1) locus. Thisgenetic locus is located on chromosome 1 and controls the pro-gression ofM. tuberculosis infection to severe and necrotic lesionsin a lung-specific manner: C3HeB/FeJ (C3H) mice carrying thesusceptible sst1 allele develop TB pneumonia with strong in-flammatory responses with exudation throughout the lung andearly onset of massive necrosis, whereas C57BL/6J (B6) micecarrying the resistant sst1 allele develop smaller, interstitialgranulomas without necrotic lesions that control bacterial multi-plication. C3H.B6-sst1 congenic mice carrying the (B6-derived)

resistant sst1 locus on the C3H background showed increasedsurvival after M. tuberculosis infection compared with the sus-ceptible C3H mice, but less prominently than did the resistant B6mice. Finally, M. tuberculosis–infected B6.C3H-sst1 mice, car-rying the susceptible C3H-sst1 locus on the B6 background, de-velop robust granulomas that are fenced from the healthy tissuewhere lesions contain foamy macrophages and develop late-onsetnecrosis, resembling pulmonary TB in human adults. In contrast,the B6 strain does not display this phenotype, confirming thespecific role for sst1 in the control of cell death (23).The sst1 locus carries 22 genes, 1 of which was highly ex-

pressed in M. tuberculosis–infected lungs of C3H.B6-sst1 but notof hypersusceptible C3H mice. Interestingly, the expression of thisgene, termed intracellular pathogen resistance 1 (Ipr1), decreasedM. tuberculosis multiplication in susceptible macrophages andinduced a switch from necrotic to apoptotic cell death (24). Thelack of Ipr1 expression in C3H-susceptible sst1 locus is thereforeresponsible for the development of lung-specific necrosis uponM. tuberculosis infection (25). The closest human homolog ofIpr1 is SP110b. The expression of both Ipr1 and SP110b is re-gulated by IFNs, indicating a role in immunity (26–28). Geneticassociation studies performed in West Africa identified threepolymorphisms in the SP110b gene that were associated withgenetic susceptibility to TB (29). However, a number of otherstudies performed in Ghana, Russia, South Africa, and Indonesiadid not replicate this finding (30–33). A SP110b homolog was alsoidentified in cattle, which correlated to susceptibility to Myco-bacterium avium ssp. paratuberculosis (34).These four (congenic) mouse models we have used in this study

show a spectrum of TB susceptibility that ranges from highlysusceptible (C3H) to resistant (B6) mice, with the developmentof necrotic lesions depending on the sst1 locus and the modifyinggenetic background in which the locus is expressed. This mousemodel replicates key features of human M. tuberculosis infection.In this study, we have taken advantage of this disease spectrumand 1) analyzed quantitative real-time expression patterns of allM. tuberculosis genes predicted to be the first gene in each operon,in the lungs of M. tuberculosis–infected mice, aiming to identifythe M. tuberculosis genes that are highly or differentially ex-pressed in the lung during in vivo infection (in vivo–expressedMycobacterium tuberculosis [IVE-TB] genes); 2) compared theseM. tuberculosis gene expression patterns between susceptible(B6.C3H-sst1 and C3H) and resistant (C3H.B6-sst1 and B6)mouse strains in an attempt to correlate expression patterns toinfection phenotype; and 3) selected a set of the most consistentlyexpressed M. tuberculosis genes, produced these as recombinantproteins, and analyzed their immunogenicity in tuberculin skintest (TST)+ healthy, TB-affected individuals as well as long-termLTBI (ltLTBI) as a first step toward their validation as new TBvaccine candidate Ags.

Materials and MethodsMouse strains

C3H and B6 mice were purchased from The Jackson Laboratory (BarHarbor, ME). Congenic C3H.B6-sst1 and B6.C3H-sst1 mouse strainscarrying the resistant and susceptible alleles of the sst1 locus, respectively,were generated as previously described (24, 35). Briefly, an ∼20-cMsegment of chromosome 1, containing the sst1 locus, was introgressed inthe opposite background strain via $10 backcrosses. Mice were bred andhoused under specific pathogen-free conditions at the Harvard MedicalSchool of Public Health.

Bacterial strains

M. tuberculosis suspensions were used as previously described (36). Inshort, M. tuberculosis (Erdman strain; Trudeau Institute, Saranac Lake,

1660 NOVEL M. TUBERCULOSIS AGS EXPRESSED IN PULMONARY INFECTION

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Page 4

NY) cultures were grown to midlog phase in Middlebrook 7H9 medium(BD Biosciences, Franklin Lakes, NJ) (10% oleic acid/albumin/dextrose/catalase [OADC; Difco], 0.05% Tween 80 [Sigma-Aldrich], and 0.5%glycerol [Sigma-Aldrich]). Bacteria were washed and stored at 280˚C.Prior to infection, bacteria were thawed, sonicated, and diluted in PBSto 106 CFU/ml.

Infection of mice

Mice were infected by aerosol with 25–50 CFU M. tuberculosis usinga Madison chamber (College of Engineering shops at the University ofWisconsin, Madison, WI) with n = 2 per time point (24). B6 and C3H micewere sacrificed both 6 and 9 wk postinfection, whereas B6.C3H-sst1 andC3H.B6-sst1 mice were sacrificed 9 and 6 wk postinfection, respectively.For the reactivation model, B6 and B6.C3H-sst1 mice were infected i.v. viathe tail vein with 5 3 104 CFU M. tuberculosis per mouse as previouslydescribed (23). Twelve weeks after challenge the mice were given isoni-azid (INH) supplied via the drinking water (10 mg/100 ml) for 90 d. Micewere sacrificed 8 wk after INH treatment withdrawal.

Genome-wideM. tuberculosis transcription profiling via a two-step multiplex real-time RT-PCR

Quantification of M. tuberculosis mRNA gene expression was performedas previously described (15, 37, 38). The protocol is based on first-strandcDNA synthesis and controlled multiplex amplification of cDNAs, whichis followed by individual real-time PCR (TaqMan) quantification of am-plified cDNAs in a 384-well format using a LightCycler 480.

Total M. tuberculosis RNA was isolated from the infected mouse lungtissue by homogenization in TRIzol and bacillary disruption by beadbeating (MP Biomedicals, Solon, OH). Total RNA was isolated usingRNeasy columns (Qiagen, Valencia, CA). RNA was precipitated, cleanedwith two consecutive off-column RQ1 DNase digestions (Promega, Madi-son, WI), and resuspended in 50 ml RNase-free water (Applied Biosystems/Ambion, Austin, TX).

cDNA synthesis was performed using 50 ng total RNA, which wasseparated in reverse transcriptase (RT)+ and RT2 reactions to control forDNA contamination. Exo-resistant random primer (0.5 ml), 1 ml 10 mMdNTPs, and nuclease-free water was added and incubated for 3 min at70˚C in a thermal cycler. Subsequently, 4 ml 53 Maxima RT buffer, 0.5 mlRiboLock RNase inhibitor, 0.5 ml Maxima RT enzyme (replaced by waterfor RT2 control samples) (all Fermentas, Glen Burnie, MD), and nuclease-free water were added and incubated at 50˚C for 1 h, 95˚C for 2 min toinactivate, and then kept at 4˚C.

The generated cDNAs were further amplified via controlled multiplexpreamplification with a mix of 2179 M. tuberculosis gene-specific prim-ers (23.8 ml primer mix: ∼50 nM per amplification reaction), 2 ml cDNA,3 ml 103 Advantage 2 buffer (Clontech, Mountain View, CA), 0.6 ml 13Advantage 2 polymerase mix (Clontech, CA), and 0.6 ml 10 mM dNTPs(Fermentas) to a volume of 30 ml (ftp://smd-ftp.stanford.edu/tbdb/rtpcr/taqman_oligos.fa) (15). Sequences and design of PCR primer/probe setsare available at: http://genes.stanford.edu/technology.php and http://www.tbdb.org/rtpcrData.shtml. A comparative control of 100 pg (2 3 104 genecopy number) genomic H37Rv DNA was also included. As an additionalcontrol, 25 M. tuberculosis reference genes were used to control for var-iation across amplification mixes. The gene primer sets were designedusing Primer Express software (PerkinElmer, Foster City, CA) to cover atleast one gene of each predictedM. tuberculosis operon. Each reaction washeated at 95˚C for 5 min, followed by 15 cycles at 95˚C for 30 s, 60˚C for20 s, and 68˚C for 1 min. Previously we have validated conditions formultiplex PCR preamplification via linearity of amplification assay usingall the genes used in the assay. We also validated all individual TaqManassays from our collection for sensitivity and linearity before we startedusing them in gene expression profiling in this study.

Individual gene transcript quantification was carried out using TaqManprimer/probe sets (Biosearch Technologies). Quantitative real-time PCRmix contained 0.07 ml preamplified cDNA, 2 ml TaqMan primer/probemix, 5 ml 23 LightCycler 480 Probes Master Mix, and 2.93 ml ProbesMaster PCR-grade water (Roche) to a final volume of 10 ml. Reactionswere heated at 95˚C for 5 min, followed by 40 cycles at 95˚C for 30 s and60˚C for 20 s. A cool down step of 40˚C for 30 s was run for one cycle.Cycle threshold values were converted to relative gene copy numbers(RGCN) based on logarithmic transformation/linear regression equationsdevised from calibration curves. The data set is available at: http://www.tbdb.org/pubdata/tbdb/publications/Raw-Data-Harvard-Mice.xls.

