Overexpression of glucose transporters in rat mesangial cells cultured in a normal glucose milieu mimics the diabetic phenotype

Overexpression of glucose transporters in rat

mesangial cells cultured in a normal glucose

milieu mimics the diabetic phenotype.

C W Heilig, … , M Zhu, P Cortes

J Clin Invest.

1995;

96(4)

:1802-1814.

https://doi.org/10.1172/JCI118226

.

An environment of high glucose concentration stimulates the synthesis of extracellular

matrix (ECM) in mesangial cell (MC) cultures. This may result from a similar increase in

intracellular glucose concentration. We theorized that increased uptake, rather than glucose

concentration per se is the major determinant of exaggerated ECM formation. To test this,

we compared the effects of 35 mM glucose on ECM synthesis in normal MCs with those of 8

mM glucose in the same cells overexpressing the glucose transporter GLUT1 (MCGT1).

Increasing medium glucose from 8 to 35 mM caused normal MCs to increase total collagen

synthesis and catabolism, with a net 81-90% increase in accumulation. MCs transduced

with the human GLUT1 gene (MCGT1) grown in 8 mM glucose had a 10-fold greater

GLUT1 protein expression and a 1.9, 2.1, and 2.5-fold increase in cell myo-inositol, lactate

production, and cell sorbitol content, respectively, as compared to control MCs transduced

with bacterial beta-galactosidase (MCLacZ). MCGT1 also demonstrated increased glucose

uptake (5-fold) and increased net utilization (43-fold), and greater synthesis of individual

ECM components than MCLacZ. In addition, total collagen synthesis and catabolism were

also enhanced with a net collagen accumulation 111-118% greater than controls. Thus,

glucose transport activity is an important modulator of ECM formation by MCs; the presence

of high extracellular glucose concentrations is not necessarily required for the stimulation

[…]

Research Article

Overexpression of Glucose Transporters

in

Rat

Mesangial

Cells

Cultured

in

a

Normal Glucose Milieu Mimics the Diabetic

Phenotype

Charles W. Heilig,* LuisA.Concepcion,* Bruce L. Riser,* Svend 0.Freytag,* Min Zhu,§and Pedro Cortes*

With thetechnical assistanceof Clare C. Hassett, JeffD.Gilbert, K. S. Sury Sastry, Kathleen 0.Heilig, and Janet M. Grondin §Division of Nephrology and Hypertension, *Department of Medicine, and tMolecular Biology ResearchProgram, HenryFordHospital,

Detroit, Michigan 48202

Abstract

An environment of high glucose concentration stimulates the synthesis ofextracellular matrix (ECM) in mesangial

cell (MC)cultures. This may result fromasimilar increase

in intracellular glucose concentration. We theorized that increased uptake, rather than glucoseconcentrationper se

is the major determinant of exaggerated ECM formation.

To test this, we compared the effects of 35mM glucose on ECMsynthesisin normalMCswith thoseof 8mM glucose in the same cells overexpressing the glucose transporter

GLUT1 (MCGT1). Increasing medium glucose from 8 to

35 mM caused normalMCstoincrease totalcollagen synthe-sisandcatabolism, witha net81-90% increasein accumu-lation. MCs transduced with the human GLUT] gene

(MCGT1) grown in 8 mM glucose had a 10-fold greater

GLUT1 protein expression anda 1.9, 2.1, and 2.5-fold

in-creasein cellmyo-inositol,lactate

production,

and cell

sorbi-tol content,respectively,ascomparedtocontrolMCs

trans-duced with bacterial

f8-galactosidase (MCLacZ).

MCGT1

also demonstrated increased glucose uptake

(5-fold)

and

increasednetutilization

(43-fold),

andgreater

synthesis

of

individual

ECM components than MCLacZ. In

addition,

total collagensynthesis andcatabolismwerealsoenhanced with a net collagen accumulation 111-118% greater than controls. Thus, glucose transport activity is an

important

modulatorofECM formationby

MCs;

the presence of

high

extracellular glucose concentrations is not

necessarily

re-quired forthestimulation of matrix

synthesis. (J.

Clin. In-vest.1995.

96:1802-1814.)

Keywords:diabetic

nephropathy

*glomerulosclerosis *

hyperglycemia

* extracellularmatrix

* collagen metabolism

Introduction

The renalglomerularlesion of human andexperimentaldiabetes mellitus is characterizedbyglomerular hypertrophy (1, 2)and

Part of this work waspresented in abstract form at the 27th annual meeting of the American Society of Nephrology, 26-29 October 1994, inOrlando, FL.

AddresscorrespondencetoCharles W.Heilig,University of

Roches-terMedicalCenter, Division of Nephrology, 601 Elmwood Ave., Box

675, Rochester,NY.14642. Phone:716-275-3660;FAX:716-442-9201. Receivedfor publication26October1994 andacceptedinrevised

form15 June 1995.

the deposition of extracellular matrix in the form of diffuse thickening of the peripheral basement membrane and mesangial expansion (3). The progressive accumulation of matrix in the mesangial areas, and the associated encroachment on neigh-boring capillaries with loss of filtration surface area, is consid-ered as themain structural lesion responsible for the relentless decline inglomerular function (3, 4). There is persuasive evi-dence that thiscritical change may be the result of an altered mesangial cell

(MC)'

metabolism involving the extracellular matrix. MCs in tissue culture synthesize proteoglycans, fibro-nectin, laminin, thrombospondin, and various forms of colla-gens, primarily type IV and type I (5-7). Therefore, a meta-bolic derangement of these cells in diabetes resulting in the

excessive formationanddeposition of these matrix components isalikelydeterminant of mesangial expansion and glomerulo-sclerosis.

The knowledge on how extracellular matrix synthesis is controlled in MCs is only fragmentary. It is known that the

synthetic activity may be stimulated by very diverse factors, notablythemechanicalstrain inducedby distending forces dur-ingglomerular hypertension (8) and the action of

TGF-,31

(9).

Indiabetes,anobvious injurious alteration could be the contin-ued presence ofanabnormally highconcentration of extracellu-larglucose.Recentevidence in humans(10, 11) has confirmed early findings in animal studies (12-14) indicating that strict control ofglycemia with insulin administration or successful

pancreastransplantation maydelaytheonsetand slow the pro-gression of the characteristic mesangial matrix expansion. In

addition, it has been amply demonstrated that MCs in tissue culture increase theproductionoffibronectin, laminin, collagen

type IVaswellasmRNAlevelsfor thesematrixproteinswhen incubated in the presence ofsupraphysiological concentrations ofglucose (7, 15-18). Althoughit has beensuggestedthat the increase in matrixsynthesismay bepartially duetotheosmolar effect causedbyhigh glucoseconcentrations (18),mostof the evidence accumulated thus far indicates that thechange is

re-latedtothe metabolism ofglucose (16, 19, 20). Glucoseenters

MCs by a facilitated diffusion process which is independent

of insulin action (21). It has, therefore, been proposed that

intracellular concentrations ofglucose may approach those in

theextracellular environment in diabetes(21).High

intracellu-lar concentrations ofglucose may then increase extracellular

matrix formation by activating the polyol pathway, inducing myo-inositol depletion, increasing nonenzymatic

glycosylation

ofproteins,orby generationof the second messengers inositol

1. Abbreviations used in thispaper: MC, normal ratmesangial cell; MCGT1, transduced rat MC overexpressing GLUTI transporter;

MCLacZ,transduced MCoverexpressing 3-galactosidase. J.Clin. Invest.

©The American Societyfor ClinicalInvestigation,Inc. 0021-9738/95/10/1802/13 $2.00

triphosphate and diacylglycerolfollowedby transcriptional

acti-vation of extracellular matrix genes(20, 22).

We hypothesized that the glucose-induced stimulation of

extracellular matrix formation by MCsmay notnecessarily

re-quire thepresence ofsupraphysiological concentrations of

glu-cose if there is anincreased transport andthus, excessive

bio-availability of this hexoseas substrateformetabolism. Except forunique tissues in which glucose is concentrated byanactive

process involving Na+/glucose cotransporters, i.e.,renal proxi-mal tubule and intestinal epithelial cells, glucose enters cells

by the passive process of facilitated diffusion. In thisprocess, specific integral membrane proteins, identified as the GLUT

family, transport glucose down a concentration gradient.

Prelim-inary work on the identification of glucose transportersinMCs

suggests that GLUT1 may be the preponderant isoform (23, 24). This transporter, consideredtoberesponsiblefor

constitu-tivetransport, is known to be regulatablein some tissues. It is

also the mostubiquitously distributed of the GLUT isoforms in vivo, and is expressed in virtuallyall cultured cells (25).The

high affinity of GLUTI for glucose (26, 27) ensures thatthis

transporter functions at, or close to,its

Va,

under normal physi-ological and diabetic hyperglycemicconditions. InMCs,

satura-tion of glucoseuptake has been reportedatthe 30-35mMlevel (21). Therefore, enhanced glucose uptake through GLUTI may possibly be achieved by raising itsextracellular concentration up to - 30mM, by increasing the intrinsic activityofthe

trans-porter (28), or through stimulation of GLUTI expression and/ ortranslocation from intracellular sitestotheplasma membrane (29). However, it has not been determinedinMCsifenhanced GLUTl-mediated transport per se is associated with an

in-creased utilization of substrate, i.e., whether transport activity may be animportant modulator of glucose metabolism.

