Activated lymphocytes as a metabolic model for carcinogenesis.

0.49 MB·English

by  Macintyre, Andrew N

MacintyreandRathmellCancer&Metabolism2013,1:5
Cancer &
http://www.cancerandmetabolism.com/1/1/5
Metabolism
REVIEW Open Access
Activated lymphocytes as a metabolic model for
carcinogenesis
Andrew N Macintyre and Jeffrey C Rathmell*
Abstract
Metabolic reprogramming is a key event intumorigenesis to supportcell growth, and cancer cells frequently
become both highlyglycolytic and glutamine dependent.Similarly, T lymphocytes (T cells) modify their metabolism
after activationby foreign antigens to shift from an energetically efficient oxidative metabolism to a highly
glycolytic and glutamine-dependent metabolic program. This metabolic transition enables T cell growth,
proliferation, and differentiation.Inboth activated T cellsand cancer cells metabolic reprogramming is achieved by
similar mechanismsand offers similar survival and cell growth advantages. Activated Tcells thus presenta useful
model with which to study thedevelopmentof tumor metabolism. Here, wereview themetabolicsimilarities and
distinctions between activated T cells and cancer cells, and discuss both the common signaling pathways and
master metabolic regulators that lead to metabolic rewiring.Ultimately, understanding how and whyT cells adopt
a cancer cell-like metabolic profile may identify new therapeutic strategies to selectively target tumor metabolism
or inflammatory immune responses.
Keywords: Cancer, Lymphocyte, Metabolism, Aerobic glycolysis
Review ineachcelltype.Thesesimilaritiesmakelymphocytemeta-
The mid-twentieth century has been described as the bolicreprogrammingausefulmodelwithwhichtodiscover
‘goldenageofintermediarymetabolism’[1],withthework howandwhytumorsrewiretheirmetabolism.Thepurpose
of Krebs, Lippman, Crane and others greatly advancing ofthisreviewistohighlightanddiscussthesimilaritiesand
our understanding of cellular metabolic pathways. In the distinctions in how T cells and tumor cells solve similar
past decade interest in cell metabolism has been rej- metabolicproblems.
uvenated in several fields, especially cancer biology and
lymphocyte immunology. In cancer biology, this renais-
Tlymphocyteactivation:akeylifestyleswitch
sancehasbeendrivenbyevidencethatcancermetabolism
Because of its inherently destructive nature, the immune
presents an underexploited therapeutic target. Immuno-
system must be maintained in a quiescent state. To pro-
logists have been drawn to metabolic studies with the
videprotectionfrom pathogens,however,itmustremain
realizationthatthemetabolismofTlymphocytes(Tcells)
capable of rapid responses and effector function. This
isintimatelytiedtoimmunity[2].Functionally,Tcellsand
challenge is solved with a diverse pool of naïve lympho-
tumorshavelittleincommon;theformerprotectsagainst
cytes that can quickly activate to produce a large, clonal
invasive pathogens, the latter is a diseased tissue charac-
pool of rapidlyproliferating effector Tcells. NaïveTcells
terized by the accumulation of abnormal cells. However,
express near-unique Tcell antigen receptors (TCR) that
bothTcells and cancer cells have strong proliferative sig-
are randomly generated through V(D)J recombination
nals and undergo metabolic reprogramming during their
and pre-selected to recognize foreign antigens presented
respective life cycles, and there exist clear functional and
on major histocompatibility complexes (MHC). These
mechanisticsimilaritiesbetweenthereprogrammingevents
naïve cells continually circulate the blood and lymphatic
system sampling MHC-peptide complexes. Upon en-
*Correspondence:[email protected] counter with an antigen-presenting cell (APC) and cog-
DepartmentofPharmacologyandCancerBiology,Departmentof
nate antigen, the T cell ceases to migrate, forming a
Immunology,SarahW.StedmanNutritionandMetabolismCenter,Duke
University,Durham,NC27710,USA prolonged contact with the APC. This induces sustained
©2013MacintyreandRathmell;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsofthe
CreativeCommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,
distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.
MacintyreandRathmellCancer&Metabolism2013,1:5 Page2of12
http://www.cancerandmetabolism.com/content/1/1/5
signaling through theTCR and other co-receptors, indu- not directly generate ATP, it is inexorably linked to
cing T cell activation, proliferation and differentiation OXPHOS, providing several metabolic inputs to drive
into effector cells. These effectors rapidly accumulate ATP production. In addition, intermediate metabolites
and migrate to sites of inflammation, ultimately clearing from both the TCA cycle and glycolysis can be used as
theinvader[3]. carbon sources for catabolic pathways producing choles-
Activation therefore simultaneously places Tcells under terol, lipids, ribose, and other biosynthetic molecules
several types of stress: they must proliferate rapidly; they (Figure 1) [4]. Resting or non-proliferative cells often
must synthesize large amounts of effector proteins; and rely on mitochondrial lipid β-oxidation. Proliferative
they have to prepare to enter a heterogeneous and poten- cells, in contrast, generally decrease lipid oxidation and
tially hypoxic, nutrient poor environment. Each of these instead conservelipids tosupportcellgrowth[5].
stressors has a significant metabolic aspect reminiscent of For mammalian cells that lack significant intracellular
the classic cancer metabolism paradigm: proliferation and nutrient stores, extracellular glucose uptake represents a
anabolismrequireenergy,biosyntheticbuildingblocksand major carbon and energy source. Glucose is transported
reducing equivalents, while nutrient stress and hypoxia through facilitative glucose transporters and phosphory-
bothpotentiallylimitmetabolicfluxbyrestrictingmetabol- lated by hexokinases to initiate metabolic pathways and
ite access and oxygen. With similar metabolic demands prevent its exit. Glucose-6-phosphate (G6P) is primarily
andstresses,itisnotsurprisingthatthesediversecelltypes metabolized through glycolysis or the pentose phosphate
respondbyadoptingasimilarmetabolicprofile. pathway (PPP). Glycolysis provides a small net ATP gain
perglucosemoleculeconsumedandyieldspyruvatewhich
Acommonmetabolicsolution:aerobicglycolysis cantheneitherbe:i)reducedtolactatebylactatedehydro-
Three metabolic pathways are central to ATP produc- genase (LDH), concomitantly restoring NADH to NAD+,
tion in proliferative lymphocytes and cancer cells: gly- ii) converted to alanine by alanine aminotransferase, si-
colysis, the tri-carboxylic acid (TCA) cycle and oxidative multaneously converting glutamine to α-ketoglutarate, or
phosphorylation (OXPHOS). While the TCA cycle does iii) converted to acetyl-CoenzymeA (acetyl-CoA) in the
Lactate Glucose
Extracellular
GLUT
MCTs
family
Intracellular
Glucose
Pentose Phosphate
Hexokinase Pathway
Glucose-6-Phosphate CO
2
s
si
y
ol NTPs RNA
c
Gly Triose Phosphates RPhiboosspehates
ADP dNTPs DNA
ATP
Lactate Pyruvate Fatty Acid
Lactate NADPH NADP+ Synthesis
dehydrogenase CO
2
Fatty Acids Structural
Acetyl-CoA Cholesterol Lipids
HO + ATP NADH
2 TCA
cycle
O + ADP NAD+
2
TCA Cycle & Oxidative
Phosphorylation CO2
Figure1Majormetabolicfatesofglucoseinhighlyproliferativecells.GlucoseistakenintothecellbyGLUTfamilytransportersandthen
phosphorylatedbyhexokinases,trappingitwithinthecellasglucose-6-phosphate(G6P).G6Pcanbecatabolizedviaglycolysisorusedasacarbon
donorforthesynthesisofribosesviathepentosephosphatepathway(PPP).CatabolizedG6PgeneratespyruvateplussmallquantitiesofATP,with
muchoftheresultantpyruvatebeingconvertedtolactatebylactatedehydrogenaseandthensecretedthroughmono-carboxylictransporters(MCT).
