RESEARCH ARTICLE◥

INFLAMMATION

mTOR regulates metabolic adaptationof APCs in the lung and controls theoutcome of allergic inflammationCharles Sinclair,1*† Gayathri Bommakanti,1* Luiz Gardinassi,2 Jens Loebbermann,1

Matthew Joseph Johnson,1,3 Paul Hakimpour,1 Thomas Hagan,1

Lydia Benitez,1 Andrei Todor,2 Deepa Machiah,4 Timothy Oriss,5 Anuradha Ray,5

Steven Bosinger,6 Rajesh Ravindran,1‡ Shuzhao Li,2 Bali Pulendran1,7,8§

Antigen-presenting cells (APCs) occupy diverse anatomical tissues, but their tissue-restricted homeostasis remains poorly understood. Here, working with mouse models ofinflammation, we found that mechanistic target of rapamycin (mTOR)–dependentmetabolic adaptation was required at discrete locations. mTOR was dispensable fordendritic cell (DC) homeostasis in secondary lymphoid tissues but necessary to regulatecellular metabolism and accumulation of CD103+ DCs and alveolar macrophages in lung.Moreover, while numbers of mTOR-deficient lung CD11b+ DCs were not changed, they weremetabolically reprogrammed to skew allergic inflammation from eosinophilic T helpercell 2 (TH2) to neutrophilic TH17 polarity. The mechanism for this change wasindependent of translational control but dependent on inflammatory DCs, whichproduced interleukin-23 and increased fatty acid oxidation. mTOR therefore mediatesmetabolic adaptation of APCs in distinct tissues, influencing the immunologicalcharacter of allergic inflammation.

Afundamental property of the immune sys-tem is its ability to sense pathogens throughpattern recognition receptors. Emergingevidence demonstrates that immune cellsalso sense environmental perturbations that

induce cellular stress and metabolic changes(1, 2). However, the mechanisms underlying suchenvironmental sensing are poorly understood.The amino acid sensor GCN2 plays a role inprogramming dendritic cells (DCs) to respondto viral vaccination (3, 4) and controls intestinal

inflammation by suppressing inflammasome ac-tivation in the gut (5). In addition to GCN2, themechanistic target of rapamycin (mTOR), a cen-tral orchestrator of cellular metabolism, regu-lates DC function (6).mTOR functions in two supramolecular com-

plexes termed mTORC1 and mTORC2, engagedto differing degrees by diverse stimuli, includingToll-like receptor signaling, growth factors, andnutrient availability. mTOR signaling affects di-verse downstream cellular processes, includingcell cycle progression, cell growth, differentia-tion, survival, and metabolism (7). AlthoughmTOR is required for effector T cell expansion(8) and germinal center B cell responses (9), itrestrains T memory generation (10). Within DCs,mTOR promotes type I interferon production byplasmacytoid DCs (11, 12) but limits proinflam-matory cytokine production by classical DCs (13).Moreover, mTOR promotes steady-state DC ac-cumulation downstream of homeostatic cytokinesignaling (14–16). The precise functions of mTORhave been proposed to be DC subset dependent(17), but the consequence of mTOR signaling inDCs and the effect on immune responses in vivoare poorly understood.

ResultsmTOR promotes lung APC homeostasisin vivo

We bred mice bearing floxedMtor with a CD11c-Cre deleter strain (mTORDAPC), resulting in spe-cific genetic ablation, reduced protein abundance,

and loss ofmTOR pathway activity in CD11c+ cells(fig. S1, A to D). mTOR inhibition augments DCproinflammatory cytokine production (13, 17),which we observed in mTORDAPC bone marrow(BM)–derived DCs and primary splenic DCs (fig.S1E), confirming functional deletion of mTOR.DCs can be subdivided into CD8a+/CD103+

and CD11b+ subsets. Globally, DCs require Flt3Land granulocyte macrophage colony–stimulatingfactor (GM-CSF) signals for homeostatic main-tenance (18, 19), both of which activate mTOR(14). In addition, antigen-presenting cells (APCs)occupy distinct microenvironments wheremTORmay exhibit site-specific functions (15, 20).mTORDAPC ablation did not exert global effectson APCs but instead revealed defects in APCaccumulation at specific anatomical sites. Therewas negligible requirement for cell-intrinsicmTOR in most tissues examined; however, splenicCD11b+ DCs were modestly reduced and therewas a skewed ratio of CD103+ to CD11b+ DCs inthymus (fig. S2). We also confirmed a substan-tial reduction in Langerhans cell frequency (15)(fig. S3). However, the most striking effect ofmTORDAPC deletion was observed in the lung.Here, CD103+ DC and CD11c+ alveolar macro-phage (AM) populations were significantly re-duced (Fig. 1A). Mixed BM chimeras showedthat homeostatic defects were cell intrinsic andadditionally revealed a minor impairment inCD11b+ lung DC accumulation (Fig. 1B).mTOR may simply regulate numerical accu-

mulation of APCs or, alternatively, may be centralto controlling lung APC programming. Transcrip-tomes derived from mTORDAPC lung APCs ver-sus wild-type (WT) controls clustered distinctlyby principal components analysis (PCA), whereasmTORDAPC splenic populations clustered closelywith theirWTcounterparts (fig. S4). Previous studieshave identified core sets of lineage-specifyinggenes, enriched in either CD11b+ lung APCs orAMs, relative to other APC subsets (21, 22).mTORDAPC AMs exhibited a broad loss of AMmolecular identity (fig. S5A). Rather than disrupt-ing the CD11b+ APC cellular identity, mTOR ab-lation skewed the balance of CD11c+CD11b+ lungAPCs toward a macrophage/monocytic composi-tion (fig. S5B) (22), typified by reduced heat-stableantigen (HSA), Flt3, and a trend of lower Irf4expression, with elevatedMerTK and CD64 tran-scription (Fig. 1C). Using an alternative gatingscheme (22), we found that CD11c+MHC-II+Sirpa+