To test ourhypothesis,wecomparedtheproduction of extra-cellular matrixinrat MCs exposed tohigh glucose concentra-tions with that inthesamecellsexposedtophysiological levels of extracellular glucose, but overexpressing GLUTI protein.

The results hereinreported differentiate the effects of extracellu-lar glucose concentration from those of enhanced intracellular glucose availability and utilization on extracellular matrix

for-mation, providing new insights into the pathogenesis of the

glomerularlesionof diabetes.

Methods

Materials. Thepurified extracellular matrixcomponentsutilizedas

stan-dards includedratcollagen type I (Upstate Biotechnology Inc., Lake

Placid,NY),murinecollagen type IV, murine laminin(both from Col-laborativeResearch,Inc.,Bedford,MA) andratfibronectin (Chemicon International,Inc.,Temecula, CA).The antibodies used werepolyclonal

anti-ratcollagentypeI,anti-ratfibronectin,and anti-mousecollagen

type IV(Chemicon International, Inc.),andanti-murinelaminin (Col-laborative Research,Inc.). Thepolyclonal,rabbit anti-ratGLUTI anti-bodyused in theidentification ofGLUTI wasgeneratedtospecifically

reactwith a 13aminoacid carboxy terminalpeptide of this transporter isoform. The latter wereobtained from East Acres Biologicals

(South-bridge,MA).Theaffinity-purified, goat anti-rabbit IgG conjugated to 4-nmgold particles and the silver enhancement system IntenseSE'

used for the localization ofGLUTI atthelight microscopic level, were obtainedfrom Amersham Life Sciences Co. (Little Chalfont, United

Kingdom). A monoclonal rabbit anti-rat Ig was used in the

immu-noblottinganalysis of

61-tubulin

(Sigma Chemical Co., St. Louis, MO). The human GLUTI cDNA(vectorpSPGT) (30) waskindly provided

by Dr. M. Mueckler (Department of Cell Biology and Physiology,

WashingtonUniversitySchool of Medicine, St. Louis MO). The fibro-blast packaging cell line ICREwaskindlyprovided by Dr. R. Mulligan

(Whitehead Institute Biomedical Research, Massachusetts Institute of Technology, Cambridge, MA) and maintained in tissue cultureas pre-viouslydescribed (31).All MCtissue culturemedia werebasedon a

special RPMI-1640 formulation lacking glucose, proline, and glutamine (Gibco-93-5044EA; Gibco Laboratories, Grand Island, NY) buffered with 25 mM Hepes. ProlineL-[14C (U)], 286 mCi/mmol; proline

L-[2,3,4,5,-3H], 112 Ci/mmol; hydroxyproline L-4-[3H (G)], 5.5 Ci! mmol; inulin [3H (G)], 257 mCi/gram; 2-deoxy[1-3H]glucose, 30.6 Ci/mmolwereallpurchased fromNewEngland NuclearResearch

Prod-ucts (DuPontCo., Wilmington, DE). The purity oftheradioisotopic

internalstandards usedinthequantitation of proline and hydroxyproline

wasestablishedbefore theirusebychromatographic analysis (see

be-low).Additivestotissueculture mediawerecell culture-testedquality (Sigma ChemicalCo). High purity CollagenaseVII(Sigma Chemical Co.)wasusedin theprotein-digestionassays.The bacterialmyo-inositol

dehydrogenase used in the measurement ofmyo-inositolwasobtained from Sigma Chemical Co.The columnsused for HPLCwere4.6 mm

x 25 cmUltrasphere ODS, 5-amparticle size (BeckmanInstruments

Inc., San Ramon,CA).

Tissueculture. MCs wereobtained fromourclonedline (16KC2) derived fromoutgrowths ofratglomeruli (Charles River Laboratories, Cambridge, MA) and previously characterized by us (32). In brief,

thesecells demonstrateafusiformor stellateappearance, intracellular

fibrils,anabilitytogrow inmediumlacking d-valine, growthinhibition whenculturedinthepresenceof heparinormitomycin, amarked in-crease inguanosine 5;5'-cyclic monophosphate contentupon exposure toatrial natriureticpeptide, andthe presenceof densecytoplasmic

im-munochemical stainingforcollagen types I and IV, fibronectin,laminin,

andthrombospondin.Inaddition,these cells express theThy-i antigen and form "cell hillocks" containing dense extracellular matrixin post

confluent cultures.Thesecharacteristics have been retainedonrepeated

passage.Finally,wehaverecently shown that these cells demonstrate the samehigh sensitivitytophorbol-stimulated neutrophiladhesion and

lysis (33)as doearlypassage MCs, indicatingcontinued and like

ex-pression of essential cell surfacereceptors, includingthosefor CDll/

CD18 molecules.

Exceptwhereindicated, MCswereseeded (10,000cells/cm2)into

8-cm diameterplastic dishesor2.5cmdiameter six-wellplates(Corning GlassWorks,Corning NY) andgrown in themedium describedabove to whichpenicillin, streptomycin, 20% Nu-Serum(Collaborative Research

Inc.), 8mMglucose, 2.05 mMglutamine, and an amount of proline to

provideafinal concentration of183

_ttM

(including proline contained in

Nu-Serum),wasadded.Considering the high concentrations ofglucose

commonly occurringin adiabetic milieu (20-35mM), aglucose con-centration of 8 mM was consideredas normal. Lower concentrations

of glucosewere not used because MCs show deficientgrowth when maintainedinthephysiological concentration of5 mM (7). Exceptfor

studies to determinegrowthrates,experimentswereterminated 7 d after

seeding,when cultures hadjustreachedconfluency.Growth rates were determined in cells seeded in 8-well, 0.79cm2 glass chamber slides at

adensity of 12,600 cells/cm2. Beginningon day 1, and on alternate

daysthereafter,cells werecounted in four separate wells at each time

period.

Preparation of infective virus and transduction ofratMCs. The procedures used were similar to those previously reported (31). Gene transductionswerecarriedoutusingthepWZLneoMoMuLVretroviral vector(AriadPharmaceuticalCo.,Cambridge, MA).Thisnew vector

containsaninternalribosomalentry site from theencephalomyocarditis

viruswhich allows for translation of the GLUT] (GTI)andneomycin

phosphotransferase(neoR)geneproductsfrom thesameRNAtranscript

(Fig. 1). This property implies that cells acquiring resistance to the

neomycin analog G418 will also express the GLUTI gene product.

Transcriptionof thediscistronicproviralRNAis driven from the

B E

A I } _

B _

Figure1.Diagram of the pWZLneo-based retroviral expressionvectors. The pWZLneovector(A) containsaninternal ribosomalentrysite (IRES) from the encephalomyocarditis virus. It also contains theneoR

gene(NEO) coding for resistancetoG418. The humanGLUTI cDNA isspliced into the BamHl [B]/EcoRl [E] multiple cloning region of thevector(B). To produceacontrolvector,the bacterialLacZ cDNA (LacZ)was similarly spliced into pWZLneo in place of GLUTI (C). Arrowsindicate the transcription initiation site.

The gene which encodes bacterial

_{f3-galactosidase}

(LacZ) or the cDNAencoding human GLUT] was spliced into the multiple cloning

site ofpWZLneo (Fig. 1). Once this was completed, pWZLneoLacZ

andpWZLneoGTI wereusedtotransfectthe ICRE fibroblast packag-ing cell lineusing acalciumphosphate precipitationmethod(34). After incubation of the transfected packaging cells for2 d,the supernatants were removed, filtered through a0.45-kim syringefilter and stored at

-80'Cuntilused for transduction of MCs.

Normal MCs(16KC2)insubconfluent culturesweretransducedby

exposurefor 2 htothe virus-containingsupernatant in thepresenceof

8 jig/mlPolybrene (SigmaChemicalCo.).Cellswerethenwashed and

cultured for2 d in the growth medium described above before their

selection. Selection ofstablytransduced cells was carriedoutbytwo successiveincubations, first,inamediumcontaining0.25mg/mlof the

neomycin analog G418(SigmaChemical Co.)andthen, ina medium

in which the G418 concentration wasdoubled. A survivingclone

ex-posed to pWZLneoLacZ was selected as a transduced MC control

(MCLacZ) duetoitshighlevel of 13-galactosidase expression according tothechromogenic dye X-gal (5-bromo-4-chloro-3-indolyl /3-D-galacto-pyranoside) test (35). A second surviving clone exposed to pWZLneoGTIwasselectedas atransduced MCoverexpressingGLUTI

accordingtothe levels of thisproductasdeterminedby immunoblotting (see below).