Correction for some biological heterogeneity between the different miceand mouse strains such as differences in bacterial load was not possiblebecause these were inherent to the extensive differences in genetic TBsusceptibility.

IVE-TB gene selection procedure

First, genes were selected that were expressed in one data set (.1000RGCN) but not in the other data set (+/2), thus selecting M. tuberculosisgenes that are differentially expressed owing to genetic host susceptibilityand/or infection phenotype variations.

The second approach was to select genes that were highly expressed inboth data sets from two different mouse strains in the chosen comparison(+/+), selecting for M. tuberculosis genes that are expressed independentof the genetic makeup of the host. For this selection, the RGCN data wereranked from the highest to the lowest value and overlapping genes wereselected from the top 100 highest expressed genes of both data sets. Thisnumber of 100 genes was arbitrarily chosen to limit the number of can-didates to be analyzed further.

The third and last approach included genes that were expressed in dataset 2 (.1000 RGCN) but not in data set 1 (2/+), following the same ra-tionale as approach 1. Hence, approaches 1 and 3 included differentiallyexpressed genes. There was no number restriction limit (because fewergenes were identified compared with the second approach [+/+]) (Figs. 1, 2).

Recombinant proteins

Recombinant proteins were produced from the selected M. tuberculosisgenes as described previously (39). Briefly, M. tuberculosis genes wereamplified by PCR from genomic H37Rv DNA and cloned by Gatewaytechnology (Invitrogen, Carlsbad, CA) in a bacterial expression vectorcontaining a histidine tag at the N terminus. Vectors were overexpressed inEscherichia coli BL21 (DE3) and purified. Size and purity of recombinantproteins were analyzed by gel electrophoresis and Western blotting with ananti-His Ab (Invitrogen) and an anti–E. coli polyclonal Ab (gift of StatensSerum Institute, Copenhagen, Denmark). Rv2380c, Rv2435c, andRv2737c proteins were prepared as two or three recombinant proteinfragments owing to their large sizes (C, middle [M], and N termini). En-dotoxin contents were ,50 IU/ mg as tested using a Limulus amebocytelysate assay (Cambrex, East Rutherford, NJ). All recombinant proteinswere tested in lymphocyte stimulation assays to exclude Ag nonspecificT cell stimulation and cellular toxicity using PBMC of in vitro–purifiedprotein derivative of M. tuberculosis–negative healthy Dutch donors(Sanquin Blood Bank, Leiden, The Netherlands) (12, 40–42). Purifiedprotein derivative of M. tuberculosis was purchased from Statens SerumInstitute.

FIGURE 1. Overview of IVE-TB gene selection procedure. (A) The M.

tuberculosis RGCN profiles of each mouse model were independently

compared with each other. (B) Three gene selection procedures were used

to select genes for each comparison: genes that were expressed in data set

1 (.1000 RGCN) but not in data set 2 (+/2); genes highly expressed in

both data sets (top 100 highest expressed genes in both models and select

overlapping genes) (+/+); and genes not expressed in data set 1 but

expressed in data set 2 (.1000 RGCN) (2/+) (see also Materials and

Methods).

The Journal of Immunology 1661

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Page 5

M. tuberculosis lysate

M. tuberculosis H37Rv organisms were grown to a late log phase ina culture flask at 37˚C and collected in a V-bottom tube. The pellet waswashed twice with PBS and heat killed for 30 min at 80˚C. The cellsuspension was subsequently collected in a BioSpec tube containing 0.1-mm glass beads. The bacteria were disrupted using a MiniBeadBeater(BioSpec Products). Concentrations of lysates were measured using theBCA assay (Thermo Scientific Pierce).

Study subjects

One hundred thirty-three donors were selected that responded to M.tuberculosis–purified protein derivative by TST (weak positive, 5–11mm [n = 26]; positive, $12 mm [n = 90]; not determined [n = 17],ranging from 6 to 32 mm [average, 16 mm]) and that had documentedexposure to a TB index case (n = 63) and/or had traveled to high en-demic countries (n = 76). TST+ individuals entered a follow-up studyand at recruitment a QuantiFERON-TB Gold In-Tube test (QFT-GIT)was performed (Cellestis, Carnegie, VIC, Australia). The test wasconsidered positive when there were $0.3 IU/ml. Blood samples werecollected by venipuncture. M. tuberculosis–unexposed donors wereincluded as healthy control (HC) donors (n = 11). PBMC from TST+

donors and treated TB patients (n = 7) were used in lymphocyte stimu-lation assays. PBMC from ltLTBI (41, 42) (n = 6) were used for poly-chromatic flow cytometry assays. Informed consent was obtained prior tovenipuncture. The study protocols (P07.048 and P027/99) were approvedby the Institutional Review Board of the Leiden University Medical Centerand the Regional Committees for Medical and Health Research Ethics inNorway.

Whole blood assay

Blood was diluted 1:10 in AIM-V medium (Invitrogen, Breda, TheNetherlands), incubated in 48-wells plates (Nunc, Roskilde, Denmark), andcultured with or without recombinant protein (10 mg/ml), PHA (2 mg/ml),orM. tuberculosis lysate (5 mg/ml) at 37˚C, 5% CO2. After 6 d supernatantwas harvested and stored at 220˚C for use at a later stage.

Lymphocyte stimulation test

PBMC (1.5 3 105/well) were cultured in triplicate in 96-well round-bottom plates (Nunc) and incubated with or without protein (10 mg/ml) inIMDM (Life Technologies/Invitrogen) containing 10% pooled humanserum (Invitrogen) at 37˚C in 5% CO2. After 6 d, supernatants wereharvested, pooled, and stored at 220˚C for future use in IFN-g ELISAs.

IFN-g ELISAs

IFN-g concentrations in supernatants were measured with a standardELISA technique (U-CyTech, Utrecht, The Netherlands). ELISA sampleswere tested in duplicate and the assay was performed according to themanufacturer’s guidelines. Detection limit of the assay was set arbitrary at20 pg/ ml for whole blood assay (WBA) and 100 pg/ml for a lymphocytestimulation test.

Flow cytometric analysis

PBMC (1–2 3 106/tube) were thawed and rested overnight and subse-quently stimulated for 16 h with protein (10 mg/ml) in the presence ofcostimulatory Abs anti-CD28 and anti-CD49d (Sanquin and BD Bio-sciences, respectively). Brefeldin A (3 mg/ml; Sigma-Aldrich) was addedafter the first 4–6 h. Cells were stained with Live/Dead fixable violet deadcell stain (ViViD; Invitrogen) to discriminate between live and dead cellsaccording to manufacturer’s instructions. Cells were stained for 1 h at 4˚Cwith the following surface markers: anti-CD3 PE-Cy5 (BD Biosciences),anti-CD4 Texas Red (Caltag), and anti-CD8 V500 (BD Biosciences).Additionally, anti-CD14 Pacific Blue and anti-CD19 Pacific Blue (bothInvitrogen) were included to select for CD142 and CD192 live cells. In-tracellular staining was performed with anti–IFN-g Alexa 700 (BD Phar-mingen), anti–TNF-a PE-Cy7 (BD Biosciences), anti–IL-2 PE (BDPharmingen), and CD154 allophycocyanin-Alexa 780 (eBioscience) usingthe ADG Fix&Perm kit (An Der Grub Bio Research, Vienna, Austria).Data were acquired on a BD LSRFortessa (BD Biosciences) and analyzedusing FlowJo version 7.6.5 (Tree Star, Ashland, OR). Single CD142/CD192/CD3+ live cells were gated to analyze CD4+ and CD8+ cytokineresponses. Final Ag-specific CD4+ and CD8+ T cell subset populations allcontained at least 200 events. For comparative purposes, medium back-ground values were subtracted for each response in each donor.

Statistical analysis

Statistical analysis was performed using GraphPad Prism (version 5.1). AMann–Whitney U test was used to analyze 1) the difference between cu-mulative IFN-g responses for the ESAT6/CFP10 hybrid (E/C)+, E/C2, andHC individuals and 2) the difference between PHA-induced IFN-g re-sponses measured in E/C+ and E/C2 donors. A p value #0.05 was con-sidered significant.

ResultsIdentification of IVE-TB genes in the lungs of geneticallyresistant and/or susceptible mice

To start identifying novel candidate M. tuberculosis Ags in anunbiased and M. tuberculosis genome-wide fashion, we analyzedthe gene expression patterns of 2170 M. tuberculosis genes, mostof which represent the first gene of each predicted M. tuberculosisoperon, in the lungs of four different M. tuberculosis–infectedmouse strains (B6, C3H, C3H.B6-sst1, and B6.C3H-sst1) thatshow a spectrum of TB susceptibility (Table I). The RGCN weredetermined using quantitative PCR (15). This allows for absolutequantization of the level of transcripts per sample, because thedata are normalized against a standard reference gene numbercopy (as described in Materials and Methods).In this IVE-TB gene selection screen we used the following

criteria to select candidate genes for further analysis. First, we usedthe most strongly upregulated M. tuberculosis genes in all four an-alyzed mouse models. This group of genes includes genes that areexpressed independently of the host susceptibility background.Second, we used M. tuberculosis genes specifically upregulated ineither B6, C3H, B6.C3H-sst1, or C3H.B6-sst1. The expression ofthese genes is influenced by the host genetic background and maytherefore include genes whose expression is associated to partic-ular TB disease characteristics such as granuloma formation and

FIGURE 2. Selection of IVE-TB genes. Numbers of M. tuberculosis

genes obtained after each comparison of two different mouse strains for

every approach described in Fig. 1 are shown in (A)–(G).

