otto warburg’s contributions to cancer metabolism

7/28/2019 Otto Warburgs contributions to cancer metabolism

1/13

Es ware mglich, die gesamte Geschichte der Biochemie an Otto Warburgs werk aufzuzeigen.(It would be possible to illustrate the entire history of biochemistry with the work of Otto Warburg.)(Adolf F. J. Butenandt, 19701

Otto Warburg(FIG. 1) was one of the first true interdisci-plinary scientists. Warburg, who spent his entire careerin Germany, pioneered work onrespiration and photo-synthesis during the early twentieth century. Duringthe 1910s, it was thought that the energy-yieldingreactions necessary for the growth of cancer cellswere lipolysis and/or proteolysis2. However, Warburgfocused onglycolysis and showed that all of the cancercells he investigated exhibit a reversedPasteur effect (the inhibition of fermentation by O2 . In other words,cancer cells produce lactic acid from glucose evenunder non-hypoxic conditions3, an observation thathas come to be known as the Warburg effec t 4 (whichis not to be confused with the other Warburg effect:the inhibition of photosynthetic CO2 fixation by O2(REF. 5) ). With few exceptions, Warburgs findings werepublished in German-language journals, and duringthe latter part of the twentieth century, with the post-Second World War relocation of scientific primacy toEnglish-language institutions and the blossoming of the field of molecular biology, Warburgs contribu-tions became largely disregarded. The discovery inrecent decades of a connection between oncogenes andmetabolic processes has led to a renaissance of interest

in Warburgs work today 6, although his findings andconclusions are often misinterpreted. The semanticsof Warburgs report that the respiration of all cancercells is damaged7 continues to be debated, because theexperiments by Warburg and his co-workers, and thoseof contemporary investigators, indicate that such aconclusion is erroneous.

In this Review, we describe the historical contextof Warburgs investigations of lactic acid productionby cancer cells and explore the impact of his work on our current conceptual framework of cancer cellmetabolism.

Warburgs lifeThe details of Warburgs life and personality havebeen gleaned from biographies written by Krebs8,9,Werner 1,10, Hxtermann and Sucker11 and Koepcke12.Otto Heinrich Warburg was born 8 October 1883 inFreiburg im Breisgau. His father, Emil Warburg, wasone of the most eminent physicists of his time13 andwas revered by young Otto. As was common amongprofessors families, the Warburgs resided at Emilsinstitute, first at the University of Freiburg and laterin Berlin, to allow him to concentrate on research.Thus Otto was raised in an academic environment Ottos sister Lotte claimed that Papa weiss nichteinmal, wo Mamas Schlafzimmer ist! (Papa doesnteven know where Mamas bedroom is!12. Warburgslife and his academic achievements are summarizedin the TIMELINE.

*Institute of InorganicChemistry, Swiss Federal Institute of Technology,CH8093 Zurich, Switzerland.Department of Medicine,

Johns Hopkins UniversitySchool of Medicine, Baltimore,Maryland 21212, USA.Correspondence to W.H.K.and C.V.D.emails: [email protected];[email protected]:10.1038/nrc3038Published online 14 April 2011

RespirationThe metabolic process bywhich energy is produced inthe presence of O 2 through the

oxidation of organiccompounds (typically sugars)to CO 2 and H 2O by glycolysis,the citric acid cycle andoxidative phosphorylation.

Otto Warburgs contributions tocurrent concepts of cancer metabolismWillem H. Koppenol*, Patricia L. Bounds* and Chi V. Dang

Abstract | Otto Warburg pioneered quantitative investigations of cancer cell metabolism, aswell as photosynthesis and respiration. Warburg and co-workers showed in the 1920s that,under aerobic conditions, tumour tissues metabolize approximately tenfold more glucose tolactate in a given time than normal tissues, a phenomenon known as the Warburg effect.However, this increase in aerobic glycolysis in cancer cells is often erroneously thought tooccur instead of mitochondrial respiration and has been misinterpreted as evidence fordamage to respiration instead of damage to the regulation of glycolysis. In fact, manycancers exhibit the Warburg effect while retaining mitochondrial respiration. We re-examineWarburgs observations in relation to the current concepts of cancer metabolism as beingintimately linked to alterations of mitochondrial DNA, oncogenes and tumour suppressors,and thus readily exploitable for cancer therapy.

REVIEWS

NATURE REVIEWS| CANCER ADVANCE ONLINE PUBLICATION| 1

Nature Reviews Cancer | AOP, published online 14 April 2011; doi:10.1038/nrc3038

2011 Macmillan Publishers Limited. All rights reserved
mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]


2/13


3/13

Habilitation

A quasi-independentpostdoctoral appointment thatis required for furtheracademic advancement inGerman-speaking countries.

Citric acid cycleA cyclic series of eightenzymatic reactions that occurin the mitochondrial matrixand that convert acetyl CoAderived from carbohydrates,fatty acids and amino acids toCO2 and H 2O; also known asthe tricarboxylic acid (TCA)cycle or Krebs cycle.

to attach his name to every publication emanatingfrom his institute. He was vigorous and arrogant in hisopposition to scientists who questioned his findings,even indulging in unscientific emotional attacks withinscientific reports1, and he was criticized by Krebs forhis tendency towards polemics9. Warburg preferredto employ instrument makers to whom he taught bio-chemistry and from whom he tolerated no argument.He enjoyed working with his hands and was a firmbeliever in quantitative methods. He continually soughtmeans to improve quantification in biological research:he invented the use of thin tissue slices for physiol-ogy research14, improved manometric techniques11 tomeasure changes in pressure accompanying cell andtissue processes14,15, and is credited as the inventor of the single-beam spectrophotometer11. These contribu-tions were pivotal to his research on metabolism andcancer physiology, which are described in a collectionof his early works15. These publications were ground-breaking because Warburg used quantitative physicalchemical approaches to investigate the rapid growthof cancer cells.

Formulation of the Warburg hypothesisWarburg studied and conducted research during agolden age of biochemical discovery (TIMELINE). Hecould trace his scientific lineage to Adolf von Baeyer

(who won the Nobel Prize in Chemistry in 1905 , andwas thus part of a scientific family that includes a dozenNobel laureates1,16.