Northernanalyses.A standard method withslightmodificationswas

used (36). In brief, total RNA was obtained from cultures of MC

usingacommerciallyavailable kit basedontheguanidiniumandphenol extraction method (RNA Stat-60; Tel-Test Inc., Friendswood, TX). RNAwasdenatured inglyoxal/DMSOand20-,ugsamplesloaded into

individual lanes ofa 1% agarosegel made with 10 mM sodium phos-phate buffer. Electrophoretic separationwascarriedoutinacirculating

buffergelbox (Hoefer Super Sub; Hoefer Scientific Instruments, San

Francisco CA). Gels were then stained with ethidium bromide,

de-stained,andphotographed.RNAintegritywasconfirmedby inspection of theribosomal RNA bands. Gelswere blottedtoGenescreen

mem-branes(DupontNEN Research Products, Boston, MA)and the RNA

immobilized by UV irradiation. Blots were then prehybridized, and probedfor theGLUTIisoform,typeIcollagen,typeIVcollagen, fibro-nectin,and thehouskeepinggenej3-tubulinusingtherespectivecDNA's

(1.66-kb humanGLUT], 1.5-kbproal(I), 0.7-kbproal(IV), 0.5-kb

fibronectin)The latterwere32P-labeledbythe Random Hexamer

Prim-ing method (PRIME-1 kit; Sigma Chemical Co.). Afterexposure to Kodak XAR-5 film(EastmanKodakCo., Rochester, NY)for 3-14d, the autoradiograms were analyzed by optical scanning densitometry

(ScanMaster3+;HowtekInc., Hudson, NH) usingthe National

Insti-tutesof Health (NIH) Image gel plotting software(NIH Image 1.52;

Natl. Technical InformationService, Springfield, VA).Relative

quanti-ties of GLUTI mRNA in MCLacZ and MCGT1 cells were compared after normalization to mRNA for the housekeeping gene/-tubulin.

Immunoblotting of GLUT]. Immunoblot analysis was carried out

according to methods previously described for the study of glucose transporterisoformswith minormodifications(29). 50

_,g

of solubilized protein samples were separated by SDS-PAGE and electrophoretically

transferred to Hybond-ECL nitrocellulose membranes (Amersham Corp., Arlington Heights,IL). Asprimary antibody, the rabbit anti-rat

GLUT1 antibody described abovewasused. The secondary antibody was a horseradish peroxidase anti-rabbit-Ig conjugate (Amersham Corp.). Antigenswereidentified byachemiluminescenceassaybased on the luminol reaction(ECL WesternBlotkit,Amersham Life Sciences; AmershamCorp.). Immunoblotting of/3-tubulinwas used as a confir-matory methodto assuretheequal sample loading between gel lanes.

Identification of GLUT1bands wasconfirmedby preadsorption of

anti-GLUTI antiserum with 25

_pg/ml

ofthepurifiedGLUT1 peptide.

Measurement of production of extracellular matrix components; ELISA. In experiments inwhich specific extracellular matrix compo-nents werestudied, productionwasquantifiedasthe total amount

accu-mulated in the tissue culture medium during24h of incubation.Atthe

startof thisperiod, growth mediumwaschangedtoone inwhich

Nu-Serumwasreplaced by1%FCS.The presenceof Nu-Serumorgreater

concentrations of FCS were foundto result inincreased backgrounds andreducedsensitivity of theELISA.Removal of all serum during the

collectionperiod resultedinlowrecovery rates(3%orless) of added purified extracellular matrix components. Under theconditions used, recoveries were between 62 and96% of added purified extracellular

matrix components (ECM).The amountofspecific ECM components

secreted into the culture medium was quantified by ELISA, using a modificationofaprocedurepreviouslydescribed(8). Samples of

cul-ture medium(50-100

_p1)

wereadded in triplicate to wells ofa 96-well ELISA plate (Falcon Labware,Lincoln Park,NJ) and incubated

for 18hat4°C. Purified matrixcomponents,dilutedinthesamemedium,

wereadded (0.5ng- 1 ag/well)to each assay plateasstandards. At theendofthis incubationperiod, the mediumwas removed,and the unoccupiedsitesblocked bya2-h treatmentwith5%nonfat dry milk (Carnation Co.,Los Angeles, CA) in PBS containing 0.05% Tween.

Wells were then washed and incubated for 3h with 100

_MI

of rabbit antisera forspecific extracellularmatrix components. Allantiserawere

tested for specificity, beforetheir use, by immunoblotting, with and

without blocking, using the extracellular matrix standards described above.After extensivewashingof thewells,anenzyme-linked alkaline phosphatase-labeledgoat anti-rabbitIgG (Organon Teknika,Durham,

NC) was added and the plates were incubated for anadditional 3-h period. Thiswasfollowed by extensive washingand the additionofa

phosphatase substratesolution(Sigma Chemical Co.). Color intensity

wasmeasured withaTitertek MultiscanMCC/340 (Flow Laboratories, Inc., McLean, VA), and results analyzed in a curve-fitter computer program(Interactive Microware Inc.,StateCollege, PA).

Study of collagen metabolism. Methods previously described (8)

wereusedwith modifications.The culture medium waschanged24 h

beforethestartoftheradiolabeling periodto amediumlackingproline

(except for that contained in Nu-Serum which resulted inafinalproline

concentration of 40

_piM).

Preliminary experiments demonstrated no

difference in growth rates between cells cultured in 40 .tM or0.174 mMprolineover a14-dperiod. Radiolabelingwascarriedoutby incuba-tion for 48or72 h inanidenticalmedium,butcontaining0.15 mM

/3-aminopropionitrile, 210

ttM

ascorbic acid and 183 mM ['4C]proline (82.3mCi/mmol).Previousexperimentsbyusdemonstrated that[14C]

-proline incorporationintocollagenincreaseslinearlyovera72-hperiod

ofradiolabeling (8)All tissue culture wells were supplementedevery 24 htoprovide 140

AM

fresh ascorbic acid. When thesp actofthe cellularproline endogenouspoolwasdetermined,2.5_YCiof[3H]inulin/ ml wasadded 15 min before the end of theradiolabelingperiodas an extracellular fluid marker, and the contents of each well thoroughly

mixed. At the termination of the radiolabeling period, medium was

perchloric acidpouredontothe celllayer. The cell layerwasnotrinsed

to removeresidualmedium, in ordertoavoid losses of intracellular free proline.Inexperiments done in 6-well plates, the media and cell layers

of 6 wells were pooled as onesample for analysis.

Total protein contained in the medium samples was precipitated

in 75% ethanol at -50C, andthe supernatant analyzed for [3H]inulin concentration. After the addition of 89

liCi

[3H]proline as internal standard, medium supernatants were filtered in

Centricon'-3

filters (Amicon Co.,Danvers,MA)andamino acidswereseparated by solid

phase extraction using AG5OW-X8 (He) columns (Poly-Prep&; Bio Rad Laboratories, Richmond, CA) and 6N NH40H as eluant. After NH40H removal undervacuum(Speed-Vac concentrator; Savant In-struments Inc., Farmingdale, NY), purified amino acids were

resus-pended in 0.1 N HCl for the subsequent determination of ['4C]-hydroxyproline, total proline, prolinesp act,and calculation of the pro-lineand['4C]proline/[3H]inulinratios.

Net collagen accumulation in the medium was estimated bytwo

independent methods.Thefirstmeasurementwasobtained accordingto

the14Cincorporation intoprotein-associatedhydroxyproline (8).Inthis method,themedium protein precipitatewashydrolyzedundervacuum

with 6NHClat 1 0C for 18 h and aminoacidsseparated asabove by solid phase extraction after the addition of 3.32

tiCi

[3H]

-hydroxyprolineasaninternal standard. These purified amino acidswere

subsequently analyzed formeasurementof'4Cincorporationintoproline andhydroxyproline. The second methodwasbasedontheamountof total '4Cincorporated intocollagenase-digestible protein (37). Inthis method, after completion of the radiolabeling period, 1 mlof medium

wasmixed with 330 pl ofaproteinase inhibitor solution (providingper

milliliter:3 pmol PMSF, 0.1mmolEDTA, 40,umolN-ethylmaleimide).

Mediumproteinwasprecipitatedand thepelletwashedfivetimes with

cold 10% TCA. This precipitate wasthenresuspended in 1 NNaOH,

incubated for 10min at 37°C, andthe solution neutralized with 1 N

HCl. After adjusting the pHto7.5 with1 NTrisbuffersolution, PMSF

andN-ethylmaleimidewereaddedinthesame amountsasbefore, and CaCl2 addedtoprovideafinal 5 mMsolution.Forenzymatic digestion,

thesamplewasdividedintotwoequal portions and 140 U/ml of

colla-genase addedtoone of them, while theother wasusedas acontrol. Afterincubation for2 hat37°C, the undigested proteinwasremoved by precipitationwith10%TCAand 0.5% tannic acid.Finally, the '4C

radioactivityin the supernatantsandprotein precipitatewas measured andtheradiolabelincorporation intocollagenase-digestibleand colla-genase-resistant protein determined from the difference between the

treated andnontreated samples.

