1220 Volume I . Number I I ‘ 1991

Effect of myo-lnositol on Cell Proliferation and CollagenTranscription and Secretion in Proximal Tubule CellsCultured in Elevated Glucose1

Fuad N. Ziyadeh, M.D.,2 David A. Simmons, M.D., Edward R. Snipes, M.D., and Stanley Goldfarb, M.D.

F.N. Ziyadeh, ER. Snipes, S. Goldfarb, Renal-Electrolyte

Section, Department of Medicine, University of Penn-

sylvania School of Medicine, Philadelphia, PA

D.A. Simmons, Diabetes Section, Department of Med-

icine, University of Pennsylvania School of Medicine,

Philadelphia. PA

(J. Am. Soc. Nephrol. 1991; 1:1220-1229)

ABSTRACTTubulointerstitial lesions often develop in the kidneys

of patients and experimental animals with diabetes

mellitus. In an in vitro model of diabetic renal dis-ease, it has been previously demonstrated in thislaboratory that elevated glucose levels stimulate

procollagen transcription and secretion in proximal

tubule cells in culture while inducing cellular hyper-

trophy and reducing cellular proliferation. Previous

experiments in other tissues have suggested thatmyo-inositol supplementation, probably by reversing

a disturbance in cell myo-inositol metabolism re-

lated to increased activity of the polyol pathway,reverses the effects of glucose on cell function. Wetested the effect of myo-inositol supplementation onproximal tubule cells in culture in the presence of

elevated medium glucose level. Incubation in 450mg/dL of glucose media reduced cell proliferation;

450 mg/dL of glucose plus myo-inositol (800 �M)increased proliferation, returning the value to thatseen in cells incubated in 100 mg/dL of glucose.Incubation in 450 mg/dL of glucose media increased

type IV and type I procollagen mRNA levels andpeptide secretion rates compared with those seenin cells incubated in medium containing 100 mg/dL

of glucose. This glucose-induced stimulation of pro-

collagen mRNA levels and procollagen secretionwas not observed when the elevated glucose me-

, Received October 26, 1990. Accepted January 29. 1991.2 Correspondence to Dr. F.N. Ziyadeh, 700 CRB, Renal-Electrolyte Section, Uni-

versity of Pennsylvania School of Medicine, 422 Curie Boulevard, Philadelphia.

PA 19104-6144.

1046-6673,10101 1-1220$03.00/0Journal of the American Society of NephrologyCopyright © 1991 by the American Society of Nephrology

dium was supplemented with 800 �M myo-inositol.

On the other hand, myo-inositol supplementation didnot prevent the glucose-induced cellular hypertro-

phy: there was no reduction in the increased Ieucineincorporation and cellular protein content. Cell in-

cubation in 450 mg/dL of glucose media did notlead to a measurable decrease in total cellular myo-

inositol. However, myo-inositol supplementation of

the medium resulted in a rise in cellular myo-inositollevels. These data suggest that some effects of 450mg/dL of glucose medium on proximal tubule cellsare mediated by alterations in cell myo-inositol me-tabolism. myo-lnositol supplementation reverses the

glucose-induced reduction in cell proliferation andthe increase in procollagen gene transcripts and

peptide secretion. The role of these effects in thepathogenesis of the tubulointerstitial renal lesions ofdiabetes remains to be defined.

Key Words: Extracellular matrix, diabetes, diabetic nephrop-

athy, basement membrane. type I collagen. type IV colla-

gen, kidney hypertrophy

I n diabetic nephropathy, renal failure is correlatedwith the development of interstitial fibrosis (1 ,2)

as well as with the more widely recognized prognes-sive mesanglal expansion in the glomerulus (2).Renal fibrogenesis is a complex process involving theinteraction of resident cells and circulating cells andthe influence of local and systemic humoral factors.The contributions of the renal tubule epithelial cell

and the renal fibroblast to this interstitial process indiabetes are unknown. Previously, by using a cellline derived from mouse cortical proximal tubules(MCT), we have found that high medium glucoseInhibits cell proliferation, induces cell hypertrophy,

and stimulates the transcription and secretion ofbasement membrane procollagen (type IV) and Inter-stitial procollagen (type I) (3).

Increased activity of the polyol pathway has nowbeen shown to mediate many of the effects of highextracellular glucose levels through an alteration incellular myo-inositol metabolism (4). This, in turn,induces altered function in peripheral nerve (5), pig-mented epithelium of the retina (6), and neovasculartissue (7) in early hyperglycemia. Winegrad et at. (4.8)

Ziyadeh et al

Journal of the American Society of Nephrology I 221

have postulated that in these tissues, reduction in aspecific, labile cellular pool of myo-inositol, which isin equilibrium with extracellular myo-inositol, re-sults from activation of the polyol pathway. Depend-ing on the type of tissue, several disturbances inspecific cellular function may then become manifest(4,8-1 1). In streptozotocin-induced diabetes in therat, we have recently shown that the early glomerular

hyperfiltration has also been prevented by feeding

the animals a diet supplemented with myo-inositol(12).

In the current studies, we examined the effect of

myo-inositol supplementation on the response ofproximal tubule cells to high-glucose medium. Ourresults suggest that alterations in cell myo-inositol

metabolism could underlie a number of the previ-ously described effects of high ambient glucose onproximal tubule cells.

