Distribution of glucose transporter messenger RNA transcripts in tissues of rat and man

Distribution of glucose transporter messenger

RNA transcripts in tissues of rat and man.

J S Flier, … , A L McCall, H F Lodish

J Clin Invest.

1987;

79(2)

:657-661.

https://doi.org/10.1172/JCI112864

.

We used the complementary DNA for the human hepatoma Hep G2 glucose transporter to

determine the distribution of glucose transporter messenger RNA (mRNA) in rat and human

tissues. Under stringent hybridization conditions, a single 2.8-kilobase (kb) transcript is

seen in all rat and human tissues examined. The mRNA is most abundant in brain, and is

especially enriched in the brain microvascular fraction. The mRNA abundance in rat muscle

and fat is 5% that in brain. Rat liver (both adult and fetal) and human liver have very little

2.8-kb mRNA, but it is abundant in cultured human fibroblasts and EB virus-transformed

lymphoblasts. The same size mRNA is present in leg muscle of two type II diabetic patients.

A very homologous glucose transporter mRNA is expressed in both insulinsensitive and

-insensitive tissues of rat and man. Hepatocytes, which have abundant glucose transport,

may express a homologous but nonidentical glucose transporter.

Research Article

Find the latest version:

http://jci.me/112864/pdf

Rapid Publication

Distribution

of Glucose

Transporter Messenger

RNA

Transcripts

in

Tissues of Rat and

Man

Jeffrey S.Flier,**MikeMueckler,1 Anthony L.McCall,11andHarveyF. Lodish*¶

*The Whitehead Institute forBiomedicalResearch, 'Department ofBiology,MassachusettsInstitute_{ofTechnology,}

Cambridge, Massachusetts 02139; tThe Charles A. Dana Research Institute and the Harvard-Thorndike Laboratory

ofBeth Israel Hospital, Department ofMedicine, Beth Isreal Hospital and Harvard Medical School, Boston, Massachusetts 02215;

IDepartment

ofCellBiologyandPhysiology, Washington UniversitySchoolofMedicine,St. Louis, Missouri63110;

IlEvans

MemorialDepartment ofMedicine and DivisionofDiabetes and Metabolism, University Hospital, andDepartment ofPhysiology, BostonUniversityMedicalCenter,Boston,Massachusetts 02118

Abstract

We used the complementary DNA for the human hepatoma Hep G2glucosetransportertodetermine the distribution ofglucose transportermessengerRNA(mRNA)inratand humantissues. Understringenthybridization conditions, asingle 2.8-kilobase (kb) transcript is seenin allratand human tissues examined. The mRNA ismostabundant in brain, and is especially enriched inthe brain microvascular fraction.

The mRNA abundance inratmuscle and fatis5% thatin brain.Rat liver (both adult and fetal) and human liver havevery little 2.8-kb mRNA, but it is abundantincultured human fibro-blasts and EBvirus-transformed lymphoblasts.Thesamesize mRNAispresent inlegmuscle oftwotypeIIdiabeticpatients. Averyhomologous glucosetransporter mRNAisexpressed inbothinsulin-sensitive and -insensitive tissuesofratandman.

Hepatocytes, which have abundant glucose transport, may ex-press ahomologous but nonidentical glucosetransporter.

Introduction

Glucose homeostasis inmammalsrequires both the uptake of extracellularglucose into tissues suchasmuscle, fat, and liver

aswellasthe mobilizationof glucose from hepatocytes intothe

circulation. The molecular basis for cellular glucosetransport

hasbeen intensively investigated, and thisprocess appearstobe mediated byone or moreglucose-specifictransportproteins (1). Although glucosetransport has been studied inmany tissues, the bestcharacterizedglucosetransportprotein isthefacilitated diffusionglucosecarrierof the human erythrocyte,a55,000-kD integral membrane glycoprotein (2-4). Recently, Mueckler et

al.usedarabbit antibody directed against thisproteintoobtain

acomplementaryDNA(cDNA) clone for the glucosetransporter

Addressreprintrequests toDr.Flier, DiabetesUnit, BethIsrael Hospital,

330 Brookline Ave., Boston, MA02215.

Receivedfor publication 18 July 1986 and in revisedform1October

1986.

from a lambda gt- 11 cDNA library prepared from theHep-G2 human hepatoma cell line (5). Theamino acidsequence deduced from the cDNA clone agreedwith thepartial sequencesavailable for theerythrocytetransporter(5), suggestingthat human eryth-rocytesand hepatocytes, two tissues in which glucose transport

is insensitivetoinsulin,mayhave thesametransporterspecies.