Immediately after addition of0.2 Nperchloric acid,celllayers were

scraped, briefly homogenizedinthecold and theprecipitatesand acid-soluble supernatantsseparated by centrifugation.To measure theproline endogenous pool, thesesupernatantswereneutralizedat40C with 1 N

KOHtopH 7.0 andtheconcentration of [3H]inulin determined before theadditionof 45

_ACi

of [3H] prolineas aninternalstandard. Theamino acids containedinthis acid-soluble cellextractwereseparated by solid phase extractionasabove, and lyophilized beforemeasurementof cell layer-associated free['4C]hydroxyproline, freeproline, and prolinesp

act.Because valuesfor theratiosproline/[3H]inulin and

['4C]proline/

[3H]inulin in the medium from the same sample were known, the

amountofproline and[14C]proline contributed by residual mediumin

the celllayer could be estimatedinindividualsamples accordingto the

amountof[3H]inulin measuredinthe celllayer acid extract,asdone

in previous studies (8). The cell layer's perchloric acid precipitate was lipid-extracted and consecutively subjected to alkaline and acid

hydrolysisfor the measurement of total RNA, DNA, and the separation ofprotein (38). The final protein precipitate was hydrolyzed as de-scribed above,[3H]hydroxyprolineinternalstandard added, and amino acidspurified and separated for the quantitation of 14Cincorporation intoprolineandhydroxyproline.

Measurementof 2-deoxyglucose uptake rates and kinetics. The up-take ofglucosewas determined by using the nonmetabolizable analog 2-deoxy-D-[1- 3H]glucose according to a modification of the method of

McClainetal. (39). MCswereseeded in 35mmdiameter wellsat a

density of 42,000 cells/cm2 and allowedtoattach for 2 h. After removal of the medium and rinsing with PBS, cultureswereincubated in

glucose-free PBS for30min and then, this buffer solution replacedwith one containing 0.1

jCi/ml

oftheradiolabeledanalog (3.27nM). 1 mlof thissolutionwasaddedper welland thesampleswereincubated for5 min. Afterthis, theunincorporated radioisotope wasrapidly removed by washing the cell layer with cold PBS and cellswereharvested for countingbytrypsinization.Todetermine the kinetics of glucose uptake for thetwodifferent celltypes,similarexperiments tothose described abovewerecarriedout using 13 different media concentrations of

D-glucose between 0 and 24mM.Resultswereexpressedpermilligram proteinasdeterminedinparallel culture plates. Lineweaver-Burk double reciprocal plotswereused for the calculation ofKmand

Vmax.

Chromatography. Amino acidswereanalyzedastheir

precolumn-dansylated derivatives byreversephase HPLCaspreviously described (8).Inbrief, derivatizationwascarriedoutatroomtemperatureatpH 9.0 ina 3.5/1 molarratio of5-dimethylaminonaphthalene-l-sulfonyl

(dansyl) chloride/amino acids for 20h.Analyseswereperformed using

a Beckman344 HPLC(Beckman Instruments) and 0.05Mmonosodium phosphate/acetic acidbuffer, pH 7.0, as the initial eluant andacetonitrile

asthe final eluant. Sample sizewas46 and 355 nmol ofamino acid residues for analysis of samples fromsupernatantsandprotein

precipi-tates,respectively. Alineargradient between 10and 25% acetonitrile

at 1.5ml/min flowrateresulted inoptimal separation of hydroxyproline

andprolinein46min. The columneffluentwasmonitoredfor fluores-cence (Spectroflow 980 fluorescence detector; Applied Biosystems,

Ramsey,NJ) at350nmexcitation and 470nmemissionwavelengths, and0.3ml fractionscollected formeasurementof 3H and 14Ccontent. The recovery of the[3H]prolineandhydroxyprolinewas43-77% and 70-96%, respectively.

All radioactivity measurements were carried out using Optiphase HisaveII (LKB Scintillation Products, Loughborough, United King-dom)asscintillatorin a3-channelliquid scintillationcounterproviding quench compensation (Beckman LS-3801; Beckman Instruments).

Immunogold labeling. The presence of cell-associatedGLUT1 was studiedby light microscopic examination of immunogold silver-stained samples of acetone-fixed MC cultures (40). Cellswereseeded in 0.79

cm wells at 12,600 cells/cm2. At 5 d of

growth

the medium was

aspiratedand thecell layer washed with PBSfollowed byfixationin acetonefor 10min. After airdrying,thespecimenswereimmersedin

PBS for 20min, placedin 1% BSAfor 1 h at room temperature and thenincubatedwith theanti-GLUTlantibody. After extensive washing

inPBS,thegold-conjugated secondary antibodywas applied for 2 h at

25°Cand the specimens washed againin PBS. Finally, samples were

treatedwith glutaraldehydeand thelabeling enhancedwiththe

Amer-sham silver solution following the manufacturers instructions. Light

microscopic examination was made in Mayer's hematoxylin-counter-stained specimens.

Chemicalmeasurements.Myo-Inositolwasmeasured spectrophoto-metrically by followingthe reductionofNADduringtheinositol

dehy-drogenase reaction (41). D-Sorbitolwasanalyzed bymodificationofa

colorimetricmethod(42)based on thesorbitoldehydrogenase reaction

and the NADH-induced reduction of iodonitrotetrazolium chloride (Test-combination D-sorbitol/Xylitol; Boehringer Mannheim Biochemi-cals, Indianapolis, IN).Lactatewasmeasuredaccordingto the NADH

formed duringthe lactate dehydrogenase reaction utilizing a

commer-cially available kit(SigmaDiagnostics Lactate; Sigma Chemical Co.).

D-glucosewasmeasured by acolorimetricmethodbasedontheglucose

oxidase-peroxidase reaction (glucose procedure no. 510 kit; Sigma

ChemicalCo.). Proteinwasmeasuredby themethod ofLowry using

BSAasthe standard.

RNAwas measuredby the orcinol reaction forquantitation of its ribosecontent.With this method 1 qg of yeast RNA(TypeI; Sigma ChemicalCo.)isequivalentto0.6ttgof ribose(43). DNAwas

mea-suredbyits ultravioletabsorptionbya2wavelengthratio method(44)

1 2 3 4 5 6 7 Figure 2. Ethidium bro-mide-stained agarose gel

Kb electrophoresis bands

de-picting

theconstruction

94_ ofthepWZLneovector.

InlaneIis the uncut

do-

6.6-norvector pSPGT. In

44 lane2, thisvector has

beencut with BamH1 to 2.3 remove the 2.6 kb

2.0-

_GLUTI

cDNAinsert. In lane 3, EcoRl digestion

linearizesthe vector. Lane4 contains the

XHindIIIsize markers.

Lane 5 demonstrates the final uncut product

pWZLneoGLUTl

re-sulting from splicing of the GLUT1 cDNA into the BamHl cloning site of pWZLneo.Lane6demonstrates the product ofaBamH1 digestof

pWZLneoGLUTltoconfirm thepresenceof the2.6 kbGLUT1 insert. The 5'-+_{3' orientation of}GLUT1 _{inside thepWZLneo}vectoris confirmedinlane 7, where digestion withEcoRlresultsinthe expected 6.0 kb and2.2kb bands.

standard. Total amino acids were quantified by a modifiedninhydrin method (45) using L-leucineasthestandard.

Expression of results and statistical analyses. Depending on the type of experiment, results were expressed as perunitprotein inthe celllayeror aspercellorDNA, andpresentedasmeans±SEM. With themethods utilized, the cellularcontentofDNA was27.4±1.74pg.

Theoptical density of the bands in the immunoblotting analyseswas

expressed in arbitrary units and thefinal resultspresented aspercent

change from controlvalues. Thenetcollagenaccumulation in the

me-dium andin thecell layerwasexpressed accordingtothe amountof [14C]prolineincorporated intoprotein-associated [14C]hydroxyproline,

while theincorporation intoprotein-associated[14C]prolinewas consid-eredas anindex of totalproteinsynthesis.Inthealternatemethod,net

collagen accumulation in the medium was measured asthe total 14C radioactivity incorporated intocollagenase-digestible protein and total protein synthesisasthe`4Cradioactivity incorporatedinto collagenase-resistant protein. Collagen breakdownwasestimatedasthetotalnewly formed freehydroxyproline,i.e.,thesumof the mediumandcelllayer

[14C]hydroxyproline detectedas afreeamino acid. Totalcollagen

syn-thesiswasquantifiedasthetotal ['4C ]hydroxyproline formed, freeor

protein-associated,in the whole tissueculturesample.Allresultswere

adjusted forrecoveryratesofpureradiolabeled internal standards.

Incor-porationvalueswereindividually corrected in each sampleforthe sp

actof themediumfreeprolineatthecompletionofthelabeling period, andpresentedasnanomoles ofproline incorporatedper 24 hof radiola-beling. Differences between groups were evaluated using Student's t testfornonpaired samplesand the distribution oftinatwo-tailedtest.

Since previous studies demonstrated thatcollagen synthesis may be inversely relatedtotissue culture cell density in subconfluent cultures (8), significantdifferences inincorporationresultswere confirmedby analysisofcovariance.Inthis analysis,toremove the effect of

differ-encesincell content, the totalamountof DNA in thesamplewasadded

as aregressoraffectingthedependentvariable(incorporation).

