Exp. Eye Res. (1996) 63, 649–659

Glucose Transporter 1 Expression is Enhanced During Corneal

Epithelial Wound Repair

HIROSHI TAKAHASHI, AUDREY E. KAMINSKI JAMES D. ZIESKE*

Schepens Eye Research Institute and the Department of Ophthalmology, Harvard Medical School,

Boston, MA, U.S.A.

(Received Columbia 11 March 1996 and accepted in revised form 22 May 1996)

Corneal epithelial wound healing involves a number of metabolically demanding processes such as cellmigration and proliferation. The energy for those processes is known to be provided by glycolysis. Thus,it was hypothesized that migrating epithelium would require high levels of glucose to provide a substratefor glycolysis. It is well established that glucose is transported into virtually all mammalian cells by thefacilitated glucose transport protein, glucose transporter (GLUT). We sought to investigate the expressionof one of the isoforms of glucose transporter, GLUT1, in corneal epithelium after epithelial debridementin the rat. Three-millimeter debridement wounds were made on central rat cornea and allowed to healfrom 1 hr to 21 days. To quantitate changes in GLUT1 mRNA and protein levels, whole cornealepithelium was harvested and analysed by reverse transcription-polymerase chain reaction, northernblot, and western blot analysis. GLUT1 protein localization was also observed immunohistochemically.Expression of GLUT1 protein rapidly increased following wounding and was 2±4-fold higher than controlat 4 hr post debridement. GLUT1 protein levels continued to increase even after epithelial wound closure(24 hr) and peaked at 2 days post debridement, 5±8-fold higher than control. The increase in GLUT1protein levels coincided with enhanced GLUT1 mRNA levels (3±7-fold higher than control at 4 hr postdebridement). Immunofluorescence microscopy showed increased binding of anti-GLUT1 concentrated inthe membranes of the basal cells from limbus-to-limbus until 24 hr post wounding. After 24 hr, bindingremained enhanced in the wound area, while binding to the limbal basal cells returned to the controllevel. In conclusion, the expression of GLUT1 mRNA and protein are rapidly enhanced after wounding.This may allow the increased transport of glucose, providing the metabolic energy necessary for cellmigration and proliferation. # 1996 Academic Press Limited

Key words : corneal epithelial wound healing; glucose transporter ; glucose transporter mRNA;glycosylation; cell migration.

1. Introduction

Corneal epithelium undergoes continuous cell turn-

over by replacing terminally differentiated apical cells

with younger cells that proliferate and migrate from

the limbal basal zone (Thoft and Friend, 1983). These

homeostatic events are highly metabolic and depend

on glucose as their main energy source. In the cornea,

glucose is supplied from aqueous humor or glycogen

stores in the epithelium (Friend, 1979). In addition, it

has been demonstrated that anaerobic glycolysis is the

principal metabolic pathway of glucose in the corneal

epithelium (Gottsch et al., 1986; Aguayo et al., 1988),

like epidermis (Johnson and Fusaro, 1972), and is

essential for cell migration (Kuwabara, Perkins and

Cogan, 1976). It has previously been reported that

α-enolase, a glycolytic enzyme, is highly expressed in

the limbal basal cells that are postulated to be stem

cells of corneal epithelium (Zieske, Bukusoglu and

Yankauckas, 1992). We have also reported the

enhanced expression of α-enolase during wound heal-

ing (Chung, Bukusoglu and Zieske, 1992), with levels

remaining elevated up to four weeks after wounding.

* For correspondence at : Schepens Eye Research Institute, 20Staniford Street, Boston, MA, 02114, U.S.A.

These results may reflect an important role of glycoly-

sis, not only in cell migration, but also in proliferation

and differentiation. If there is an increase in glycolytic

rate, it would require increased levels of glucose, the

substrate for glycolysis. This prompted us to in-

vestigate an essential protein required for movement

of glucose into the cell, glucose transporter (GLUT).

GLUT is a facilitated transport protein that delivers

glucose into virtually all mammalian cells. To date,

seven isoforms have been identified and named based

on the order of cloning (GLUT1–7) (Gould and

Holman, 1993). These isoforms vary in their tissue

specificity, affinity for glucose, and as to whether their

expression is inducible by insulin (Gould et al., 1993).

Among these isoforms, GLUT1 (50–55 kDa) is the

most abundant and is expressed at high levels in

erythrocytes and brain (Bell et al., 1993). GLUT1 is a

glycoprotein and has been shown to exhibit differing

amounts of N-linked glycosylation, giving rise to a

50–55 kDa form and a higher molecular weight form

of 60–65 kDa. The core protein is approximately

40 kDa. It is suggested that N-glycosylation is im-

portant for GLUT1’s transport activity (Feugeas et al.,

1990; Asano et al., 1991, 1993; Masumi, Akamatsu

and Kitagawa, 1993). Interestingly, in addition to its

ability to transport glucose, GLUT1 contains a water-

0014–4835}96}120649­11 $25.00}0 # 1996 Academic Press Limited

650 H. TAKAHASHI ET AL.

filled channel that spans the plasma membrane, which

may serve as a water channel (Fischbarg et al., 1990).