Incontrast toerythrocytesandhepatocytes,therateofglucose

transport in tissues such as muscle (6) and fat (7) is strongly

influencedbyinsulin. Itis unknown, however,whether differ-encesbetweentissuesinthesensitivityof theirglucosetransport systemstoinsulinaredue, entirelyorin part,todifferencesin thestructure of theglucosetransporter.As an initialapproach tothis question, we undertookto examine the tissue-specific expression of mRNAencoding orhomologous totheHep-G2

glucose transporter in various tissues ofratandman.

Methods

RNAisolation.RNA wasisolated from fresh tissueorcultured cellsby the guanadinium thiocyanate-CsCl technique (8). Where indicated, polyadenylatedRNAwasobtainedbyoligodeoxythymidylatecellulose chromatography (9).Isolatedrathepatocyteswereprepared bythe col-lagenasetechnique (10),and calf brain microvesselswerepreparedas described byMcCalletal.(11).

Northerngels.RNA waselectrophoresedon1.2%formaldehyde aga-rosegels(12), blotted and fixedontonylon filters, and then hybridized

toeither cDNAorantisenseRNAprobes. The cDNA probe isamixture oftwo [32P]nick-translated glucose transportercDNAfragments. The

twofragmentsare450 basepair (pGT25S)and2,400basepair (pGT25L) EcoRIfragments that together contain nearly a full-length copy of the

mRNA(5). Nick translation was performed using a nick translation kit (Amersham Corp., Arlington Heights, IL) according to instructions specified by the manufacturer.

Hybridizationwith the cDNA probe was carried out at

420C

in a solution comprised of 50%formamide,5XSSPE (0.9 M NaCI, 5 mM EDTA, and 50 mM NaH2PO4, pH 7.4), 0.2% sodium dodecyl sulfate,

0.1% eachof bovine serumalbumin, polyvinylpyrolidine, and Ficoll,

anddenatured,shearedsalmonspermDNA(200

,g/ml).

Theprobe was included at I0'cpm/ml. Afterbeing washedin0.1 X SSPEat 50°C (threewashesof 30 min each), theblot wasexposedtoKodakXAR-5

filmat-_{70°C for varying time periods as indicated in figure legends}

(intensifying screen, Cronex Lightening Plus, DuPont, Wilmington, DE). RNAprobe.TheHep-G2 glucose transporter cDNA was subcloned into theBamHIsite of the pGEM plasmid (Promega Biotech, Madison, WI) and theantisense RNA was synthesized using T7 RNA polymerase and[32P]GTP(410Ci/mmol,spact)asdescribed by the manufacturer.

The labeled RNA wasseparated from unincorporated GTP by G-50

GlucoseTransporter MessengerRNALevels 657

J.Clin. Invest.

© The American Society for Clinical Investigation, Inc. 0021-9738/87/02/0657/05 $ 1.00

Sephadex chromatography. Hybridization to the RNA probe wascarried

out at650Cin the same hybridization solution as described above except that theformamide concentration was varied from 50 to 35%. Filters were washed in0.1 X SSPE at 650C(three washes of 30 min each) followed by autoradiography.

Results

Probing Northern blotsofpolyadenylated(Fig. 1) or total (Fig. 2) Hep-G2RNAunderstringent conditions withfull-length

Hep-G2 glucose transporter cDNAyieldedasingle2.8-kilobase (kb) transcript. The same size, single 2.8-kb transcript was detected in blotsof total cellularRNApreparedfrom rat heart, fat, brain, andliver(Fig. 1). When normalizedtotheamountof total cel-lularRNAadded, message abundance was comparable in rat heart andfat RNA, and was 5-10-fold greater in brain RNA. The strongsignal in brain appears to be in part a result of a very abundant transporter mRNA in brain microvessels (Fig. 3). Thus, whensimilar amounts of RNA from calf brain cortex and mi-crovesselsderived from this cortex are probed on Northern blots, the2.8-kb transportermRNAismuch more abundantin vessels thanin whole cortex.