Theremainingpyruvateisconvertedtoacetyl-CoenzymeA(acetyl-CoA)bypyruvatedehydrogenaseandusedeitherasfuelforATPproductionviathe
tri-carboxylicacid(TCA)cycleandoxidativephosphorylationorconvertedtofattyacidstogeneratestructurallipids.Atvariouspointsduringglycolysis
andtheTCAcyclereactionintermediatescanberemovedtoprovidecarbonforaminoacidbiosynthesis(notshown).
MacintyreandRathmellCancer&Metabolism2013,1:5 Page3of12
http://www.cancerandmetabolism.com/content/1/1/5
mitochondriatobeoxidizedviatheTCAcycle,generating glucose uptake up to forty- or fifty-fold and to secrete
large amounts of ATP via OXPHOS (respiration). Most most of the glucose-liberated carbon as lactate [14]
non-proliferatingcellsutilizethislatterpathwaywhenoxy- (Figure 2). In parallel, lymphocytes increase oxygen con-
genisavailableinaprocesstermedthePasteureffect. sumption by around 60% [15-19]. These data have subse-
Not all cells, however, exhibit the Pasteur effect and quently been confirmed using purified Tcell populations
cease lactate production under aerobic conditions. In the stimulated with antibodies that trigger the TCR complex
early 20th century, Otto Warburg observed that many and associated co-receptors [20,21]. Importantly, this in-
tumorcellsandtumorsectionscontinuedlactatesecretion crease in aerobic glycolysis precedes and has been shown
in the presence of oxygen [6]. This metabolic program is to be essential for the growth and proliferation of stimu-
referredtoasaerobicglycolysis,differentiatingitfromthe latedTcells[21-23].
obligatory fermentation of glucose to lactate that occurs Cancer cells and Tcells are not metabolically unique,
under anaerobic conditions where no oxygen is available and the induction of aerobic glycolysis has also been
to fuel OXPHOS. Warburg postulated that the switch to- reported during proliferation of other non-transformed
wards aerobic glycolysis arose from faults in respiration cells. For example, a similar phenotype is also observed
and that such defects were the primary cause of cancer in both intestinal cells and fibroblasts during logarithmic
[6,7].Whilehisobservationsstand,hisproposedmechan- growth [4,24]. However, few other cell types have shown
ismforaerobicglycolysishasnowlargelybeendiscounted such a distinct and acute induction of aerobic glycolysis
following studies demonstrating that cancer cells often from a near proliferative and metabolic standstill. Tcell
have grossly normal respiratory function [8-10] and, in- activation, therefore, provides a unique model to explore
deed, canexhibit elevated ratesofrespiration [11].Never- how andwhymetabolicrewiringoccurs incancer cells.
theless,mitochondrialmutationsareassociatedwithsome
cancers and the relationships between aerobic glycolysis, Aerobicglycolysissupportsrapidproliferation
mitochondrialfunctionandtumorigenesis remain contro- The metabolic needs ofTcells change dramatically upon
versial[12]. activation.Beforeencounteringpathogens,restingTcells
Similar to his observations of aerobic glycolysis in can- require only sufficient energy to support basal cellular
cercells,in1958Warburgalsofoundthatstimulatedleu- needs and replacement biosynthesis. After activation, T
kocytes become highly glycolytic [13]. Subsequent reports cells undergo a transient period with little cell growth
in the 1970s to 1990s, using lectin-stimulated rat thymo- andthenbegintorapidlygrowanddivide.Tcellsspecific
cytes and lymphocytes, also showed lymphocytes become for a given MHC-antigen complex are rare [25,26], so
glycolyticuponactivation.Together,thesestudiesdemon- clonal expansion must rapidly expand these small popu-
strated thatrestinglymphocytesobtainmostoftheir ATP lations of hundreds of cells to the tens or hundreds of
byOXPHOSofglucose,aminoacids,andlipids.However, millions of cells necessary for protection. Remarkably,
withinhoursofstimulation,lymphocytesbegintoincrease activated T cell doubling times of 4 to 6h have been
T cell receptor and
Nave T cell costimulatory Activated T cell
signals
Glucose Glutamine Fatty Acids Glucose Glutamine
ADP ADP
Glycolysis Glutaminolysis ß-oxidation Glycolysis Glutaminolysis
ATP ATP
Lactate
H2O + ATP
TCA
cycle
O2+ ADP
CO2 H2O + ATP Fatty Acids
TCA DNA
cycle Biosynthesis RNA
O2+ ADP CAmhoinleos taecriodls
CO2
Figure2Tcellactivationresultsinmetabolicreprogramming.NaïveTcellshaveanoxidativemetabolism,usingglucose,glutamine,and
fattyacidsasfuelsources.ThemajorityofATPisgeneratedviaoxidativephosphorylation.FollowingactivationbystimulationoftheTcell
receptorandco-receptors,thecellsadoptametabolicprofilethatresemblesthemetabolismofmanycancercells,consuminglargequantitiesof
bothglucoseandglutaminebutperformingrelativelylittleoxidativephosphorylation.Themajorityofglucose-derivedcarbonissecretedas
lactate,withtheremainderbeingusedforbiosynthesis.
MacintyreandRathmellCancer&Metabolism2013,1:5 Page4of12
http://www.cancerandmetabolism.com/content/1/1/5
observed in vitro [27], with even faster doubling rates are used for the generation of amino acids, lipids, chol-
reportedinvivo[28,29].ActivatedTcells,therefore,have esterol and nucleotides. A major function of aerobic gly-
a tremendous need for both ATP [30] and biosynthetic colysis, therefore, is to provide sufficient intermediates
capacity to synthesize new proteins, lipids, and nucleic to fuel biosynthesis for proliferation and growth. Indeed,
acids. increased glucose uptake can enhance T cell responses
While a hallmark of cancer is cell cycle deregulation, and growth in vivo as mice transgenically overexpressing
there is little propensity for tumor cells to adopt increas- the glucose transporter GLUT1 in Tcells accumulate ef-
ingly rapid rates of cell division like activated T cells. fector T cells with age [22,41] and GLUT1 overexpres-
Indeed, the majority of cells within a solid tumor may be sion is correlated with poor prognosis in a variety of
in a state of G1 cell cycle arrest [31]. Extensive clinical cancers[42].