In ocular tissues, including corneal epithelium, several

investigators have demonstrated GLUT expression,

and GLUT1 is considered to be the principal isoform

(Harik et al., 1990; Gherzi et al., 1991; Kaulen et al.,

1991; Takata et al., 1991; Mantych, Hageman and

Devaskar, 1993; Kumagai, Glasgow and Pardridge,

1994; Takagi et al., 1994). The main interest in those

reports, however, is related to the endothelial barrier

function; there have been no reports about alterations

in GLUT1 expression during epithelial wound healing.

It is well established that several growth factors or

oncogenes stimulate GLUT1 mRNA transcription

(Birnbaum, Haspel and Rosen, 1987; Flier et al.,

1987; Rollins et al., 1988; Hiraki, Rosen and

Birnbaum, 1988; Kitagawa, Masumi and Akamatsu,

1991) and that GLUT1 protein expression is enhanced

when cells are starved for glucose (Haspel et al.,

1986). These findings may reflect the critical role of

GLUT1 in cells that are mitotically active and require

an increased level of glucose. Since wound healing

involves both proliferative and metabolically demand-

ing processes, we hypothesized that GLUT1 protein

levels would be enhanced after wounding and that

this increase would occur in association with increased

GLUT1 mRNA.

In the current study, we demonstrate alterations in

GLUT1 mRNA and protein expressions quantitatively,

and show GLUT1 protein localization immunohisto-

chemically during corneal wound healing in the rat.

2. Materials and Methods

Animal and Wound Models

Adult Sprague-Dawley rats were used, and all

procedures conformed to the ARVO Statement for the

Use of Animals. Rats were anesthetized by sub-

cutaneous injection of rodent anesthesia cocktail

containing ketamine, xylazine, and acepromazine,

with topical administration of proparacaine hydro-

chloride. Then, a central area was demarcated with a

3-mm trephine, and the epithelium within the area

was removed with a small scalpel, leaving an intact

basement membrane (Zieske and Gipson, 1986).

Wounds were allowed to heal in vivo. Rats were killed

with an overdose of sodium pentobarbital.

Isolation of Total RNA

At the time points of 1, 2, 4, and 24 hr after

wounding, rats were killed and whole corneal epi-

thelium from limbus-to-limbus was removed with a

small scalpel and immediately frozen in liquid ni-

trogen. Whole corneal epithelium scraped from un-

wounded rats was used as a control. Epithelium from

ten eyes was used for each time point. Total RNA was

isolated from samples by the acid guanidinium

thiocyanate-phenol-chloroform extraction method

(Chomczynski and Sacchi, 1987), using an RNA

Isolation Kit (Stratagene, La Jolla, CA, U.S.A.). The

time course experiments were performed four times.

Reverse Transcription-Polymerase Chain Reaction (RT-

PCR)

The complementary DNA (cDNA) was prepared

with 0±5 µg of total RNA in a final volume of 10 µl

using the Reverse Transcription System (Promega,

Madison, WI, U.S.A.). Reverse transcription was

performed at 42°C for 15 min in a mixture containing

10 m Tris–HCl, pH 8±8, 50 m KCl, 5 m MgCl#,

1 m each dNTP, 0±5 µg oligo (dT)"&

primer, and

15 U of AMV reverse transcriptase. Two microliters

of the product was amplified in a reaction mixture

(total volume 100 µl) containing 10 m Tris–HCl,

pH 8±3, 50 m KCl, 1±5 m MgCl#

and 0±2 m of

each dNTP using the PCR Core Kit (Boehringer

Mannheim, Indianapolis, IN, U.S.A.) with a final

concentration of 2±5 U Taq polymerase and 20 nM

of the specific primers for GLUT1 (Wadhwani et al.,

1993). Samples were denatured for 1±5 min. at 94°C,

followed by 25 PCR cycles of denaturation for 1±5 min

at 94°C, annealing 1 min at 56°C, and extension for

1 min at 72°C. The termination was performed at

72°C for 7 min. Twenty microliters of the PCR

fragment was then resolved on a 1±2% agarose gel

containing 0±5 µg ml−" ethidium bromide. The gel was

photographed and then analysed using scanning

laser densitometry (Molecular Dynamics, Model 300A,

Sunnyvale, CA). PCR amplifications using glyceralde-

hyde 3-phosphate dehydrogenease (G3PDH) primers

and β-actin primers were performed using the same

conditions described above, except for annealing

temperatures (Table I). The RT-PCR experiments were

repeated five times.

The specificity of the GLUT1 PCR product was

checked for the presence of a restriction site for Ava II

known to be present in the sequence of GLUT1 cDNA.

The restriction enzyme produced two fragments of

approximately 110 bp and 180 bp, verifying that the

PCR product corresponds to GLUT1 cDNA (data not

shown).