Surprisingly, and insharpcontrast tothefindingwith Hep-G2 human hepatomacells, therewasvery littlehybridization

of thisprobetototalRNAisolatedfrom normal rat liver(Fig.

1). LiverRNAdisplayed only 2-5% relative abundance of this

transcriptcomparedwith fatorheart RNA. When polyadenyl-ated RNAfromratliverorisolatedrathepatocyteswasprobed, however, adefinite 2.8-kb signal was observed, indicating that there is some expression of a highly homologous oridentical

mRNAspeciesin hepatocytes(Fig.3). Theminimalexpression of thismRNAspeciesin normal adultratand humanliver, as

opposed to HepG2 hepatoma cells, raised the possibility that expression of the transporter mRNA might be greater in fetal liver cells. However, when RNA from fetal and neonatal rat liver was probed on Northern blots with the Hep G2 glucose transportercDNA, there was little or no hybridizationsignalat high stringency, just as was seen in adult liver (Fig. 3).

Several humantissues were also available for study. As with ratRNA, theHep-G2 cDNA probe detected a very weak 2.8-kbsignal when a large amount of normal human liver RNA was analyzed by Northern blotting (Fig. 2). In contrast,Hep-G2 cells, cultured skin fibroblasts, and EB virus-transformed lymphocytes

containedanabundant 2.8-kb RNAspeciesthathybridized with the cDNA probe(Fig. 2). Normal human muscle was not avail-able for study, but leg muscle from twopatients with type II diabetes was available. The same size 2.8-kb transcript was seen athigh stringencyin RNAfrombothofthese(Fig.2).

To further investigate thefindingthatRNA from normal ratand human liverhybridized weakly to the Hep-G2 cDNA probe, we prepared a32P-labeledantisenseRNAand used this toprobe Northern blots ofpolyadenylatedRNAfrom ratand humanliver and rat adipocytes at varying hybridization strin-gencies (data not shown). Athigh stringency ofhybridization

(50%formamide,65°C) andwashing (0.1 XSSPE, 65°C)asingle

2.8-kb transcript was detected in all samples, and the relative abundancewasconsistentwithpreviousstudiesusingacDNA probe. When hybridization stringency was reduced moderately bydecreasingtheformamide concentration from 50to35%,a numberof new transcripts were detected in normal human liver RNA.These were bothlarger and smaller than the 2.8-kb tran-script, andtheircombinedintensityexceeded that of the 2.8-kb

transcript. Discussion HEP G2 HI

2

2.8kb-Figure1.Detection ofgluc (HEP) G2cells and rattissi

;:AT PAT RPAIKI I IVFR

*Ars

1 I

*

Ann

U U

Ul-

VQn

We have used theglucose transporter cDNA from

Hep-G2

hu-Opg

7pg

20pg

_3O1P9

manhepatoma cells toassessthe expression of homologous glu-cosetransportermRNAspeciesinvariousratand humantissues,

andhave made several significant observations. First, there is

likely to beconsiderable sequence homology between the rat and humanglucose transporters since mRNAtranscriptsof the same size and abundance aredetected intissuesof thesetwo

speciesunderstringenthybridizationconditionsusingthehuman cDNA

probe.'

Previousdataontheapparent molecularsizeof

the transporter(13, 14),aswellas ontheaffinityofthe transporter

forbothcytochalasinB(15, 16)andanti-transporterantibodies

(17)arealsoconsistentwith closesimilaritybetween thehuman andrattransporters. Second, different tissuesexpressmarkedly differentlevels ofglucosetransportermRNA.Of the tissues that wehaveexamined, ratbrain has the greatest abundance. This maynotbesurprising, giventhewell-knownfactthat thecentral nervoussystemishighly dependentuponacontinuous supply

of glucoseto meetitsmetabolic needs(18). Among the

adap-tations that havedeveloped is a veryhigh rate ofglucose

ex-traction bythebrain, aprocess that appearstoinvolve ahigh

osetransportermnRNAinhepatocyte rate of

glucose

transport across the blood-brain barrier

(19).

ues

bh

Northerngelblot analysisusingthe Consistent with these facts, our studies have revealed that the

I.__ ws--_.j . e._- .,- ,gt

cDNAprobeasdescribed in Methods. The Hep-G2lanecontained 1

.ugof polyadenylated RNA,and rattissuelanescontainedthe indi-catedamountof total cellular RNA. The autoradiogramwasexposed for 18h at-70°C.Thisexperimentisrepresentative ofthree addi-tional experiments, each of whichusedindependently extractedRNA

from different animals.