HSA+ DCs were reduced in mTORDAPC mice (Fig.1D) due to cell-intrinsic defects in accumulation(Fig. 1E). Thus,mTORplays a central, cell-intrinsicrole in APC homeostatic accumulation and subsetcomposition in the lung.Steady-state DC homeostasis depends on cell

development, proliferation,migration, and deathrates. Ablation or inhibition of mTOR increaseddeath rates of lung APCs (Fig. 2A and fig. S6, Aand B), whereas precursor frequency, prolifera-tion rates, andmigrationwere unaffected (fig. S6, Cto G). There was no evidence that GM-CSF,M-CSF,or Flt3L expression was perturbed in mTORDAPC

mice, excluding a role for other cell-extrinsic ho-meostatic factors (fig. S6H). mTOR associates with

RESEARCH

Sinclair et al., Science 357, 1014–1021 (2017) 8 September 2017 1 of 8

1Emory Vaccine Center, Emory University, 954 GatewoodRoad NE, Atlanta, GA 30329, USA. 2Department of Medicine,Emory University, Atlanta, GA 30329, USA. 3Department ofPediatrics, University of Minnesota, Minneapolis, MN 55455,USA. 4Yerkes Molecular Pathology Core Laboratory, YerkesNational Primate Research Center, 954 Gatewood Road NE,Atlanta, GA 30329, USA. 5Division of Pulmonary, Allergy andCritical Care Medicine, Department of Immunology,University of Pittsburgh Asthma Institute at University ofPittsburgh Medical Center (UPMC), University of PittsburghSchool of Medicine, Pittsburgh, PA, USA. 6Primate GenomicsCore, Yerkes National Primate Research Center, 954Gatewood Road NE, Atlanta, GA 30329, USA. 7Department ofPathology and Laboratory Medicine, Emory University,Atlanta, GA, USA. 8Institute for Immunity, Transplantationand Infection, Department of Pathology, Department ofMicrobiology and Immunology, Stanford University, Stanford,CA 94305, USA.*These authors contributed equally to this work. †Present address:AstraZeneca, Cancer Research UK (CRUK) Cambridge Institute, Li KaShing Center, University of Cambridge, Robinson Way, CambridgeCB2 0RE, UK. ‡Present address: Genomics Institute of the NovartisResearch Foundation, 10675 John Jay Hopkins Drive, San Diego, CA92121, USA. §Corresponding author. Email: bpulend@emory.edu;bpulend@stanford.edu

on March 6, 2019

http://science.sciencem

ag.org/D

ownloaded from

two molecular complexes (mTORC1/2). Whereasmice lacking the critical mTORC1 protein Raptorphenocopiedhomeostatic defects seen inmTORDAPC

lung, APCs were unaffected in mTORC2-deficientRictorDAPC mice (fig. S7A). mTORC1 signaling tar-gets the translational regulator S6K1/2 and deac-tivates the translational repressor 4E-BP1/2 (fig.S7, B and C) (7). Curiously, S6k1−/−S6k2−/− miceexhibited normal lung APC homeostasis, and ge-netic deletion of 4ebp1/2 did not rescue the sur-vival of mTORDAPC DCs (fig. S7, D and E). Thus,mTORC1 controls APC cell survival through atranslation-independent mechanism.

APCs in the lung display a distinctivemTOR-dependent metabolic signature

The pronounced phenotype of APCs in mTORDAPC

lung relative to other tissues was striking, partic-ularly considering that development of CD103+

DCs in both lung and other tissues is regulatedby the common transcription factor Batf3 (23).We hypothesized that the lung microenviron-mentmay imprint a distinctivemetabolic pheno-typewith heightened sensitivity tomTORablation.To investigate this, comprehensivemetabolic pro-filing was performed on lung and splenic APCsfrom WT mice. Lung and spleen APCs weremetabolically distinct (Fig. 2B), exhibiting dif-

ferences in amino acid and nucleic acid meta-bolic intermediates (fig. S8, A and B), which werealso reflected within observed transcriptionalsignatures (fig. S8, C and D). We next examinedmTORDAPCmetabolomes, and the degree ofmeta-bolic perturbation inmTOR-deficientAPC subsetscorrelated with the extent of their homeostaticdisruption (Fig. 2C). Despite the skewing in thephenotypic composition of the lung CD11b+ DCs(Fig. 1D), there was a minor perturbation in themetabolome (Fig. 2C). Pathway analysis of dif-ferentially abundant metabolites revealed thatseveral amino acid metabolic pathways weredisrupted in mTORDAPC APCs (Fig. 2, D and E).mTORDAPC AMs also exhibited alterations inlipid metabolite expression and correspondingchanges in expression of the sterol regulatorybinding protein (Srebp1/2) target genes (24)(fig. S9). Srebp1 orchestrates lipid anabolismand is a downstream target of mTOR, whichdirectly links mTOR deficiency and metabolicperturbations in lung APCs (25, 26). To test thisexperimentally, we administered the Srebp1/2inhibitor fatostatin, which substantially decreasedAM frequency, a phenotype that was recapitu-lated by rapamycin treatment (fig. S10A). How-ever, CD103+ DCs remained unaltered by eitherinhibitor, suggesting that inhibitor administration

was insufficient to entirely phenocopy mTORDAPC

mice, or that the Srebppathways are not critical forCD103DC survival. Finally,we investigated the roleof another mTOR target, the glycolytic regulatorHif1a, observing that lung APC homeostasis wasindependent of Hif1a (fig. S10B). Thus, lung APCshave distinct mTOR-dependent metabolic sig-natures, and there appears to be a direct linkbetweenmTORand Srebp-dependent lipidmetab-olism in AMs.