1662 NOVEL M. TUBERCULOSIS AGS EXPRESSED IN PULMONARY INFECTION

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Page 6

necrosis. The obtained RGCN of all four mouse models wereindependently compared with each other as visualized in Fig. 1A.All four mouse models were infected with a low-dose aerosol M.tuberculosis inoculum. Additionally, a subset of B6.C3H-sst1was infected using a previously described TB relapse model tostudy features of gene expression during M. tuberculosisreactivation (23).Every individual M. tuberculosis gene’s expression data (e.g.,

B6 versus C3H, B6 versus B6.C3H-sst1) was subjected to threedifferent selection approaches as indicated in Fig. 1B and de-scribed in detail in Materials and Methods. The gene selectionresults for each comparison are visualized in Fig. 2 and Supple-mental Table I. The results shown in Fig. 2 were then used toselect IVE-TB genes whose expression was associated with par-ticular disease characteristics as indicated in Fig. 3. These in-cluded 1) M. tuberculosis genes highly expressed independentlyof host genetic background (expressed in all four mouse models;these M. tuberculosis genes included esxA encoding ESAT6 andother esx genes); 2)M. tuberculosis genes expressed in associationwith necrosis (expressed in both C3H and B6.C3H-sst1, but not inB6 or C3H.B6-sst1); 3) M. tuberculosis genes expressed in asso-ciation with severe necrotic infection or susceptibility (expressedin C3H, but not in B6, C3H.B6-sst1, or B6.C3H-sst1); 4) M. tuber-culosis genes expressed in association with dense granuloma devel-opment (only expressed in B6.C3H-sst1 but not in C3H, B6, or C3H.B6-sst1); 5) M. tuberculosis genes expressed in association withdiffuse granuloma development (expressed in C3H but not in B6.C3H-sst1, B6, or C3H.B6-sst1); 6) M. tuberculosis genes expressedin association with resistance (expressed in every more resistantmouse strain per comparison); 7) M. tuberculosis genes expressedin association with low inflammation (expressed in B6, but not inC3H.B6-sst1); 8) inflammation (expressed in C3H.B6-sst1, but notin B6); and 9) relapse (expressed in i.v. M. tuberculosis–infected,INH-treated B6.C3H-sst1, but not low-dose aerosol–infected B6.C3H-sst1). An overview of the resulting IVE-TB genes is pre-sented in Supplemental Table II.

Immunogenicity of newly identified IVE-TB Ags

Further down selection of IVE-TB genes. The goal of the aboveselection of IVE-TB genes was to identify potentially interestingnew vaccine candidate Ags. Thus, we next determined their im-munogenicity. To this end, a number of M. tuberculosis genes werefurther selected that were either present in more than one group orwere among the top number of genes in the IVE-TB selectionsperformed (Table II). Some M. tuberculosis genes identifiedusing our unbiased genome-wide approach were also identifiedin previous studies as environmental stress induced proteins(7, 14, 17).A subsequent literature search revealed that almost all further

selected IVE-TB proteins (14 of the 16) have been identified pre-viously inM. tuberculosis proteomic studies, confirming the protein

expression of the M. tuberculosis genes identified in our study(44–57) (Table III). Indeed, we observed that M. tuberculosis–

Table I. Mouse strains, genetic background, and TB infection phenotypes

Mouse StrainaGenetic

Background Sst1 Allele Inflammation Lung Necrosis Clinical TB Correlate

C3H C3H Susceptible +++ + (Early) Caseous pneumoniaC3H.B6-sst1 C3H Resistant +++ 2 Progressive interstitial

granuloma without necrosisB6.C3H-sst1 B6 Susceptible ++ + (Late) Caseous granulomaB6 B6 Resistant + 2 Chronic persistent interstitial

granuloma without necrosis

+, ++, and +++ indicate intensity of inflammation.aPichugin et al. (23) and Pan et al. (24).

FIGURE 3. Selection of IVE-TB genes associated with particular TB

disease characteristics. Flowchart of analysis to identify IVE-TB genes re-

lated to TB disease phenotypes is shown. Letters in graphs refer to the spe-

cific selections indicated in Fig. 2. M. tuberculosis genes highly expressed

independent of host genetic background are presented in Fig. 2A (b), 2B (e),

2C (h), 2D (k), 2E (n), and 2F (q).M. tuberculosis genes highly expressed in

association with necrosis are presented in Fig. 2A (c), 2B (f), 2D (j), and/or

2F (p). M. tuberculosis genes highly expressed in association with severe

necrotic infection or susceptibility are presented in Fig. 2A (c), 2D (j), and/or

2C (i). M. tuberculosis genes highly expressed in association with dense

granuloma are presented in Fig. 2C (g), 2B (f), and/or 2F (p).M. tuberculosis

genes highly expressed in association with diffuse granuloma are presented

in Fig. 2C (i), but also Fig. 2A (c) and 2D (j). M. tuberculosis genes highly

expressed in association with resistance are presented in one or more of the

following selections: Fig. 2A (a), 2B (d), 2D (l), 2E (m) (although not related

to the susceptible sst1 locus), 2F (r) and/or 2G (s). M. tuberculosis genes

highly expressed in association with low inflammation are presented in Fig.

2E (m). M. tuberculosis genes highly expressed in association with inflam-

mation are presented in Fig. 2E (o). M. tuberculosis genes highly expressed

in association with relapse-associated genes presented in Fig. 2G (u).

The Journal of Immunology 1663

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Page 7

infected C3H mice recognized most of the selected IVE-TB Ags

as measured by Ag-specific IFN-g production by splenocytes

(Supplemental Fig. 1). Furthermore, we analyzed the conserva-

tion of these IVE-TB proteins using protein BLAST searches on

different M. tuberculosis strains as well as other mycobacterial

species. This showed that the IVE-TB protein sequences are

strongly conserved among all tested M. tuberculosis, M. bovis,

and Mycobacterium africanum strains. Strong conservation wasobserved also for other mycobacterial strains (Table IV).

Recognition of IVE-TB proteins by TST+ donors. To assess theimmunogenicity of the 16 selected IVE-TB proteins, recombinantproteins of the IVE-TB genes were generated and analyzed for theinduction of IFN-g responses in 133 TST+ individuals. Approxi-

mately one third of the TST+ individuals responded to M. tuber-

Table II. Predicted function and classification of selected IVE-TB genes

Gene Name IVE-TB Selectiona Selectionsa Functionb Categoryc Classification Reference

Rv1284 Rv1284 High expression b + e + h + k+ n + q + t

Conserved hypothetical protein(carbonic anhydrase)

7 EHR/starvation 14, 17, 43

Rv2380c mbtE High expression b + e + h + k+ n + q + t

Peptide synthetase mbtE 1 21

Rv3515c fadD19 High expressed sst1s h + q + t Probable fatty acid-CoA ligasefadD19 (involved in lipid

degradation)

1 EHR 17, 21

Rv0079 Rv0079 Necrosis and sni c + f + j + p Hypothetical protein 10 DosR 7, 9, 12, 20, 21Rv2324 Rv2324 sni and relapse c + j + u Probable transcriptional regulator,

asnC family9 EHR 17

Rv2737c recA Necrosis and sni c + f + j + p Recombination protein recombinaseA (recA; M. tuberculosis

recA intein)

2

Rv2838c rbfA sni, diffuse granuloma,and relapse

c + i + j + u Probable ribosome-binding factor A(P15B protein)

2

Rv3420c rimI sni, diffuse granuloma,and relapse

c + i + j + u Ribosomal-protein-alanineacetyltransferase rimI

2

Rv2034 Rv2034 Inflammation o arsR type repressor protein 9 EHR/starvation 14, 17, 20Rv1956 Rv1956 Resistance, diffuse

granuloma, lowinflammation, and relapse

d + i + j + m+ u

Transcriptional regulator (possibleantitoxin; TA operon with Rv1955)

0 EHR/starvation 14, 17, 21

Rv2225 panB Dense granuloma f + g + p Probable 3-methyl-2-oxobutanoatehydroxymethyltransferase (panB)

7 21

Rv2465c rpiB Dense granuloma f + g + p Isomerase (ribose 5-phosphateisomerase)

7 EHR 17, 20

Rv2982c gpsA Resistance, inflammation,and relapse

l + o + r + u Probable glycerol-3-phosphatedehydrogenase (gpdA2)

1

Rv3353c Rv3353c Relapse u Conserved hypothetical protein 10Rv1363c Rv1363c Resistance s Possible membrane protein 3Rv2435c Rv2435c Resistance and

inflammationl + o Probable cyclase (adenylate or

guanylate cyclase)7

aIn reference to Figs. 2 and 3.bAvailable at: http://genolist.pasteur.fr/TubercuList/ and www.tbdb.org.cTubercuList functional classification codes are available art: http://genolist.pasteur.fr/tuberculist.sni, Severe necrotic infection.