studies have shown that although cell cycle length in Rapid glucose uptake fuels both glycolysis and the PPP,
tumors is more diverse than non-cancerous tissue, the each of which provides numerous metabolites to support
medianS-phaselengthacrossalltumortypesisaround10 cellgrowth.Glycolysisisamajorsourceofserinesynthesis
h [32] and, counter-intuitively, there is no clear relation- as well as pyruvate that can either be converted to lactate
shipbetweenproliferativeabilityandtumoraggressiveness to replenish NAD+ or can be transported into the mito-
[33]. Still, proliferation of cancer cells must exceed cell chondria to enter theTCA cycle as acetyl-CoA. From the
death toallowtumorgrowth.Thus, withthe exceptionof TCAcycle,citratecanexittothecytosoltoprovideabasis
an alternate glycolytic pathway in which tumor cells may for lipid synthesis [21,43]. Under hypoxic conditions, glu-
bypass pyruvate kinase to convert phosphoenol pyruvate tamine can undergo reductive carboxylation to provide a
topyruvate,andyieldnonetgainofATP[34],activatedT reverseflowthroughtheTCAcycleasasourceoflipogen-
cellsand tumorcellsharness aerobic glycolysistoprovide esisinbothcancercellsandinCD8+Tcells[44].Notably,
ATPandbiosynthesisforproliferation. both tumor cells [45] and lectin-stimulated lymphocytes
[46,47] perform extensive de novo synthesis of lipids, and
Advantagesofaerobicglycolysis:rapidATPproduction only limited lipid β-oxidation. In addition to de novo lipo-
In contrast to OXPHOS, glycolysis is energetically ineffi- genesis, aggressive cancer cell lines and primary tumors
cient,theoreticallyyieldingonlytwomoleculesofATPper alsoperformextensivelipidremodeling,inpartduetoele-
glucose molecule consumed compared to up to thirty-six vated monoacylglycerol lipase activity [48]. Tumor lipid
if fully oxidized. This is not a trivial issue as cancer cells metabolismcanbefurtherenhancedbyAkt-drivenexpres-
havebeenshowntopossessadditional,unusedrespiratory sionofthelow-densitylipoproteinreceptor(LDLR),which
capacity [8,35,36]. Thus, cancer cells do not increase gly- increasescholesterolintakeandpromotescellgrowth[49].
colysis solely because their capacity for OXPHOS is sa- The relative importance of each of these pathways to
turated. Rather, aerobic glycolysis and basal OXPHOS lymphocytelipidmetabolismhasyettobedetermined.
provide sufficient energy to support the cell survival and The PPP provides nicotinamide adenine dinucleotide
growthdemandsofcancercellsandactivatedTcells. phosphate (NADPH) reducing potential and generates ri-
One energetic advantage of adopting aerobic glycolysis bose sugars that can be directed into TCA cycle inter-
as a primary metabolic program is the speed at which mediatesand into purine, pyrimidineand aromaticamino
ATP can be regenerated. While OXPHOS yields more acid synthesis pathways. The PPP are strongly induced in
ATP than glycolysis, there is a trade-off between yield Tcell activation [21] and can be important in cancer; in-
and speed [37,38]. Indeed, as described by Koppenol and deed U-C14 glucose tracer experiments have suggested
Bounds [39], Warburg and colleagues observed this phe- that in some tumor types over 80% of the nucleotides in
nomena as early as 1923, reporting that for every one DNA and RNA are synthesized from glucose-derived car-
molecule of glucose oxidized by respiration, twelve are bon[50,51].UpregulationofthePPPisfacilitated,inpart,
metabolized by glycolysis. Increased glycolysis can boost byincreasedenzymeexpression.ActivatedTcellsincrease
ATP production rate by two-thirds, provided cells are expressionofPPPenzymesandhighlevelsofPPPenzyme
not concerned with efficiency. While wasteful, therefore, activity have been reported in metastatic tumor cells [52].
the speed of aerobic glycolysis offers a selective advan- For example, glioblastoma expression of the transketolase
tage both to tumor cells competing against other cells TKTL1,thekeyenzymelinkingthePPPtoglycolysis,dir-
within the same environment [37,40], and to Tcells ra- ectlycorrelateswithtumorseverityintheclinic[53].
cingtosuppressinvadingpathogens. NADPH is a critical reducing agent in the synthesis of
fatty acids and cholesterol as well as maintaining cellular
Advantagesofaerobicglycolysis:biosynthesis redox status and control reactive oxygen species (ROS)
Beyond ATP production, glycolysis and the TCA cycle producedbyOXPHOS[54].WhilesomedegreeofROSis
form the nexus for many biosynthetic processes. Carbon beneficial for bothTcell activation [55] and tumor devel-
intermediates derived from glycolysis and theTCA cycle opment [56], excessive ROS leads to oxidative organelle
MacintyreandRathmellCancer&Metabolism2013,1:5 Page5of12
http://www.cancerandmetabolism.com/content/1/1/5
damage and the induction of apoptosis. Strategies that glucose. Following T cell activation, hexokinase activity
drive cancer cells to increase the OXPHOS-glycolysis increases significantly [71] and Tcells undergo a switch in
ratio, for example by increasing pyruvate dehydrogenase HK isozyme expression from HKI to HKII [72,73]. While
activity to drive mitochondrial conversion of pyruvate to both HKI and HKII both feature two potential catalytic
acetyl-CoA, decrease both proliferation and growth [57]. domains,inHKIoneoftheseisnon-functional,thusHKII
Similarly, glucose restriction of activated lymphocytes has a higher Km for both glucose and ATP compared to
inducesanincreaseinOXPHOS,adropinglycolysis,and HKI [74]. Second, signals from theTCR and co-receptors
an inhibition of proliferation [20,58]. In proliferating cells drive HKI and HKII to bind mitochondria at porin (ATP-
efficient OXPHOS should, therefore, be balanced by high exporting) complexes [75]. This close coupling of HK and
PPPfluxtopreventoverloadingthedemandforNADPH. mitochondria provides HKII with access to a large pool
ofATP.
Advantagesofaerobicglycolysis:adaptationto Following lectin stimulation, lymphocytes also switch
theenvironment expression of other glycolytic isozymes. This includes in-
Glycolysis and the TCA cycle are amphibolic and supply duction of pyruvate kinase M2 (PKM2), LDH-A4, and
both ATP and intermediates to multiple pathways to po- enolase I [21,73]. These changes in expression are asso-
tentially support cells under stress conditions. Indeed, we ciatedwithincreasesinmaximalglycolyticenzymeactivity
have shown that high rates of glycolysis can be protective [16,72],andtherelievingofallostericinhibitionthatwould
against apoptosis [59,60]. A high rate of metabolic flux otherwise limit glycolytic flux. One example of this is the
makes it thermodynamically less costly to redirect inter- regulation of the glycolytic enzyme 6-phosphofructo-1-
mediates down different pathways, that is, high metabolic kinase (PFK1), a key regulatory enzyme in glycolysis
flux permits rapid rerouting of metabolites [61-63]. This (Figure3).PFK1isallostericallyinhibitedbyATPandallos-
controlsensitivitymaypermitafasterresponsetospecific terically activated by fructose-2,6-bisphosphate (F26P2).
nutrient deprivation as cells enter potentially nutrient- F26P2isgeneratedbythebifunctionalenzyme6-phospho-
poorenvironments.Thismayexplainwhytherateofglu- fructo-2-kinase/fructose-2,6-bisphosphatase (PFKFB), and
cose consumption in both activated T cells and many in naïve lymphocytes PFKFB isoform 2 predominates.
tumortypesappearsinexcessofthatrequiredtomeetei- However, following activation T cells express large quan-
therthebiosyntheticorenergeticdemandsofthecell[64]. tities of PFKFB isoform 3 [76,77]. PFKFB3 has a very low
Further, glycolysis is not oxygen dependent, and so phosphataseactivitycomparedtoPFKFB2[78],andsothis
adopting a glycolytic metabolism can prepare cells for isozyme switch enhances PFK1 flux by both increasing
entry or survival in a hypoxic environment. Even after F26P2anddepletingATP.
vascularization, solid tumors feature extensive hypoxic Cancer cells also show a general increase in glycolytic
domains [65]. Similarly, lymph nodes [66], spleen [67], enzyme activity and expression of specific isozymes. This
tumors, dermal/surgical wounds [68] and other regions includes expression of HKII, LDH-A and PFKFB3
frequented by activated lymphocytes contain extensive [52,79,80]. Tumor cells express PKM2, but there is now
areas of low oxygen tension. Adaption of a highly glyco- strongevidencethatthisislargelyinthemetabolicallyin-
lytic metabolism with low oxygen dependency may help active, dimeric form, rather than the active tetramer [81].