Ablation of mTOR in APCs reprogramsallergic inflammation in the lung

Wenext investigatedwhether lung immunity wasfunctionally affected in mTORDAPC mice. Micewere intranasally challenged with house dust mite(HDM) allergen, which is an established murinemodel for asthma (27, 28) (Fig. 3A). Despite re-duced frequencies of select lung APC subsets,humoral responses in mTORDAPC mice were un-expectedly increased (Fig. 3B and fig. S11A). Sur-prisingly, we observed a striking reduction in Thelper cell 2 (TH2) cells producing interleukin-4

+

(IL-4+) and IL-5+/IL-13+ and enhanced frequen-cies of TH17 cells in the lung and bronchoalveolarlavage (BAL) (Fig. 3C and fig. S11B).TH2-polarized allergy is typified by robust

recruitment of eosinophils, whereas IL-17 can

Sinclair et al., Science 357, 1014–1021 (2017) 8 September 2017 2 of 8

Fig. 1. Critical require-ment for mTOR signalingto maintain lung innateimmune homeostasis.Lungs from naïve WT(black) and mTORDAPC (red)mice were analyzed by flowcytometry. (A) Bar chartsshow mean frequencies ±SEM of indicated lung APCpopulations; contour plotsshow representative flowcytometric data. Red arrowsindicate gating hierarchies.(B) Mixed chimeras wereestablished by reconstitutingB6 CD45.1+/CD45.2+ hostswith donor mTORDAPC

(CD45.2) and B6 CD45.1 BMat a ~3:1 ratio. Bar chartshows the percentage con-tribution to chimerism ofmTORDAPC-derived donorcells after exclusion ofresidual host cells. Gray barsindicate control (CD11c–

populations); red bars indi-cate CD11c+ APCs. (C) Barcharts show normalized readcount for indicated genes inWT (black) and mTORDAPC

(red) within the CD11c+CD11b+

lung DC population.(D) Contour plot showsexpression of Sirpa and HSA.Pie chart shows average fre-quencies of indicated populations from one representative experiment (n = 4). (E) Bar chart shows the chimeric contribution of mTORDAPC-derived donor cells.Gray bars denote control CD11c– populations; red bars indicate CD11c+ APCs. Data represent three or more independent experiments.

0

500

1000

1500

2000

0

200

400

600

800

0

500

1000

1500

2000

0

1000

2000

3000

4000

0

500

1000

1500

B-cell

s

CD11c-C

D11b+

CD11b+ D

C

CD103+ D

C AM0

20

40

60

80

100

CD

11c

MHC-IIC

D11

c

Siglec-F

CD45+

(AM excluded)

CD11chi

MHC-II+

CD45+

WT mTOR APC

% C

him

eris

m

Lung

CD

103

CD11b

9.1 1.8

98.090.8

1.0 1.1

31.4 9.1

63.7 81.2

WTmTOR APC

Alveolarmacrophage

0

5

10

15

% o

f CD

45

+

0.0

0.1

0.2

0.3

0.4

CD103 + DC0.0

0.5

1.0

1.5

2.0

CD11b + DC

*******

****

********

Sirp

HSA

WT mTOR APC

24.5 5.7

38.5 27.131.2 62.1

Sirp +HSA-

Sirp +HSA+

Sirp -HSA+

B-cell

s

HSA-Sirp

a+

HSA+Sirp

a+

HSA+Sirp

a-0

20

40

60

80

100

****

***

n.s.

% C

him

eris

m

**

MerTK CD64HSA Flt3 Irf4

Nor

mal

ized

re

ads

*

n.s.**p=0.5

RESEARCH | RESEARCH ARTICLEon M

arch 6, 2019

http://science.sciencemag.org/

Dow

nloaded from

enhance neutrophils. Neutrophilia occurs in ~50%of human asthma cases, correlating with steroidresistance and worsening prognosis (29, 30). Re-markably, we observed pronounced neutrophiliain mTORDAPC lung and BAL (Fig. 3D). SkewedTH2 to TH17 modality was further observed inan alternative papain-induced model (fig. S12),whereas an anatomically discrete TH17-dependentexperimental autoimmune encephalomyelitismodel was not significantly affected by mTORdeficiency in APCs (fig. S13). TH17 allergic pheno-types were driven by mTORC1 ablation (fig. S14A)but were independent of Hif1a (fig. S14B) orSrebp1/2 (fig. S14C). mTORC1 thus regulatesinnate programming of T helper responses in

lung, controlling the balance between eosino-philic and neutrophilic mediators of allergicinflammation.

mTOR-deficient CD103+ DCs and AMsare dispensable for neutrophilic allergy

We next determined the APC subset responsi-ble for promoting TH17/neutrophilic allergy inmTORDAPC mice. CD103+ DCs have importantantiviral functions due to a superior ability tocross-present antigens (31). Consistent with this,mTORDAPC mice exhibited impaired CD8+ T cellresponses to live-attenuated influenza vaccinechallenge (Fig. 4A). If CD103+ DCs restrain TH17/neutrophilic allergy, then we should observe a

similar phenotype inBatf3−/−mice, which lack atranscription factor required for CD103+ DC de-velopment (Fig. 4B). However, Batf3−/−mice didnot exhibit differential severity or modality ofinflammation (Fig. 4, C to E), suggesting thatimpaired development of CD103+ DCwas not nec-essary to promote neutrophilic inflammation.AMs reside in the lung alveolar lumen and

play a primary role in phagocytizing particulatesand turning over lung surfactant. Defects in AMresult in accumulated lung surfactant protein,termed pulmonary alveolar proteinosis (PAP)(32). Histological analysis of naïve mTORDAPC

lung showed airspace filling by eosinophilicmaterial (Fig. 5A), elevated mucus (Fig. 5B),

Sinclair et al., Science 357, 1014–1021 (2017) 8 September 2017 3 of 8

PC1

–300200

1000

100200

300400

500600

PC2

400300

200100

0100

200300

PC3

200

100

0

100

200

300

400

*+,-."1##8 2 34566,."1##8 2

CD11b + - 77 metabolites

CD103 + - 1026 metabolites

Xenobiotics metabolism

Vitamin E metabolismUrea cycle / amino group metabolism

Tyrosine metabolism Tryptophan metabolism

Linoleate metabolismLeukotriene metabolism

Glyoxylate and Dicarboxylate metabolismGly, Ser, Ala, Thr metabolism

Butanoate metabolism

Bile acid biosynthesisAspartate and asparagine biosynthesis

TCA cycle

WT

mTO

RAPC

WT

mTO

RAPC

WT

mTO

RAPC

WT vs KOCD11b

WT vs KOCD103

WT vs KOAM

Lung CD11b+

Lung CD103+

Lung AMSpleen CD8a+

Spleen CD11b+

2 3 4 2 3 4 2 3 4

-Log10 p

WTmTOR APC5.7 3.5

3.7

21.3

37.1

31.5

43.5

13.6

35.13.1

LungCD8+ T-cell

ex vivo

Splenic CD8 +

Splenic CD11b+

LungCD11b+

LungCD103+

LungAM

4.0

9.3

L/D Aqua

-1.5 1.50

Row Z-score

AM - 178 metabolites

Fig. 2. mTOR regulates lung APC survival via a metabolic mechanism.Lung from WT (black) or mTORDAPC (red) mice was examined by flowcytometry and by cellular metabolomic profiling. (A) Histograms showrepresentative staining of Live/Dead Aqua on indicated ex vivo populationsfrom WT (gray) and mTORDAPC (red) lung. (B) PCA plot shows distinctclustering of lung and splenic APC metabolomes from WTmice. (C) Heatmaps show differentially abundant metabolite features between WTandmTORDAPC subsets.The metabolite numbers include redundant mass spectral