Table III. Identification of IVE-TB proteins in M. tuberculosis

GeneProteinName

ProteinIdentification

Protein Location

Essential In Vitro Essential In VivoMembrane/Lipid

Associated Cytosol CF WCL

Rv0079 Rv0079 49, 51, 52 X No (53) Yes (57)Rv1284 Rv1284 45–48, 51 X X X Yes (53) Yes (54)Rv1363c Rv1363c 45, 46 X X No (53)Rv1956 Rv1956 46 X No (53, 55)Rv2034 Rv2034 45, 46 X No (53)Rv2225 panB 46, 52 X X Yes (53)Rv2324 Rv2324 45, 46 X X No (53)Rv2380c mbtE 51, 52 X X Probably yes (55)Rv2435c Rv2435c No (53)Rv2465c rpiB 44–48 X X X No (53)Rv2737c recA 52 X Slow grow mutant (53)Rv2838c rbfA 46 X NTRv2982c gpsA 45, 46, 50, 51 X X No (53)Rv3353c Rv3353c No (53)Rv3420c rimI 52 X No (53)Rv3515c fadD19 46, 49, 50 X X No (53, 55); yes (56)

Note that the locations of the proteins indicated may not be exclusive given limitations and difficulties in annotating exact protein localization (46).CF, Culture filtrate; NT, not tested; WCL, whole cell lysate.

1664 NOVEL M. TUBERCULOSIS AGS EXPRESSED IN PULMONARY INFECTION

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Page 8

culosis–specific E/C protein (32%, 61–9980 pg/ml). ESAT6 and

CFP10 M. tuberculosis–specific proteins are frequently used as

immunodiagnostic Ags to identify M. tuberculosis exposure (12).

One third of our population also responded to these diagnostic Ags

(ESAT6, CFP10 complemented with TB7.7 [p4] peptides) using

the commercial QFT-GIT. The responses in the QFT-GIT corre-

lated very well with responses to our E/C protein measured in

WBA (84%). Division of the group into WBA E/C+ and WBA

E/C2 donors showed that 74% of the E/C+ donors responded to

M. tuberculosis lysate (87–3112 pg/ml) and 40% to Ag85B Ag

(97–735 pg/ml) (Fig. 4A). Importantly, high levels of recognition

of the newly identified M. tuberculosis proteins were seen within

this population, with responses ranging from 10 to 60% of the

donors. To identify the Ags that are well recognized, we first ex-

amined which ones were recognized by $25% of the E/C+ donors.

From all 20 proteins and protein fragments (derived from the 16

selected Ags) tested, 13 were recognized by $25% (n $ 3 of 10

and $11 of 43) of the donors. Six of 13 Ags induced only inter-

mediate levels of IFN-g (60–600 pg/ml) after stimulation, whereas

high levels of IFN-g were produced after stimulation with

Rv1363c (102–1053 pg/ml), Rv1956 (100–1861 pg/ml), Rv2034

(77–1256 pg/ml), Rv2324 (63–2130 pg/ml), Rv2380c-C (80–1159

pg/ml), Rv3353c (70–899 pg/ml), and Rv3420c (88–660 pg/ml),

representing the seven best recognized proteins.Of the TST+ E/C2 donors still 47% responded to M. tubercu-

losis lysate and 20% to Ag85B (Fig. 4B). Additionally, HC were

analyzed for possible IVE-TB responses (Fig. 4C). Most impor-

tantly, limited responses to no responses against the IVE-TB

proteins were detected in the TST+ E/C2 and HC groups, dem-

onstrating clear specificity of recognition, presumably linked to

M. tuberculosis exposure.To further verify that the IVE-TBAgs are specifically recognized

by E/C+ donors, the cumulative IFN-g response to the 20 testedAgs per individual was calculated, as described before (9). A highlysignificant difference between the E/C+ and E/C2 population wasobserved (p , 0.0001). As expected, a significant difference wasalso observed between E/C+ and HC donors (p = 0.049) (Fig. 5),confirming the association between Ag recognition and E/C testpositivity. Interestingly, of the seven best recognized proteins,Rv3420c was the most discriminatory between the E/C+ ($25%)and E/C2 group (#1%) (p, 0.0001) and was not recognized by HCdonors (p = 0.016), suggesting a possible role as M. tuberculosis–specific biomarker Ag in addition to ESAT6, CFP10, and TB7.7.

Recognition of IVE-TB proteins by PBMC from TB patients. We

next investigated whether TB patients could also recognize theseAgs. PBMC from WBA E/C+ TST+ donors (Fig. 6A) and TBpatients (Fig. 6B) were therefore stimulated with the 20 proteinsand protein fragments and IFN-g production was measured. HighIFN-g responses (123–3391 pg/ml) to the IVE-TB Ags were ob-served in the PBMC cultures from the WBA E/C+ TST+ pop-ulation, confirming the results obtained in the WBA assay above.Nine of the tested Ags were recognized by $50% of the donorsand eight Ags by 38% of the donors. Only one protein fragmentwas not recognized in this assay (Rv2380c-M). In contrast, sevenof the tested IVE-TB Ags did not induce detectable IFN-g pro-duction in PBMC from TB patients, whereas most of the IVE-TBAgs induced only low levels of IFN-g compared with the WBA E/C+

TST+ individuals (107–1825 pg/ml). Only two Ags induced highlevels of IFN-g in the TB patients. Additionally, only one Ag wasrecognized by 50% of the TB patients, whereas the remainder of

the Ags were recognized by relatively fewer TB patients (14–

43%). Thus, IVE-TB Ags seem to be less immunogenic in TB

patients than in TST+ individuals.Table

IV.

IVE-TBprotein

sequence

identity

amongmycobacterial

strains

Mycobacterial

Species

Strain

Rv0079

Rv1284

Rv1363c

Rv1956

Rv2034

Rv2225

Rv2324

Rv2380c

Rv2435c

Rv2465c

Rv2737c

Rv2838c

Rv2982c

Rv3353c

Rv3420c

Rv3515c

tuberculosis

Erdmann

100

100

100

100

100

100

100

100

100

100

100

100

100

100

100

100

tuberculosis

H37Rv

100

100

100

100

100

100

100

100

100

100

100

100

100

100

100

100

tuberculosis

C100

100

100

100

99

100

100

100

100

100

99

100

99

100

100

94

tuberculosis

Haarlem

100

100

100

100

99

ND

100

100

100

100

100

100

99

100

100

99

tuberculosis

F11

100

100

100

100

99

100

100

100

100

100

100

100

99

100

100

100

tuberculosis

CDC1551

100

100

100

100

99

100

100

100

100

100

100

100

99

100

100

100

africanum

GM041182

100

100

100

100

100

100

100

100

99

100

99

100

99

100

100

99

bovis

AF2122/97

100

100

99

100

100

100

100

100

100

100

100

100

99

100

100

100

bovis

BCG

100

100

100

100

100

100

100

100

100

100

100

100

99

100

100

100

avium

Paratuberculosis

K-10

ND

80

74

ND

41

88

91

65

65

91

96

77

84

50

82

90

avium

104

ND

80

74

ND

41

87

89

65

65

91

96

77

40

51

82

90

ulcerans

Agy99

62

85

70

ND

39

86

83

43

60

90

96

81

86

65

82

90

smegmatis

Mc2

155

33

79

58

ND

39

78

81

66

59

84

93

76

75

ND

67

83

marinum

M63

85

72

ND

39

86

84

43

81

91

96

81

87

64

82

90

leprae

TN

ND

ND

ND

ND

ND

82

ND

ND

32

89

61

80

77

ND

72

ND

LocalBLASTprotein

sequence

identities

areavailable

at:http://www.tbdb.org

andhttp://www.ncbi.nlm

.nih.gov/blast/.ND,Notdetected.

The Journal of Immunology 1665

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Page 9

T cell responses toward IVE-TB Ags in long-term latent M.

tuberculosis–infected individuals. Because the IVE-TB Ags were

strongly recognized by TST+ individuals, we subsequently ana-lyzed the immune responses toward the seven best recognized Agsin more detail using PBMC from donors that had been exposed toM. tuberculosis decades ago, but had never developed TB despitethe lack of any preventive treatment, designated ltLTBI (41, 42). Ofadditional importance, the availability of several vials of PBMCalso allowed more detailed cell subset analysis.Interestingly, high frequencies of TNF-a– and IL-2–producing

CD4+ T cells were observed after stimulation with the IVE-TB Ags,whereas only intermediate frequencies of IFN-g–producing CD4+

T cells were detected (Fig. 7A). In contrast, high frequencies ofIFN-g–producing as well as TNF-a+ CD8+ T cells were present inthese donors, whereas fewer IL-2+ T cells were detected comparedwith CD4+ T cells. Besides IFN-g, TNF-a, and IL-2, also the Ag-induced CD4+ T cell activation marker CD154 (58) was expressed.More detailed analysis of the multifunctional Th1 responses

among CD4+ and CD8+ T cells showed that CD154+CD4+ T cells

were mostly TNF-a+/IL-2+ and TNF-a+ (Fig. 7B). Furthermore,

intermediate frequencies of IFN-g+/TNF-a+/IL-2+ CD154+CD4+

T cells were detected. Finally, a CD154+ population was detected,

producing none of the IFN-g, TNF-a, and IL-2 cytokines. In-

triguingly, the same pattern was observed for every IVE-TB Ag

or E/C control Ag. Furthermore, interindividual variation of Ag

recognition was observed. Remarkably, few TNF-a+/IL-2+ CD8+

T cells were detected compared with TNF-a+/IL-2+ CD4+ T cells.