METHODS

Cell Culture

Details of cell isolation and characterization are aspreviously described (1 3). The cell line designated

MCT (�mouse cortical tubule”) is derived from micro-dissected proximal tubule segments from kidneys of

normal SJL mice and is stabilized In long-term cul-tune by virus transformation with a nonreplicating,noncapsid-fonming strain of simian virus 40. Multi-ple morphological and functional criteria demon-

strate that this cell line exhibits characteristics con-sistent with differentiated proximal tubule epithelialcells ( 1 3). MCT cell growth can be maintained in

serum-free media for several days. and the cells elab-orate many extracellulan matrix proteins, predomi-nantly laminin, procollagens type IV and V. and, toa lesser extent, procollagens type III and I.

Culture Media

The cell culture medium used is Dulbecco’s modi-fled Eagle medium (DMEM; GIBCO Laboratories,

Grand Island, NY), containing a physiologically nor-mal D-glucose concentration of 100 mg/dL, i.e., 5.5mM, or a high D-glucose concentration of 450 mg/dL,i.e. , 25 mM. The concentration of myo-lnosltol in this

standard medium is 40 zM, close to that in blood ofnormal animals (5). The level of myo-inositol in thetest medium Is then varied between 40 and 1,200�M. In all of the experiments in which DMEM is used,

the media are supplemented with streptomycin (200gg/mL). penicillin (192 �g/mL), glutamine (2 mM).and transferrin (1 ,zg/mL) (3).

The MCT cells are passaged every 48 to 72 h andare carried in DMEM supplemented with 1 0% mac-tivated fetal calf serum. The cells are Incubated In ahumidified atmosphere of 5% carbon dioxide at 37#{176}C.

Thymidine Incorporation Assays

Thymidine incorporation is carried out by standardtechniques as previously described (3). Cells areplated in 96 microtiter wells in serum-free DMEM.After 48 h, the media are removed and the cells are

allowed to grow for an additional 48 or 72 h in varioustest media. The serum-free media tested contain 100or 450 mg/dL D-glucose with variable concentrationsof myo-inositol ranging between 40 and 1 ,200 �M.

During the last 6 h of culture, the cells are pulsedwith [3H�thymidine (5 Ci/mmol; Amersham Corp.,

Arlington Heights, IL) by the addition of 1 �Ci perwell. At the end of this period, the media are removedand the cells are released with trypsin-EDTA mixture(Flow Laboratories, Inc. , McLean, VA) and are col-lected with a cell harvester (Brandel, Gaithersburg,MD) onto glass-microfiber filter paper (934-AH:

Whatman, England). The radioactivity is assayed bycounting filters in scintillation cocktail for �3-emis-sions. �3Hjthymidine incorporation into cell DNA is

taken as an index of cell proliferation and is ex-pressed in counts per minute per well. Each expeni-

mental condition Is tested In three to six replicatewells, and the mean is taken to represent an individ-

ual experiment.

Radioimmunoassay of Procollagen Secretion

Details of the procedure are described elsewhere(3, 1 3). MCT cells are plated in 24-well plates, andgrowth is rested in serum-free DMEM for 24 to 48 h.

The media are then switched to fresh DMEM contain-Ing low- or high-glucose concentrations with or with-out myo-inositol supplementation. To measure pro-collagen secretion in the supernatants, the media aresupplemented with 25 ,�g/mL of L-ascorbic acid and25 �zg/mL of /3-aminopropionitnile, a cross-linking in-hibitor. After another 48- or 72-h period. superna-tants are removed from each plate, transferred intoseveral U-bottom polyvinyl chloride 96-microtiter

wells (Costar, Cambridge, MA), and frozen at -20#{176}Cfor subsequent solid-phase radloimmunoassay of sol-uble procollagens. Standard collagens type I and typeIV are dispensed into polyvinyl chloride wells at var-

bus concentrations for comparison with superna-tants from MCT cells. The wells are dried overnightat 4#{176}C,blocked with 4% BSA, and incubated with1 : 1 00 dilution of affinity-purified polyclonal rabbit

anti-type I and anti-type IV collagen antibodies (giftsof Dr. Antonio Martinez-Hernandez, Thomas Jeffer-son University. Philadelphia. PA). Binding of theseprimary antibodies is measured by using ‘251-labeledanti-rabbit antibodies (10� cpm/well) in 4% BSA.Binding of supernatants is compared with knownstandards as described previously (13).

Proximal Tubule Cell Effects of myo-lnositol

I 222 Volume I ‘ Number I I ‘ 1991

Protein Synthesis

The methods for leucine incorporation and cellularprotein content have been described previously (3).

For leucine incorporation, cells cultured in 24-wellplastic plates are used. After a 48-h quiescence, thecells are exposed for 48 h to various media and arepulsed during the last 1 4 h with 2 �zCi/mL of [3H)leucine (70 Ci/mmol; Amersham). After the media

are removed, the monolayers are washed with ice-cold phosphate-buffered saline (PBS), precipitated in

5% trichioracetic acid for 10 mm, and then washedwith ice-cold deionized water to remove the aqueousphase of the label. The acid-precipitated monolayersare then solubilized with 0.5 N NaOH-0. 1 % Tritonx- 1 00 and counted for radioactivity. [3Hlleucine in-corporation is expressed per 1 06 cells and is taken asan index of total protein synthetic rate. Cell numberis determined simultaneously in replicate wells byusing a Coulter cell counter.

mRNA Isolation and Measurement

The methods are essentially as described previ-ously (3). For each experiment, 1 X 1 06 to 1 .6 x 106

MCT cells are plated into each of four or more 75-cm2 plastic flasks. After the cells are grown for 48 hin serum-free DMEM containing glucose (1 00 mg/

dL), the media are replaced by fresh DMEM contain-ing either 1 00 or 450 mg/dL of glucose. The final

myo-inositol concentration is either 40 or 800 �M.