peaks. (D) Dot graph showsmetabolic pathways distinctly perturbed bymTORdeficiency, with amino acid metabolism pathways highlighted in red. Dotsize relates to the number of differentially expressed metabolites. (E) Networkanalysis reveals a higher representation of tentative amino acid metaboliteswithin the mTORDAPC CD103+ DC subset. Each link is from a known metabolicreaction.The network and tentative metabolite annotation was produced bythe Mummichog software. Data represent three or more independentexperiments or two to three biological replicates.

RESEARCH | RESEARCH ARTICLEon M

arch 6, 2019

http://science.sciencemag.org/

Dow

nloaded from

accumulation of lung surfactant protein (Fig.5C), and increased debris and “foamy” macro-phages in BAL (Fig. 5D). Increasedproteinaceousmatter was also evident in lung preparations byflow cytometry (Fig. 5E). These data representhallmark features of PAP disease (33), represent-ing a first link between mTOR signaling andPAP. We next determined whether functionalloss of AMs also contributes to neutrophilic lunginflammation. Using a previously described pro-tocol (34), we observed ~75% depletion of AMs3 days after a single intratracheal (i.t.) dose ofclodronate liposomes but not control liposomes(Fig. 5F). After HDM challenge, humoral immu-nity was enhanced when AMs were depleted, inconcordance with the phenotype observed inmTORDAPC mice (Fig. 5G). However, TH17 po-

larization and neutrophilia were not present un-der these conditions (Fig. 5, H and I), consistentwith observations that AMs cannot present anti-gen (28). Thus, absence of AM likely contributesto an enhancedmagnitude of humoral immunityin mTORDAPC mice but cannot account for TH17/neutrophilic allergy.

mTOR deficiency reprogramsinflammatory CD11b+ DCs to promoteneutrophilic allergy

CD11b+ DCs directly promote TH2 allergy afterHDM insult (28, 35) but can also prime TH17immunity when activated by alternative stimuli(22, 36). To address whether mTORDAPC CD11b+

DCs promoted neutrophilic allergy in a domi-nant manner, we first generated mixed chime-

ras where donor WT andmTORDAPC bone marrowwas injected at a 1:10 ratio to allow repopula-tion of CD103+ DC and AM populations by WTcells, whereas ~80% of CD11c+CD11b+ DCs re-mained mTOR-deficient (fig. S15, A and B). HDMpromoted elevated humoral responses and pro-nounced neutrophilia in chimeras (fig. S15, Cand D). Moreover, primed CD11b+ DCs frommTORDAPCmice directly promoted TH17/neutrophilicallergy in WT recipients (fig. S16). Thus, neutro-philic inflammation was dominantly conferredby mTORDAPC CD11b+ DCs.It was recently shown that mice deficient in

Kruppel-like factor 4 (Klf4) have reduced fre-quencies of the Sirpa+HSA+ subset in lung CD11b+

DCs and display impaired TH2 responses to HDM(37).We thus investigated howmTORDAPC CD11b+

Sinclair et al., Science 357, 1014–1021 (2017) 8 September 2017 4 of 8

d0, 2, 4 d7, 9, 11 d14, 15, 16

Bleed:d16

Analysis:d17

1

10

100

1000

10000

100000

1

10

100

WTN

aive

mTOR

APCNaiv

eWT

mTOR

APC

HDM-specific Ig G1

A.U

.

ng/m

L

Total IgE

0.0

0.5

1.0

1.5

2.0

2.5

0.0

0.5

1.0

1.5

MedLN

IL-1

7A+

%of

CD

4

0

1

2

3

Lung

0

5

10

15

20

0

5

10

15

20

0

10

20

30

IL-4

+%

of C

D4

IL-1

7A+

%of

CD

4IL

-4+

%of

CD

4

IL-1

7+

%of

CD

4IL

-4+

%of

CD

4

CD

44

IL-4

4.9

WT mTOR APC

0.9

CD

44

IL-17A

3.5 19.1

BAL

CD

44

IL-4

WT mTOR APC

CD

44

IL-17A

26.1 5.4

7.1 33.7

WT

WTmTOR APC

mTOR APC

WTN

aive

mTOR

APCNaiv

eWT

mTOR

APC

0

20

40

60

80

0

10

20

30

Sig

lec-

F

Ly-6G

73.1

WT mTOR APC

41.6

Lung

BAL

0

20

40

60

80

100

Eos

inop

hils

% o

f CD

11b+

Eos

inop

hils

% o

f CD

11b+

Neu

trop

hils

% o

f CD

11b+

Neu

trop

hils

% o

f CD

11b+

0

5

10

15

20

25

Sig

lec-

F

Ly-6G

8.5 27.2

97.1

1.1

59.2

35.9

**** *

*

****

*

****

***

****

*****

**** *

25 µg House dust mite extract. i.n.

Fig. 3. mTOR ablation in APCs results in neutrophilic lung inflamma-tion. House dust mite (HDM) extract was administered to WT (black)or mTORDAPC (red) mice. (A) Diagram shows the allergen administration andanalysis schedule. i.n., intranasal. (B) Scatter plots show serum antibodytiters in indicated mouse strains. (C) Bar charts show the frequency of

CD4+IL-4+ TH2 cells or CD4+IL-17A+ TH17 cells in indicated organs. Contourplots show representative flow cytometric data. (D) Bar charts show thefrequency of CD11b+Siglec-F+ eosinophils or CD11b+Ly-6G+ neutrophils.Pseudocolor plots show representative flow cytometric data. Data representthree or more independent experiments.