IFN-g+/TNF-a+ CD8+ T cells were the most prominent population

present, followed by TNF-a+ CD8+ T cells. Also, intermediate

IFN-g+/TNF-a+/IL-2+ and IFN-g+ CD8+ T cells were observed.

Again, as mentioned for CD4+ T cells, the same patterns were

observed for every Ag within the CD8+ T cell population as well

as interindividual variation of Ag recognition.The integrated median fluorescence intensity (iMFI) was cal-

culated to determine the quantitative contribution of cytokines

produced by the different multiple and single cytokine producing

CD154+/CD4+ and CD8+ T cells (Fig. 8). IFN-g+/TNF-a+/IL-2+

CD154+/CD4+ T cells had the highest iMFI, which gradually

declined for double producing and single IFN-g+ CD154+/CD4+

T cells. IFN-g+/TNF-a+ CD8+ T cells contributed the most to

IFN-g production, directly followed by the IFN-g+/TNF-a+/IL-2+

CD8+ T cells. IFN-g+/TNF-a+ CD8+ T cells are also the main

contributors for TNF-a, whereas IFN-g+/TNF-a+/IL-2+ CD8+

T cells showed a higher IL-2 iMFI. TNF-a and IL-2 iMFI were

also the highest for TNF-a+/IL-2+ CD154+/CD4+ T cells, followed

by the IFN-g+/TNF-a+/IL-2+ CD154+/CD4+ T cells. Thus the

TNF-a+/IL-2+ CD4 and IFN-g+/TNF-a+ CD8 T cells contribute

strongly to the production of Th1 cytokines, followed by the

triple-positive T cells. Single cytokine–producing cells only showed

a relatively minor contribution.In conclusion, seven of the identified IVE-TB Ags are strongly

immunogenic, triggering specific and high cellular immune

responses in E/C+ TST+ individuals and long-term ltLTBI indi-

viduals, but not in E/C2 TST+ individuals, healthy mycobacterial

naive individuals, and TB patients. The strong IVE-TB responses

that were measured in the ltLTBI group were identified as IFN-g+/

TNF-a+ CD8+ T cells and TNF-a+/IL-2+ CD4+ T cells, which

were the most prominent contributors to the produced cytokines,

followed by triple-positive T cells.

DiscussionUsing quantitative genome-wide M. tuberculosis transcriptionalprofiling, we have identified a series of M. tuberculosis genes that

FIGURE 4. IFN-g responses to IVE-TB Ags in E/C+ (A), E/C2 (B), and

HC donors (C). A total of 43 E/C+ donors (A), 90 E/C2 donors (B), and 11

HC donors (C) were analyzed for their IFN-g WBA responses to Ags and

controls; the Ags wereM. tuberculosis Rv1284, Rv1956, Rv2034, Rv2324,

Rv3353c, and Rv3420c. Ten E/C+ donors (A), 36 E/C2 donors (B), and 9

HC donors (C) were also analyzed for responses to M. tuberculosis Ags

Rv0079, Rv1363c, Rv2225, Rv2380c, Rv2435c, Rv2465c, Rv2737c,

Rv2828c, Rv2982c, and Rv3515c IFN-g. The proportion of responders for

each Ag is indicated at the top of the graph. For comparative purposes,

medium background values were subtracted for each response in each

donor. Horizontal bars represent the median IFN-g responses. The dotted

line indicates the cut-off value for positivity, arbitrarily calculated as 3 3medium value.

1666 NOVEL M. TUBERCULOSIS AGS EXPRESSED IN PULMONARY INFECTION

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Page 10

are expressed during in vivo M. tuberculosis infection in the lungsof resistant and susceptible mice, which we term IVE-TB. Most ofthe genes identified have previously been found to be expressed inthe M. tuberculosis proteome, and thus encode bona fide M. tu-berculosis proteins. This is further supported by their immuno-genicity profiles, as many of these proteins triggered IFN-gproduction in human WBA and lymphocyte stimulating assaysin M. tuberculosis ESAT6/CFP10-responsive patients, but not

in ESAT6/CFP102 TST+ individuals, HC donors, or TB patients.This is particularly relevant in the case of Rv2435c and Rv3353c,as their protein products have not been identified yet; however,IFN-g responses were demonstrated in E/C+ TST+ individuals,indirectly showing that these M. tuberculosis proteins are presentedto the human immune system during mycobacterial infection.Many of the IVE-TB genes we identified have been described

previously in relationship to the adaptative response of M. tu-

FIGURE 5. Cumulative IFN-g responses induced

by IVE-TB Ags, calculated per individual in the E/C+,

E/C2, and HC groups. Cumulative IFN-g responses

to all 20 IVE-TB protein and protein fragments in E/

C+ (n = 43), E/C2 (n = 90), and HC donors (n = 11).

Squares indicate cumulative IFN-g response of all 20

IVE-TB Ags, and circles indicate cumulative IFN-g

response of Rv1284, Rv1956, Rv2034, Rv2324,

Rv3353c, and Rv3420c Ags. Horizontal bars rep-

resent the median cumulative IFN-g responses.

FIGURE 6. PBMC IFN-g responses toward

IVE-TB Ags in TB patients and WBA E/C+

TST+ donors. PBMC of WBA E/C+ TST+

donors (n = 8) (A) and TB patients (n = 7) (B)

were stimulated with IVE-TB Ags and control

conditions for 6 d. Levels of IFN-g were mea-

sured and medium background values were

subtracted for each response in each donor for

comparative purposes. The proportion of re-

sponders for each Ag is indicated at the top of

the graph. Horizontal bars represent the median

IFN-g responses. The dotted line indicates the

cut-off value for positivity, arbitrarily set at 100

pg/ml.

The Journal of Immunology 1667

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Page 11

berculosis to environmental stress conditions, especially thosethat M. tuberculosis likely encounters during host infection. Weidentified one M. tuberculosis DosR regulon–encoded gene(Rv0079) (7) as well as six genes that are part of the EHR regulon(Rv1284, Rv1956, Rv2034, Rv2324, Rv2465, and Rv3515) (17).Three of these have also been described as starvation/nutritionalstress genes (Rv1284, Rv1956, and Rv2034) (14). This functionof the IVE-TB genes in responding to host-induced stressconditions during in vivo pulmonary infection enhances thebiological plausibility of our findings and lends validity to ourapproach.Of further interest, nine of the M. tuberculosis genes identified

in this study have not been described previously in relationship toM. tuberculosis host infection, although some of their functionshave been linked to possible adaptation to in vitro host-induced

stress conditions (Supplemental Table III). Several of these geneshave a role in metal transport, metalloregulatory transcriptionalregulation, or represent metalloenzymes. Furthermore, genes wereidentified that play a role in lipid metabolism. This is in agreementwith the documented shift toward using fatty acids as an alter-native carbon source instead of carbohydrates under nutrient-limiting conditions. Altogether, many of the IVE-TB genes wehave identified appear to be related to the adaptation of M. tuber-culosis to environmental stress conditions encountered in the host.Of additional importance, the identification of these genes in ourin vivo model supports previous findings mostly obtained inin vitro models by showing that they are induced during pulmonaryM. tuberculosis infection in vivo. On a cautionary note, however,our data do not allow us to discriminate whether the observeddifferential M. tuberculosis gene expression patterns are cause or

FIGURE 7. Polychromatic flow cytometric analysis of IVE-TB–specific T cell responses in long-term latentM. tuberculosis–infected individuals. PBMC

from long-term ltLTBI (n = 6) were stimulated for 16 h with the seven best recognized Ags as determined in Fig. 4. Frequencies of IFN-g–, TNF-a–, IL-2–,

and CD154-producing CD4+ and CD8+ T cells were determined (A). Subsequently, “multifunctional” responses were determined by analyzing combi-

nations of IFN-g, TNF-a, IL-2, and CD154 responses for CD4+ T cells and IFN-g, TNF-a, and IL-2 responses for CD8+ T cells. Results for two rep-

resentative IVE-TB Ags are shown (Rv2034 and Rv3420c) (B). For comparative purposes, medium background values were subtracted for each response in

each donor. Horizontal bars represent the median frequency of cytokine-producing CD4+ or CD8+ T cells.

1668 NOVEL M. TUBERCULOSIS AGS EXPRESSED IN PULMONARY INFECTION

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Page 12

consequence of variations in host susceptibility (background and/or sst1 locus).As already mentioned, M. tuberculosis gene expression profiling

has been performed in the past, mostly focusing on in vitro–culturedbacteria grown under a variety of different conditions. Subsequentwork assessed M. tuberculosis gene expression profiles followinginfection of host cells (8, 19, 43), and some recent studies haveanalyzed M. tuberculosis gene expression patterns also in vivo (20,59). Ward et al. (59) showed that there was little overlap in theM. tuberculosis genes reported to be expressed in different studiesreporting onM. tuberculosis intracellular infection, likely as a resultof methodological differences. Nonetheless, the two studies Wardet al. described (8, 60) indicate that similar functional categoriesof M. tuberculosis genes are expressed during intracellular infec-tion. In line with this notion, when comparing our data to previousreports there are few overlapping individual M. tuberculosis genes,but we nevertheless do identify genes with previous describedfunctional categories. These differences are probably due to dif-ferences in selection criteria, in experimental settings such as in-fection route, and the specific mouse models we have used, whichhave not been analyzed previously.Despite these differences, several of our selected IVE-TB genes

do overlap with M. tuberculosis genes identified in other studiesas indicated in Table II. The M. tuberculosis gene Rv2225 whoseexpression was TB granuloma associated was also significantlyexpressed in the artificial granuloma model of Karakousis et al.