In his earlier embryological investigations of seaurchin eggs, Warburg had observed a rapid increase inO2 uptake and subsequent rapid cell division upon fer-tilization17, and he postulated that cancer tissues mightalso take up more O2 than normal tissue. To address thishypothesis, Warburg used his improved manometrictechnique14,18 (FIG. 3) to measure O2 consumption in thintissue slices metabolizing glucose:

C 6 H 12 O 6 + 6O 2 6CO 2 + 6H 2 O (1

The Warburg manometer was also used to measure CO2 emission, which is equivalent to lactic acid production,from bicarbonate-containing buffers:

CH 3 CHOHCOOH + HCO 3

CH 3 CHOHCOO + H 2 O + CO 2

(2

Warburg and co-workers discovered that FlexnerJobling rat liver carcinoma does not take up more O2 than normal liver tissue, but that, even in the presenceof O2, such tissue produces lactic acid. This indicatesthe processing of glucose by lactic acid fermentation,bypassing the entry of pyruvate into thecitric acid cycle (respiration 18. As already mentioned, normal tissue wasknown to exhibit the Pasteur effect that is, to stopproducing lactic acid in the presence of O2. Human car-cinomas (from throat, intestine, skin, penis and nosealso demonstrated lactic acid production19,20.

Seigo Minami, an academic guest at the KWI forBiology, reported that although the respiration of FlexnerJobling rat liver carcinoma tissue slices is 20%less than that of normal tissue, which could be attributedto the presence of necrotic cells, approximately tenfoldmore glucose was metabolized than could be accountedfor by respiration. Minami confirmed Warburgs mano-metric lactic acid analysis by chemical means19, andWarburg subsequently determined that the amountof lactic acid produced by cancer cells is two orders ofmagnitude higher than that produced by normal tissue20.

With these methods, Warburg and co-workersdetermined how O2 affects glycolysis and defined theMeyerhof quotient as the molar ratio of the O2 con-sumed to the difference in lactic acid production underanaerobic conditions compared with aerobic condi-tions that is, a measure of the amount of O

2required

to convert one lactic acid molecule to glucose20. Fromexperiments with thin tumour tissue slices(FIG. 3), they determined a Meyerhof quotient of 1.3, which wasequivalent to that determined previously for normaltissues. As such, they concluded that respiration incancer tissue is normal but inadequate to prevent theformation of lactic acid. It should be noted that, inthe experiments performed in the presence of O2, glucosewas present in excess at all times, and the thickness of thetissue slices was limited to


4/13

Aerobic glycolysisThe enzymatic transformationof glucose to pyruvate in thepresence of O 2; see glycolysis.

that rates of respiration for ascites cells were comparableto those of muscle and yeast cells24,25, thus the enhancedproduction of lactic acid was not at the cost of respira-tion. Weinhouse26 also reported that cancer cells exhibitnormal rates of respiration and described Warburgscontentions as hypothesis based on essentially fallaciousreasoning, but his account was dismissed by Warburg7 and Burk and Schade27.

The lactic acid levels of mouse carcinoma and rat sar-coma tumoursin vivo were reported by Cori and Cori28 in 1925 to be very much lower than the levels observed inthe in vitro experiments of Warburg and co-workers29,30.Cori and Cori31further showed that the blood drawn froma vein exiting a Rous sarcoma tumour implanted on onewing of a chicken contained significantly less glucose andmore lactic acid than blood passing through the tissues of the corresponding normal wing, and they concluded that,in vivo, the excess lactic acid production in tumours iswashed out by the blood flow through the tissue. In similarexperiments on rats, Warburg and co-workers29,30reportedarterial and venous plasma levels of glucose and lactic acidin healthy organs compared with those in Jensens sarco-mas transplanted into the stomach; the glucose content of the veins from control organs was 218% less than thatof the arteries, compared with a 4770% drop acrossthe tumours. Arterial versus venous levels of lactic acidfrom tumours indicate that, on average, 66% of the glu-cose consumed is converted to lactic acid, whereas healthy organs produced no net lactic acid. Because cancer cellsrecycle lactic acid under aerobic conditions32, the lacticacid levels recorded in thein vivo experiments may belower than the actual levels produced by tumours29,30.The glucose and O2 concentration gradient acrosstissue decrees that the metabolism of tumour cells closer tothe arterial blood is more like that of in vitro tissue slices,whereas the metabolism of cells deeper in the tumour islimited by diffusion. Thus, thein vitro experiments betterreflect thein vivo conditions of cells close to the meta-bolic supply side of glucose and O2. Warburg attempted toaddress the influence of glucose and O2 supply to tumourcellsin vivo29,30, and concluded that it is difficult to inhibitthe growth of tumours in living animals through themanipulation of metabolic substrates.

Warburg and co-workers had expected that the O2 consumption of rapidly dividing cancer cells wouldbe greater than that of normal differentiated tissue, asoccurs in embryonic cells. The Meyerhof quotients of approximately 12 for thin slices of both normal andcancerous tissues20 indicate that O2 consumption (thatis, respiration by cancer tissues is the same as that of normal cells. Warburg believed respiration to be funda-mentally more complex than glycolysis and, therefore,more vulnerable to injury:

The origin of cancer lies in the anaerobic metaboliccomponent of normal growing cells, which ismore resistant to damage than is the respiratory component. Damage to the organism favours thisanaerobic component and, therefore, engenderscancer.33

Crabtree34 concurred in 1929: Warburg postulatesa disturbance of respiration as being the fundamentalcause of the development of aerobic glycolysis . Warburgreasoned that, since the increased production of lacticacid by cancer cells is not nullified by higher O2 con-sumption, respiration must be damaged33. Today, weunderstand that the Meyerhof quotient, as defined by Warburg, erroneously links respiration too intrinsi-cally to lactate production; further, Warburgs reasoningabout respiration that higher rates of respiration couldreduce the production of lactic acid35 is incorrect.Sonveauxet al.32 recently showed that normoxic cancercells metabolize lactic acid but anaerobic cells do not.This finding may explain Warburgs observation thatoxygenated tumour cells appear to produce less lacticacid (FIG. 3).