both tumors and lymphocytes survive and proliferate Inmanytumor cellsPKM2activityis further inhibited by
duringlowoxygenavailability. direct tyrosine phosphorylation and by the binding of
phosphotyrosine containing peptides, both of which re-
Commonmechanismsdriveglycolyticreprogrammingin strict cofactor binding. Reduced PKM2 activity enhances
Tcellsandtumors aerobic glycolysis and tumor growth [82,83]. Cascades of
Transporterexpressionandizozymeswitching tyrosine phosphorylation are central to T cell activation;
A limiting step in glucose metabolism is the rate at which however, it has yet to be determined if these cascades re-
glucosecanbecapturedandtrappedwithinthecell.There sult in PKM2 inhibition. Recent reports in tumor cells
are two major glucose transporter families, the Na+/glu- havedemonstratedthatPKM2canbeselectivelydegraded
coselinkedtransporter(SGLT)symporters,andtheGLUT in an acetylation-dependent fashion at times of high
familyofpassivetransporters.FourteenmammalianGLUT glucose availability [84], allowing additional glycolytic
familytransportershavebeenidentified[69]andthemajor intermediates to be used for biosynthesis. Phosphoenol-
glucose transporters in lymphocytes appear to be GLUT1 pyruvatefluxthroughanon-ATPgeneratingpathwaymay
andGLUT3,theexpressionlevelsofwhichincreasesignifi- then sustain rapid pyruvate generation while preventing
cantlyfollowingactivation[70].Facilitateddiffusionofglu- ATP-driven feedback inhibition of glycolysis [34]. This
cose by the GLUTs requires a glucose gradient across the regulatory loop for PKM2 may represent a further mech-
extracellular membrane. This so-called glucose sink is anism to maintain high rates of glycolytic flux, but this
maintained by hexokinase phosphorylation of intracellular hasyettobeexaminedinactivatedlymphocytes.
MacintyreandRathmellCancer&Metabolism2013,1:5 Page6of12
http://www.cancerandmetabolism.com/content/1/1/5
Activated T cell
Nave T cell
or cancer cell
Glucose Glucose
ATP ADP ATP ADP
PFKFB2 PFKFB3
Fructose 6-phosphate Fructose 2,6-bisphosphate Fructose 6-phosphate Fructose 2,6-bisphosphate
ATP ATP
Phosphofructokinase-1 Phosphofructokinase-1
ADP ADP
Fructose 1,6-bisphosphate Fructose 1,6-bisphosphate
Pyruvate Lactate Pyruvate
Lactate
dehydrogenase
ADP ADP
TCA TCA
cycle cycle
ATP ATP
Figure3Glycolyticisozymeswitchingpromoteshighratesofglycolysis.ActivatedTcells,cancercellsandotherhighlyproliferativecells
expressdifferentglycolyticisozymesincomparisontoquiescentcells,increasingglycolyticflux.Onekeystepinglycolysisisthephosphorylation
offructose6-phosphatebyphosphofructokinase-1(PFK-1).PFK-1isallostericallyactivatedbyfructose2,6-bisphosphateandallostericallyinhibited
byATP.BothactivatedTcellsandtumorcellsexpressisoform3ofthebifunctionalenzyme6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase
(PFKFB).Incontrast,naïveTcellsexpressPFKFBisoform2.PFKFB3differsfromPFKFB2inthatithaslowphosphataseactivity,leadingtothe
accumulationoffructose2,6-bisphosphateandlocalizeddepletionofATP.ThisresultsinincreasedPFK-1activityandhigherratesofglycolysis.
Beyondglucosemetabolism:glutamine glutamate flux. Also, expression of pyruvate carboxylase
Glutamine has multiple metabolic fates, being used for canallowglucose-derivedpyruvatetoconverttooxaloace-
ATPregeneration,anaplerosisoftheTCAcycle,andredox tate to support the TCA cycle and maintain export of
regulation. Within the cell glutamine is readily converted citrate for lipid synthesis through anapleurosis [100].
to glutamate by glutaminase. Glutamate is used together Giventhesepotentialdifferences,activatedTcellsmayrep-
with cysteine and glycine to generate glutathione, is used resent a better metabolic model for primarily glutamine-
for lipid synthesis through reductive carboxylation under dependenttumors.
hypoxia,andisamajornitrogendonorduringpurineand
pyrimidine synthesis. Naïve lymphocytes utilize glutamine Commonsignalingeventsdrivemetabolic
as a primary oxidative fuel for ATP generation. Following reprogramming
Tcell activation, cMyc greatly increases the expression of The cancer metabolism phenotype is ultimately initiated
glutaminolysis enzymes and the rate of glutamine uptake by oncogenic signaling events that induce metabolic gene
[15,21].Afterconversiontoglutamate,glutamatedehydro- expressionandstimulateaerobicglycolysis.Importantly,T
genase generates α-ketoglutarate to support the TCA cell receptor and co-receptor engagement are now well
cycle.Notably,whiletheearlystagesoflymphocyteactiva- understood and activate many of these same signaling
tion do not require glutamine, subsequent proliferation pathways (see Smith-Garvin et al., 2009, for a detailed
and the expression of effector cytokines following TCR review [101]). Briefly, the TCR is associated with several
stimulation correlate directly with glutamine availability CD3accessorychainsandwhentheTCRisengaged,tyro-
[85-87], and there is clinical evidence to suggest that glu- sine phosphorylation of accessory chains recruits kinases
tamine availability can be a limiting factor in lymphocyte and scaffold proteins. This recruitment, along with co-
activationduringinflammatoryresponses[88-90]. stimulation, triggers localized stimulation of three signal-
Manytumortypes exhibithighrates ofglutamine con- ing pathways: calcium flux, MAPK (ERK/p38) signaling,
sumption relative to non-transformed cells [91-93]. Can- and phosphatidylinositol-3,4,5-trisphosphate (PI(3,4,5)P3)
cers driven by oncogenic cMyc, for example, become signaling. Autocrine and paracrine cytokine signaling
highly dependent on glutamine [94,95] and can be ex- loops induce further PI(3,4,5)P3 and MAPK activation,
quisitely sensitive to glutamine deprivation [96]. Other along with JAK/STAT signaling. Notably, several of the
tumors,however,canexhibitlittlesensitivitytoglutamine downstream targets of these pathways regulate key
deprivation [93,97-99]. This resistance to glutamine metabolic regulators, with mutations in components
deprivation may relate to the induction of glutamine syn- downstream of these pathways strongly implicated in
thase in these cells, and so although less dependent on oncogenesis. Identifying the specific signaling pathways in
exogenous glutamine, they still exhibit high rates of activated Tcells that induce metabolic reprogramming is
MacintyreandRathmellCancer&Metabolism2013,1:5 Page7of12
http://www.cancerandmetabolism.com/content/1/1/5
thereforeinformativeinidentifyingtheoncogenesinvolved alsoplayakeyroleinenhancingHIF1αtranscriptionalabil-
indrivingthesameprocessesintumors. ity, by enhancing HIF1α interactions with transcriptional
co-factors[114].