Results

Gene construction and characterization ofLacZ and GLUT] expression vectors. Construction of the pWZLneo expression vectoris depicted in Fig. 2. TheGLUT1 cDNA was first

re-moved from thevectorpSPGT by cutting withBamHland then

1 2 Figure3. Northern

anal-ysis of transduced

mes-angial cells.Total RNA

wasisolated from

con-proviral RNA fluent cultures of

28S-

- poviral RNA

MCLacZ

(lane 1)

or

MCGT1 (lane2) cells 7

Glut1 - dafterseeding.The

18S- Northernblotwasprobed

witha32P-labeledhuman

GLUT1 cDNAfragment

todemonstrateGLUT1

mRNA. Theexogenous

GLUT1 mRNAderived frompWZLneoGLUTl inMCGT1cells is included in the 5.6 kb band

ofproviralRNA.The endogenousGLUT1 mRNA ofMCGT1 cells

(Glut1)appears atthe 3.0kb level, althoughit is obscured by the

intensesignal fromtheproviralRNA. MCLacZ cells contain only the

3.0 kb endogenousGLUTI mRNA.

spliced into the BamHl cloning siteof pWZLneo. Digestion withEcoRi confirmed the 5' -+3' orientation of the insert.

Expression of GLUT] mRNA and GLUTI protein. The G418-resistantLacZ transduced MCs expressed large quantities of

P-galactosidase

asshown by the X-gal staining test. In

addi-tionthesecells, grown in 8mMglucose, expressed substantial

amounts of GLUTI mRNA (Fig. 3). The level of

GLUTI

mRNA increased18-foldinMCGT1cultured in the same condi-tions. MCLacZ alsoexpressedthe glucose transporter protein,

as demonstrated byimmunogold localization and immunoblot

analysis (Figs. 4 and 5). As compared to these controls,

MCGT1 cells cultured in the same glucoseconcentration

dem-onstrateda10-foldenhancedexpression of GLUTI proteinthat was evident inimmunoblotting analyses(Figs.5 and 6) and in

immunogold studies (Fig. 4).The latter alsodemonstratedthat this changewas generalized and ofa similarmagnitudein all the MCGT1 cells. In addition, this overexpression was still present at similar levels after 3 moin cultures maintained in the same normal glucose concentration, as demonstrated by

immunoblot analysis (datanotshown).

Glucose transport and kinetics. Preliminary experiments demonstrated that, under the conditions selected, 2-deoxyglu-coseuptake increasedlinearlywith time overthe first 10 min ofincubation. AscomparedtotheMCLacZcontrols,the5-min

uptake of the glucose analog in MCGT1 cells was markedly augmented (Fig. 7), suggesting a greatly enhanced entry of glucose in cells overexpressing GLUT1. In support of these

findings, kinetic analyses revealed a 4.3-fold higher

V.,

in MCGT1 vsMCLacZ cells (P < 0.001) (Fig. 8). In addition, the Km values for the rat (MCLacZ) and human

(MCGTl)

transporters were similar (P > 0.2) and within the expected

range of values (Fig. 8).

Cellgrowth. Exposureof normal MCs to 35 mM

glucose

for7dresulted inamoderatelydiminished

proliferating activity

and slight cell hypertrophy, as shown by a 17%

significantly

lower DNAcontentandan8%significantly higherRNA/DNA ratio in cells cultured inhigh glucose concentration

(Table I).

Changes similar to these, but

greatly exaggerated,

were ob-served inMCGT1 cells cultured in normal

glucose

concentra-tions. Atthe end of the same observation

period,

cultures of

Figure 5. GLUTI expression intransduced mesangialcells.Duplicate immunoblot analyses of50-pgprotein samples, obtainedfrom the cell

layers of confluent cultures,areshown. GLUTIprotein is demonstrated

inMCLacZ and MCGT1 cellsassingle bands ofverydifferentintensity migratingat48 kD.

hanced glucose transport, albeit these changes were greatly

magnified.

Metabolic characteristics. To determine if the enhanced

glucose uptakeinducedbytheoverexpressionof GLUTI trans-porterwasalso associated withanincreased metabolism of the

hexose,lactate and sorbitolcontentswere measuredas indices

of substrateutilization. Under the same conditions of normal

glucose concentrationasabove,lactate release into the medium

as well as that associated with the cell layer were 2.5- and 2.2-foldgreater,respectively, inMCGT1 cultures than in their

MCLacZ counterparts (Table II). In similar experiments, cell sorbitolcontentwasalso increased 2.1-fold in MCGT1 cultures

(Table II). Interestingly, this sorbitol accumulation was also

associated with a significantly increased content of cell

myo-inositol (Table II). These findings suggest that an increased

glucose transport inMCs is linked to the greatermetabolism

of this sugar,atleast via theglycolytic and polyol pathways. Productionof extracellular matrixcomponents. The

secre-tion into the medium of specific extracellular matrix

compo-nents was studied in MCGT1 cells to establish whether

en-hanced glucose transport, albeit in an environment of normal

glucose concentration, couldeffectively stimulate the synthesis

of the maincomponents of mesangial matrix. As comparedto

their MCLacZ controls, MCGT1 cells secreted significantly

Figure 4.Immunogold-silver labeling of cell-associated GLUTI in transducedmesangial cells. The cell layer of MCLacZ (middle panel)

orMCGT1 cultures(lower panel)wereincubated withspecific anti-GLUT1 antibodyornonimmuneserum(MCGT1 cells,toppanel)as a control. Thelevel ofGLUTI expressionisdemonstratedaccordingto theintensity of the brown silver staining. Original magnification

was40.

a 38% higher RNA/DNA ratio than their MCLacZ controls

(Table I).The lowerproliferative activityof the MCGT1

cul-tures ascomparedto MCLacZ cultureswas alsosuggested by

the differentshape of the growth curvesobtained over a 15-d

periodofculture (Fig. 9). Therefore, the inhibition of

replica-tion andhypertrophic effects associated withexposure of

nor-malMCstohigh extracellularglucose concentrationswerealso

present under normal glucose concentration in cells with

en-1500

-oC:

:

0

e0

2*:

O1C

Ed,

1000

-500

-MCLacZ (Control)

Figure6.GLUT1 protein

contentintransduced mesangial cells. The in-tensityofthe GLUTI

bandsobtainedinthe

im-munoblot analyses of MCLacZ and MCGT1

cellswerequantified by optical scanning densi-tometryand results

nor-malizedtotheamountof

_6-tubulin. Resultsfor

MCGT1 cells in three

separateanalysesare

presentedasthemean percentchangeover con-MCGT1 trol MCLacZ values,

with SEM indicated.

GLUT1

46kD

MCLacZ MCGT1 MCLacZ MCGT1

T

MCLacZ MCGT1

Figure 7. Uptake of 2-deoxyglucosein

trans-duced mesangial cells.

Studies werecarriedout incultures preincubated

in aglucose-freebuffer

solutionbeforethe addi-tionof

2-deoxy-D-[1-3H]glucose. Results are presented as the mean±SEM.

morecollagentypeI, collagen type IV,fibronectin,andlaminin (Fig. 10).

Northern analysesforindividual matrix components

demon-strated fibronectinmRNAas asingleband andcollagensIand IV astheircharacteristic doublets(46, 47). Differences in the secretion of extracellular matrix components likely resulted

fromincreasedsynthesis in MCGT1 cellsbecause their respec-tive mRNAs were elevated 43-80% as compared with their MCLacZ controls (Fig. 11).

Collagenmetabolism in conditionsof high glucose

concen-trationandhighglucosetransportactivity.To

analyze

indetail

howcollagenmetabolism may be altered

by

the presence ofhigh

glucose concentrationsorbythe enhanced

glucose

transport, the synthesis and catabolism of collagen was studied in normal MCs cultured in 8or 35 mMglucose and in transduced MCs cultured in 8 mM glucose. The incorporationrate of radiola-beled amino acid precursor into

protein

is

strongly

influenced by changes in the sp act of its endogenous

pool.

Therefore,

KM-3=JJ | Figure 8.Kinetic

analy-sesof2-deoxyglucose

o

uptake

in transduced

mesangial cells. Studies

werecarried out as in

Fig.7 exceptthat media

totalglucose concentra-tion was varied between 0 and 24 mM. Each data

pointrepresents the mean valueof six separate de-o _{terminations. Results}

presented are the com-binedvalues from three separateexperiments. Li-* neweaver-Burkdouble

reciprocal plots were cal-culatedtodeterminethe

*'8 ; Km and Vm. values for 0.8 1

each of the two cell

;e)-1 types.

Tablel. GrowthofMC and MCGTI or MCLacZ*

DNA_(jig) RNA(Isg)/DNA(mg)

MC8 mMglucose 96.29±5.17 136.3±2.22

MC 35 mM glucose 84.00±1.67t 146.8±3.5§

MCLacZ 8 mMglucose 91.32±1.57 151.8±2.7

MCGT1 8 mMglucose 60.87±1.7211 210.2±1.711

*_{Resultswereobtainedin cultures at 7 d of growth. To compare the}

relative effects of 35mMglucose and GLUTI overexpression, experi-mentsdemonstrating similar DNAinthe MC control and MCLacZ groupswereselected. Values are mean±SEM of six samples in each

group. tP =0.047; §P=0.031,significant from MC, 8 mMglucose.