Is respiration damaged?The observations that cancer cells simultaneously oxi-dize and ferment glucose has engendered confusion overthe role of respiration in the Warburg effect, particularly as Warburg misinterpreted his own early observationsand promoted the erroneous idea that damaged respira-tion is thesine qua non that causes increased glucose fer-mentation in cancers. Thein vitro findings of Warburg

Timeline | Significant events in Warburgs life and relevant discoveries in cancer cell metabolism biochemistry

1860 1865 1869 1877 1883 1895 1901 1903 1905 1906 1911 1912 1914 1918 1922 1929

Pasteurdiscoversfermentation

Otto Warburg isborn in Freiburg(8 October)

Meischerdiscovers DNAin cell nuclei

Warburg beginsstudying chemistryin Freiburg

Warburg beginsstudying medicinein Berlin

Warburg becomesDoctor of Medicine

Warburgvolunteers formilitary service

Warburg is appointedAssoc. Prof. of Physiology in Berlin

Mendel publishestheory of genetics

Bernard observesconversion of glucose to lactic acid

Warburg familyrelocates to Berlin

Warburgcompletes hisPh.D. dissertation

Warburg tranfers toBerlin and joinsEmil Fischers group

Warburg beginsHabilitation inphysiology

Warburg iswounded andresumes research

ATP isdiscovered

Compiled from information in REFS 1, 911 . Red boxes refer to events in Warburgs life; black boxes refer to milestones in cancer metabolism research.

R E VI E W S

4 | ADVANCE ONLIN E PUBLICATION www.nature.com/reviews/cancer



5/13

Oxidative phosphorylation(OXPHOS). A metabolicprocess that occurs inmitochondria. It producesenergy in the form of ATP fromADP and inorganic phosphate,and is driven by a protongradient generated by thereactions of the citric acid cycle.

and co-workers20 show that, in the time required forcancer tissue under normoxic conditions to completely metabolize one molecule of glucose to yield 36 moleculesof ATP, ten more glucose molecules(FIG. 3) are convertedto 20 molecules of lactic acid to yield, at one ATP per lac-tic acid, an additional 20 molecules of ATP. Under anoxicconditions, cancer cells convert 13 glucose molecules to26 lactic acid and 26 ATP; thus, in the time it takes a nor-mal cell to produce 36 ATP from one glucose, the aerobiccancer cell produces 56 ATP from 11 glucose, whereasthe anoxic cancer cell generates 26 ATP from 13 glu-cose36. When Warburg and co-workers determined lacticacid levels, they found that the tumour removes 70 mgglucose and releases 46 mg of lactic acid per 100 ml of blood29,30, which, by our reckoning, corresponds to 10%more ATP produced by cancer cells than by normal cells.Recentin vitro data on glucose uptake and lactic acidrelease by human glioblastoma LN18 cells show a similar13% increase in ATP production37.

In 1956, Warburg reiterated the respiration of allcancer cells is damaged7, even though findings fromhis own laboratory 18 and those of others24,26 indicatedotherwise. In the second collection of his work pub-lished in 1962(REF. 35) , Warburg attempted to clarify andmodulate his classifications of cancer cells as well as to justify the conclusions he had drawn from his own work,admitting that the description based on insufficientrespiration had led to unfruchtbaren Kontroversen(fruitless controversy . Today, we understand that therelative increase in glycolysis exhibited by cancer cellsunder aerobic conditions was mistakenly interpreted asevidence for damage to respiration instead of damage tothe regulation of glycolysis.

Mitochondrial defects and the Warburg effectOver the past two decades, the discoveries of oncogenesand tumour suppressor genes have created a paradigmin which cell-autonomous genetic alterations were per-ceived as the sole driving force for neoplastic transfor-mation38,39and oncogenic alterations of cell metabolismwere considered as epiphenomena. However, with thediscoveries of oncogenic mutations in mitochondrialmetabolic enzymes, such as fumarate hydratase (FH ,

succinate dehydrogenase (SDH and isocitrate dehydro-genase 2 (IDH2 , it is now untenable to deny the role of metabolism in tumorigenesis40,41.

Warburg reasoned that respiration must be damagedin cancers because high levels of O2 are unable to sup-press the production of lactic acid by cancer cells42. So,are mitochondrial defects sufficient and necessary fortumorigenesis? Although the observations of Chanceand Weinhouse2426 negated Warburgs contention of mitochondrial defects in cancers, many studies overthe past several decades have documented oncogenicnuclear and mitochondrial DNA mutations in proteinsinvolved in respiration.

The metabolic profiles of chromaffin tissues, fromwhich paragangliomas and phaeochromocytomas arise,must somehow be amenable to tumorigenesis by muta-tions in these tumour suppressoroxidative phosphorylation(OXPHOS proteins. Mutations linked to hereditary paragangliomas and phaeochromocytomas in nucleargenes that affect mitochondrial respiration have beenfound in all four subunits (SDHA, SDHB, SDHC andSDHD of the SDH complex41. Mutations in SDH5,which is involved in the assembly of SDHD into thecomplex, were also recently documented in hereditary paragangliomas43 rare tumours that are not associatedclinically with more commonly occurring cancers. Thissuggests that these germline mutations are insufficientto promote commonly occurring epithelial cancers.Intriguingly, mutations of FH, which is involved in thecitric acid cycle downstream of SDH, result in familialleiomyoma, renal cell carcinoma (RCC and uterinefibroids. Mutations of SDH and FH promote increasedlevels of succinate and fumarate, which inhibit prolylhydroxylases that are responsible for the O2-dependentmodification of hypoxia inducible factor 1 (HIF1and its degradation. Therefore, even in the presence of normal levels of O2, these mutations are thought to con-stitutively increase production of HIF1 to levels thattrigger tumorigenesis44. In this regard, prolyl hydroxy-lases (particularly PHD2 confer the Pasteur effect by mediating the degradation of HIF1 in the presenceof O2 (REFS 45,46) . Specifically, HIF1, a heterodimercomprising HIF1 and HIF1 (also known as ARNT ,

1931 1937 1940 1943 1944 1945 1950 1953 1961 1963 1965 1970 1984 1989

Warburg establishes theKaiser Wilhelm Institute(KWI) for Cell Physiologyin Berlin-Dahlem

Glycolysispathway isformulated

Krebs discoversthe citric acid cycle

The KWI isevacuated toLiebenberg

DNAestablishedas geneticmaterial

The KWIreopens inBerlin-Dahlem

Jacob and Monodpropose theoperon model of gene control

Watson andCrick deducethe double helixstructure of DNA

Geneticcodeelucidated

First tumoursuppressor genedescribed

The KWI in Berlin-Dahlemis occupied by theAmerican army

The KWI for CellPhysiology becomes aMax Planck Institute

Warburg receivesHonoured Citizenof Berlin award

Firstoncogeneidentified

Warburg diesat 87 in Berlin(1 August)

Glycolytic enzymesidentified as targetsof oncogene products

Warburg is awarded the NobelPrize in Physiology or Medicine

R E VI E W S




6/13


7/13

a b

Anaerobic glycolysisThe enzymatic transformationof glucose to pyruvate in theabsence of O 2 ; see glyco lysis.