PI3K,PTEN,AktandmTORC1 HIF1α is not strongly expressed in normal tissues
PI(3,4,5)P3 is generated by phosphatidylinositol-3-kinase under normoxic conditions and presents a potential
(PI3K) and depleted by phosphatases such as the tumor therapeutic target to selectively suppress tumor glucose
suppressor, PTEN (phosphatase and tensin homologue metabolism. In support of this strategy, several studies
deleted on chromosome 10). Both sides of this signaling have reported that HIF1α null tumor xenografts show
equilibrium can impact cancer, as activating PI3K and dis- reduced growth, while overexpression of xenograft
rupting PTEN mutations frequently promote constitutive HIFα promotes increased growth [115]. Curiously, and
signaling through PI(3,4,5)P3-dependent pathways [102]. in contrast to these data, HIF1α−/− Tcells exhibit nor-
Of the downstream targets for PI(3,4,5)P3 signaling, the malproliferativeandinitialmetabolicresponsestoTCR
bestdescribedisAkt,anestablishedmetabolicregulatorin and co-receptor stimulation [116,117]. Instead, the im-
both tumors and lymphocytes. In hematopoietic cells and pact of HIF1α loss is only apparent when activated
naïve Tcells, the expression of a constitutively active Akt HIF1α−/− T cells are subsequently skewed to different
leads to increased GLUT1 surface localization, improved cell fates. HIF1α−/− CD4+ T cells are unable to form
coupling of HKII to the mitochondria and increased rates interleuken-17 (IL-17) producing T helper cells, which
of glycolysis [20,103,104]. Similarly, in tumor models Akt are highly glycolytic. Instead, HIF1α−/− Tcells become
drives cells towards aerobic glycolysis and makes cells immunosuppressive regulatory T cells in which lipid
highlydependentonexogenousglucoseforsurvival[105]. metabolism, not glycolysis, is the major metabolic pro-
Akt promotes aerobic glycolysis by direct phosphoryl- gram [41,117]. TheroleofHIF1αinmetabolicregulation
ation and activation of glycolytic enzymes, such as PFK2 is therefore limited during Tcell activation. Determining
[106],byphosphorylationofTBC1D1/4toregulateGLUT1 the signaling context by which T cell skewing directs
trafficking, and by regulating several transcription factors HIF1αregulationofmetabolismmay,however,beinform-
(reviewed in detail by Manning and Cantley, 2007) [107]. ativeindetermininghowHIF1αfunctionsintumors.
Further, Akt is able to activate mTORC1 (mammalian
target of rapamycin complex 1) via phosphorylation of up- JAK/STATsandthePIMkinases
stream regulators PRAS40 and TSC2. mTORC1 is a key Tcellactivationinducedmetabolismismaintainedbysus-
driverofanabolicmetabolism.Indeed,activatingthePI3K/ tained signaling from IL-2 and other cytokines acting on
Akt pathway can be considered a key regulator of glucose common gamma chain (γc) cytokine receptor complexes.
metabolism in bothTcells and cancer [108]. Inhibition of ThiseffectisinpartmediatedbydirectandSTAT5driven
this pathway in Tcells is potently immunosuppressive and PI(3,4,5)P3/Akt signaling [118,119]. However, additional
leads to generation of tolerant or regulatory Tcells rather STAT driven, Akt-independent, signaling events also play
than effectors. Given the frequency of cancer-associated a role. Of note, JAK/STAT3 signaling in lymphocytes
mutations in this pathway, delineating how PI(3,4,5)P3 sig- inducestheexpressionofthePIMfamilyofkinases,which
naling leads to metabolic reprogramming in lymphocytes themselvescanpromoteglycolyticmetabolism[120].
mayprovideauniqueopportunityto understandtheregu- PIM kinases are constitutively active [121] and are po-
lationofcancermetabolism. tent oncogenes, being induced by, and synergizing with,
the transcription factor cMyc in several cancer types
MAPKandHIF1α [122]. In addition, persistent STAT3 signaling is common
The multifactorial roles of the mitogenic ras-MAPK sig- in many tumor types. While oncogenic STAT3 mutations
naling pathways in cancer have been extensively reviewed have not been reported, aberrant STAT3 signaling can
recently [109-111]. MAPK have multiple roles in meta- arise from inactivation of the STAT-suppressing suppres-
bolicregulationinbothtumors[112]andduringTcellac- sor of cytokine signaling (SOCS) proteins or by elevated
tivation[71,87].One mechanistic role ofrecent interestis activation of JAKs [123]. The γc-receptor-directed main-
MAPK regulation of hypoxia inducible factor 1α (HIF1α). tenance of activated Tcell metabolism, therefore, poten-
HIF1αisaheterodimerictranscriptionfactorthatinduces tiallypresentsausefultoolwithwhichtostudytheroleof
geneexpressioninresponsetohypoxia.HIF1αinducesthe STAT-driven, PIM-mediated, regulation of metabolism.
expression of many glycolytic genes, and HIF1α can be a Unfortunately, the PIMs share substrate specificity with
key mediator of the Pasteur effect in normal cells [113]. Akt[120],andareinhibitedbytheclassicalPI3Kinhibitor
HIF1αproteinlevelsareelevatedwithouttheneedforhyp- LY294002, a compound historically used to study Akt
oxia by PI(3,4,5)P3 signaling through mTOR and other function [124]. The specific role of PIM kinases in meta-
pathways. Activated T cells and many tumor cells, there- bolic reprogramming is hence unclear. Studies of acti-
fore,canexhibitelevatedlevelsofHIF1α.MAPK,however, vated,PIM-nullTcells[125]mayhelpresolvethisissue.
MacintyreandRathmellCancer&Metabolism2013,1:5 Page8of12
http://www.cancerandmetabolism.com/content/1/1/5
CalciumsignalingandAMPK TCellmetabolicreprogrammingisbothtransientand
Immediately after TCR activation there is a coordinated reversible
flux of calcium from intracellular stores and also an in- Following activation,Tcells can differentiate into effector,
crease in mitochondrial calcium uptake [126]. These regulatory and memory Tcells that have differing meta-
changes stimulate the calcium-activated mitochondrial bolic profiles [2,117,134]. Activated Tcells are, therefore,
dehydrogenases that drive the TCA cycle [127]. In metabolically flexible and not fixed into a specific meta-
addition, calcium flux downstream of the TCR causes a bolicprogram.Unlikecancercellswithspecificoncogenic
short term phosphorylation of AMP activated protein mutations, T cell metabolism is dependent on signaling
kinase (AMPK) [128], a master metabolic regulator that pathwaystriggeredbythelocalenvironment.Indeed,even
promotes catabolic pathways when the ATP-AMP ratio onceTcellfunctionalandmetabolicfatehasbeendefined
falls. AMPK is activated by binding of AMP and when thereisadegreeofreversibilityandplasticity,forexample,
phosphorylated by CaMKKβ or the tumor suppressor lipid-dependent regulatory T cells can be redirected to
LKB1 [129]. While the metabolic impact of AMPK acti- form highly glycolytic, IL-17-producing cells by altering
vation by the TCR has yet to be fully defined, calcium- the cytokine environment [41,135]. In contrast, tumor
inducedAMPKactivityduringTcellactivation mayhelp cells are largely fixed on one metabolic route that is dic-
to promote aninitial phase ofoxidative and ATP-gener- tatedbyirreversiblemutationsinupstreamsignalingpath-
ating metabolism. This could prepare T cells to enter a ways. Thus, cancer cells have less metabolic flexibility
rapid growth phase and to resist the stress of nutrient- than Tcells and the response of each cell type to inhib-
deficient conditions. The latter role may be particularly ition of specific metabolic pathways may lead todistinctly
important as AMPK-null T cells show only a limited differentoutcomes.