After 72 h, the cells are released with a trypsmn-EDTAmixture, washed twice in PBS, and counted. The cells

are then lysed with 0.5% Nonidet P-40 in Tnis-EDTAbuffer. Cell equivalents (0.2 X 1 06) of cytoplasmicRNA are denatured into a solution of 40% Formalinin 1 2X SSC (SSC is 0. 15 M sodium chloride and 0.0 15M sodium citrate, pH = 7.0) at 65#{176}Cfor 1 5 mm andthen centrifuged. and the supernatants are spottedonto ZetaBind filters (CUNO Laboratory Products,Meniden, CN) by using a hybnidot manifold (BethesdaResearch Laboratories, Inc. , Gaithersburg. MD). The

blots are warmed overnight at 42#{176}Cand are thenprehybnidized for at least 1 6 h in fluid containing 0.5M phosphate buffer (pH = 7.2), 1 50 �g/mL of poly

(A), 1 50 ,�g/mL of salmon sperm DNA, 7% sodiumdodecyl sulfate, 1 % BSA, and 1 mM EDTA at 42#{176}C.The blots are then hybridized with 32P-labeled cDNAprobes (SA, 0.5 x iO� to 1 .0 x iO� cpm4tg) for 16 hat 45#{176}C.The cDNA probes used are a 0.85-kilobase(kb) XhoI fragment containing the promoter-proxi-mal exon encoding murine a2(I) procollagen from pAZ1002 (14), a 0.66-kb PstI fragment encoding murine

�(IV) procollagen from pcI5 (15), and a 1 .3-kb PstIfragment encoding rat glyceraldehyde phosphate de-hydrogenase (GAPDH) from pBR322 (16). The probes

are labeled with (32PJdCTP (>3,000 Ci/mmol) by ran-dom hexamer priming (17). The filters are washed 15

mm twice in 3x SCC-0.5% socium dodecyl sulfate at55#{176}Cand then autoradiographed (Kodak XAR-5 film)with intensifying screens at -70#{176}Cfor 1 to 3 days(GAPDH) or 10 to 20 days (procollagens). Relative

mRNA levels are quantitated by densitometnic scan-ning (Hoefer Instruments, San Carlos, CA) followed

by area integration. Northern blot analysis demon-strated that MCT cells express the mature mRNAmessage for the cDNA probes used in this study.

Measurement of Cellular myo-lnositol Content

myo-Inositol levels are measured by a previouslyreported method (8). MCT cells are cultured as de-scribed for dot-blot mRNA experiments. After growthfor 48 or 72 h, cell monolayers are washed threetimes with PBS and then the cells are lysed with ice-cold deionized water. A known amount of alpha-

methyl mannose pyranoside is added to the flask

before lysis as an internal standard for subsequentchromatographic assays. Cell lysates are centrifugedat 4#{176}C.and the supennatants are collected and storedat -20#{176}Cfor polyol measurements. Barium zinc fil-trates of lysed cells are prepared by the method ofSomogyi (1 8) and lyophilized. The trimethylsilyl de-nivatives are extracted with hexane and injected onpacked columns of 3% SE-30 on 80/100 GasChromQ (Supelco) in a Hewlett-Packard 5890 gas chromat-ograph by using flame ionization detectors and amodel 3392 integrator. Levels of myo-inositol are

expressed in milligrams of cellular protein as deter-

mined by a modification of the Lowry assay.

Statistics

For each individual experiment, the mean of repli-cate determinations is calculated. The data are pre-sented as means ± SE with N indicating the number

of different experiments. Statistical analysis be-tween two groups is performed by using paired orunpaired t test, as appropriate. Differences amongseveral groups are tested by using analysis of van-ance. A value of P < 0.05 is considered to representa significant difference.

RESULTS

[3H�thymidine Incorporation

Our previous studies showed that [3Hjthymidmneincorporation into DNA was significantly inhibitedwhen medium glucose was elevated, and inhibitionwas most pronounced after a period of 48 or 72 h ofexposing the cells to 450 mg/dL of glucose (3). In thecurrent studies, the response to changes in myo-inositol content of the media was also assessed. Inexperiments not shown, we first demonstrated thatmyo-mnositol supplementation increased thymidine

A 48 hours

*

60’

50

�30.

a- 20C)

10

0-Glucose

mg/dl

H100 450

B 72 hours

30

25

�20

�15

�10

5

0Glucosemg/dI

100 450

Ziyadeh et al

Journal of the American Society of Nephrology I223

incorporation only in media containing 450 mg/dL ofglucose. This effect was dose dependent in the rangebetween 40 and 1 ,200 �zM. The effect at 800 �zM wasmaximal with no further significant stimulation athigher myo-inositol levels. This concentration was

therefore used in subsequent studies.The results shown in Figure 1 summarize the find-

ings in two different sets of experiments. Whetherthe test period is for 48 (Figure 1 A) or 72 h (Figure1B), the supplementation of the media with 800 �zM

myo-inositol leads to a restoration of �:3HJthymidifleincorporation in 450 mg/dL of glucose to values seenin 1 00 mg/dL of glucose media. Note that supplemen-

tation with myo-inositol of the media containing 100mg/dL of glucose does not change cell proliferationsignificantly.

Procollagen Secretion

As seen in Figure 2, secretion of types IV and Iprocollagen is stimulated by 450 mg/dL of glucosemedia as we have previously found (3). When 800 �zM

myo-inositol is added to the media, there is a reduc-tion in the high-glucose-induced stimulation of pro-collagen secretion. On the other hand, procollagensecretion in cells exposed to low-glucose medium isnot significantly altered by myo-inositol supplemen-tation of the growth medium.