Deregulated glycolysis and the Warburg effectThe Warburg effect (aerobic glycolysis could arise frommtDNA mutations and defective respiration; however, asdiscussed, aerobic glycolysis can occur concurrently withmitochondrial respiration. Hence, if the Warburg effect is

evident in cancers with ongoing respiration, what are themechanisms underlying enhanced conversion of glucoseto lactic acid even in the presence of adequate O2?

All major tumour suppressors and oncogenes haveintimate connections with metabolic pathways6064 (FIG. 4). Some of the earliest evidence for links betweenoncogenes and aerobic glycolysis is the stimulation of glucose uptake by activated RAS and the ability of SRCto phosphorylate a number of glycolytic enzymes infibroblasts65,66. SRC was later implicated in the activationof HIF1, which induces glycolysis, but this link appearsto be dependent on cell type67,68. The first documenteddirect mechanistic link between an activated oncogene

and altered glucose metabolism was the transcriptionalactivation of LDHA by the oncogenic transcription factorMYC(FIG. 5a) , which later proved to activate most gly-colytic enzyme genes as well as glucose transporters6971.Pyruvate kinase M2 (PKM2 , which converts phospho-

enolpyruvate to pyruvate, favours aerobic glycolysis incellular transformation compared with PKM1, whichis encoded by alternative splicing of thePK mRNA72,73.MYC induces the splicing factors that produce PKM2,further underscoring the role of MYC in aerobic glyco-lysis74. MYC and HIF1 share many target glycolyticenzyme genes; however, whereas the normal role of HIF1 is to induceanaerobic glycolysis , MYC can stimulateaerobic glycolysis, as shown when it is overexpressedin vivo in transgenic cardiomyocytes70,75.

The AKT oncogenes, which are frequently activated downstream of PI3K, enhance glycolysis throughactivation of hexokinase 2 and phosphofructokinase 1

Figure 3 | The reaction vessel for tissue slices developed by Otto Warburg and representative data. a | Thereaction vessel used by Warburg and co-workers 14 to measure O 2 uptake or lactic acid production consisted of achambered trough in which a tissue slice (S), cut with a razor blade, was mounted on a glass needle (N, fixed to the bottomof the main chamber) and submerged in 0.5 ml Ringer solution. The vessel was closed with a paraffin-coated ground glass joint (H) attached to tubing that connects to a Barcroft manometer. The solid glass bulb (G) served as a handle to facilitatefitting the glass joint, and additions to the reaction trough were made through port T (sealed with a glass stopper duringmeasurements). For measurements of O 2 uptake (which registered as pressure decreases over time), 0.1 ml of 5%potassium hydroxide solution was added to chamber E to absorb CO 2. Lactic acid production was measured as pressureincreases due to CO 2 emission from the Ringer solution, which, for these experiments, contained 24 mM NaHCO 3 (REF. 14) .O2 uptake and/or CO 2 release were measured at 37.5 C for 0.51 hour. Warburg

18 calculated that, to avoid anaerobiosis inthe centre, the tissue thickness must be smaller than 8c oDA

1, where c o is pO 2, D is the diffusion coefficient of O 2 (1.4 105cm 3 O2 per cm

2 tissue at 38 C 129) and A is the O 2 consumption of the tissue; this corresponds to a tissue sample0.20.4 mm in thickness and 25 mg in weight. b | Results obtained using the apparatus in a from experiments withFlexnerJobling rat carcinoma tissue at 37.5 C, 0.2% glucose 20, at pH 7.41 (not 7.66 as indicated 130), in which the respiration

(per mg of dried tissue) was 7.2 mm3

O2 per hour (0.28 mol per hour). The volume of CO 2 driven out of the Ringer solutionby lactic acid during respiration in the presence of O 2 was 25 mm 3 per hour (0.93 mol per hour), and in the presence of N 2 the volume was 31 mm 3 per hour (1.22 mol per hour) (values in parentheses calculated for this Review). The uptake of 0.28 mol O 2 per hour implies that 0.047 mol glucose is oxidized to H 2O and CO 2 (see equation 1 (respiration)). The CO 2 produced during the aerobic and anaerobic experiments corresponds to 0.93 mol and 1.22 mol lactic acid (see equation2 (glycolysis)), respectively, or 0.46 and 0.61 mol glucose, respectively. Thus, in tumour cells in the presence of O 2, tentimes more glucose is used for glycolysis than for respiration. Image is reproduced, with permission, from REF. 15 (1926)Springer Science+Business Media.

R E VI E W S




8/13

Pyruvate

Acetyl CoAOxaloacetate

Malate Citrate

IDH Isocitrate

-ketoglutarate

Fumarate

Succinate

FH

SDH

Glutamate

Glutamine

GS

-catenin

NF-B

Rho-GTPase

p53

p53

TIGAR

GLS2

GSH

ASCT2

MYC

RAS

Glucose

AKT

VHL

PDK1

GLS

NADPHand ribose

Lactate

HIF1

OXPHOS

PGC1

H2O2 and O 2

SCO2

GLUT

HK2

PGAM2

PKM2

Pyruvate

LDHA

(PFK1; also known as PFKM and PFK2 (also knownas PFKFB3 and recruitment of glucose transporters tothe cell surface37,76,77 (FIG. 5a) . Although AKT functionsindependently of HIF1 to induce aerobic glycolysis78,79,it can also increase the activity of HIF1, furtherenhancing induction of glycolysis80. Ectopic expressionof AKT or MYC induces aerobic glycolysis in FL5.12pre-B cells but, unlike MYC, AKT does not increasemitochondrial function81. Intriguingly, aerobic gly-colysis in early passage human breast cancer cells isassociated with elevated HIF1 or MYC but not activated AKT82. Hence, it is likely that the cellular contextand the range of cancer-specific mutations are impor-tant for the metabolic manifestations of activated