RESEARCH | RESEARCH ARTICLEon M

arch 6, 2019

http://science.sciencemag.org/

Dow

nloaded from

lung DCs differentially responded to allergy.HDMwas instilled i.t., andmice were examinedafter 3 days. mTORDAPC CD11b+ migratory DCsaccumulated with enhanced frequency in medias-tinal lymph node (MedLN) (fig. S17A), and cellularinflammationwas enhanced in the lung (fig. S17B).During allergic airway inflammation, a distinctCD11b+ Csf-1–dependent inflammatory DC popu-lation develops from monocytes (28, 38). Inflam-

matory DCs could be identified in allergic lungbased on their surface CD11c+MHC-II+CD11b+

phenotype and Csf1 dependence (fig. S18, A andB) (28) andwere elevated inmTORDAPCmice (Fig.6A). Lung inflammatory DCs were confirmed tobe distinct from lung-resident interstitial macro-phages, AMs, and classical DCs. They exhibitedlower expression of F4/80, higher expression ofMHC-II and CD11c compared to interstitial mac-

rophages (fig. S18C), lower Siglec-F expression thanAMs (fig. S18D), and high expression of CD64 andLy6C relative to resident CD11b+ DCs (fig. S18E).ComparedwithWT,mTORDAPC inflammatoryDCshad higher expression of lineage-associatedmarkers CD64 and Ly6C and elevated CD80and CD86 costimulatorymolecule expression (fig.S18, E and F), consistent with a more progressivedifferentiation/activation status. To directly testthe requirement for inflammatory DCs in TH17/neutrophilic inflammation, we ablated this pop-ulation by administering aCsf1r neutralizing anti-bodies, which reverted mTORDAPC allergy to TH2/eosinophilic modality (Fig. 6B and fig. S19).Given the observed mTOR-dependent meta-

bolic phenotypes in steady-state lung APCs, weasked whether there was a metabolic basis forthe TH17 allergic phenotype. Despite minimaldifferences between resting classical CD11b+ DCsin lungs of WT and mTORDAPC mice, there was astriking perturbation in the metabolome of in-flammatory DCs (fig. S20). In addition to differ-ences in amino acid metabolism found in otherDC subsets, we detected elevated fatty acid (FA)metabolites (Fig. 6C). mTOR inhibition in BMDCscan enhance fatty acid oxidation (FAO) and there-by regulate DC life-span/cytokine expression pro-files (39, 40). However, a physiological role forthis mechanism in primary DCs has not beendescribed. To directly investigate whether ele-vated FAO underlies enhanced inflammatoryDC activity, we treated mTORDAPC mice withetomoxir, an inhibitor of the rate-limiting FAOenzyme carnitine palmitoyltransferase 1 (Cpt1)(41). Etomoxir significantly reduced the activa-tion phenotype of mTORDAPC inflammatory DCswithout affecting WT (Fig. 6D) and also reducedTH17 polarization in mTORDAPC lung (Fig. 6E).These observations are consistent with a modelin which mTOR deletion in inflammatory lungDCs increases FAO to promote TH17-polarizingactivity.Finally, we investigated whether dysregulated

cytokine production by inflammatory DCs influ-ences allergic polarity. The proinflammatory cy-tokines IL-6 and IL-1b function upstream of IL-23to promote a TH17-polarizing cascade. Analysisof CD11b+ DCs in lung and MedLN 3 days afteradministration of HDM revealed evidence ofelevated IL-12/23p40, IL-6, and IL-1b, all ofwhich are associated with TH17 polarization andneutrophilia (fig. S21, A and B) (29). IL-23 neu-tralization reverted lung neutrophilia toward aneosinophilic phenotype (Fig. 6F), concomitantwith restoration in T cells of TH2 polarity (fig.S21C), identifying this cytokine as a key immu-nological player in mTORDAPC-dependent neu-trophilic inflammation.

Discussion

Here, we demonstrate a striking heterogeneity inthe metabolic profiles of APCs in distinct tissuesand reveal a key role for mTOR in programmingthe homeostasis and function of APC subsets inthe lung. In the steady state, mTORDAPC miceexhibited tissue-dependent phenotypes with re-duced numbers of AMs and CD103 DCs in the

Sinclair et al., Science 357, 1014–1021 (2017) 8 September 2017 5 of 8

WTBatf3-/-

IFN

+ %

of C

D8+

PMA + Ionomycin

NP366-374

MedLN Lung

IFN

+ %

of C

D8+

WT mTOR APCMedLN Lung

CD

44

IFN

18.8

WT mTOR APC

22.013.3 12.4

CD

44

IFN

0.1 1.3<0.1 0.5

****

0

5

10

15

20

25

MedLN

****

0

5

10

15

20

25

Lung

0.0

0.5

1.0

1.5

0

0

1

1

0.0

0.5

1.0

1.5

*

*ns

1

10

100

1000

10000

100000

0.1

1

10

100

WT

Batf3-/-

WT

Batf3-/-

A.U

.