(21). This strengthens their association with host granuloma for-mation. Our in vivo pulmonary TB granuloma-associated M. tu-berculosis–expressed genes did not overlap with the granuloma-associated genes or macrophage-associated genes described byRamakrishnan and colleagues (61, 62) for M. marinum, whichmight be due to differences between the mycobacterial speciesstudied. Moreover, several of the IVE-TB genes we identified tobe highly expressed have also been described previously, includingRv0467 (icl), which encodes an enzyme in the glyoxylate path-way, which is important for M. tuberculosis persistence of M.tuberculosis (63, 64), and Rv0991c, which is part of the so-calledin vivo–expressed genomic island (20).The new M. tuberculosis Ags we have identified in this study

may represent interesting targets for vaccination, as they areexpressed during M. tuberculosis infection in the (geneticallysusceptible) lung, which we consider a critical parameter for ap-propriate Ag selection. Moreover, successful vaccine Ags shouldbe conserved between multiple M. tuberculosis strains. All proteinsequences examined were conserved among the tested M. tuber-culosis strains. Additionally, for almost all IVE-TB genes multipleproteome studies have documented their expression as proteinsin M. tuberculosis (Table III). A subset of the analyzed IVE-TBproteins was shown to be strongly immunogenic as judged by Th1responses in WBA, lymphocyte stimulation assays, and poly-chromatic flow cytometry. Indeed, the highest IFN-g responseswere identified within the E/C+ population of our TST+ cohort,

FIGURE 8. iMFI of IVE-TB–specific CD154+CD4+ and CD8+ T cell subsets in long-term latent M. tuberculosis–infected individuals. iMFI values for

IFN-g, TNF-a, and/or IL-2 were calculated via multiplication of CD154+CD4+ and CD8+ T cell subset frequency by their MFI. Six ltLTBI donors were

analyzed. For comparative purposes, medium background iMFI values were subtracted for each response in each donor. Light gray boxes represent CD154+

CD4+ T cell responses and dark gray boxes CD8+ T cell responses. Lines within boxes represent the medians. The lower boundary of the box represents the

25th percentile and upper boundary the 75th percentile. Whiskers extend to the lowest and highest values.

The Journal of Immunology 1669

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Page 13

whereas no differences in mitogen-induced responses were seen.No responses were seen in M. tuberculosis nonresponder healthyindividuals, suggesting that T cell recognition of IVE-TB Ags isindeed Ag specific and is correlated with M. tuberculosis exposurebased on TST and QFT-GIT conversions. Interestingly, TB patientsshowed relatively low recognition of the IVE-TB Ags, suggestingthat they did not develop strong Th1 immunity against these Ags.Importantly, IVE-TB Ag-specific responses could be detected in

ltLTBI, which have been exposed to M. tuberculosis many yearsago and never developed TB symptoms despite not having hadpreventive treatment. The most prominent T cell subsets withactivity against IVE-TB Ags included IFN-g+/TNF-a+ CD8+

T cells and TNF-a+/IL-2+ CD154+CD4+ T cells. Thus, CD8+

T cells were the major contributors of IFN-g production. Inter-estingly, also a population of Ag-specific–activated CD154+CD4+

T cells was observed that did not produce IFN-g, TNF-a, or IL-2.These cells may exert alternative functions, which could includeIL-17 production, immune regulation, or yet other functions,which need further study. Finally, we previously reported multi-functional CD4+ and CD8+ T cell responses toward resuscitationpromoting factor and DosR proteins and showed that IFN-g+/TNF-a+ CD8+ T cells were also the most prominent subset in theresponse to these Ags, suggesting that the development of specificdifferential T cell subsets may be unrelated to the nature of thespecific protein Ag involved.CD8+ T cells are activated upon recognition of epitopes pre-

sented via MHC class I molecules, indicating that the Ags arepresent and processed via the canonical cytosolic pathway or viaalternative (e.g., cross-priming) pathways (65). Both CD4+ andCD8+ T cells are important in M. tuberculosis control, and CD4+

and CD8+ T cell–deficient mice, for example, have increasedsusceptibility to M. tuberculosis (66). CD4+ T cells were recentlyshown to play an important (IFN-g–independent) role in the in-direct activation of IFN-g+ CD8+ T cells (67). In any case, our dataobtained in the ltLTBI individuals show that the M. tuberculosisAg-specific CD4+ and CD8+ T cells recognizing IVE-TB Ags mustbe long lived.The immunogenicity of some of the IVE-TB Ags has been an-

alyzed previously. The immunogenicity of the DosR Rv0079 proteinwas analyzed in TST+ (endemic) individuals as well as (cured) TBpatients (9, 12, 68). In these studies Rv0079 protein was recognizedby a minority of individuals, in agreement with our results in thisstudy. The immunogenicity of EHR and starvation Ags Rv1284 andRv1956 was previously analyzed in M. bovis–exposed cattle (17,69). Rv1284 was one of the five best recognized Ags, whereasRv1956 was also highly recognized. In contrast to the responsesobserved in M. bovis–exposed cattle, Rv1284 was moderately rec-ognized in our study, whereas Rv1956 was better recognized.In conclusion, by combining M. tuberculosis genome-wide

transcriptional profiling in the lungs of infected mice with strik-ingly differing host susceptibility backgrounds, we have identifiedM. tuberculosis genes that are specifically expressed in resistantor susceptible animals during pulmonary infection. These genesreveal a signature of the M. tuberculosis stress response in vivodepending on the genetic host background and host susceptibility.From these genes we selected 16 proteins, of which proved to behighly immunogenic in E/C+ TST+ donors and ltLTBI andtherefore represent interesting TB vaccine candidate and possiblyTB biomarker Ags (70).

AcknowledgmentsWe thank Louis Wilson for the production of theM. tuberculosis lysates, as

well as all blood donors who participated in this study.

DisclosuresT.H.M.O. is coinventor of an M. tuberculosis latency Ag patent, which is

owned by Leiden University Medical Center. The other authors have no

financial conflicts of interest.

References1. World Health Organization. 2011. The Global Plan to Stop TB 2011–

2015. Available at: http://www.stoptb.org/assets/documents/global/plan/TB_GlobalPlanToStopTB2011-2015.pdf

2. Trunz, B. B., P. Fine, and C. Dye. 2006. Effect of BCG vaccination on childhoodtuberculous meningitis and miliary tuberculosis worldwide: a meta-analysis andassessment of cost-effectiveness. Lancet 367: 1173–1180.

3. Fine, P. E. 1995. Variation in protection by BCG: implications of and for het-erologous immunity. Lancet 346: 1339–1345.

4. Hesseling, A. C., B. J. Marais, R. P. Gie, H. S. Schaaf, P. E. Fine, P. Godfrey-Faussett, and N. Beyers. 2007. The risk of disseminated bacille Calmette-Guerin(BCG) disease in HIV-infected children. Vaccine 25: 14–18.

5. Ottenhoff, T. H., F. A. Verreck, E. G. Lichtenauer-Kaligis, M. A. Hoeve,O. Sanal, and J. T. van Dissel. 2002. Genetics, cytokines and human infectiousdisease: lessons from weakly pathogenic mycobacteria and salmonellae. Nat.Genet. 32: 97–105.

6. van de Vosse, E., M. A. Hoeve, and T. H. Ottenhoff. 2004. Human genetics ofintracellular infectious diseases: molecular and cellular immunity againstmycobacteria and salmonellae. Lancet Infect. Dis. 4: 739–749.

7. Voskuil, M. I., D. Schnappinger, K. C. Visconti, M. I. Harrell, G. M. Dolganov,D. R. Sherman, and G. K. Schoolnik. 2003. Inhibition of respiration by nitricoxide induces a Mycobacterium tuberculosis dormancy program. J. Exp. Med.198: 705–713.

8. Schnappinger, D., S. Ehrt, M. I. Voskuil, Y. Liu, J. A. Mangan, I. M. Monahan,G. Dolganov, B. Efron, P. D. Butcher, C. Nathan, and G. K. Schoolnik. 2003.Transcriptional adaptation of Mycobacterium tuberculosis within macrophages:insights into the phagosomal environment. J. Exp. Med. 198: 693–704.

9. Leyten, E. M., M. Y. Lin, K. L. Franken, A. H. Friggen, C. Prins, K. E. vanMeijgaarden, M. I. Voskuil, K. Weldingh, P. Andersen, G. K. Schoolnik, et al.2006. Human T-cell responses to 25 novel antigens encoded by genes of thedormancy regulon of Mycobacterium tuberculosis. Microbes Infect. 8: 2052–2060.

10. Roupie, V., M. Romano, L. Zhang, H. Korf, M. Y. Lin, K. L. Franken,T. H. Ottenhoff, M. R. Klein, and K. Huygen. 2007. Immunogenicity of eightdormancy regulon-encoded proteins of Mycobacterium tuberculosis in DNA-vaccinated and tuberculosis-infected mice. Infect. Immun. 75: 941–949.