oncogenes such as AKT.Activated RAS was initially linked to increased cel-lular glucose transport, but recent studies indicate thatthe role of RAS in cancer metabolism is more complex.It was recently reported that depriving colon carci-noma cells of glucose increases the mutation rate of RAS, which, thus activated, facilitates glucose importthrough induction of GLUT1 (also known as SLC2A1 ,an important glucose transporter83. In a multistep,multigene transformation of human breast epithelialcells, it was documented that the initial transforma-tion of normal epithelial cells by viral oncogenes andtelomerase reverse transcriptase is associated with

increased mitochondrial function; with activated KRASas the final reaction step in this model, the transformedcells exhibit the Warburg effect through high conver-sion of glucose to lactate84. It is notable that activatedRAS has been proposed to induce MYC activity andenhance non-hypoxic levels of HIF1, although the pre-cise mechanisms remain to be established85,86. Hence,RAS could mediate its effects on metabolism throughHIF1 or MYC(FIG. 5a) .

Because HIF1 appears at the crossroads of multipleoncogenes that can stabilize HIF1 under non-hypoxicconditions, it is not surprising that HIF1 also has a piv-otal role in the manifestations of tumour suppressors(FIG. 5a) . For example, the von Hippel-Lindau (VHL

tumour suppressor protein, which normally mediatesproteasomal degradation of HIF1, is lost in RCCs,which results in elevated non-hypoxic expression of HIF1 and HIF287. In RCCs, MYC appears to col-laborate with activated HIF2 to confer tumorigenicity,whereas HIF1 appears to be expressed in RCCs onlywhen HIF2 is expressed, suggesting a potential tumoursuppressive function of HIF1. Other tumour suppres-sor genes and proteins have also been implicated asmodulators of HIF1, and thereby might contribute tothe Warburg effect; for example, HIF1-mediated geneexpression is facilitated by loss of thePTEN tumour sup-pressor gene88. The association of the tumour suppressor

Figure 4 | The regulation of metabolism in cancer. Oncoproteins and tumour suppressors (shown in red) are intimatelylinked to metabolic pathways through transcriptional or post-transcriptional regulation of metabolic enzymes; arrows inbold depict the conversion of wild-type to mutant ( ) tumour suppressors or mutant activated oncogenes, presumably bymutational oxidative DNA damage. ASCT2, ASC-like Na +-dependent neutral amino acid transporter 2 (also known as ATB(0)and SLC1A5); FH, fumarate hydratase; GLS, glutaminase; GS, glutamine synthetase; GLUT, glucose transporter; GSH,glutathione; HIF1, hypoxia-inducible factor 1; HK2, hexokinase 2; IDH, isocitrate dehydrogenase; LDHA, lactatedehydrogenase A; NF- B, nuclear factor- B; OXPHOS, oxidative phosphorylation; PDK1, pyruvate dehydrogenase kinaseisoform 1; PGAM2, phosphoglycerate mutase 2; PGC1 , peroxisome proliferator-activated receptor- , co-activator 1 ;PKM2, pyruvate kinase M2; SDH, succinate dehydrogenase; TIGAR, tumour protein 53-induced glycolysis and apoptosisregulator; VHL, von Hippel-Lindau tumour suppressor.

R E VI E W S

8 | ADVANCE ONLIN E PUBLICATION www.nature.com/reviews/cancer



9/13


10/13

transcriptional regulation, whereby wild-type p53 stim-ulates mitochondrial respiration and suppresses glyco-lysis(FIG. 5a) . Activation of SCO2 (which regulates thecytochromecoxidase complex by p53 increases the effi-ciency of mitochondrial respiration94. Conversely, p53suppression of phosphoglycerate mutase 2 (PGAM2and activation of tumour protein 53-induced glycolysis andapoptosis regulator (TIGAR , which has 2,6-fructosebisphosphatase activity and depletes PFK1 of a potentpositive allosteric ligand, suppresses glycolysis andfavours increased NADPH production by the pen-tose phosphate pathway 91,95. Hence, loss of p53 func-tion induces aerobic glycolysis, presumably throughincreased PGAM and PFK activities.

Cancer metabolism unanticipated by WarburgAlthough oncogenic alteration of metabolism generally involves the Warburg effect, the enhanced flux of glucoseto lactate is insufficient to promote cell replication61. Cellsare largely comprised of protein and ribonucleic acid,and so are too complex to be supported by a simple glu-cose carbon skeleton; hence, other metabolic pathwaysmust also be stimulated to provide the building blocksfor cell replication. Although previously implicated inthe literature9698, the contribution of glutamine to ana-bolic carbons and building blocks of a growing cell hasbeen rediscovered and only recently fully appreciated. Infact, citric acid cycle intermediates in proliferating cellsare hybrid molecules of glucose and glutamine carbons,with glutamine entering the citric acid cycle throughconversion to glutamate by glutaminase (GLS and thento -ketoglutarate by either glutamate dehydrogenaseor aminotransferases99. Furthermore, proliferating cellsgenerate waste and toxic by-products, the removal of which is necessary for cancer cells to maintain redoxhomeostasis and continue replicating effectively 84.

MYC has been documented to induce genes involvedin mitochondrial biogenesis and glutamine metabolism70,specifically those for expression of glutamine transportersand GLS, resulting in increased flux of glutamine carbonsthrough the citric acid cycle100,101 (FIG. 5b) . Thus, over-expression of MYC in cancer cells renders them sensitiveto glutamine withdrawal102. The ability of MYC to induceboth aerobic glycolysis and glutamine oxidation providescancer cells with ATP, carbon skeletons and nitrogenfor nucleic acid synthesis, and hence with the ability toaccumulate biomass. Activated Rho-GTPase-mediatedtransformation is dependent on increased GLS activ-

ity, which appears to be modulated by activated nuclearfactor-B (NF-B ; chemical inhibition of GLS dimin-ishes transformation by both Rho-GTPase and MYC,showing that key metabolic nodal points can be affectedby different oncogenes103 (FIG. 5b) . Activated RAS wasalso recently shown to rely on mitochondrial func-tion for cellular transformation, particularly throughincreased glutamine metabolism, which suggests thatthe multifaceted roles of oncogenes in metabolism arecontext dependent104.