metabolic phenotype under nutrient-rich conditions, but
fail to respond to metabolic stress in vitro [130]. In vivo, ActivatedTcellsarenottumorigenic
nutrients are potentially limiting in lymph nodes or Despite the metabolic and other similarities between
inflamed tissues, and TCR-induced activation of AMPK stimulatedTcellsandacancercellundergoingaerobicgly-
may be important to maintain ATP levels and maximize colysis,activatedTcellsarenotcancerous.Instead,follow-
survival, so that Tcells can proceed to a later phase in ing clearance of an infection the vast majority of activated
which AMPK activity is reduced and rapid cell growth Tcellswilldieduetoactivation-inducedcelldeathorfrom
begins. cytokineneglect.BothactivatedTcellsandtumorcellsare
Although misregulation of calcium signaling can be kept alive by a precarious balance of pro- and anti-
important in tumorigenesis [131], direct regulation of apoptotic BH3 domain-containing proteins. In lympho-
tumor metabolism by calcium has not been studied in cytes this balance is maintained by cytokine signaling
detail.Indeed,theroleofAMPKincancermetabolismis through Akt and other pathways, and, in addition, by
still controversial. While LKB1 has an established role as glycolytic flux [136-139]. Within tumors this balance is
a tumor suppressor, LKB1 has avariety of substrates and maintainedbothbyglycolyticfluxandoncogenicsignaling.
how LKB1 tumor suppression relates to AMPK activa- Understanding how activated Tcells die following the loss
tion is unclear. AMPK activation has been proposed as of glycolytic flux and cytokine signals may provide insight
being anti-tumorigenic, as it suppresses cell cycle pro- intohowanti-metaboliteskill,orfailtokill,cancercells.
gression and can oppose Akt activity by suppressing
mTORC1 [132]. Recent data, however, indicates that Tumorcellsaremetabolicallyandgeneticallydiverse
transient AMPK activation in response to energy stress It is becoming evident that while the phenomena of aer-
can promote tumor survival by maintaining NADPH obic glycolysis is common to many tumors, different
homeostasis [133]. Understanding how AMPK activation cancer cells, potentially even within the same tumor, are
supports activated T cells in vivo in times of metabolic metabolically diverse. Even within cell lines established
stress may provide new clues as to the role of AMPK in from the same type of tumor there exists significant
tumormetabolism. metabolic variation [140,141]. This heterogeneity can be
representative of cancer stage or subtype, as in prostate
LimitationsofTcellsasamodelfortumormetabolism and breast cancer. Given the strong dependence of T
Metabolic reprogramming in activated Tcells is a useful cells on glutamine, activated T cells represent a better
model to study the metabolic changes that occur during model for glutamine-addicted tumors, for example those
tumorigenesis. Indeed, many of the pathways are similar driven by oncogenic Myc [21,95], than more glucose
and approaches to disrupt cancer metabolism can also dependent tumors, for example those driven by Met
be quite immunosuppressive. However, the two systems [141]. More importantly, activated Tcells themselves be-
havesomesignificantdifferencesthatmayprovideuseful come metabolically diverse as they differentiate into spe-
insightinto novelanti-cancertherapies. cific effector or regulatory subsets [41]. These T cell
MacintyreandRathmellCancer&Metabolism2013,1:5 Page9of12
http://www.cancerandmetabolism.com/content/1/1/5
differentiation pathways are regulated by specific signal- Acknowledgements
ing events and it will be interesting to determine if dis- WethankthemembersoftheRathmelllaboratoryforinsightfuldiscussions
andcomments.Funding:NationalInstitutesofHealthGrantsR01AI063345
tinct Tcell subtypes may represent specific cancer types
andR01HL108006(toJCR);theAmericanAsthmaFoundation(JCRisthe
or stages. This is an important consideration as the sen- BernardOsherFellowoftheAmericanAsthmaFoundation);theLupus
sitivity of tumor cells to metabolic inhibitors varies de- ResearchInstitute(JCR);andtheLeukemiaandLymphomaSociety(JCR).
pendingontheoncogenesinvolved[142].
Received:31July2012Accepted:4October2012
Published:23January2013
Conclusions
Cancer cells and activated Tcells adopt comparable meta- References
1. KhoranaHG:ChemicalBiology;SelectedPapersofH.GobindKhorana.
bolic profiles to cope with similar environmental and pro-
Singapore:WorldScientificPublishingCo;2000.
liferative stressors. Given that both T cell activation and 2. GerrietsVA,RathmellJC:MetabolicpathwaysinTcellfateandfunction.
tumorigenesis often resort to the same signaling pathways TrendsImmunol2012,33:168–173.
3. AndersonG,LanePJ,JenkinsonEJ:Generatingintrathymic
to induce this metabolic rewiring,Tcell activation offers a
microenvironmentstoestablishT-celltolerance.NatRevImmunol2007,
useful model with which to study the mechanics of meta- 7:954–963.
bolic reprogramming. While cancer metabolism is inher- 4. LuntSY,VanderHeidenMG:Aerobicglycolysis:meetingthemetabolic
requirementsofcellproliferation.AnnuRevCellDevBiol2011,27:441–464.
entlymorediverseandsusceptibletoselectivepressures,T
5. DeberardinisRJ,LumJJ,ThompsonCB:Phosphatidylinositol3-kinase-
cells have the significant advantage in a laboratory setting dependentmodulationofcarnitinepalmitoyltransferase1Aexpression
of beingquiescentandnon-cyclingpriortoactivation,aid- regulateslipidmetabolismduringhematopoieticcellgrowth.JBiolChem
2006,281:37372–37380.
inginthedelineationofcellsignalingandcellcycleeffects.
6. WarburgO,PosenerK,NegeleinE:Ueberdenstoffwechseldertumoren.
The aerobic glycolysis and glutamine dependency of InOnMetabolismofTumors.EditedbyWarburgO.London:Constable;1930.
cancer cellshavebeen identifiedas potentialnoveltargets 7. WarburgO:Ontheoriginofcancercells.Science1956,123:309–314.
8. FantinVR,St-PierreJ,LederP:AttenuationofLDH-Aexpressionuncovers
for cancer therapy, and so developing an improved un-
alinkbetweenglycolysis,mitochondrialphysiology,andtumor
derstanding of how these metabolic programs arise is of maintenance.CancerCell2006,9:425–434.
clinicalimportance.However,giventheclosesimilaritybe- 9. ZuXL,GuppyM:Cancermetabolism:facts,fantasy,andfiction.
BiochemBiophysResCommun2004,313:459–465.
tween activated T cell and tumor metabolic reprogram-
10. VaupelP,MayerA:Availability,notrespiratorycapacitygovernsoxygen
ming, consideration must be given to the impact drugs consumptionofsolidtumors.IntJBiochemCellBiol2012,44:1477–1481.
targetingthesepathwayswillhaveonTcells.Tcellmetab- 11. WeinhouseS:Onrespiratoryimpairmentincancercells.Science1956,
124:267–269.
olism and Tcell survival are intertwined, and the loss of
12. SeyfriedTN,SheltonLM:Cancerasametabolicdisease.NutrMetab(Lond)
anti-tumor T cells may negate many of the benefits of 2010,7:7.
drugs targeting tumor metabolism. This is especially sig- 13. WarburgO,GawehnK,GeisslerAW:[Metabolismofleukocytes].
ZNaturforschB1958,13B:515–516.
nificantinthecontextofrecentdataindicatingthatmeta-
14. ColoffJL,MasonEF,AltmanBJ,GerrietsVA,LiuT,NicholsAN,ZhaoY,
bolic suppression of activating Tcells skews them toward WoffordJA,JacobsSR,IlkayevaO,GarrisonSP,ZambettiGP,RathmellJC:
an immunosuppressive phenotype, which may suppress Aktrequiresglucosemetabolismtosuppresspumaexpressionand
preventapoptosisofleukemicTcells.JBiolChem2011,286:5921–5933.
anti-tumorimmuneresponses[41].
15. CalderPC:Fuelutilizationbycellsoftheimmunesystem.ProcNutrSoc
1995,54:65–82.