ng/m

L

0

2

4

6

8

IL-1

7A+ %

of C

D4

WT Batf3-/-

CD

103

CD11b

CD

8

CD11b

Lung

Spleen

6.3

69.1

0.7

79.0

23.7 6.0

67.3 86.6

0.0

0.5

1.0

1.5

2.0

2.5

IL-4

+ %

of C

D4

Lung

BAL

0

20

40

60

80

0

10

20

30

0

20

40

60

80

100

0

5

10

15

20

25

Eos

inop

hils

%

of C

D11

b+

Neu

troph

ils

% o

f CD

11b+

Eos

inop

hils

%

of C

D11

b+

Neu

troph

ils

% o

f CD

11b+

IL-1

7A+ %

of C

D4

IL-4

+ %

of C

D4

0

10

20

30

0

5

10

15

HDM-specific IgG1Total IgE

*

Fig. 4. mTOR-dependent CD103+ DCs are dispensable for neutrophilic lung inflammation butrequired for antiviral immunity. (A) Flumist Quadrivalent vaccine (25 mL) was administeredintranasally to WT (black) or mTORDAPC mice (red). Vaccine response was examined 7 days afterinfection. Bar charts show the frequency of CD8+IFNg+ cells in indicated organs. Contour plotsshow representative flow cytometric data. (B to E) HDM was chronically administered to WT(black) and Batf3−/− (gray) mice. (B) Contour plots show the absence of CD103+ and CD8a+ DCsin lung and spleen of Batf3−/− mice. (C) Serum enzyme-linked immunosorbent assays (ELISAs)showing total IgE or HDM-specific IgG1 antibody titers. (D) Bar charts show TH2 (IL-4+) andTH17 (IL-17A+) frequencies in lung and BAL. (E) Bar charts show frequencies of eosinophils andneutrophils in lung and BAL. (A) Data represent three or more independent experiments. [(B) to(E)] n = 5 mice per group.

RESEARCH | RESEARCH ARTICLEon M

arch 6, 2019

http://science.sciencemag.org/

Dow

nloaded from

lung and skin, but not elsewhere, unlike moreglobal phenotypes observed when GM-CSF orFlt3l signaling is impaired (14, 18, 38). Theseeffects were mediated by Raptor-dependentmTORC1 signaling but, surprisingly, were inde-pendent of translational regulation of mRNA, thecanonical signaling pathway downstream ofmTOR. Instead, our data reveal a role for Srebp1/2 signaling downstream of mTOR in homeosta-sis of lung APCs in the steady state (fig. S10). Theselective effects of mTOR deficiency in CD103DCs in the lung and skinwere surprising, becausethe development of this DC subset in diverse tis-sues is controlled by the transcription factor Batf3.Thus, these results suggest that the homeostasis

of DCs in tissues is orchestrated by the superim-position of ametabolic programon the transcrip-tional network that regulates development.In addition to these effects in the steady state,

mTOR deficiency also altered the character ofallergic inflammation induced by lung CD11b+

DCs, skewing it toward the TH17/neutrophilicphenotype. This phenotype closely parallels as-pects of human disease, in which neutrophiliaassociates with higher clinical severity andworseprognosis in asthma (29). Interestingly, the al-lergic phenotype observed in mTORDAPC micewas associated with elevated immunoglobulinE (IgE). Such elevation in serum IgE is alsoobserved in neutrophilic asthma patients (42)

and can be enhanced by IL-23 (43), which isconsistent with the IL-23 dependence of themTORDAPC neutrophilic inflammation pheno-type. In normal circumstances, mTOR signal-ing in activated BMDCs suppresses oxidativephosphorylation and is in part responsible for aswitch to glycolytic metabolism, which rapidlydepletes intracellular energy reserves and as-sociates with a shortened cellular life span (1).In contrast, mTOR inhibition allows BMDCs tosupply their energetic demands through a combi-nation of glycolysis and FAO, increasing cellsurvival and proinflammatory phenotypes (39).We observed elevated FAO and increased pro-duction of several proinflammatory cytokines in

Sinclair et al., Science 357, 1014–1021 (2017) 8 September 2017 6 of 8

Fig. 5. mTOR ablation inAMs results in PAP andenhances the antibodymagnitude but not modal-ity of allergy. (A toC) Images show histo-pathological sections of WTor mTORDAPC lung stainedwith (A) hematoxylin andeosin, (B) PAS, (C) immuno-histochemical antibodiesagainst surfactant protein D.(D) BAL cellular fractionswere prepared by cytospinand analyzed after Diff-Quikstaining. Slides are shown at[(A) to (C)] x200 magnifica-tion, where scale bars repre-sent 100 mm and (D) x400magnification, where scalebars represent 50 mm.(E) Pseudocolor plots showthe frequency of Live/DeadAqua+ FSClo (dead cell/debris) events as determinedby flow cytometry. (F toI) HDM was administered toindicated groups for 17 days.Mice were treated with indi-cated clodronate or controlliposome on days –3, 3, and10. (F) AM frequency on day0 before HDM administration.(G) Serum ELISAs showingtotal IgE or HDM-specific IgG1antibody titers. (H) Barcharts show the frequency ofIL-4+ and IL-17+ T cells inindicated groups and organs.(I) Bar charts show the fre-quency of eosinophils andneutrophils in indicatedgroups and organs; pseudo-color plots show representa-tive data. Data represent[(A) to (D)] three or moreslides from two or morebiological replicates, [(E) to(I)] two or more representa-tive experiments.

WT mTOR APCL/

D A

qua

FSC

24.7 52.8

WT mTOR APC

0

2

4

6

8

WT

WT

+ Con

trol

WT

+ Clod

.

**

WT mTOR APC

AM

% C

D45

+

WT + Control WT + ClodronatemTORAPC

0.0

0.1

0.2

0.3

0.4

0.5

Lung

BAL

IL-4

+ %

of C

D4

IL-4

+ %

of C

D4

0

10

20

30

40

*******

0

5

10

15

20

IL-1

7+ %

of C

D4

IL-1

7+ %

of C

D4

0

10

20

30

40

*** **

* *

0

20

40

60

Eos

inop

hils

% C

D11

b+

0

20

40

60

80

100 *** ***

0

10

20

30

40

Neu

trop

hils

% C

D11

b+N

eutr

ophi

ls %

CD

11b+

0

10

20

30

40

50

** ***

Eos

inop

hils

% C

D11

b+

******

Sig

lec-

F

Ly-6G

mTOR APC WT + Clod.

Sig

lec-

F

Ly-6G

11.4

18.6

30.9

32.0

74.3

5.3

23.6

7.9

100

1000

10000

100000

1

10

100

ng/m

L

AU

Total IgE HDM-specific IgG1

WT

+ Con

trol

WT

+ Clod

.

WT

+ Con

trol

WT

+ Clod

.