The mutant -catenin (CTNNB1 oncogene increasesglutamine synthetase (GS expression in liver cancers105 (FIG. 5b) ; GS produces glutamine from glutamate and

ammonia, hence its expression renders cancer cellsindependent of extracellular glutamine, although GSappears to be decreased overall in hepatocellular carci-noma (HCC , whereas GLS is elevated106,107. The HCCsubtype with GS expression portends a more favour-able clinical outcome108. These collective observa-tions suggest that GS expression in some liver cancersreflects the expression of GS that is required in normalliver cells for ammonia detoxification and glutamineproduction109.

GLS2is transactivated by p53 and is normally expressedin the liver110112 (FIG. 5b) . By contrast, GLS is ubiquitously inducible. The increased conversion of glutamine to gluta-mate by GLS2 is thought to increase the production of glu-tathione, which in turn attenuates metabolic by-productssuch as hydrogen peroxide. Hence, beyond the Warburgeffect, p53 plays a key part in redox homeostasis throughstimulation of NADPH synthesis by the pentose phos-phate pathway and stimulation of glutathione synthesisthrough increased GLS2 expression.

Other alterations favouring oncogenesis includereceptor tyrosine kinase activation, such asERBB2 (also known asHER2 amplification in breast cancer;ERBB2 can suppress apoptosis resulting from celldetachment from other cells or the substratum(anoikis in mammary spheroid cultures, in which cen-tral mammary epithelial cells that are detached fromsurrounding cells have diminished glucose uptake andundergo apoptosis113. It was observed that anoikis isassociated with increased oxidative stress that inhibitsfatty acid oxidation, resulting in a bioenergetic deaththat can be rescued by expression of ERBB2, whichstimulates glucose uptake, NADPH production by thepentose phosphate pathway and fatty acid oxidation,and this consequently diminishes oxidative stress. Therole of fatty acids as bioenergetic substrates for canceris not well understood and deserves more attention.

PerspectivesAlthough normal cells experience the enhanced aerobicglycolysis of the Warburg effect114,115, there is one dis-tinct metabolic difference between normal and cancercells that renders cancer cells addicted to the Warburgeffect. Normal cells, by virtue of multiple feedback andfeedforward regulatory loops, undergo quiescence whendeprived of nutrients even in the presence of growthfactors. By contrast, oncogenic stimulation of cellgrowth and proliferation induces both biomass accu-

mulation (such as increased ribosome biogenesis andlipogenesis and nutrient uptake. When bioenergeticdemand is balanced by anabolic supply, cancer cellsgrow and proliferate. However, oncogenic deregulationof biomass accumulation for cell proliferation creates anincreased, sustained bioenergetic demand that addictscancer cells to an adequate anabolic supply. In thisregard, the Warburg effect, in addition to contributingto enhanced lactic acid production, serves to provideanabolic carbons for fatty acid synthesis60. For exam-ple, MYC-induced ribosome biogenesis and biomassaccumulation sensitizes MYC-transformed cells to bio-energetic cell death triggered by glucose or glutamine

R E VI E W S

10 | ADVANCE ONLINE PUBLICATION www.nature.com/reviews/cancer



11/13

deprivation, much like yeast mutants that have constitu-tively deregulated ribosome biogenesis102,116,117. This piv-otal conceptual framework of bioenergetic supply anddemand suggests that cancer cells are addicted to theWarburg effect, and nutrient deprivation should triggeran autophagic response, which, if unsustainable, wouldresult in cancer cell death118. Hence, targeting metabo-lism for cancer therapy holds promise for new classesof anti-neoplastic drugs119,120.

The microenvironmental niches in which cancer cellslive are heterogeneous because of ineffective tumour vascularization121; as such, the genomic and metabolicnetworks of cancer cells are disrupted not only by cell-autonomous genetic mutations but also by hypoxia122.Indeed, it was demonstrated recently that hypoxictumour cells extrude lactate, which is subsequently recy-cled to pyruvate for use in mitochondrial OXPHOS by respiring stromal or tumour cells32,121,123,124.

The concepts for cancer cell metabolism framed by Warburg 90 years ago have undergone substantial revi-sion. Taken together, the progress made in the twenty-

first century towards understanding the Warburgeffect reveals that genetic alterations of oncogenesand tumour suppressors tend to increase the conver-sion of glucose to lactate, but glucose is insufficientfor cancer cell growth and proliferation. Furthermore,accelerated cancer cell metabolism also produces morewaste, such as lactate, superoxide and hydrogen per-oxide, for extrusion or neutralization125,126. However,the addiction of cancer cells to the Warburg effectfor biomass accumulation can be exploited by thera-peutic approaches that uncouple bioenergetic supply from demand or inhibit elimination of metabolic wasteproducts. The Warburg effect itself involves high lev-els of aerobic glycolysis catalysed by pivotal enzymethat are therapeutically accessible to small drug-likeinhibitors that could be aimed at primary and meta-static tumours and monitored in patients by means of metabolic imaging. As such, we are poised to witnessthe clinical benefits of Warburgs contributions in thenext 5 to 10 years, almost 100 years after his initialobservations103,127,128.

1. Werner, P. Ein Genie irrt seltener. Otto HeinrichWarburg. Ein Lebensbild in Dokumenten(Akademie

Verlag, Berlin, 1991).2. Warburg, O. ber den Stoffwechsel der Carcinomzelle.

Klin. Wochenschr. 4 , 534536 (1925).3. Keilin, D.The History of Cell Respiration and

Cytochrome(Cambridge Univ. Press, Cambridge,1970).

4. Racker, E. Bioenergetics and the problem of tumorgrowth. Am. Sci. 60 , 5663 (1972).In this paper, Racker coins the term the Warburg effect in describing his own hypothesis on theorigins of tumour growth.