Abbreviations 16. GreinerEF,GuppyM,BrandK:Glucoseisessentialforproliferationand
acetyl-CoA:acetyl-CoenzymeA;AMPK:AMPactivatedproteinkinase; theglycolyticenzymeinductionthatprovokesatransitiontoglycolytic
APC:antigen-presentingcell;F26P2:fructose-2,6-bisphosphate;G6P:glucose- energyproduction.JBiolChem1994,269:31484–31490.
6-phosphate;HIF1α:hypoxiainduciblefactor1α;HK:hexokinase;JAK:Janus 17. RoosD,LoosJA:Changesinthecarbohydratemetabolismof
kinase;LDH:lactatedehydrogenase;LDLR:low-densitylipoproteinreceptor; mitogenicallystimulatedhumanperipherallymphocytes.II.Relative
MAPK:mitogen-activatedproteinkinase;MCT:mono-carboxylictransporters; importanceofglycolysisandoxidativephosphorylationon
MHC:majorhistocompatibilitycomplexes;mTORC1:mammaliantargetof phytohaemagglutininstimulation.ExpCellRes1973,77:127–135.
rapamycincomplex1;NADPH:nicotinamideadeninedinucleotide 18. ButtgereitF,BrandMD,MullerM:ConAinducedchangesinenergy
phosphate;OXPHOS:oxidativephosphorylation;PFK1:6-phosphofructo-1- metabolismofratthymocytes.BiosciRep1992,12:381–386.
kinase;PFKFB:6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase;PI(3,4,5) 19. HedeskovCJ:Earlyeffectsofphytohaemagglutininonglucose
P3:phosphatidylinositol-3,4,5-trisphosphate;PI3K:phosphatidylinositol-3- metabolismofnormalhumanlymphocytes.BiochemJ1968,110:373–380.
kinase;PKM2:pyruvatekinaseM2;PPP:pentosephosphatepathway; 20. FrauwirthKA,RileyJL,HarrisMH,ParryRV,RathmellJC,PlasDR,ElstromRL,
PTEN:phosphataseandtensinhomologuedeletedonchromosome10; JuneCH,ThompsonCB:TheCD28signalingpathwayregulatesglucose
ROS:reactiveoxygenspecies;SGLT:sodium/glucoselinkedtransporter; metabolism.Immunity2002,16:769–777.
SOCS:suppressorofcytokinesignaling;STAT:signaltransducerandactivator 21. WangR,DillonCP,ShiLZ,MilastaS,CarterR,FinkelsteinD,McCormickLL,
oftranscription;TCA:tri-carboxylicacid;TCR:Tcellantigenreceptor; FitzgeraldP,ChiH,MungerJ,GreenDR:ThetranscriptionfactorMyc
TKTL1:transketolase1;γc:commongammachain. controlsmetabolicreprogramminguponTlymphocyteactivation.
Immunity2011,35:871–882.
22. JacobsSR,HermanCE,MaciverNJ,WoffordJA,WiemanHL,HammenJJ,
Competinginterests
RathmellJC:GlucoseuptakeislimitinginTcellactivationandrequires
Theauthorsdeclarethattheyhavenocompetinginterests.
CD28-mediatedAkt-dependentandindependentpathways.JImmunol
2008,180:4476–4486.
Authors’contributions 23. ChamCM,DriessensG,O'KeefeJP,GajewskiTF:Glucosedeprivation
ANMandJCRwrotethemanuscript.Bothauthorsreadandapprovedthe inhibitsmultiplekeygeneexpressioneventsandeffectorfunctionsin
finalmanuscript. CD8+Tcells.EurJImmunol2008,38:2438–2450.
MacintyreandRathmellCancer&Metabolism2013,1:5 Page10of12
http://www.cancerandmetabolism.com/content/1/1/5
24. MunyonWH,MerchantDJ:Therelationbetweenglucoseutilization,lactic 48. NomuraDK,LongJZ,NiessenS,HooverHS,NgSW,CravattBF:
acidproductionandutilizationandthegrowthcycleofLstrain Monoacylglycerollipaseregulatesafattyacidnetworkthatpromotes
fibroblasts.ExpCellRes1959,17:490–498. cancerpathogenesis.Cell2010,140:49–61.
25. ObarJJ,KhannaKM,LefrancoisL:EndogenousnaiveCD8+Tcell 49. GuoD,ReinitzF,YoussefM,HongC,NathansonD,AkhavanD,KugaD,
precursorfrequencyregulatesprimaryandmemoryresponsesto AmzajerdiAN,SotoH,ZhuS,BabicI,TanakaK,DangJ,IwanamiA,GiniB,
infection.Immunity2008,28:859–869. DejesusJ,LisieroDD,HuangTT,PrinsRM,WenPY,RobinsHI,PradosMD,
26. MoonJJ,ChuHH,PepperM,McSorleySJ,JamesonSC,KedlRM,JenkinsMK: DeangelisLM,MellinghoffIK,MehtaMP,JamesCD,ChakravartiA,
NaiveCD4(+)Tcellfrequencyvariesfordifferentepitopesandpredicts CloughesyTF,TontonozP,MischelPS:AnLXRagonistpromotes
repertoirediversityandresponsemagnitude.Immunity2007,27:203–213. glioblastomacelldeaththroughinhibitionofanEGFR/AKT/SREBP-1/
27. LeeWT,PasosG,CecchiniL,MittlerJN:Continuedantigenstimulationis LDLR-dependentpathway.Cancerdiscovery2011,1:442–456.
notrequiredduringCD4(+)Tcellclonalexpansion.JImmunol2002, 50. BorosLG,PuigjanerJ,CascanteM,LeeWN,BrandesJL,BassilianS,YusufFI,
168:1682–1689. WilliamsRD,MuscarellaP,MelvinWS,SchirmerWJ:Oxythiamineand
28. LawrenceCW,BracialeTJ:Activation,differentiation,andmigrationof dehydroepiandrosteroneinhibitthenonoxidativesynthesisofribose
naivevirus-specificCD8+Tcellsduringpulmonaryinfluenzavirus andtumorcellproliferation.CancerRes1997,57:4242–4248.
infection.JImmunol2004,173:1209–1218. 51. HoreckerBL,DomagkG,HiattHH:AcomparisonofC14-labelingpatterns
29. YoonH,KimTS,BracialeTJ:ThecellcycletimeofCD8+Tcells indeoxyriboseandriboseinmammaliancells.ArchBiochemBiophys
respondinginvivoiscontrolledbythetypeofantigenicstimulus. 1958,78:510–517.
PLoSOne2010,5:e15423. 52. BoardM,HummS,NewsholmeEA:Maximumactivitiesofkeyenzymesof
30. ButtgereitF,BurmesterGR,BrandMD:Bioenergeticsofimmunefunctions: glycolysis,glutaminolysis,pentosephosphatepathwayandtricarboxylic
fundamentalandtherapeuticaspects.ImmunolToday2000,21:192–199. acidcycleinnormal,neoplasticandsuppressedcells.BiochemJ1990,
31. TayDL,BhathalPS,FoxRM:QuantitationofG0andG1phasecellsin 265:503–509.
primarycarcinomas.AntibodytoM1subunitofribonucleotidereductase 53. VolkerHU,HagemannC,CoyJ,WittigR,SommerS,StojicJ,HaubitzI,Vince
showsG1phaserestrictionpointblock.JClinInvest1991,87:519–527. GH,KammererU,MonoranuCM:Expressionoftransketolase-like1and
32. RewDA,WilsonGD:Cellproductionratesinhumantissuesandtumours activationofAktingradeIVglioblastomascomparedwithgradesIIand
andtheirsignificance.PartII:clinicaldata.EurJSurgOncol2000,26:405–417. IIIastrocyticgliomas.AmJClinPathol2008,130:50–57.