IIP< 0.0001, significantfrom MCLacZ.

initialexperimentsweredonetoevaluate the effect of medium glucose concentration on medium and endogenous pool proline sp act. At the completion of the incubation period, medium proline sp act was 88 and 94% of the initial value in me-diacontaining 8 and 35 mMglucose,respectively. The

differ-ence between these two groups was significant (8 mM,

133,717±6,068, n = 6; 35 mM, 149,000±3,810 dpm/nmol pro-line, n = 6, P < 0.0001). These changes in medium proline sp act were mirrored by those occurring in the cellular

endoge-nouspool of proline. Thus,atthe end of the incubationperiod

prolineendogenous pool sp act was also significantly lower in samples incubated in 8 mM glucose (8 mM, 87,468±5,389, n = 6; 35 mM, 102,650±7,516 dpm/nmol proline, n = 6, P

=0.0004). Since calculation of incorporation results according tothe sp act of proline in the endogenous poolorin the incuba-tion media did not alter the differences between groups, all results were expressed according to the final sp act of free proline in the sample's incubation medium.

In normal MCs, exposureto 35 mM glucose for aperiod of 12 d induced a 69% increase in collagen synthesis (Table HI). This change was associated with an 80-90% greater net accumulation of newly formedcollagenin themedium,as mea-sured by two independent radiolabeling methods. Therewasan inverse relationship between theamountof DNA in the sample andcollagen accumulation (P = 0.0005). When the effect of different DNAcontent wasremoved by analysis of covariance,

12

-0

-U U

8-

4-O

A--I

_"

T/5

/

I

I I . _I

0 4 8 12

Time of Culture (days)

Figure9.Growthrateof

transduced mesangial cells.Cells were counted atdifferent growth peri-ods in continuous

MCLacZ(o) and MCGT1 (*) cultures.

I Datapoints_representthe

6 _{mean±SEM of four} sam-ples.

1-1

I,

._

0

2

0

4)

1

0;

Q

04

4)

2.0

1.5

-

1.0-0.5

-

0-* MCTG1, Vmax=852,

OMCLacZ, Vna =196,

._

E

S 30 0

co E

E ₂₀

10 '

0 0.2 0.4 0.6 1/ S(mM D-Glucos

Table II. Metabolic Characteristics of MCGTJ orMCLacZ Grown inthe Presence ofNormal Glucose Concentrations*

MCLacZ MCGT1

n=4 n=4

Lactate production

(mmollmg proteinper72 h)

Medium 4.16±0.57 10.58±0.85t

Cell 1.66±0.10 3.66±0.21§

Cellular sorbitolcontent 6.54±1.26 14.07±1.33* (nmollmg protein)

Cellularmyo-inositolcontent 18.4±1.41 35.65±2.02§

(nmol/mg protein)

* Results wereobtainedinculturesgrown toconfluencyin8mM

glu-cose.Valuesaremean±SEM. tP<0.005; 'P<0.0001.

the differences between groups were still significant (P

=0.0001).Inaddition,theaccumulationofcollagenin thecell

layer (much lowerthan into the medium dueto the presence of/3-aminopropionitrile)wasalso increasedby68%. The

incre-mentincollagen synthesis caused by glucosewasalso

associ-ated with a59% greater catabolism. Although the fraction of thecollagenproducedundergoing catabolismwassignificantly lower inhigh glucoseconcentration cultures, comparison of the magnitude of the changes in synthesis and catabolism reveal that themain causefornetcollagenaccumulationwasenhanced

formation.

The increase in collagen accumulation in the incubation medium coincidedwithastimulation in overallproteinsecretion

asmeasuredby thetworadiolabelingmethods(Table III)

(col-lagenase-resistant protein: 8 mM, 42.6±1.4, n = 8; 35 mM,

79.5±3.9 nmol proline/mg DNA/24 h, n = 8, P < 0.0001).

However, the change incollagen formation was significantly greater than thatfortotalprotein(Table III).

At thecompletion ofthelabeling period, medium proline sp act did not differ in the two groups of transduced cells cultured in 8 mM glucose (MCLacZ, 205,288±3,030, n = 6;

Collagen Collagen Fibronectin Lainin

TypeI TypeIV

-Figure10. Secretion into

themedium of

extracel-lularmatrixcomponents

intransducedmesangial

cells. Theamounts

se-creted in24 hby

MCGT1 cultures and their MCLacZcontrols

arepresented. Valuesare

means±SEM,n= 6. P

< 0.0001.

FN

CoI.NV

Col.l

A.

B._

1 2 3 4 5 6

D

a

cL

0

0

FN Co. IV Co. MC LacZ

(Control) MCGT1

Figure 11. Northern analysis of individual extracellular matrix compo-nentsin transducedmesangial cells. 20ggof totalRNAisolatedfrom MCLacZ(lanes 1, 3,and5)orMCGT1 (lanes 2, 4, and 6),were

loadedtoeachlane forelectrophoretic separation. Blotswereprobed forfibronectin(FN), typeIVcollagen (Col. IV)andtypeIcollagen (Col. I)using their respective32P-labeledcDNAs(A).The bargraph indicatesquantitation of Northern analyses for matrixcomponentsby optical scanning densitometry,withresults normalizedtothe amount

ofmRNAfor the housekeepinggene,B-tubulin(B).Results forMCGT1

cellsinthreeseparateanalysesarepresentedasthemeanpercentof controlMCLacZvalues, with SEM indicated.

MCGT1,

200,230+3,290,

n = 6).MCGT1 cells demonstrated

a 109% increase in total collagen synthesis, associated with

a 111-117% greater net accumulation, as compared to their MCLacZ controls (Table IV). As in experiments in normal mesangial cells, there was a significant effect of the sample's DNAcontenton thecollagen accumulated (P = 0.005). After this effect was eliminated in an analysis of covariance, the

differencesbetween groupswerestillsignificant (P = 0.015). Thisaugmentedcollagen accumulation in the medium was part ofanoverall enhancement inproteinsynthesis as suggested by results from the two methods used in this study (Table IV)

(collagenase-resistant protein: MCLacZ, 45.6±13.3, n = 6;

MCGT1, 82.8±10.7 nmol proline/mg DNA/24 h, n = _{6, P}

= _{0.055). Nevertheless,} as shown above in normal cells, the

synthesis of collagen was particularly stimulated in MCGT1 cells (Table IV). The amountof collagen accumulated in the cell layerwas also increased in MCGT1 cells by 64% of the value for MCLacZ cells. Also, as shown in normal MCs, the greater collagen synthesisinMCGT1 cellswasassociatedwith amarked increaseincollagen catabolism. However, the fraction of collagen produced which was catabolized was similar in both types of cells, therefore, the net collagen accumulation demonstrated inmedium and celllayer ofMCGT1 cultureswas

fully attributable to anincreased rateofsynthesis.

10-ra0) 0

u8

Table III. CollagenMetabolism in RatMesangial Cells Exposed to High Glucose Concentrations*

8 mM

Glucose 35mMGlucose

n=8 n=8

Totalcollagen synthesis (nmolProincorporated intototal Hyp) 69.10±1.58 117.11±2.76t Medium collagen accumulation(nmolProincorporatedinto protein-associated Hyp) 32.28±0.65 58.33±1.58t Mediumcollagenaccumulation (nmolProincorporated into collagenase-sensitiveprotein) 46.94±0.79 89.37±2.78t

Medium total protein accumulation (nmol ProincorporatedintoProtein) 166.6±10.5 250.8±11.2$

Mediumfractional collagen accumulation (collagen formationaspercentage of total protein) 16.47±0.66 18.96±0.39§ Cell layer collagen accumulation (nmolProincorporatedinto protein-associated Hyp) 0.766±0.046 1.286±0.0911

Totalcollagencatabolism (nmol Proincorporatedintofree Hyp) 36.05±1.14 57.49±1.140t

Fractionalcollagen catabolism (percentage of total collagencatabolized) 52.11±0.63 49.10±0.461

* Resultswereobtained after48 hof incubation with 183

_ILM

radiolabeled proline. Incorporation data were corrected for the media specific

radioactivity of theprecursorand expressedpermilligram of DNA/24h.Values aremean±SEM.Pro,proline; Hyp,4-hydroxyproline. *_{P < 0.0001;}

§P= 0.0056; 11P=0.0002; 1P= 0.0017.