TYPE

11

HEP G2

LIVER

DIABETIC

MUSCLE

1Opg

_35pg

lOpg

f.t..

2.8kb-A

GT

-II_

SKIN

FIBRO

-BLASTS

15pg

LYMPHOID

LINE

15pg

Figure 2. Detection ofglucose transporter mRNA in total cellu-lar RNAfrom Hep G2 cells and humantissuesorcultured cells by Northern gel blot analysis

us-ing the cDNA probeasdescribed in Methods.Leg musclewas ob-tained fromtwopatients with type II diabetes whorequired amputation for gangrene of the foot. Skinfibroblast cultureswere

establishedfromapunchbiopsy of forearmskin fromahealthy donor, and the lymphoid line is

anEBvirus-transformed human lymphoid line. The autoradio-gram wasexposed for 24 hat -700C.

B

C

BRAIN

BRAIN

VESSELS

LIVER

ISOL.

HEP

FETAL LIVER

day

-3-2-1+1+4

TUB

Figure 3. Detection of glucose transporter mRNA in various tissues by Northern gel blot analysis using cDNA or cRNA probes as described in Methods. In A, total cellular RNA from whole calf brain (27 ug) or calf brain microvessels (12

,g)

wasprobed with either glucose trans-porter cDNA(i.e., GT) or rat tubulin cDNA (TUB). The GT

autora-diogramwasexposedfor4 hand TUBwasexposedfor 1 h. InB,rat

liver (5Mg)orisolated(ISOL.)rathepatocyte(HEP) (4 Mg) polyade-nylatedRNA wasprobedwith theglucosetransporter cRNA.

Expo-sure wasfor 70 h.InC,16-21-dfetalliver RNA(22Mug)wasprobed with either GT(72-hexposure)orTUB(2-hexposure)cDNA.

abundance of glucose transporter mRNA is severalfold higher inbrainmicrovessels than inextractsof whole brain.

After the brain, the primary tissues with the greatest abun-dance of glucose transporter mRNA transcripts appear to be fat and muscle. These two tissues are important for in vivo glucose homeostasis andtheirglucose transport systemsare wellknown tobe sensitivetoinsulin. Thatahighly homologoustransporter mRNAspeciesis present in erythrocytes, where transport is in-sensitive to insulin, and in fat and muscle, where transport is insulin sensitive, suggests that the insulin-responsive transporter is similar, if not identical, to the insulin-insensitive erythrocyte transporter. This would imply the existence of an additional molecular component that would be responsible for conferring insulin sensitivity to the transport process. However, the fact that hybridization to the same size mRNA occurs under stringent conditions does not itself constitute proof that themRNAspecies

encoding glucosetransportersininsulin-sensitive and -insensitive tissuesareidentical.

SI

nucleaseprotectionstudies and/or cloning andsequencing ofthetransporter cDNAs from eachtissue would

be necessaryfor definitive information on this subject. The extremely low abundanceofmRNA in normal rat and humanliver that is capable of hybridizing to the Hep-G2 glucose transporter cDNAproberaisesanumberofquestions. Could it be thatthe observed weaksignalreflects RNAfromcells other thanhepatocytesthat make up10%of the liver mass, as suggested

by Birnbaumetal.(20)? Thisreasonablesuggestionisunlikely, givenourobservationthatRNAfrom isolatedhepatocytesgives

asignalasstrong as that seen withRNA from total liver. Thus, hepatocytes do appeartohaveahomologousoridenticalmRNA

species. Why istheabundance of thismRNA solow?Previous

studiesonintactliver(2

1)

andonisolatedhepatocytes (22)reveal that hepatocytes have a very active glucose-facilitated uptake system that is similar in magnitude tothat of the erythrocyte (whennormalized to cellular waterspace).