***

**

***

RESEARCH | RESEARCH ARTICLEon M

arch 6, 2019

http://science.sciencemag.org/

Dow

nloaded from

mTORDAPC lung. Indeed one of these cytokines,IL-23, plays an obligate role in TH17/neutrophilicinflammation. Together, our study reveals a keyrole for mTOR in programming the homeostasisand function of APCs in the lung. In the steadystate, mTOR supports accumulation of lung sub-sets by promoting lipid synthesis (anabolism)(fig. S22A). In the context of allergic inflamma-tion, mTOR restrains catabolic processes suchas FAO, preventing excessive inflammation (fig.S22B). The results herein identify tissue-restrictedmetabolic adaptations of APCs as a critical deter-minant of immune regulation, raising the prospectof metabolic reprogramming of such tissue-resident APCs for therapeutic benefit in variousinflammatory disorders.

REFERENCES AND NOTES

1. E. J. Pearce, B. Everts, Nat. Rev. Immunol. 15, 18–29(2015).

2. B. Pulendran, Annu. Rev. Immunol. 33, 563–606 (2015).3. T. D. Querec et al., Nat. Immunol. 10, 116–125 (2009).4. R. Ravindran et al., Science 343, 313–317 (2014).5. R. Ravindran et al., Nature 531, 523–527 (2016).6. N. Sukhbaatar, M. Hengstschläger, T. Weichhart, Trends

Immunol. 37, 778–789 (2016).7. M. Laplante, D. M. Sabatini, Cell 149, 274–293 (2012).8. F. J. Dumont, M. J. Staruch, S. L. Koprak, M. R. Melino,

N. H. Sigal, J. Immunol. 144, 251–258 (1990).9. R. Keating et al., Nat. Immunol. 14, 1266–1276 (2013).10. K. Araki et al., Nature 460, 108–112 (2009).11. R. Colina et al., Nature 452, 323–328 (2008).12. W. Cao et al., Nat. Immunol. 9, 1157–1164 (2008).13. T. Weichhart et al., Immunity 29, 565–577 (2008).14. T. Sathaliyawala et al., Immunity 33, 597–606 (2010).15. B. Kellersch, T. Brocker, Blood 121, 298–307 (2013).16. S. P. Nobs et al., Immunity 43, 674–689 (2015).17. M. Haidinger et al., J. Immunol. 185, 3919–3931 (2010).18. H. J. McKenna et al., Blood 95, 3489–3497 (2000).19. D. Vremec et al., Eur. J. Immunol. 27, 40–44 (1997).20. S. P. Nobs et al., Immunity 43, 674–689 (2015).21. E. L. Gautier et al., Nat. Immunol. 13, 1118–1128 (2012).22. A. Schlitzer et al., Immunity 38, 970–983 (2013).23. B. T. Edelson et al., J. Exp. Med. 207, 823–836 (2010).24. D. Eberlé, B. Hegarty, P. Bossard, P. Ferré, F. Foufelle,

Biochimie 86, 839–848 (2004).25. T. Porstmann et al., Cell Metab. 8, 224–236 (2008).26. M. Laplante, D. M. Sabatini, Curr. Biol. 19, R1046–R1052

(2009).27. S. Saglani et al., Am. J. Respir. Cell Mol. Biol. 41, 281–289

(2009).28. M. Plantinga et al., Immunity 38, 322–335 (2013).29. J. F. Alcorn, C. R. Crowe, J. K. Kolls, Annu. Rev. Physiol. 72,

495–516 (2010).30. J. Douwes, P. Gibson, J. Pekkanen, N. Pearce, Thorax 57,

643–648 (2002).31. J. Helft et al., J. Clin. Invest. 122, 4037–4047 (2012).32. J. F. Seymour, J. J. Presneill, Am. J. Respir. Crit. Care Med. 166,

215–235 (2002).33. P. L. Shah, D. Hansell, P. R. Lawson, K. B. Reid, C. Morgan,

Thorax 55, 67–77 (2000).34. S. A. Mathie et al., Allergy 70, 80–89 (2015).35. Y. Gao et al., Immunity 39, 722–732 (2013).36. E. K. Persson et al., Immunity 38, 958–969 (2013).37. R. Tussiwand et al., Immunity 42, 916–928 (2015).38. M. Greter et al., Immunity 36, 1031–1046 (2012).39. E. Amiel et al., J. Immunol. 193, 2821–2830 (2014).40. E. Amiel et al., J. Immunol. 189, 2151–2158 (2012).41. S. M. Houten, R. J. Wanders, J. Inherit. Metab. Dis. 33,

469–477 (2010).42. M. Milovanovic, G. Drozdenko, C. Weise, M. Babina, M. Worm,

J. Invest. Dermatol. 130, 2621–2628 (2010).43. J. Peng, X. O. Yang, S. H. Chang, J. Yang, C. Dong, Cell Res. 20,

62–71 (2010).

ACKNOWLEDGMENTS

We thank V. Bliss and P. Sharma for histological support, G. Tharpand N. Patel for genomics analysis, V. Tran for metabolomics,B. Cervasi and K. Gill for fluorescence-activated cell sorting

Sinclair et al., Science 357, 1014–1021 (2017) 8 September 2017 7 of 8

0

2000

4000

6000

8000

Lung

CD

86 M

FI

CD86

**

Saturated fatty acids beta-oxidation

Purine metabolism

Omega-3 fatty acid metabolism

Unsaturated fatty acid -oxidation

Met, Cys metabolism

Linoleate metabolism

Gly, Ser, Ala, Thr metabolism

Fatty acid oxidation

De Novo fatty acid synthesis

Carnitine shuttle

Arg, Pro metabolism

0

10

20

30

40

Infla

mm

ator

y D

C

% o

f C

D45

+

***

WT

mTOR

APC

WT

naive

0

10

20

30

40

BALLung

0

20

40

60

80

100

0

10

20

30

40

0

50

100

Neu

troph

ils %

of C

D11

b+

Eos

inop

hils

% o

f CD

11b+

Neu

troph

ils %

of C

D11

b+

mTOR APC + Isotype

*****

****

mTOR APC + CSF1r

Tetradecanoyl–CoA

Lauroyl–CoA

3–Oxotetradecanoyl–CoA

trans-Tetradec–2–enoyl–CoA

Palmitoylcarnitine

−! " !#$%&'−()$*+

Metabolite WT mTORΔAPC

Urea group/amino group metabolism

+ +

mTOR APC

7.317.3

Lung

Control Etomoxir

CD

44

IL-17A

0

5

10

15

20

*

Inflammatory DC

n.s.