5. Turner, J. S. & Brittain, E. G. Oxygen as a factorin photosynthesis. Biol. Rev. 37 , 130170(1962).

6. Pedersen, P. L. The cancer cells power plants aspromising therapeutic targets: an overview.

J. Bioenerg. Biomembr. 39 , 112 (2007).7. Warburg, O. On respiratory impairment in cancer

cells. Science 124 , 269270 (1956).8. Krebs, H. Otto Heinrich Warburg. 18831970. Biogr.

Mems Fell. R. Soc. 18 , 628699 (1972). An excellent English-language biography of OttoWarburg.

9. Krebs, H. Otto Warburg, Zellphysiologe, Biochemiker,Mediziner (Wissenschaftliche Verlagsgesellschaft mbH,Stuttgart, 1979).

10. Werner, P. Otto Warburg. Von der Zellphysiologie zur Krebsforschung (Verlag Neues Leben, Berlin,1988).

11. Hxtermann, E. & Sucker, U. Otto Warburg (BSB B. G.Teubner Verlagsgesellschaft, Leipzig, 1989).

12. Koepcke, C.Lotte Warburg (Iudicium, Mnchen,2000).

13. Einstein, A. Emil Warburg als Forscher. Naturwiss. 10 ,823826 (1922).

14. Warburg, O. Versuche an berlebendem Carcinom-Gewebe (Methoden). Biochem. Zeitschr. 142 ,317333 (1923).

15. Warburg, O. ber den Stoffwechsel der Tumoren. Arbeiten aus dem Kaiser WilhelmInsti tut fr Biologie BerlinDahlem(Julius Springer, Berlin,1926).

16. Sri Kantha, S. The question of nepotism in the awardof Nobel prizes: a critique of the view of Hans Krebs.Med. Hypotheses 34 , 2832 (1991).

17. Warburg, O. Notizen zur Entwickelungsphysiologie desSeeigeleies. Arch. f. d. ges. Physiol. 160 , 324332(1915).

18. Warburg, O. Verbesserte Methode zur Messung der Atmung und Glykolyse.Biochem. Zeitschr. 152 ,5163 (1924).

19. Minami, S. Versuche an berlebendemCarcinomgewebe. Biochem. Zeitschr. 142 , 334350(1923).

20. Warburg, O., Posener, K. & Negelein, E. ber denStoffwechsel der Carcinomzelle. Biochem. Zeitschr. 152 , 309344 (1924).In this landmark paper, Warburg and co-workersreported quantitative descriptions of respirationand lactic acid production measured by manometry in a variety of normal, embryonic and canceroustissues.

21. Freyer, J. P., Tustanoff, E., Franko, A. J. & Sutherland,R. M. In situ oxygen consumption rates in V-79multicellular spheroids during growth. J. Cell. Physiol. 118 , 5361 (1984).

22. Braun, R. D. & Beatty, A. L. Modeling of oxygentransport across tumor multicellular layers. Microvasc.Res. 73 , 113123 (2007).

23. Warburg, O. & Hiepler, E. Versuche mit Ascites-Tumorzellen. Z. Naturforsch. 7b , 193194 (1952).

24. Chance, B. & Castor, L. N. Some patterns of therespiratory pigments of ascites tumors in mice.Science 116 , 200202 (1952).

25. Chance, B. & Hess, B. Spectroscopic evidence of metabolic control. Science 129 , 700708 (1959).

26. Weinhouse, S. On respiratory impairment in cancercells. Science 124 , 267269 (1956).

27. Burk, D. & Schade, A. L. On respiratory impairment incancer cells. Science 124 , 270272 (1956).

28. Cori, C. F. & Cori, G. T. The carbohydrate metabolismof tumors. I. The free sugar, lactic acid, and glycogencontent of malignant tumors. J. Biol. Chem. 64 ,1122 (1925).

29. Warburg, O., Wind, F. & Negelein, E. ber denStoffwechsel der Tumoren in Krper. KlinischeWochenschrift 5 , 829832 (1926).

30. Warburg, O., Wind, F. & Negelein, E. The metabolismof tumors in the body. J. Gen. Physiol. 8 , 519530(1927).Reference 29 is an important work in which the

consumption of glucose and production of lacticacid in tumours transplanted in rats weremeasured directly and compared with thecorresponding metabolism by normal tissues.Reference 30 is an English translation of reference 29.

31. Cori, C. F. & Cori, G. T. The carbohydratemetabolism of tumors. II. Changes in the sugar,lactic acid, and co-combining power of bloodpassing through a tumor. J. Biol. Chem. 65 ,397405 (1925).

32. Sonveaux, P. et al. Targeting lactate-fueled respirationselectively kills hypoxic tumor cells in mice. J. Clin.Invest. 118 , 39303942 (2008).

33. Warburg, O. ber den heutigen Stand desCarcinomproblems. Naturwiss. 15 , 14 (1927).

34. Crabtree, H. G. Observations on the carbohydratemetabolism of tumors. Biochem. J. 23 , 536545(1929).

35. Warburg, O. in New Methods of Cell Physiology Applied to Cancer, Photosynthesis, and Mechanism of XRay Action. Developed 19451961 (ed. Warburg, O.)631632 (Interscience Publishers, New York, 1962).This book contains reprints of Warburgs work onboth cancer cell metabolism and photosynthesispublished from 19451961. Most contributions arein German, but some are in English, including athree-part forum On respiratory impairment incancer cells that appeared in Science in 1956. In thepenultimate chapter of the book, Warburg reviseshis classification of cancer cells as cells in whichrespiration is insufficient rather than impaired.

36. Koppenol, W. H. & Bounds, P. L. The Warburg effectand metabolic efficiency: re-crunching the numbers.Science [online], http://www.sciencemag.org/content/324/5930/1029/reply#sci_el_12397?sid=398be983-ebcd-4502-817d-e3f931b9bc37 (2009).

37. Elstrom, R. L. et al. Akt stimulates aerobic glycolysis incancer cells. Cancer Res. 64 , 38923899 (2004).

38. Varmus, H., Pao, W., Politi, K., Podsypanina, K. & Du, Y. C. Oncogenes come of age. Cold Spring Harb. Symp.Quant. Biol. 70 , 19 (2005).