Type I Procollagen and Type IV ProcollagenmRNA Isolation and Measurement

We have previously shown by Northern hybridiza-

tion that the mature mRNA for type I procollagen andtype IV procollagen is expressed in MCT cells (unpub-lished observations) and that the effect of 450 mg/

dL of glucose media to increase mRNA levels for typeI procollagen and type IV procollagen is at least inpart expressed at the level of gene transcription (3).The latter was concluded on the basis of the resultsof nuclear run-off assays which measure the rate ofgene transcription. In the study presented here, we

examined whether the effects of myo-inositol supple-mentation on procollagen secretion represented a

change in message expression. To assess this, wemeasured steady-state mRNA levels. As shown inFigure 3, 450 mg/dL of glucose led to an increase insteady-state mRNA levels for type IV procollagen and

type I procollagen, consistent with our previous find-ings (3). Again, myo-inositol supplementation re-

duced the levels of mRNA which are expressed in thecells exposed to the high glucose media. The resultswere similar in each of three different experiments(Figure 3). This response suggests an action of myo-inositol either on gene transcription or on the stabi-lization of mRNA levels. It is clear, however, that themyo-inositol-induced reduction in the increased

Figure 1. Effect of myo-inositol supplementation on prolifer-ation of proximal tubule cells in culture. Tritiated thymidineincorporation (mean ± SE) was measured in two differentsets of experiments, one carried for 48 h (N = 10, panel A)and the other for 72 h (N = 9, panel B) in cells exposed toserum-free DMEM containing either 100 or 450 mg/dL ofglucose. The basal myo�inositoI concentration in the me-dium was 40 pM (open bars). In paired experiments, thiswas increased to 800 pM (hatched bars). At either timeperiod, thymidine incorporation was significantly less inhigh- versus low-glucose when medium mya-inositol was 40pM (P < 0.05). When medium myo�inositoI was 800 pM,there was no significant difference in incorporation in cells

grown in high- versus low-glucose media.

*

0

,(zE0>

Cs0

Glucose(mg/dl)

I

100

a(lV) a (I)1 2

100 450 100 450

Glucose myo-lnositol

(mg/dI) (jiM)

100 40 #{149}#{149}#{149};�100 800 #{149}.�.#{149} �.

A

100�

80#{149}C0

��60

� ‘�s 40

� 20

0�

Glucose(mg/dl)

B

�s .�.200

�

� %�0�- as

450

C

300�

250�

100 450

Proximal Tubule Cell Effects of myo-lnositol

1224 Volume I ‘ Number I I ‘ 1991

Glucose(mg/dI)

Figure 2. Effect of myo-inositol supplementation on procol-lagen secretion in proximal tubule cells. Cells were culturedfor 48 or 72 h in media containing 100 or 450 mg/dL ofglucose. Radioimmunoassay of procollagens was per-formed on conditioned supernatants, and data (mean ±

SE) were expressed per 10� cells. When medium myo�ino-sitol was 40 �M (open bars), procollagen secretion wassignificantly higher in the cells exposed to the high- versusthe low-glucose concentration (‘P < 0.05 for type IV pro-collagen, N = 7, panel A; ‘P < 0.05 for type I procollagen,N= 10, panel B). When medium myo-inositol was increased

to 800 MM (shaded bars), procollagen secretion in the cellsexposed to the high-glucose concentration was signifi-cantly reduced, to levels which were not different from

those of cells exposed to the low-glucose concentration.

450 40 #{149}#{149}S.

450 800 � ...

a1(IV) a2(I)

Figure 3. Effect of myo�inositol supplementation of thegrowth medium on steady-state mRNA levels of procolla-gens in proximal tubule cells. Cells were cultured for 72 hin media containing 100 or 450 mg/dL of glucose; the myo-inositol concentration in the medium was 40 �M (open bars)or 800 MM (shaded bars). Dot-blot RNA hybridization fol-lowed by densitometric scanning of autoradiographs weredone as described in Methods. The data are the means ofrelative mRNA level from three different experiments, eachperformed in two or more replicates. The level of procolla-gen message expression in the control cells (100 mg/dL ofglucose, 40 pM myo-inositol) was assigned a relative valueof unity. Note that mRNA levels for aj(IV) and a2(l) procol-lagens are augmented in high-glucose media when themyo-inositol concentration is 40 but not 800 �M. Autoradi-ographs of representative blots, in duplicate, are providedin the lower panel.

mRNA levels seen in the high-glucose media pro-vides, at least in part, for the concomitant reductionIn the glucose stimulation of procollagen secretion.Additional effects to modulate mRNA degradationwere not addressed in this study. Levels of GAPDH

mRNA were not altered by the concentrations ofglucose or myo-inositol in the growth medium (datanot shown).

Protein Synthesis

We previously determined that elevated glucosemedia stimulated protein synthesis in proximal tu-

* C

Glucose ioo 450mg/dl

Ziyadeh et al

Journal of the American Society of Nephrology 1225

bule cells (3). This, in part, may also account for theglucose-induced stimulation of procollagen synthesis

and secretion. In this study, we tested whether theeffect of myo-inositol supplementation to reduce glu-

cose-induced procollagen secretion was reflective of

a generalized effect on total cellular protein synthe-sis. The results shown in Figure 4 confirm our pre-vious studies that elevated glucose levels stimulate

ItHileucine incorporation (3). However, myo-inositolsupplementation has no effect on this response. Wepreviously demonstrated that raisingglucose concen-tration in the culture medium also increases the totalprotein content of proximal tubule cells (3). Alongwith the observed glucose-induced stimulation of leu-cine incorporation and the increase in cell size, asdemonstrated by flow cytometry (3), the increase in

total cellular protein content provided the evidencefor glucose-induced cellular hypertrophy. In this

study, we tested whether the hypentrophic responseto glucose could also be prevented by myo-inositolsupplementation. In nine different experiments, wefound that raising the medium glucose from 1 00 to