Medium was regularly changed at 48-h intervals during the period of cell growth. Subsequently, during the 3-d proline

radiolabelingperiod, withoutreplenishmentof the medium, the MCGT1 cells demonstrated a 42-fold greater netglucose utiliza-tion(TableIV), which caused a decrease in themedium glucose

concentration from 8 mM to values - _{5 mM.}

Discussion

Inthis work wehave demonstratedthat MCsoverutilizing

glu-cose synthesizeand accumulate increased amounts of

extracel-lular matrixeven inthe absenceof elevated extracellularglucose

concentrations. Clearly, it can be concluded from this finding thatchanges directly relatedtothe presence ofahigh glucose concentration, i.e., extracellular

hyperosmolarity

and abnor-mally increased transmembrane gradients of glucose, are not

necessarily requiredfor the excessive formation ofextracellular

matrixbyMCs in adiabetic milieu. Thisstudysuggests,instead,

that the relevantfactorislinkedtometabolic

changes occurring

duringtheoverutilization ofglucosewhich follows its enhanced

uptake. MCGT1 cells demonstrated a markedly increased

trans-port ofa glucose analog which was due to an increased

V,.

whileaffinity of the transporter remained unchanged. The

mea-sured Km in MCGT1 cells and in their MCLacZ controls of 3.1-3.7 mM is consistent with the values of 1-7 mM determined by similar 2-deoxyglucose uptakes in multiple tissues forGLUTM

(48, 49). In conditions of 8 mM glucose concentration, this

transporteris,thus, fully saturated and any increases in uptake

areexpectedtobe mediated by the up-regulation of transporters. This suggeststhat under conditions of high glucose

concentra-tions exaggerated glucose uptake and increased extracellular matrix synthesis may be related to increased expression of func-tional GLUT1.

The increased glucose uptake in MCGT1 cells was also accompanied byahighnetutilization of glucose andan exag-gerated formationof lacticacidand sorbitol. In addition, itwas

also associated with the accumulation ofmyo-inositol.

There-fore, alterations related to myo-inositol depletion may be

ex-cludedascausative factors for theglucose-stimulated extracel-lular matrix formation.

Table IV. CollagenMetabolism and Glucose Utilization inMCGTIorMCLacZExposedtoNormalGlucose Concentrations*

MCLacZ MCGT1

n=6 n=6

Totalcollagen synthesis (nmolProincorporatedinto totalHyp) 102.33±1.66 214.37±4.49*

Mediumcollagenaccumulation (nmolProincorporatedintoprotein-associated Hyp) 23.37±0.41 49.35±2.04*

Mediumcollagenaccumulation (nmolProincorporatedintocollagenase-sensitiveprotein) 39.66±2.94 86.06±6.01*

Medium totalproteinaccumulation(nmolProincorporatedintoprotein) 131.71±5.54 238.16±8.69*

Mediumfractional collagen accumulation(collagenformation as percentage of totalprotein) 15.14±0.43 17.16±0.20w

Celllayer collagenaccumulation(nmolProincorporatedintoprotein-associated Hyp) 1.419±0.046 2.323±0.113*

Totalcollagencatabolism (nmol Pro incorporated into free Hyp) 77.53±1.32 162.70±3.12t

Fractional collagen catabolism(percentage of total collagencatabolized) 75.73±0.25 75.92±0.58

Netglucoseutilization

(Amol)

2.69±6.51 116.20±5.98*

Thefacilitativetransporters involved in the

energy-indepen-dent uptake of glucose comprise a group of integral membrane

proteins, GLUTI -GLUT5, andGLUT7,whichareencodedby

separate genes. These proteins transport glucose with different efficiencies and kinetics (27). GLUT6 isapseudogene, GLUT7

functionsintheendoplasmicreticulummembrane,andGLUT5 is primarily an intestinal fructose transporter. The remaining GLUTisoforms, involved in the cellular transport ofglucose,

areexpressed differently within tissues demonstrating distinct metabolism of this hexose, suggesting a close link between

specific transporters andthe handling ofglucose through spe-cific metabolic pathways(26).Theinsulin-regulatableGLUT4

isoform has been identified in MCs (50), however, its func-tional roleremainsin doubt sinceglucosetransport and extracel-lular matrix synthesis in these cellsdo not appearto be influ-enced by insulin (20, 21). In this study, weconfirmprevious

observations in tissue cultures and in renal histological speci-mens (23, 24)demonstrating the presence ofGLUTI inMCs. To enhance glucoseuptakeweaugmentedthe transport ca-pacity by increasing the number of transporters. To this end,

wetransduced theneoR geneencoding neomycin

phosphotran-sferase and human GLUTI or bacterial LacZ (ascontrol) ina

cloned line ofrat MCs. The resultant MCGT1 cells

demon-strated amarkedincrease in GLUTI synthesisasshownbythe overexpression of GLUTI mRNA and GLUTI protein. The

overexpressionof LacZorGLUTI in theG418-resistant

surviv-ingcloneswasgeneralizedtoall cells andwellmaintainedafter multiple passages in culture. In cells suchas3T3-Ll adipocytes

in whichGLUTI intrinsicactivityappearstobemodulated,the heterologous expression of human GLUTI is also subject to

the sameinhibitory control(51). Therefore, itisexpectedthat if the activity of the endogenous GLUTI were regulated in MCs, theadditional exogenous GLUTI transportersexpressed

in ourtransduced cellswouldbeunder the same form of control. Theeffectofhigh glucoseconcentrationsonMCgrowth in tissue culture has beenvariouslyreportedasbeingneutral (16,

21) orexerting inhibitory effects (18, 52). We observed de-creasedproliferationandcellhypertrophyincultures of normal MCs exposedto 35 mM glucose, according to the DNA and RNA/DNA values obtained atthe endof theexperimental pe-riod.Similargrowthcharacteristicsweredemonstrated in a

nor-malglucose environment by MCs overexpressing the GLUTI

transporter when comparedto their LacZ-transduced controls. The mechanism by which this enhanced glucose uptake, whether elicited by increasing the extracellular concentration

orby stimulatingtransport, may affect MC growthisnotfully understood. However, it is likely that the process involves the induction of endogenous

TGF-f31

expression and/or activation

(53).Inlongtermcultures ofMCs, highglucose concentrations stimulate

TGF-/31

secretion and causesustained inhibition of cellproliferation,cellhypertrophyandincreasedprotein synthe-sis(53, 54).Thesechangesareprevented by neutralizing anti-body against

TGF-/3

andtheyare notreproducedin an hyperos-molar environment obtained by the addition ofL-glucose or mannitol (53).

We observed an increased cellular content of myo-inositol concomitant with the augmented glucose uptake and sorbitol accumulation inMCGT1 cells. This is an alteration similar to

that seen in cells exposed to high glucose concentrations in which the Va. of the Na + -dependent myo-inositol cotransporter is increased (55, 56). Althoughtheprecise mechanism for this

effect has not been elucidated, it involves stimulation of the

polyol pathway and activation of protein kinase C. These two metabolic alterations are likely to be present in our MCGT1 cells (see below).

Asshown in previous studies (7, 15-17, 21), high

concen-trations of extracellularglucoseincreasednetformation of col-lagen in the medium and in the cell layer in MC cultures. In addition, as demonstratedby others (7, 16), we identified an

enhanced synthetic rate as the major metabolic alteration

re-sponsible for the accumulationof collagen. Contrary toother studies (15), however, we have observed associated changes

in collagen catabolism and total protein synthesis. Collagen

catabolismwas acceleratedbyglucose,but this changewas of

insufficient magnitudetooffset themarkedly increased synthe-sis. In addition,protein secretion intothe culture mediumwas

also increased, although this change was of lessermagnitude

than thatfor theaccumulationofcollagen.Apossiblecausefor thesediscrepant results may be that, contrary topreviouswork, incorporation results in this study were corrected for changes

inthe sp act of the aminoacidprecursor.

The collagen metabolic changes shown in MCGT1 cells

incubated innormal glucose conditions qualitatively mirrored those observed in normal MCsexposedtohigh glucose

concen-tration. However, when comparedto theircorresponding con-trols, MCGT1 cellsexhibitedagreater increase in totalcollagen synthesis (109 vs 69%) and in collagen accumulation

(111-117%vs81-90%) than normal MCs incubated in high glucose conditions. All four of the individual matrix components

exam-ined in the culturemedium, collagenI,collagen IV, fibronectin, and laminin, were increased 2.3- to 4.3-fold over values in MCLacZ cultures. In addition, the increased mRNAs for colla-gen I, collagen IV, and fibronectin is consistent with the in-creasedsynthesisof these individual matrixcomponents. Total protein secretion wasalso enhanced, although not to the same

extent as for collagen accumulation. Therefore, in terms of

growth characteristics, myo-inositolaccumulation and collagen metabolism, MCs overexpressing theGLUTI transporter in an 8 mM glucose environment behaved like normal MCs grown in 35 mMglucose.

Recent studies have begun to unravel the mechanisms by which an increased entry of glucose into MCs may stimulate extracellular matrix formation. MCs grown in a high ambient

glucoseconcentration demonstrate activation of protein kinase Cas aresultofincreased diacylglycerol mass (21, 57). It has beenproposedthatprotein kinase C modulates activator protein 1 complex (AP-1), thetranscriptional productof jun and f os protooncogenes, which in turn, binds to specific sequences in the promoterregionsof extracellular matrix genes (58). This mechanism appearstobeoperative in vivo also, because

diacyl-glycerol mass andprotein kinase C activity are also increased inglomerulifrom diabeticratsand in isolated normalglomeruli

acutely exposedtohigh glucoseconcentrations (59). Since the change in diacylglycerol formation isthroughan enhanced de novosynthesis from glycolytic intermediates (19, 21, 57)in a

process favored by the altered cellularredox state caused by the increasedpolyolpathwayactivity,it follows that the stimu-lation of extracellular matrix synthesis requires the accelerated metabolism ofglucose.