39. Land, H., Parada, L. F. & Weinberg, R. A. Tumorigenicconversion of primary embryo fibroblasts requires atleast two cooperating oncogenes. Nature 304 ,596602 (1983).

40. Parsons, D. W. et al. An integrated genomic analysis of human glioblastoma multiforme. Science 321 ,18071812 (2008).This seminal paper reports frequent mutations in IDH1 and IDH2 in human brain cancer.

41. Bayley, J. P. & Devilee, P. Warburg tumours and themechanisms of mitochondrial tumour suppressorgenes. Barking up the right tree? Curr. Opin. Genet.Dev. 20 , 324329 (2010).

42. Nachmansohn, D. GermanJewish Pioneers in Science,

19001933 (Springer, New York, 1979).43. Hao, H. X.et al. SDH5, a gene required for flavination

of succinate dehydrogenase, is mutated inparaganglioma. Science 325 , 11391142 (2009).

44. King, A., Selak, M. A. & Gottlieb, E. Succinatedehydrogenase and fumarate hydratase: linkingmitochondrial dysfunction and cancer. Oncogene 25 ,46754682 (2006).

45. Seagroves, T. N. et al. Transcription factor HIF-1 is anecessary mediator of the Pasteur effect in mammaliancells. Mol. Cell. Biol. 21 , 34363444 (2001).

46. Kaelin, W. G. Jr & Ratcliffe, P. J. Oxygen sensing bymetazoans: the central role of the HIF hydroxylasepathway. Mol. Cell 30 , 393402 (2008).

47. Firth, J. D., Ebert, B. L., Pugh, C. W. & Ratcliffe, P. J.Oxygen-regulated control elements in thephosphoglycerate kinase 1 and lactate dehydrogenase

A genes: similarities with the erythropoietin 3 enhancer.Proc. Natl Acad. Sci. USA 91 , 64966500 (1994).

R E VI E W S


http://www.sciencemag.org/content/324/5930/1029/reply#sci_el_12397?sid=398be983ebcd4502817de3f931b9bc37http://www.sciencemag.org/content/324/5930/1029/reply#sci_el_12397?sid=398be983ebcd4502817de3f931b9bc37


12/13


13/13

115. Lemons, J. M. et al. Quiescent fibroblasts exhibit highmetabolic activity. PLoS Biol. 8 , e1000514 (2010).

116. Shim, H., Chun, Y. S., Lewis, B. C. & Dang, C. V. A unique glucose-dependent apoptotic pathwayinduced by c-Myc.Proc. Natl Acad. Sci. USA 95 ,15111516 (1998).

117. Lippman, S. I. & Broach, J. R. Protein kinase A andTORC1 activate genes for ribosomal biogenesis byinactivating repressors encoded by Dot6 and itshomolog Tod6. Proc. Natl Acad. Sci. USA 106 ,1992819933 (2009).

118. Rabinowitz, J. D. & White, E. Auophagy andmetabolism. Science 330 , 13441348 (2010).119. Le, A.et al. Inhibition of lactate dehydrogenase A

induces oxidative stress and inhibits tumorprogression. Proc. Natl Acad. Sci. USA 107 ,20372042 (2010).

120. Michelakis, E. D. et al. Metabolic modulation of glioblastoma with dichloroacetate. Sci. Transl. Med. 2 ,31ra34 (2010).

121. Dang, C. V. Rethinking the Warburg effect with Mycmicromanaging glutamine metabolism. Cancer Res. 70 , 859862 (2010).

122. Kim, J. W., Gao, P., Liu, Y. C., Semenza, G. L. &Dang, C. V. Hypoxia-inducible factor 1 anddysregulated c-Myc cooperatively induce vascular

endothelial growth factor and metabolic switcheshexokinase 2 and pyruvate dehydrogenasekinase 1. Mol. Cell. Biol. 27 , 73817393 (2007).

123. Funes, J. M. et al. Transformation of humanmesenchymal stem cells increases their dependencyon oxidative phosphorylation for energy production.Proc. Natl Acad. Sci. USA 104 , 62236228 (2007).

124. Fogal, V. et al. Mitochondrial p32 protein is a criticalregulator of tumor metabolism via maintenance of oxidative phosphorylation. Mol. Cell. Biol. 30 ,13031318 (2010).

125. Kroemer, G. & Pouyssegur, J. Tumor cell metabolism:cancers Achilles heel.Cancer Cell 13 , 472482(2008).

126. Cairns, R. A., Harris, I. S. & Mak, T. W. Regulation of cancer cell metabolism. Nature Rev. Cancer 11 ,8595 (2011).

127. Tennant, D. A., Durn, R. V. & Gottl ieb, E. Targetingmetabolic transformation for cancer therapy. NatureRev. Cancer 10 , 267277 (2010).This is a comprehensive critical Review of theemerging area of therapeutic targeting of cancermetabolism.

128. Thornburg, J. M. et al. Targeting aspartateaminotransferase in breast cancer. Breast Cancer Res. 10 , R84 (2008).

129. Krogh, A. The rate of diffusion of gases throughanimal tissues, with some remarks on thecoefficient of invasion. J. Physiol. 52 , 391408(1919).

130. Warburg, O. ber Milchsurebildung beimWachstum. Biochem. Zeitschr. 160 , 307311(1925).

AcknowledgementsWe thank John Eaton for instigating this collaborative Review.The authors work is partially funded by US National Cancer

Institute grants (C.V.D.), the Leukemia Lyphoma Society(C.V.D.) and an American Association for Cancer ResearchStand Up To Cancer translational grant (C.V.D.). We alsoacknowledge support by the Swiss Federal Institute of Technology Zurich (P.L.B. and W.H.K.).

Competing interests statementThe authors declare competing financial interests . See Webversion for details.

FURTHER INFORMA TIONChi V. Dangs homepage: http://www.myccancergene.org/

ALL LINKS ARE ACTIVE IN THE ONLINE PDF

R E VI E W S
http://www.nature.com/nri/journal/vaop/ncurrent/box/nrc3038_audecl.htmlhttp://www.myccancergene.org/http://www.myccancergene.org/http://www.myccancergene.org/http://www.myccancergene.org/http://www.nature.com/nri/journal/vaop/ncurrent/box/nrc3038_audecl.html

otto warburg’s contributions to cancer metabolism

Documents