450 mg/dL resulted in a 29 ± 10% (P < 0.025) in-crease in total cell protein content after 72 h, whichconfirmed our previous findings (3). However, whenthe high-glucose medium was supplemented with800 �M myo-inositol. total protein content did notchange significantly (2 1 0 ± 23 �g/ 1 06 cells in 800�M myo-inositol versus 207 ± 29 ,ig/106 cells in 40

�M myo-inositol). Thus, the previously observed re-duction in procollagen secretion induced by myo-

inositol supplementation in the cells exposed to ele-

vated glucose is not associated with a myo-inositol-

.� U,

#{149}0.-JoI

Figure 4. Absence of significant effect of medium myo-inositol supplementation on cell protein synthetic rate inproximal tubule cells. Cells were cultured for 48 h in serum-free media containing 100 or 450 mg/dL of glucose; mya-inositol concentration was 40 pM (open bars) or 800 pM

(shaded bars). Cells exposed to the high-glucose concen-tration exhibited significantly increased (3H)leucine incor-poration compared with cells exposed to the low-glucoseconcentration (‘P < 0.05), irrespective of the level of myo-inositol in the medium.

induced inhibition in total protein synthesis or (‘ci-

lular protein content.

Measurement of myo-lnositol Levels

Raising medium glucose concentration from 1 00 to

450 mgjdL did not cause a decrease in MCT cellular

myo-inositol content (Table 1 ) and in fact producedan increase. However, myo-inositol supplementationof the growth medium led to an approximately two-fold increase in myo-inositol content, both underconditions of physiologic (1 00 mgjdL) or elevated (450mgjdL) medium glucose concentrations.

DISCUSSION

Tubulointerstitial damage and changes in the tu-bule cells of the diabetic kidney are prominent man-ifestations of the disease, but their significance andpathogenesis remain unclear ( 1 -3, 1 9). We have pre-

viously described a cell culture system using mouseproximal tubule epithelial cells where we tested the

effects of high ambient glucose levels on growth andextracellular matrix biosynthesis (3). We found thatraising media glucose concentrations to levelsachieved In vivo in uncontrolled diabetes mellitus

inhibited proximal tubule cell proliferation and in-creased cell size, total protein synthetic rate, and cellprotein content. In addition. the biosynthesis of tworepresentative procollagen components, type IV and

type I, was also stimulated. Glucose-induced tran-

scriptional control of procollagen genes was an im-portant mechanism underlying this response. Thesedata suggested that renal tubule epithelial cells couldplay a role in the formation of tubulointerstitial le-

sions in diabetic nephropathy.In the series of studies presented here, we have

extended the previous observations on the effects ofhigh glucose on MCT cells and have examined the

TABLE I . Effect of myo-inositol supplementation on

myo-inositol levels in proximal tubule cells#{176}

Glucose(mg/dL)

myo-Inositolin Media

(pM)

myo-InositolContent of Cells

(pg/mg ofprotein)

100 40 24.5±0.3100 800 42.1 ± 0.5450 40 47.9±0.5450 800 90.2 ± 0.4

a Data are mean ± SE of cellular myo-inositol levels from three different

experiments. each performed in duplicate. See Methods for details ofgas chromatographic analysis.

Proximal Tubule Cell Effects of myo-lnositol

1226 Volume I ‘ Number I I ‘ 1991

hypothesis that the action of glucose could be the

result of alterations in cellular myo-inositol metabo-lism, as has been reported to occur in other tissuesexposed to high glucose levels (4). We found that

elevating medium myo-inositol from 40 to 800 zM

raised cell myo-inosltol content and stimulated cellproliferation, reduced type I and type IV procollagen

secretion. and normalized steady-state levels ofmRNA for the two procollagen types. The effects onproliferation, procollagen secretion, and mRNA levelswere specific for 450 mg/dL of glucose medium; noeffect of myo-mnosltol supplementation was observedIn 1 00 mg/dL of glucose media.

It should be noted, however, that not all of the

effects of high glucose can be normalized by myo-

inositol supplementation. For instance, the glucose-induced stimulation of total protein synthesis and of

cellular protein content was not inhibited by myo-

inositol supplementation. This finding also providesfor two conclusions. First, the effect of myo-inositol

on procollagen transcription and secretion is not anonspecific effect. Second, an absence of an effectof myo-inositol on total protein synthesis and cellular

protein content may also imply that the cellular hy-pertrophy observed in the cells exposed to high glu-cose levels (3) is not reversed by myo-inositol supple-

mentatlon . Because myo-mnositol supplementationleads to a dissociation of glucose-induced cell hyper-

trophy from increased extracellular matrix synthe-sis, these results also suggest that different cellularmechanisms may underlie these two effects of high-

glucose medium on proximal tubule cells.Glucose-induced inhibition of cellular proliferation

has been previously described for cultured bovineretinal pericytes (20) and human endothelial cells(21). The underlying mechanism for glucose-inducedinhibition of DNA synthesis remains to be elucidatedbut may relate to alterations in cell-cycle traversalwhich have been previously described In cultured

endothelial cells (22) and which we have also con-

firmed in MCT cells (unpublished data). Proximaltubule cells (23) and endothelial cells (24) share theproperty that glucose uptake readily occurs without