At leastthreeexplanationscouldaccount forthe apparent

discordance between the abundance of glucose transport and the relatively low level of hybridizable mRNA detected by

Northern blotsof liver RNA. Theglucosetransporterofliver cells may havealonghalf-lifewhencomparedwiththat present in other cells. If thisweretrue, inthe steadystateless mRNA would be required to generate a fixed amount of transporter

protein. Weknowofnodata that bearsonthispossibility. Itis

also

possible

thatthetranslational

efficiency

of the transporter message is'very high inliver cells; this supposition alsolacks experimental support. A third

possibility,

which we currently favor,is that hepatocytes contain an additionalglucose

trans-porter that is encoded by a homologous but distinct mRNA

species.

This

possibility

isconsistent

with

the appearance of

sev-eral additional bandson Northern blotsof liver polyA RNA when hybridization stringency is modestly reduced. It is also consistent with several observations on the

hepatic

glucose

transporter. Axelrodet al. (23)studied the rathepatic glucose

transporterusingcytochalasin B asaprobeand observedthat,

whereas the numberof glucoseinhibitable cytochalasin binding siteswassimilartothat oferythrocyte membranes, theaffinity ofthesebindingsites for

cytochalasin

was 10-fold less in hepa-tocytemembranes thanin erythrocytemembranes.Assuggested

by these workers, this could reflecttheexistenceofstructurally

distinct transporterspeciesinthesetwo tissues.Infurther support ofthispossibilityis theobservation (Cushman,S. W.,personal communication)that

antibodies-against

theerythrocyte glucose

transporterthatrecognize putativetransportersinadipocyteson

Western blots only very weakly recognize a hepatic glucose transporter species. Ultimately, validation of this concept will depend entirely upon the cloning and sequencing of a unique transporter cDNA from' hepatocytes, and attempts at this are currently in progress in our laboratories.

The final question relates to the meaning of the marked dif-ferences in expression of the Hep-G2 glucose transporter message innormal liver as opposed to'Hep-G2 cells, which are clearly hepatocytic in origin and which express a wide variety of differ-entiatedfunctionscharacteristic of hepatocytes (24). It has long beenknown thattransformedcells orprimarytumorsoften dis-playincreasedrates of glucose uptake andmetabolism(25), al-though the molecular mechanism by which these changes come abouthave been obscure. We suspect the increased abundance of glucose transporter-specific mRNA in

Hep-G2

cells is an exampleofsuch atransformation-relatedphenomenon, and our observation may representthe first indication that the molecular

mechanism forsuch anincreasein transport is an increase in thesteadystate levelofglucose transportermRNA.

Acknowledgments

Wewould liketothank Dr. Motozumi Okamoto for makingavailable theisolatedrathepatocytesand TerriWisemanfor aid in preparing the

manuscript.

Dr.Flier istherecipient ofa Research CareerDevelopment Award

from the NationalInstitutes ofHealth (NIH). This work was supported

in partfrom grantsAM-28082 (to Dr. Flier) and GM-35012 (to Dr.

Lodish) fromtheNIH, and by U. S.Public Health Servicegrants PO-1-HL-26895and5-ROI-NS22213(to Dr. McCall).

References

1.Wheeler,T.J.,and P.C. Hinkle. 1985.The glucose transporter

ofmammalian cells.Ann.Rev.Physiol.47:503-517.

2.Lienhard, G. E., J. M. Baldwin, S.A.Baldwin, and F. R. Gorga. 1983. Structure and Function ofMembraneProteins. E. Quagliarello and F. Palmieri, editors. Elsevier Biomedical Press, Amsterdam. 325-333.

3.Jones, M. N., and J. K. Nickson. 1981. Monosaccharidetransport

proteinsof the human erythrocyte membrane.Biochem.Biophys.Acta. 650:1-39.

4. Wheeler,T.J., and P. C. Hinkle. 1982. Membranes andTransport. A.N.Martonosi, editor. PlenumPublishingCorp., New York.

161-167.

5.Mueckler,M.M., C.Caruso, S.A.Baldwin,M.Panico,I. Blench,

H. R. Morris,J. W.

Allard,

0. E. Lienhard,and H. F. Lodish. 1985. Sequence andstructureofa humanglucose transporter. Science(Wash. DC). 229:941-945.