11. Schuck, S. D., H. Mueller, F. Kunitz, A. Neher, H. Hoffmann, K. L. Franken,D. Repsilber, T. H. Ottenhoff, S. H. Kaufmann, and M. Jacobsen. 2009. Iden-tification of T-cell antigens specific for latent mycobacterium tuberculosis in-fection. PLoS ONE 4: e5590.

12. Black, G. F., B. A. Thiel, M. O. Ota, S. K. Parida, R. Adegbola, W. H. Boom,H. M. Dockrell, K. L. Franken, A. H. Friggen, P. C. Hill, et al; GCGH Bio-markers for TB Consortium. 2009. Immunogenicity of novel DosR regulon-encoded candidate antigens of Mycobacterium tuberculosis in three high-burden populations in Africa. Clin. Vaccine Immunol. 16: 1203–1212.

13. Goletti, D., O. Butera, V. Vanini, F. N. Lauria, C. Lange, K. L. Franken,C. Angeletti, T. H. Ottenhoff, and E. Girardi. 2010. Response to Rv2628 latencyantigen associates with cured tuberculosis and remote infection. Eur. Respir. J.36: 135–142.

14. Betts, J. C., P. T. Lukey, L. C. Robb, R. A. McAdam, and K. Duncan. 2002.Evaluation of a nutrient starvation model of Mycobacterium tuberculosis per-sistence by gene and protein expression profiling. Mol. Microbiol. 43: 717–731.

15. Aagaard, C., T. Hoang, J. Dietrich, P. J. Cardona, A. Izzo, G. Dolganov,G. K. Schoolnik, J. P. Cassidy, R. Billeskov, and P. Andersen. 2011. A multistagetuberculosis vaccine that confers efficient protection before and after exposure.Nat. Med. 17: 189–194.

16. Lin, P. L., J. Dietrich, E. Tan, R. M. Abalos, J. Burgos, C. Bigbee, M. Bigbee,L. Milk, H. P. Gideon, M. Rodgers, et al. 2012. The multistage vaccine H56boosts the effects of BCG to protect cynomolgus macaques against active tu-berculosis and reactivation of latent Mycobacterium tuberculosis infection. J.Clin. Invest. 122: 303–314.

17. Rustad, T. R., M. I. Harrell, R. Liao, and D. R. Sherman. 2008. The enduringhypoxic response of Mycobacterium tuberculosis. PLoS One 3: e1502.

18. Gideon, H. P., K. A. Wilkinson, T. R. Rustad, T. Oni, H. Guio, R. A. Kozak,D. R. Sherman, G. Meintjes, M. A. Behr, H. M. Vordermeier, et al. 2010.Hypoxia induces an immunodominant target of tuberculosis specific T cellsabsent from common BCG vaccines. PLoS Pathog. 6: e1001237.

19. Cappelli, G., E. Volpe, M. Grassi, B. Liseo, V. Colizzi, and F. Mariani. 2006.Profiling of Mycobacterium tuberculosis gene expression during human mac-rophage infection: upregulation of the alternative sigma factor G, a group oftranscriptional regulators, and proteins with unknown function. Res. Microbiol.157: 445–455.

20. Talaat, A. M., R. Lyons, S. T. Howard, and S. A. Johnston. 2004. The temporalexpression profile of Mycobacterium tuberculosis infection in mice. Proc. Natl.Acad. Sci. USA 101: 4602–4607.

21. Karakousis, P. C., T. Yoshimatsu, G. Lamichhane, S. C. Woolwine,E. L. Nuermberger, J. Grosset, and W. R. Bishai. 2004. Dormancy phenotypedisplayed by extracellular Mycobacterium tuberculosis within artificial granu-lomas in mice. J. Exp. Med. 200: 647–657.

1670 NOVEL M. TUBERCULOSIS AGS EXPRESSED IN PULMONARY INFECTION

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Page 14

22. Young, D. 2009. Animal models of tuberculosis. Eur. J. Immunol. 39: 2011–2014.

23. Pichugin, A. V., B. S. Yan, A. Sloutsky, L. Kobzik, and I. Kramnik. 2009.Dominant role of the sst1 locus in pathogenesis of necrotizing lung granulomasduring chronic tuberculosis infection and reactivation in genetically resistanthosts. Am. J. Pathol. 174: 2190–2201.

24. Pan, H., B. S. Yan, M. Rojas, Y. V. Shebzukhov, H. Zhou, L. Kobzik,D. E. Higgins, M. J. Daly, B. R. Bloom, and I. Kramnik. 2005. Ipr1 genemediates innate immunity to tuberculosis. Nature 434: 767–772.

25. Kramnik, I. 2008. Genetic dissection of host resistance to Mycobacterium tu-berculosis: the sst1 locus and the Ipr1 gene. Curr. Top. Microbiol. Immunol. 321:123–148.

26. Grotzinger, T., K. Jensen, and H. Will. 1996. The interferon (IFN)-stimulatedgene Sp100 promoter contains an IFN-g activation site and an imperfect IFN-stimulated response element which mediate type I IFN inducibility. J. Biol.Chem. 271: 25253–25260.

27. Kadereit, S., D. R. Gewert, J. Galabru, A. G. Hovanessian, and E. F. Meurs.1993. Molecular cloning of two new interferon-induced, highly related nuclearphosphoproteins. J. Biol. Chem. 268: 24432–24441.

28. Bloch, D. B., A. Nakajima, T. Gulick, J. D. Chiche, D. Orth, S. M. de La Monte,and K. D. Bloch. 2000. Sp110 localizes to the PML-Sp100 nuclear body andmay function as a nuclear hormone receptor transcriptional coactivator. Mol.Cell. Biol. 20: 6138–6146.

29. Tosh, K., S. J. Campbell, K. Fielding, J. Sillah, B. Bah, P. Gustafson, K. Manneh,I. Lisse, G. Sirugo, S. Bennett, et al. 2006. Variants in the SP110 gene are as-sociated with genetic susceptibility to tuberculosis in West Africa. Proc. Natl.Acad. Sci. USA 103: 10364–10368.

30. Thye, T., E. N. Browne, M. A. Chinbuah, J. Gyapong, I. Osei, E. Owusu-Dabo,S. Niemann, S. Rusch-Gerdes, R. D. Horstmann, and C. G. Meyer. 2006. Noassociations of human pulmonary tuberculosis with Sp110 variants. J. Med.Genet. 43: e32.

31. Szeszko, J. S., B. Healy, H. Stevens, Y. Balabanova, F. Drobniewski, J. A. Todd,and S. Nejentsev. 2007. Resequencing and association analysis of the SP110gene in adult pulmonary tuberculosis. Hum. Genet. 121: 155–160.

32. Babb, C., E. H. Keet, P. D. van Helden, and E. G. Hoal. 2007. SP110 poly-morphisms are not associated with pulmonary tuberculosis in a South Africanpopulation. Hum. Genet. 121: 521–522.

33. Png, E., B. Alisjahbana, E. Sahiratmadja, S. Marzuki, R. Nelwan, I. Adnan,E. van de Vosse, M. Hibberd, R. van Crevel, T. H. Ottenhoff, and M. Seielstad.2012. Polymorphisms in SP110 are not associated with pulmonary tuberculosisin Indonesians. Infect. Genet. Evol. 12: 1319–1323.

34. Ruiz-Larranaga, O., J. M. Garrido, M. Iriondo, C. Manzano, E. Molina,I. Montes, P. Vazquez, A. P. Koets, V. P. Rutten, R. A. Juste, and A. Estonba.2010. SP110 as a novel susceptibility gene for Mycobacterium avium subspeciesparatuberculosis infection in cattle. J. Dairy Sci. 93: 5950–5958.

35. Yan, B. S., A. Kirby, Y. V. Shebzukhov, M. J. Daly, and I. Kramnik. 2006.Genetic architecture of tuberculosis resistance in a mouse model of infection.Genes Immun. 7: 201–210.

36. Yan, B. S., A. V. Pichugin, O. Jobe, L. Helming, E. B. Eruslanov, J. A. Gutierrez-Pabello, M. Rojas, Y. V. Shebzukhov, L. Kobzik, and I. Kramnik. 2007. Pro-gression of pulmonary tuberculosis and efficiency of bacillus Calmette-Guerinvaccination are genetically controlled via a common sst1-mediated mechanismof innate immunity. J. Immunol. 179: 6919–6932.

37. Dolganov, G. M., P. G. Woodruff, A. A. Novikov, Y. Zhang, R. E. Ferrando,R. Szubin, and J. V. Fahy. 2001. A novel method of gene transcript profiling inairway biopsy homogenates reveals increased expression of a Na+-K+-Cl2

cotransporter (NKCC1) in asthmatic subjects. Genome Res. 11: 1473–1483.38. Woodruff, P. G., H. A. Boushey, G. M. Dolganov, C. S. Barker, Y. H. Yang,

S. Donnelly, A. Ellwanger, S. S. Sidhu, T. P. Dao-Pick, C. Pantoja, et al. 2007.Genome-wide profiling identifies epithelial cell genes associated with asthmaand with treatment response to corticosteroids. Proc. Natl. Acad. Sci. USA 104:15858–15863.

39. Franken, K. L., H. S. Hiemstra, K. E. van Meijgaarden, Y. Subronto, J. denHartigh, T. H. Ottenhoff, and J. W. Drijfhout. 2000. Purification of his-taggedproteins by immobilized chelate affinity chromatography: the benefits from theuse of organic solvent. Protein Expr. Purif. 18: 95–99.