the requirement for insulin. Thus, increased activityof the polyol pathway and subsequent abnormalitiesin myo-inositol metabolism could be an importantresponse to high-glucose medium in these tissues.Although this pathway has been well described inendothelial cells (25), the presence of this system inproximal tubule cells has not been described untilvery recently. We, with MCT cells (26), and others,with a cell line derived from human proximal tubule

cells (27). have shown the induction of sorbitol for-mation In these two cell systems when incubated inhigh-glucose media. Thus, the effect which we ob-

served with myo-inositol supplementation may rep-resent an action similar to that seen In endothelial

cells and nerve cells exposed to high glucose levels

where increased extracellular myo-Inosltol specifi-cally reverses a cellular defect which results from

polyol pathway activation in those tissues (4,28).The interaction between medIum glucose concen-

tratlon and myo-Inositol In modulating cell growth

and proliferation in other cell culture systems hasbeen evaluated by several groups with varying re-sults, depending on the cell type examined. Endothe-hal cells show an absolute requirement for at least1 6 �M myo-Inositol for proliferation and demonstratemaximum response at 160 MM myo-Inosltol (21). Inthat system. myo-inosltol supplementation wasfound to enhance cell proliferation in the presence ofelevated medium glucose but the response was not

complete and the inhibItory action of glucose per-sisted (2 1). Yorek et at. have also studied the effectsof high glucose on endothelial cells (29) and neuro-blastoma cells in culture (25). In neuroblastoma cells,the content of myo-inosltol was depressed after a 2-wk Incubation In high-glucose medium, primarily the

result of a noncompetitlve inhibition In the uptake of

myo-inosltol (25). In endothelial cells, there was alsoa reduction in cellular myo-Inositol levels after pro-longed exposure to elevated glucose levels but, in thatsystem. the inhibItion was competitive (29). However,

In these two cell systems. unlike those in other ne-ports noted above (2 1 .22), a clear inhibitory action ofglucose on cell number was not found. Thus, theresults we observed on cell proliferation and the re-sponse to myo-inositol supplementatIon In MCT cells

may be applicable to some, but not all, cell culturesystems.

A characteristic feature of the diabetic state Is theincreased thickness of basement membranes suchas those lining the retinal and glomerular capillariesas well as renal tubule epithella (29-34). We havepreviously shown that proximal tubule cells exposed

to high glucose levels secrete approximately twice asmuch procollagen type IV as do cells exposed to nor-mal glucose concentratIon, and this response couldbe accounted for, at least in part, by Increased blo-synthesis of procollagen molecules (3). Similar in-creases in collagen secretion have been described forretinal penicytes (20) and endothelial cells (35) cul-tuned in high-glucose medium. In this study, we dem-onstrate that myo-inosltol supplementation of ele-vated glucose media reduces the Increased procolla-gen mRNA levels and the peptide secretory rates.

The current study also demonstrated that myo-

Inositol supplementation reduced the glucose-in-duced stimulation of mRNA levels and secretion ratesof procollagen type I, an interstitial-type extracellulanmatrix component. Interstitial fibrosis is a prominentfeature of established diabetic nephropathy (1 ,2). On

the basis of studies of kidney biopsies from patientswith diabetIc nephropathy, there is a close correla-

Ziyadeh et al

Journal or the American Society of Nephrology 1227

tion between the degree of tubulointerstitial fibrosis

and mesangial expansion ( 1 ,2). Fibrogenesis in thekidney is complex and involves the interaction of

resident cells with local and systemic humoral fac-tons (19). The proximal tubule cell is a likely pantici-

pant in the fibrogenic response of diabetic renal dis-ease. Proximal tubule cells are derived from mesen-chyme (36) and retain the capability of producingfinite, although limited, amounts of interstitial-typematrix components ( 1 3, 1 9). Even small increases in

secretory rates of interstitial collagens elaborated by

each cell could eventually contribute to the cumula-tive fibrosis as proximal tubule cells constitute the

bulk of renal cortical volume.The current results suggest that the alterations in

procollagen secretion and cellular proliferation maybe lInked to a putative defect in myo-inositol metab-olism induced by elevated extracellular glucose. The

absence of an action of myo-inositol in 1 00 mg/dL ofglucose supports this hypothesis. These results, byanalogy with those obtained in other systems such

as diabetic nerve and vascular smooth muscle, sug-gest that altered myo-inositol metabolism is a centralmechanism whereby elevated glucose leads to a func-

tional disturbance in various cell types (4).The observation that elevated extracellular fluid

glucose concentrations may cause alterations in eel-

lular function correctable by myo-mnositol supple-

mentation of the medium, but without producing adepletion in the cellular myo-inositol content, Is notsurprising. Glomerular hyperfiltration reversible bydietary myo-inositol supplementation is demonstra-ble in streptozotocin-diabetic rats ( 1 2) weeks beforethere is any detectable decrease in tissue myo-inosi-

tol content (37). Furthermore, in intact preparationsof arterial wall isolated from normal rabbits, incu-

bations in medium containing elevated glucose con-centrations result in functional alterations which are

completely reversible by myo-inositol supplementa-tion of the medium and occur In the absence of anychange in composite tissue myo-inositol content (1 1).The mechanism appears to be depletion of a small,discrete poo1 of myo-inositol which is in equilibriumwith extracellular myo-inositol and which is requiredfor the synthesis of a specific pool of phosphatidyll-nositol, whose turnover regulates a distinct compo-nent of Na�,K�-ATPase activity, probably throughactivation of a specific protein kinase C ( 1 1 ). Al-though a global decrease in polyphosphomnositidecontent or metabolism has not been reported as part

of the acute tissue response to diabetes, secondaryincreases in polyphosphoinositide metabolism havebeen reported in diabetic nerve (38) and increasedresponsiveness to agonists whose signal-transduc-

tion mechanism is via a polyphosphoinositide hy-drolysis has been reported in human (39) and exper-imental diabetes (40).