6.Wardzala,L.J., andB.Jeanrenaud. 1981. Potential mechanism of insulin actiononglucosetransport inisolatedratdiaphragm.J.Biol. Chem. 256:7090-7093.

7.Crofford,0.B.,andA.E.Renold. 1965. Glucoseuptakeby in-cubated rat epididymal adipose tissue. Characteristics oftheglucose

transport systemand action of insulin.J. Biol. Chem. 240:3237-3244. 8. Chirgwin,J. M.,A.W. Przybyla,R. J.MacDonald, andW. J. Rutter.1979. Isolation ofbiologicallyactive ribonucleicacidfromsources enrichedinribonuclease.Biochemistry. 18:5294-5299.

9.Aviv, H.,andP.Leder. 1972. Purificationofbiologically active globinmessenger RNA bychromatographyonoligothymidylate

acid-cellulase.Proc.Natl. Acad Sci.USA. 69:1408-1412.

10.Okamoto, M.,M. F.White,R.Maron,andC.R.Kahn. 1986. Autophosphorylationand kinase activity of insulinreceptorindiabetic

rats.Am. J. Physiol. 14:E542-E550.

inducedalterations of glucose metabolism in cerebral microvessel.Am. J.Physiol. 247:E462-E467.

12. Maniatis, T., E. F. Frisch, and J. Sambrook. 1982. Molecular Cloning:ALaboratory Manual. ColdSpring Harbor Laboratory, New

York.202-203.

13. Carter-Su, C., J. E.Pessin, R. Mora, W. Gitomer, and M. P. Czech. 1982.Photoaffinity labelling of the human erythrocyte D-glucose

transporter. J.Biol. Chem. 257:5419-5425.

14.Wheeler, T. J., I.A.Simpson, D. C.Sogin, P.C. Hinkle, and S.W.Cushman. 1982.Detectionoftheratadiposecellglucosetransporter

with antibody againstthe humanred cell glucosetransporter. Biochem. Biophys.Res. Commun. 105:89-95.

15.Jung, C.J., andA. L. Rampol. 1977. Cytochalasin Bbinding sites and glucosetransportcarrierinhumanerythrocyte ghosts.J.Biol. Chem. 252:5456-5463.

16. Wardzala,L. J.,S.W.Cushman, andL. B.Salans.1978. Mech-anismof insulinactiononglucosetransport in theisolatedratadipose cell. J. Biol. Chem. 253:8002-8005.

17. Allard,W. J.,and G.E.Lienhard. 1985. Monoclonal antibodies

tothe glucosetransporterfrom human erythrocytes.J.Biol.Chem. 260: 3668-3674.

18.Sokoloff, L., G. G. Fitzgerald, andE. E.Kaufman. 1977.Nutrition

and the Brain. R. J. Wurtman andJ.J.Wurtman,editors.RavenPress,

NewYork. 87-139.

19. Pardridge, W. M., and W. H.Oldendorf. 1977. Transport of

metabolic substrates throughthe blood:brain barrier.J.Neurochem.28: 5-12.

20. Birnbaum, M. J., H. C. Haspel, and0.M. Rosen.1986.Cloning

and characterizationofacDNAencoding therat brainglucosetransporter

protein.Proc.Natl. Acad.Sci. USA. 83:5784-5788.

21. Williams, T. F., J. H. Exton, C. R. Park, and D. M. Regen. 1968. Stereospecifictransportof glucose in theperfusedratliver.Am.J.Physiol. 215:1200-1209.

22. Craik, J. D., and K. R. F.Elliott. 1979. Kinetics of30methyl glucose transport in isolated rat hepatocytes.Biochem.J. 182: 503-508. 23. Axelrod, J. D., and P. F. Pilch. 1983. Unique cytochalasin B

binding characteristics ofthehepatic glucose carrier.Biochemisiry. 22: 2222-2227.

24. Knowles, B. B., C. C. Howe, and D. P. Aden. 1980. Human hepatocellular carcinoma cell linessecretethe major plasmaproteins andhepatitisBsurface antigen. Science(Wash. DC). 209:497-499.

25. Salter, D. W., and M. J. Weber. 1979. Glucose specific

cyto-chalasinBbinding is increased in chicken embryo fibroblasts transformed by Roussarcomavirus.J.Biol. Chem.254:3554-3561.