40. Lin, M. Y., A. Geluk, S. G. Smith, A. L. Stewart, A. H. Friggen, K. L. Franken,M. J. Verduyn, K. E. van Meijgaarden, M. I. Voskuil, H. M. Dockrell, et al. 2007.Lack of immune responses to Mycobacterium tuberculosis DosR regulon pro-teins following Mycobacterium bovis BCG vaccination. Infect. Immun. 75:3523–3530.

41. Commandeur, S., K. E. van Meijgaarden, M. Y. Lin, K. L. Franken,A. H. Friggen, J. W. Drijfhout, F. Oftung, G. E. Korsvold, A. Geluk, andT. H. Ottenhoff. 2011. Identification of human T-cell responses to Mycobacte-rium tuberculosis resuscitation-promoting factors in long-term latently infectedindividuals. Clin. Vaccine Immunol. 18: 676–683.

42. Commandeur, S., M. Y. Lin, K. E. van Meijgaarden, A. H. Friggen,K. L. Franken, J. W. Drijfhout, G. E. Korsvold, F. Oftung, A. Geluk, andT. H. Ottenhoff. 2011. Double- and monofunctional CD4+ and CD8+ T-cellresponses to Mycobacterium tuberculosis DosR antigens and peptides in long-term latently infected individuals. Eur. J. Immunol. 41: 2925–2936.

43. Fontan, P., V. Aris, S. Ghanny, P. Soteropoulos, and I. Smith. 2008. Globaltranscriptional profile of Mycobacterium tuberculosis during THP-1 humanmacrophage infection. Infect. Immun. 76: 717–725.

44. Malen, H., F. S. Berven, K. E. Fladmark, and H. G. Wiker. 2007. Comprehensiveanalysis of exported proteins from Mycobacterium tuberculosis H37Rv. Pro-teomics 7: 1702–1718.

45. Malen, H., S. Pathak, T. Søfteland, G. A. de Souza, and H. G. Wiker. 2010.Definition of novel cell envelope associated proteins in Triton X-114 extracts ofMycobacterium tuberculosis H37Rv. BMC Microbiol. 10: 132.

46. de Souza, G. A., N. A. Leversen, H. Malen, and H. G. Wiker. 2011. Bacterialproteins with cleaved or uncleaved signal peptides of the general secretorypathway. J. Proteomics 75: 502–510.

47. Rosenkrands, I., A. King, K. Weldingh, M. Moniatte, E. Moertz, andP. Andersen. 2000. Towards the proteome of Mycobacterium tuberculosis.Electrophoresis 21: 3740–3756.

48. Mattow, J., U. E. Schaible, F. Schmidt, K. Hagens, F. Siejak, G. Brestrich,G. Haeselbarth, E. C. Muller, P. R. Jungblut, and S. H. Kaufmann. 2003.Comparative proteome analysis of culture supernatant proteins from virulentMycobacterium tuberculosis H37Rv and attenuated M. bovis BCG Copenhagen.Electrophoresis 24: 3405–3420.

49. Wolfe, L. M., S. B. Mahaffey, N. A. Kruh, and K. M. Dobos. 2010. Proteomicdefinition of the cell wall of Mycobacterium tuberculosis. J. Proteome Res. 9:5816–5826.

50. Kruh, N. A., J. Troudt, A. Izzo, J. Prenni, and K. M. Dobos. 2010. Portrait ofa pathogen: the Mycobacterium tuberculosis proteome in vivo. PLoS ONE 5:e13938.

51. Gu, S., J. Chen, K. M. Dobos, E. M. Bradbury, J. T. Belisle, and X. Chen. 2003.Comprehensive proteomic profiling of the membrane constituents of a Myco-bacterium tuberculosis strain. Mol. Cell. Proteomics 2: 1284–1296.

52. Mawuenyega, K. G., C. V. Forst, K. M. Dobos, J. T. Belisle, J. Chen,E. M. Bradbury, A. R. Bradbury, and X. Chen. 2005. Mycobacterium tubercu-losis functional network analysis by global subcellular protein profiling. Mol.Biol. Cell 16: 396–404.

53. Sassetti, C. M., D. H. Boyd, and E. J. Rubin. 2003. Genes required for myco-bacterial growth defined by high density mutagenesis. Mol. Microbiol. 48: 77–84.

54. Sassetti, C. M., and E. J. Rubin. 2003. Genetic requirements for mycobacterialsurvival during infection. Proc. Natl. Acad. Sci. USA 100: 12989–12994.

55. Lamichhane, G., M. Zignol, N. J. Blades, D. E. Geiman, A. Dougherty,J. Grosset, K. W. Broman, and W. R. Bishai. 2003. A postgenomic method forpredicting essential genes at subsaturation levels of mutagenesis: application toMycobacterium tuberculosis. Proc. Natl. Acad. Sci. USA 100: 7213–7218.

56. Griffin, J. E., J. D. Gawronski, M. A. Dejesus, T. R. Ioerger, B. J. Akerley, andC. M. Sassetti. 2011. High-resolution phenotypic profiling defines genes es-sential for mycobacterial growth and cholesterol catabolism. PLoS Pathog. 7:e1002251.

57. Dutta, N. K., S. Mehra, P. J. Didier, C. J. Roy, L. A. Doyle, X. Alvarez,M. Ratterree, N. A. Be, G. Lamichhane, S. K. Jain, et al. 2010. Geneticrequirements for the survival of tubercle bacilli in primates. J. Infect. Dis. 201:1743–1752.

58. Frentsch, M., O. Arbach, D. Kirchhoff, B. Moewes, M. Worm, M. Rothe,A. Scheffold, and A. Thiel. 2005. Direct access to CD4+ T cells specific fordefined antigens according to CD154 expression. Nat. Med. 11: 1118–1124.

59. Ward, S. K., B. Abomoelak, S. A. Marcus, and A. M. Talaat. 2010. Transcrip-tional profiling of mycobacterium tuberculosis during infection: lessons learned.Front. Microbiol. 1: 121.

60. Rachman, H., M. Strong, U. Schaible, J. Schuchhardt, K. Hagens,H. Mollenkopf, D. Eisenberg, and S. H. Kaufmann. 2006. Mycobacterium tu-berculosis gene expression profiling within the context of protein networks.Microbes Infect. 8: 747–757.

61. Ramakrishnan, L., N. A. Federspiel, and S. Falkow. 2000. Granuloma-specificexpression of Mycobacterium virulence proteins from the glycine-rich PE-PGRSfamily. Science 288: 1436–1439.

62. Chan, K., T. Knaak, L. Satkamp, O. Humbert, S. Falkow, and L. Ramakrishnan.2002. Complex pattern ofMycobacterium marinum gene expression during long-term granulomatous infection. Proc. Natl. Acad. Sci. USA 99: 3920–3925.

63. McKinney, J. D., K. Honer zu Bentrup, E. J. Munoz-Elıas, A. Miczak, B. Chen,W. T. Chan, D. Swenson, J. C. Sacchettini, W. R. Jacobs, Jr., and D. G. Russell.2000. Persistence of Mycobacterium tuberculosis in macrophages and micerequires the glyoxylate shunt enzyme isocitrate lyase. Nature 406: 735–738.

64. Munoz-Elıas, E. J., and J. D. McKinney. 2005. Mycobacterium tuberculosisisocitrate lyases 1 and 2 are jointly required for in vivo growth and virulence.Nat. Med. 11: 638–644.

65. Harriff, M. J., G. E. Purdy, and D. M. Lewinsohn. 2012. Escape from thephagosome: the explanation for MHC-I processing of mycobacterial antigens?Front. Immunol. 3: 40.

66. Mogues, T., M. E. Goodrich, L. Ryan, R. LaCourse, and R. J. North. 2001. Therelative importance of T cell subsets in immunity and immunopathology ofairborne Mycobacterium tuberculosis infection in mice. J. Exp. Med. 193: 271–280.

67. Bold, T. D., and J. D. Ernst. 2012. CD4+ T cell-dependent IFN-g production byCD8+ effector T cells in Mycobacterium tuberculosis infection. J. Immunol. 189:2530–2536.

68. Chegou, N. N., G. F. Black, A. G. Loxton, K. Stanley, P. N. Essone, M. R. Klein,S. K. Parida, S. H. Kaufmann, T. M. Doherty, A. H. Friggen, et al. 2012. Po-tential of novel Mycobacterium tuberculosis infection phase-dependent antigensin the diagnosis of TB disease in a high burden setting. BMC Infect. Dis. 12: 10.

69. Jones, G. J., C. Pirson, H. P. Gideon, K. A. Wilkinson, D. R. Sherman,R. J. Wilkinson, R. G. Hewinson, and H. M. Vordermeier. 2011. Immuneresponses to the enduring hypoxic response antigen Rv0188 are preferentiallydetected in Mycobacterium bovis infected cattle with low pathology. PLoS ONE6: e21371.

70. Ottenhoff, T. H., and S. H. Kaufmann. 2012. Vaccines against tuberculosis:where are we and where do we need to go? PLoS Pathog. 8: e1002607.

The Journal of Immunology 1671

by guest on July 7, 2018http://w

ww

.jimm

unol.org/D

ownloaded from