The cellular mechanisms whereby the change inextracellular myo-inositol levels result in a change in

procollagen secretion and in procollagen mRNA 1ev-

els were not studied in these experiments. Important

changes, particularly in protein kinase C activity,have been a feature of several systems exposed to

high extracellular glucose levels, and such actionsmay also provide a link explaining the mechanism ofchanges in gene activation resulting from alteredglucose and myo-inositol levels (28,4 1).

In conclusion, these data extend our previous stud-ies demonstrating the effects of glucose to inducecellular hypentrophy, inhibit cell proliferation, and

stimulate procollagen types I and IV biosynthesis inproximal tubule cells. They also provide evidence for

a role of altered cellular myo-lnositol metabolism inthe response to high ambient glucose levels( 1 2,28,4 1 ), as myo-inositol supplementation corrects

some of these glucose-associated abnormalities(namely inhibition of cell prolIferation and increasedprocollagen gene transcripts and peptide secretion).

Because myo-inositol supplementation leads to a dis-soclatlon of glucose-induced cellular hypertrophyfrom increased extracellular matrix biosynthesis, ourfindings also suggest that distinct cellular mecha-nisms may underlie these two effects of high mediumglucose levels on proximal tubule cells. The mecha-

nisms of these effects and the potential implications

for the pathogenesis and therapy of diabetic neph-

ropathy remain to be understood.

ACKNOWLEDGMENTS

The authors acknowledge the Invaluable advice and support of Dr.

Albert Winegrad. These studies were supported by Research Grants

from Pfizer Inc. . the National Institutes of Health (grants DK39727

and DK39565). an NIH Training Grant (T32-AM070061, and a grantfrom the American Diabetes Association (Philadelphia. PA). Ms.

Melanie Watanabe provided expert technical assistance In the per-

formance of these studies.

REFERENCES

1 . Bader RH, Bader KE, Grund S, Markensen-Haen S. Christ H, Bohle A: Structure and func-tion of the kidney In diabetic glomenulosclerosis:Correlations between morphologic and func-tional parameters. Pathol Res Pract 1980;167:204-216.

2. Mauer SM, Steffes MW, Ellis EN, SutherlandDR. Brown DM, Goetz FC: Structural-functionalrelationships in diabetic nephropathy. J Clin In-vest 1984:74:1143-1155.

3. Ziyadeh FN, Snipes ER, Watanabe M, AlvarezR, Goldfarb S. Haverty TP: High glucose inducescell hypertrophy and stimulates collagen genetranscription in proximal tubule. Am J Physiol1990;259:F704-F7 14.

4. Winegrad A!: Banting lecture 1 986. Does a corn-mon mechanism induce the diverse complica-

Proximal Tubule Cell Effects of myo-lnositol

1228 Volume I ‘ Number I I ‘ 1991

tions of diabetes? Diabetes 1 987;36:396-406.5. Greene D, DeJesus PVJ, Winegrad A: Effects

of insulin and dietary myo-inositol on impairedmotor nerve conduction velocity in acute strep-tozotocin diabetes. J Clin Invest 1 975;55: 1326-1336.

6. MacGregor L, Matschinsky F: Altered retinalmetabolism in diabetes II. Measurement of(Na,K)-ATPase and total sodium and potassiumin individual retinal layers. J Biol Chem 1986;261:4052-4058.

7. Williamson JR, Kolo C: Granulation tissue: Anew model for studies of vascular complicationsof diabetes. Horm Metab Res Suppl 1 985; 15:27-31.

8. Simmons DA, Kern EFO, Winegrad Al, MartinDB: Basalphosphatidylinositol turnover controlsaortic Na /K� ATPase activity. J Clin Invest1986;77:503-513.

9. Simmons DA, Winegrad A!, Martin DB: Signif-icance of myo-inositol concentrations in meta-bolic regulation. Trans Assoc Am Physicians1 982;95:292-298.

1 0. Simmons DA, Kern EFO, Winegrad A!, MartinDB: Phos,phatidyllnositol turnover regulates an-tenial (Na -K”)-ATPase activity. Trans Assoc AmPhysicians 1 983;96: 10-18.

1 1 . Simmons DA, Wine�rad A!: Mechanism of glu-cose-induced (Na�, K )-ATPase Inhibition in aor-tic wall of rabbits. Diabetologia 1989:32:402-408.

1 2. Goldfarb S, Ziyadeh FN, Kern EFO, SimmonsD: Effects of polyol pathway inhibition and die-tary myo-inositol on glomerular hemodynamicfunction in experimental diabetes mellitus in therat. Diabetes 1991;40:465-471.

1 3. Haverty TP, Kelly CJ, Hines W, et at. : Charac-terization of a renal tubular epithelial cell linewhich secretes the autologous target antigen ofautoimmune experimental interstitial nephnitis.J Cell Biol 1988:107:1359-1368.

14. Liau G, Yamada Y, de Combrugghe B: Coordi-nate regulation of the levels of type III and type Icollagen mRNA in most but not all mouse fibro-blasts. J Biol Chem 1985;260:531-536.

15. Myers JC, Brinker JM, Kefalides NA, Rosen-bloom J, Wang SY, Guddas 14: Discriminationamong multiple AATAAA sequences correlateswith interspecies conservation of select 3’ un-translated nucleotides. Nucleic Acids Res 1986;14:4499-4517.

1 6. Fort P, Marty F, Piechaczyk M, et at.: Variousrat adult tissues express only one major mRNAspecies from glyceraldehyde-3 phosphate-dehy-drogenase multigenic family. Nucleic Acids Res1985; 13: 1431-1462.