33. RewDA,WilsonGD:Cellproductionratesinhumantissuesandtumours 54. MurphyMP:Howmitochondriaproducereactiveoxygenspecies.
andtheirsignificance.Part1:anintroductiontothetechniquesof BiochemJ2009,417:1–13.
measurementandtheirlimitations.EurJSurgOncol2000,26:227–238. 55. ChaudhriG,ClarkIA,HuntNH,CowdenWB,CeredigR:Effectof
34. VanderHeidenMG,LocasaleJW,SwansonKD,SharfiH,HeffronGJ,Amador- antioxidantsonprimaryalloantigen-inducedTcellactivationand
NoguezD,ChristofkHR,WagnerG,RabinowitzJD,AsaraJM,CantleyLC: proliferation.JImmunol1986,137:2646–2652.
Evidenceforanalternativeglycolyticpathwayinrapidlyproliferating 56. WeinbergF,HamanakaR,WheatonWW,WeinbergS,JosephJ,LopezM,
cells.Science2010,329:1492–1499. KalyanaramanB,MutluGM,BudingerGR,ChandelNS:Mitochondrial
35. SmolkovaK,BellanceN,ScandurraF,GenotE,GnaigerE,Plecita-HlavataL, metabolismandROSgenerationareessentialforKras-mediated
JezekP,RossignolR:Mitochondrialbioenergeticadaptationsofbreast tumorigenicity.ProcNatlAcadSciUSA2010,107:8788–8793.
cancercellstoaglycemiaandhypoxia.JBioenergBiomembr2010,42:55–67. 57. BonnetS,ArcherSL,Allalunis-TurnerJ,HaromyA,BeaulieuC,ThompsonR,
36. SchulzTJ,ThierbachR,VoigtA,DrewesG,MietznerB,SteinbergP,Pfeiffer LeeCT,LopaschukGD,PuttaguntaL,HarryG,HashimotoK,PorterCJ,
AF,RistowM:Inductionofoxidativemetabolismbymitochondrial AndradeMA,ThebaudB,MichelakisED:Amitochondria-K+channelaxisis
frataxininhibitscancergrowth:OttoWarburgrevisited.JBiolChem2006, suppressedincanceranditsnormalizationpromotesapoptosisand
281:977–981. inhibitscancergrowth.CancerCell2007,11:37–51.
37. PfeifferT,SchusterS,BonhoefferS:Cooperationandcompetitioninthe 58. BrandKA,HermfisseU:Aerobicglycolysisbyproliferatingcells:
evolutionofATP-producingpathways.Science2001,292:504–507. aprotectivestrategyagainstreactiveoxygenspecies.FASEBJ1997,
38. StuckiJW:Theoptimalefficiencyandtheeconomicdegreesofcoupling 11:388–395.
ofoxidativephosphorylation.EurJBiochem1980,109:269–283. 59. MasonEF,ZhaoY,Goraksha-HicksP,ColoffJL,GannonH,JonesSN,
39. KoppenolWH,BoundsPL:TheWarburgeffectandmetabolicefficiency: RathmellJC:Aerobicglycolysissuppressesp53activitytoprovide
re-crunchingthenumbers.Science2009,E-letter. selectiveprotectionfromapoptosisuponlossofgrowthsignalsor
40. SchusterS,KreftJU,SchroeterA,PfeifferT:Useofgame-theoretical inhibitionofBCR-Abl.CancerRes2011,70:8066–8076.
methodsinbiochemistryandbiophysics.JBiolPhys2008,34:1–17. 60. ZhaoY,ColoffJL,FergusonEC,JacobsSR,CuiK,RathmellJC:Glucose
41. MichalekRD,GerrietsVA,JacobsSR,MacintyreAN,MacIverNJ,MasonEF, metabolismattenuatesp53andpuma-dependentcelldeathupon
SullivanSA,NicholsAG,RathmellJC:Cuttingedge:distinctglycolyticand growthfactordeprivation.JBiolChem2008,283:36344–36353.
lipidoxidativemetabolicprogramsareessentialforeffectorand 61. NewsholmeEA,BoardM:Applicationofmetabolic-controllogictofuel
regulatoryCD4+Tcellsubsets.JImmunol2011,186:3299–3303. utilizationanditssignificanceintumorcells.AdvEnzymeRegul1991,
42. AirleyRE,MobasheriA:Hypoxicregulationofglucosetransport, 31:225–246.
anaerobicmetabolismandangiogenesisincancer:novelpathwaysand 62. NewsholmeEA,CrabtreeB,ArdawiMS:Theroleofhighratesofglycolysis
targetsforanticancertherapeutics.Chemotherapy2007,53:233–256. andglutamineutilizationinrapidlydividingcells.BiosciRep1985,5:393–400.
43. MichalekRD,GerrietsVA,NicholsAG,InoueM,KazminD,ChangCY,Dwyer 63. NewsholmeP,NewsholmeEA:Ratesofutilizationofglucose,glutamine
MA,NelsonER,PollizziKN,IlkayevaO,GiguereV,ZuercherWJ,PowellJD, andoleateandformationofend-productsbymouseperitoneal
ShinoharaML,McDonnellDP,RathmellJC:Estrogen-relatedreceptor-alpha macrophagesinculture.BiochemJ1989,261:211–218.
isametabolicregulatorofeffectorT-cellactivationanddifferentiation. 64. CuriR,NewsholmeP,Pithon-CuriTC,Pires-de-MeloM,GarciaC,Homem-
ProcNatlAcadSciUSA2011,108:18348–18353. de-BittencourtJuniorPI,GuimaraesAR:Metabolicfateofglutaminein
44. MetalloCM,GameiroPA,BellEL,MattainiKR,YangJ,HillerK,JewellCM, lymphocytes,macrophagesandneutrophils.BrazJMedBiolRes1999,
JohnsonZR,IrvineDJ,GuarenteL,KelleherJK,VanderHeidenMG,Iliopoulos 32:15–21.
O,StephanopoulosG:ReductiveglutaminemetabolismbyIDH1mediates 65. HelmlingerG,YuanF,DellianM,JainRK:InterstitialpHandpO2gradients
lipogenesisunderhypoxia.Nature2011,481:380–384. insolidtumorsinvivo:high-resolutionmeasurementsrevealalackof
45. OokhtensM,KannanR,LyonI,BakerN:Liverandadiposetissue correlation.NatMed1997,3:177–182.
contributionstonewlyformedfattyacidsinanascitestumor.AmJ 66. OhtaA,DiwanjiR,KiniR,SubramanianM,SitkovskyM:InvivoTCell
Physiol1984,247:R146–153. ActivationinLymphoidTissuesisInhibitedintheOxygen-Poor
46. NathanielD,MellorsA:Mitogeneffectsonlipidmetabolismduring Microenvironment.FrontImmunol2011,2:27.
lymphocyteactivation.MolImmunol1983,20:1259–1265. 67. CaldwellCC,KojimaH,LukashevD,ArmstrongJ,FarberM,ApasovSG,
47. deBittencourtPIHJr,PeresCM,YanoMM,HirataMH,CuriR:Pyruvateisa SitkovskyMV:Differentialeffectsofphysiologicallyrelevanthypoxic
lipidprecursorforratlymphocytesinculture:evidenceforalipid conditionsonTlymphocytedevelopmentandeffectorfunctions.
exportingcapacity.BiochemMolBiolInt1993,30:631–641. JImmunol2001,167:6140–6149.