2 3 4Log10 p

Eos

inop

hils

% o

f CD

11b+

0

10

20

30

40

50

BAL

Lung

0

20

40

60

80

0

10

20

30

40

0

50

100

Eos

inop

hils

%

of C

D11

b+

Neu

troph

ils

% o

f CD

11b+

mTOR APC + IL-12p70

***

***

***

**

mTOR APC + IL-12/23p40

Eos

inop

hils

%

of C

D11

b+

Neu

troph

ils

% o

f CD

11b+

Inflammatory DCWT vs mTOR APC

mTOR APC

mTOR APC + EtxrWT+EtxrWT

IL-1

7A%

of C

D4

Fig. 6. CD11b+-intrinsic mTOR signaling restrains TH17/neutrophilic polarized HDM allergic responses.HDM (75 mg) was administered i.t. to WTand mTORDAPC mice on days 0 and 2. Lungs were analyzed onday 3. (A) Bar chart shows total inflammatory DC cells in lung. (B) HDMwas administered tomTORDAPCmicefor 3 weeks, along with aCsf1r neutralizing monoclonal antibody or control. Bar charts show the frequencyof eosinophils or neutrophils in lung or BAL. (C) Dot graph shows significantly altered metabolic pathwaysbetween WTand mTORDAPC inflammatory DCs, with amino acid pathways highlighted in red and fatty acidmetabolism pathways highlighted in blue. Dot size corresponds to the number of differentially expressedmetabolites for a given pathway.The x axis shows the log10P value. Heatmap shows the relative abundanceof metabolites involved in FAO pathways, based on tentative annotations produced by the Mummichogsoftware. (D)WTandmTORDAPCmicewere treatedwith etomoxir (etxr) andHDM for 3 days. Bar chart showsCD86mean fluorescence intensity (MFI); histogramshows representative flowcytometrydata. (E)mTORDAPC

mice were treated with the Cpt1 inhibitor etomoxir and HDM for 17 days. Bar chart shows expressionof IL-17A+ Tcells in lung; contour plots show representative flow cytometric data. (F) HDM was chronicallyadministered to mTORDAPC mice for 3 weeks, along with IL-12p70 control (gray) or IL-12/23p40 (blue)neutralizing antibodies. Bar charts show the frequency of eosinophils or neutrophils in lung or BAL. Datarepresent two or more independent experiments.

RESEARCH | RESEARCH ARTICLEon M

arch 6, 2019

http://science.sciencemag.org/

Dow

nloaded from

(FACS), and H. Hartweger (The Rockefeller University) andE. Schweighoffer (Francis Crick Institute) for advice on B cell FACSstaining. We also thank C. Wongtrakool for advice and discussionon lung inflammation models and readouts. Finally, we thankPulendran laboratory members for discussion and M. Sinclair forproofreading the manuscript. This work was supported by fundingfrom Action Cycling Atlanta and from the People Programme(Marie Curie Actions) of the European Union’s Seventh FrameworkProgramme (FP7/2007-2013) (C.S.) and from the U.S. National

Institutes of Health (grants R37 DK057665, R37 AI048638, U19AI090023, and U19 AI057266) (B.P.). We are not aware of anyconflicts of interest. Additional data presented in this manuscriptare available in the supplementary materials.

SUPPLEMENTARY MATERIALS

www.sciencemag.org/content/357/6355/1014/suppl/DC1Materials and Methods

Figs. S1 to S22Table S1References (44–58)

14 September 2016; resubmitted 29 March 2017Accepted 22 June 2017Published online 10 August 201710.1126/science.aaj2155

Sinclair et al., Science 357, 1014–1021 (2017) 8 September 2017 8 of 8

RESEARCH | RESEARCH ARTICLEon M

arch 6, 2019

http://science.sciencemag.org/

Dow

nloaded from

allergic inflammationmTOR regulates metabolic adaptation of APCs in the lung and controls the outcome of

Ravindran, Shuzhao Li and Bali PulendranThomas Hagan, Lydia Benitez, Andrei Todor, Deepa Machiah, Timothy Oriss, Anuradha Ray, Steven Bosinger, Rajesh Charles Sinclair, Gayathri Bommakanti, Luiz Gardinassi, Jens Loebbermann, Matthew Joseph Johnson, Paul Hakimpour,

originally published online August 10, 2017DOI: 10.1126/science.aaj2155 (6355), 1014-1021.357Science 

, this issue p. 1014; see also p. 973Sciencemetabolic characteristics that control APC development and function.influencing the immunological character of allergic inflammation. Thus, tissues endow resident APCs with distinctiveauthors identified a central role for mTOR in mediating the metabolic adaptation of such tissue-resident APCs by by their transcriptional program, but also by their metabolic state (see the Perspective by Wiesner and Klein). Thesites have distinctive metabolic signatures and that the development and function of these cells are determined not only

demonstrate that APCs at differentet al.pathogen entry, where they act as sentinels of the immune system. Sinclair Antigen-presenting cells (APCs) are scattered throughout the body in lymphoid organs and at the portals of

Metabolic programming of tissue APCs

ARTICLE TOOLS http://science.sciencemag.org/content/357/6355/1014

MATERIALSSUPPLEMENTARY http://science.sciencemag.org/content/suppl/2017/08/09/science.aaj2155.DC1

CONTENTRELATED

http://stm.sciencemag.org/content/scitransmed/9/401/eaam9171.fullhttp://stm.sciencemag.org/content/scitransmed/8/359/359ra131.fullhttp://stm.sciencemag.org/content/scitransmed/8/359/359ra132.fullhttp://stm.sciencemag.org/content/scitransmed/8/362/362ra143.fullhttp://science.sciencemag.org/content/sci/357/6355/973.full

REFERENCES

http://science.sciencemag.org/content/357/6355/1014#BIBLThis article cites 58 articles, 20 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.Sciencelicensee American Association for the Advancement of Science. No claim to original U.S. Government Works. The title Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive

(print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

on March 6, 2019

http://science.sciencem

ag.org/D

ownloaded from