1 7. Feinberg A, Vogelstein B: A technique for radi-olabeling DNA restriction endonuclease frag-ments to high specific activity. Anal Biochem1983:132:6-13.

1 8. Somogyi M: Determination of blood sugar. J BiolChem 1 945; 160:69-73.

1 9. Ziyadeh FN, Goldfarb S: The renal tubulointer-stitium in diabetes mellitus. Kidney Int1991 ;39:464-475.

20. Li W, Shen 5, Khatami M, Rockey J: Stimula-tion of retinal capillary penicyte protein and col-lagen synthesis in culture by high-glucose con-centration. Diabetes 1984:33:785-789.

21. Lorenzi M, Toledo S: Myo-inositol enhances the

proliferation of human endothelial cells in cul-tune but fails to prevent the delay induced byhigh glucose. Metabolism 1986:35:824-829.

22. Lorenzi M, Nordberg JA, Toledo S: High glucoseprolongs cell cycle traversal of cultured humanendothelial cells. Diabetes 1 987:36: 1 26 1 -1267.

23. Tang MJ, Suresh KR, Tannen RI: Carbohydratemetabolism by primary cultures of rabbit proxi-mal tubules. Am J Physiol 1989:256:C532-539.

24. Olgemoller B, Schwaabe S, Schleicher ED, Ger-bitz KD: Competitive inhibition by glucose ofmyo-inositol incorporation into cultured porcineaortic endothelial cells. Biochim Biophys Acta1 990; 1052:47-52.

25. Yorek MA, Dunlap JA: The effect of elevatedglucose levels on myo-inositol metabolism in cul-tuned bovine aortic endothelial cells. Metabolism1989:38:16-22.

26. Snipes ER, Watanabe M, Simmons DA, Gold-farb S. Ziyadeh FN: Polyol pathway (PP) activa-tion in high glucose (HG) stimulation of procol-lagen transcription and secretion in proximaltubule cells (PTC) lAbstracti. J Am Soc Nephrol1 990; 1:642.

27. Bylander JE, Sens DA: Elicitation of sorbitolaccumulation in cultured human proximal tu-bule cells by elevated glucose concentrations.Diabetes 1990:39:949-954.

28. Greene D, Lattimer S, Sima A: Sorbitol, phos-phoinositides and the sodium-potassium ATPasein the pathogenesis of diabetic complications. NEngl J Med 1987:316:599-606.

29. Yorek MA, Dunlap JA, Ginsberg BH: myo-Ino-sitol metabolism in 4 1 A3 neuroblastoma cells:Effects of high glucose and sorbitol levels. JNeurochem 1987;48:53-61.

30. Cohen MP, Surma M: Renal glomenular base-ment membrane. In vivo biosynthesis and turn-over in normal rats. J Biol Chem 1980;225:1767-1770.

3 1 . Fabre J, Balant L, Dayer P, Fox H, Vernet A:The kidney in maturity onset diabetes mellitus:A clinical study of 5 1 0 patIents. Kidney Int1982:21:730-738.

32. Gellmann D, Pirani C, Soothill J, Muehrcke R,Kark R: Diabetic nephropathy: A clinical andpathologic study based on renal biopsies. Medi-cine 1959:38:321-367.

33. Sternberg M, Cohen-Fortere L, Peyroux J: Con-nective tissue in diabetes mellitus. Biochemicalalterations of the intercellular matrix with spe-cial reference to proteoglycans, collagens andbasement membranes. Diabete Metab 1985;11:27-50.

34. Ziyadeh FN, Goldfarb S. Kern EFO: Diabeticnephropathy: Metabolic and biochemical mech-anisms. In: Brenner BM, Stein JH, eds. The Kid-ney in Diabetes Mellitus. Contemporary Issuesin Nephrology. Vol. 20. New York: Churchill Liv-ingstone; 1989:87-113.

35. Cagliero E, Maiello M, Boeri D, Roy S. LorenziM: Increased expression of basement membranecomponents in human endothelial cells culturedin high glucose. J Clin Invest 1988:82:735-738.

36. Ekblom P, Lehtonen E, Saren L, Timpi R: Shiftin collagen-type as an early response to induc-tion of the metanephric mesenchyme. J Cell Biol1981:89:276-283.

37. Cohen M: Aldose reductase, glomenular metab-olism, and diabetic nephropathy. Metabolism

Ziyadeh et al

Journal of the American Society of Nephrology 1229

1986;35(suppl 1):55-59.38. Berti-Mattera L, Peterson R, Bell M, Eichberg

J: The effects of hyperglycemia and its preven-tion by insulin treatment in the incorporation of132P1 into polyphophoinositides and other lipidsin peripheral nerve of the streptozotocin diabeticrat. J Neurochem 1985:45:1692-1699.

39. Drury PL, Smith GM, Ferriss JB: Increased va-sopressor responsiveness to angiotensin II intype I (insulin-dependent) diabetic patients with-out complication. Diabetologia 1984:27:174-

179.40. Scarborough NL, Carrier GO: Nifedipine and

alpha adrenoceptors in rat aorta. II. Role of cx-tracellular calcium in enhanced alpha-2 adre-noceptor mediated contraction in diabetes. JPharmacol Exp Ther 1984:231:603-609.

4 1 . Lee TS, MacGregor LC, Fluharty SJ, King GL:Differential regulation of protein kinase C and(Na,K)-adenosine tniphosphatase activities by ci-evated glucose levels in retinal capillary endo-thelial cells. J Clin Invest 1989:83:90-4.