Our studies in transduced cells demonstrate that glucose

transport isanimportantmodulator in MCs forglucose

metabo-lism. In addition, the glucose-stimulated rate of extracellular matrix accumulation appears to depend to a greater degree on thecapacityto transport glucose than on the actual extracellular

concentrationof the hexose. This underlines the potential impor-tance of the regulation of GLUTI expression and activity in MCs as a determinant of extracellular matrix deposition and

mesangial expansion.Glucose flux via GLUT transporters may be regulated at thetranscriptionallevel or by altering the rates of proteinsynthesisand degradation, changes in intrinsic activ-ity, and the translocation of a vesicle-associated intracellular

pool oftransporters tothe plasma membrane (26). The latter, while being paramount in the GLUT4 activation by insulin, is of lesser importance forGLUTI-mediated transport due to the alreadypreponderant localizationof this isoform on the plasma membrane under basal conditions and its lesser translocation

efficiency (60). Alargevariety of agents regulate GLUTI ex-pression(27). Inendothelialcells and hepatocyteshypoxiaand inhibitionofoxidativephosphorylation induceGLUTI expres-sion(61,62). Themostcommonly reportedeffects of glucose have been those caused by itsdeprivationinboth insulin-respon-sive andinsulin-nonresponsive cells. These consist of changes in the transport of the hexose in association with increased

GLUTi

protein, withorwithout associated changes inGLUTI

mRNA (25). These effects are readily reversed byproviding

glucose. In contrast, in the adipose tissue and skeletal muscle

ofhyperglycemicanimals withstreptozotocin-induced diabetes, GLUTI mRNA andGLUTI protein areinappropriately

unaf-fected(25, 63).

Theregulation of GLUTI expression by growth factors is

ofparticular importanceas a

potential

elementin the pathogene-sis of diabetic

mesangial

expansion. PDGF and

TGF-P

are

knowntoenhance glucoseuptake,increase GLUTI mRNA and promote GLUTI expressionin cultures of fibroblast cell lines

(64-68). MCs,in turn, produce PDGF and TGF-/3

(67-70),

thushavinganautocrinesystemcapable of

regulating

GLUTI expression. This system may be activated in diabetes because

high glucose concentrationsincrease the MC secretion of TGF-f3 and the expression of

specific

cellular receptors for this

growthfactor(71,72). Furthermore, MCsrespondtothesame growth factors by increasing extracellularmatrix formation in vitro(72,73 ) andby inducing mesangial expansionand glomer-ulosclerosis in vivo (74, 75), but it is notknown if GLUTI overexpression participates in the mediationof this effect.

It isof interest that oral hypoglycemic agents, extensively

used in thetreatmentof Type II diabetes,arehighlyeffective in increasingGLUTI expressionandglucosetransport. Metformin and the sulfonylureas tolbutamide and tolazamide increase

GLUTI protein, GLUTI mRNA, and the translocation ofthis

transportertotheplasma membrane in L6myotubes and 3T3-L1 adipocytes (76-79). The relative effect of metformin is particularly intenseinconditions of high glucose concentrations

(76). The relevance of these observations is tempered by the caution about extrapolating resultsinvitro with mechanisms in vivo. It could be speculated that the administration of these agents may result in theparadoxical circumstance in which an

improvement in glycemia may be associated with greater

glu-coseuptake and higher risk for the development of complica-tions inGLUTl-expressingtissues. Whether this mayoccur in

MCs is uncertain. There are many known examples of the

tis-sue-specific regulationandexpression of individual GLUT

iso-forms (26, 80), and there is no present knowledge as to how this may proceed in MCs.

This work suggests that increased glucose uptake, rather than the level ofglycemia per se, may be a major metabolic

determinant in the development of mesangial expansion and

glomerulosclerosis in diabetes. If MC GLUT1 expression and

activityvaries in humandiabetes,thiscould explainthe obscure

predispositionofonlyalimited group of patients to the develop-ment of renal disease and the poor correlation between glycemic levels and progressionof nephropathy in some of these cases, even afterlongperiods of diabetes (81).

Acknowledgments

This workwassupported inpartby National Institutes ofHealth grants

K08

DK01953andRO1 DK28081 awardedtoDrs.CharlesW. Heilig

and Pedro Cortes, respectively, and byagrantfromtheJuvenile Diabetes Foundation International (#1921461) awardedto Dr. Bruce L. Riser.

References

1. Bilous, R. W., S. M. Mauer, D. E. R. Sutherland, and M. W. Steffes. 1989.Meanglomerularvolume and rate ofdevelopmentofdiabetic nephropathy. Diabetes. 38:1142-1147.

2.Seyer-Hansen, K.,J.Hansen, and H. J.G.Gundersen. 1980. Renal

hypertro-phyinexperimentaldiabetes. Amorphometricstudy.Diabetologia. 18:501-505.

3.Mauer,S.M.,M. W.Steffes,E. N.Ellis,D. E. R.Sutherland,D. M.Brown, and F. C. Goetz. 1984.Structural-functional relationshipsindiabeticnephropathy.

J.Clin. Invest.74:1143-1155.

4.Steffes,M.W.,R.W.Bilous,D. E. R.Sutherland,and S. M. Mauer. 1992. Celland matrix components of theglomerularmesangiumin type I diabetes. Diabetes.41:679-684.

5.Yaoita, E.,K.Oguri,E.Okayama,K.Kawasaki,S.Kobayashi,I.Kihara,

and M.Okayama. 1990. Isolationandcharacterization ofproteoglycans synthe-tized by cultured mesangial cells.J.Biol. Chem. 265:522-531.

6.Scheinman,J.I.,D. M.Brown,and A. F. Michael.1978.Collagen synthesis byhumanglomerularcellsin culture. Biochem. Biophys.Acta.542:128-136.

7.Ayo,S.H.,R. A.Radnik,J. A.Garoni,W.F.GlassII,andJ.I.Kreisberg. 1990.High glucosecauses an increase inextracellular matrix proteinsincultured mesangial cells.Am.J.Pathol. 136:1339-1348.

8.Riser,B.L.,P.Cortes,X.Zhao,J.Bernstein,F.Dumler,and R. G. Narins. 1992.Intraglomerularpressure andmesangial stretching stimulate extracellular matrix formationintherat.J. Clin. Invest.90:1932-1943.

9.Okuda, S.,L. R.Languino,E.Ruoslahti, andW. A.Border. 1990. Elevated expression of transforminggrowthfactor /3 andproteoglycan productionin

experi-mentalglomerulonephritis.Possiblerole inexpansion ofthemesangial

extracellu-larmatrix. J. Clin. Invest.86:453-462.

10.Bilous,R.W., S. M.Mauer, D. E. R. Sutherland,J. S.Najarian,F.C.

Goetz,and M. W. Steffes. 1989. The effectsofpancreastransplantationonthe

glomerularstructure of renalallograftsinpatientswithinsulin-dependentdiabetes.

N.Engl.J.Med. 321:80-85.

11.Barbosa, J.,M. W.Steffes,D. E. R.Sutherland,J. E.Connet,K. V.Rao,

and S. M. Mauer. 1994. Effect ofglycemiccontrolonearlydiabetic renal lesions. A5-year randomized controlledclinical trial ofinsulin-dependent diabetic kidney transplant recipients. J. Am. Med. Assoc.272:600-606.

12.Rash,R. 1979.Prevention of diabeticglomerulopathyinstreptozotocin

diabeticratsbyinsulin treatment. Themesangial regions. Diabetologia. 17:243-248.

13. Steffes,M.W.,D.M. Brown,J. M. Basgen, and S. M. Mauer. 1980. Amelioration ofmesangialvolume and surface alterationsfollowingislet trans-plantationin diabetic rats. Diabetes.29:509-515.

14.Petersen, J.,J. Ross,and R. Rabkin. 1988. Effect of insulintherapyon

established diabeticnephropathyinrats.Diabetes. 37:1346-1350.

15.Ayo,S.H.,R. A.Radnick,W. F. GlassH,J. A.Garoni, E.R.Rampt, D.R.Appling,and J.I.Kreisberg. 1991. Increased extracellular matrixsynthesis and mRNA inmesangialcells grown inhighglucosemedium.Am. J.Physiol. 260(RenalFluidElectrolytePhysiol.29):F185-Fl91.

16. Danne,T.,M.J.Spiro,and R. G.Spiro. 1993. Effect ofhigh glucoseon

type IVcollagenproduction byculturedglomerular epithelial, endothelial, and

mesangialcells. Diabetes. 42:170-177.

17.Haneda, M.,R.Kikkawa,N.Horide,M.Togawa,D.Koya,N.Kajiwara,

A.Ooshima,and Y.Shigeta.1991. Glucose enhances type IVcollagenproduction