Northern Blot Analysis

Twenty micrograms of total RNA from each sample

was electrophoresed on a 1% agarose–formaldehyde

gel and transferred to a nylon membrane (GeneScreen

Plus, NEN-DuPont, Boston, MA, U.S.A.) by capillary

action. The cDNA probe for GLUT1 was prepared by

RT-PCR as described above. The products were

resolved on a 1% agarose gel, excised from the gel and

purified using the QIAquick Gel Extraction Kit (Qiagen

Inc., Chatsworth, CA, U.S.A.). The cDNA probe for β-

actin was also made from PCR products. cDNA probes

were labeled with $#P-dCTP by random priming

method and hybridized in buffer containing 5¬saline-

EXPRESSION OF GLUCOSE TRANSPORTER 1 651

T I

Oligonucleotide primer sequences

Primer Oligonucleotide sequenceAnnealing

temperatureFragment

size Reference

β-actin 5«-TTGTAACCAACTGGGACGATATGG-3«5«-GATCTTGATCTTCATGGTGCTAGG-3«

60°C 764 bp Clonetech Laboratories,Palo Alto, CA, U.S.A.

G3PDH 5«-ACCACAGTCCATGCCATCAC-3«5«-TCCACCACCCTGTTGCTGTA-3«

52°C 452 bp Clonetech Laboratories,Palo Alto, CA, U.S.A.

GLUT1 5«-GCCTGAGACCAGTTGAAAGCAC-3«5«-CTGCTTAGGTAAAGTTACAGGAG-3«

56°C 292 bp Wadhwani et al., 1993

sodium phosphate-ethylenediaminetetraacetic acid

(EDTA), 50% formaldehyde, 5¬Denhardt’s solution,

1% sodium dodecyl sulfate (SDS), 10% Dextran sulfate

and 100 µg ml−" salmon sperm DNA at 42°C over-

night. The membrane was washed twice with 2¬saline sodium citrate (SSC) for 15 min at room

temperature, twice with 2¬SSC-1% SDS for 15 min

at 60°C, and twice with 0±1% SSC-0±1% SDS for

15 min at 55°C. The membrane was placed against

KODAK Biomax MR film and allowed to expose

overnight. Band densities were compared by image

analysis software, NIH Image version 1.56b9. The

northern blot experiments were repeated four times.

Western Blot Analysis

At the time points of 2, 4, 6, 24 hr; 2, 7 and 14 days

after wounding, rats were sacrificed and whole corneal

epithelium was collected as described. The epithelium

removed to make the 3 mm wound was used as

control. Epithelium from five eyes were used for each

time point. Epithelial samples were solubilized in 1%

SDS buffer containing protease inhibitors (phenyl-

methylsulfonylfluoride—100 µg ml−",andaprotinin—

63 µg ml−"), and centrifuged at 15000 g for 20 min

at 4°C. The protein concentration of the resulting

supernatant was measured using a Bio-Rad Protein

Assay kit (Bio-Rad Laboratories, Hercules, CA, U.S.A.).

Each sample, containing 25 µg of total protein, was

then analysed by 10% SDS-polyacrylamide gel electro-

phoresis and transferred to nitrocellulose paper. The

resulting supernatant was then analyzed by 10%

SDS-polyacrylamide gel electrophoresis (PAGE) and

transferred to nitrocellulose paper. After incubation

with a blocking solution containing 5% non-fat dry

milk in Tris-buffered saline with 0±05% Tween-20, the

paper was incubated with a 1:5000 dilution of

rabbit anti-GLUT1 (rat brain) antibody (Chemicon,

Temecula, CA) for 45 min, and then incubated for

30 min with horseradish peroxidase labeled goat anti-

rabbit IgG at room temperature. Antibody binding

was detected by chemiluminescence detection using

LumiGLO Substrate Kit (KPL, Gaithersburg, MD,

U.S.A.). Band densities were compared by image

analysis software as described. The time course

experiments were repeated five times.

Digestion of Glycoprotein with Glycosidase F

Since GLUT1 protein is known to show different

molecular weights depending on its extent of glyco-

sylation (Feugeas et al., 1990; Asano et al., 1991,

1993; Masumi et al., 1993), digestion of the GLUT1

protein with N-glycosidase F was performed to remove

N-linked oligosaccharides. At the 24 hr time point,

epithelium was collected and homogenized in the

buffer containing 10 m EDTA, 0±75% NP-40, 10 m

β-mercaptoethanol, 10 m sodium phosphate,

pH 7±4. After centrifugation at 15000 g for 20

minutes, 25 µl of supernatant containing approxi-

mately 30 µg protein was incubated with 0±6 units of

N-glycosidase F (Boehringer Mannheim, Indianapolis,

MN, U.S.A.) for 18 hr at 37°C and then analysed by

SDS-PAGE and western blot. As a control, 25 µl of

supernatant was incubated without enzyme under the

same conditions.

Immunohistochemistry

At the same time points as those used for western

blot analysis, as well as 21 days post debridement, rats

were killed. Whole eyes were enucleated and frozen in

Tissue Tek II OCT compound (Miles Inc., Elkhart, IN,

U.S.A.). At least three eyes were examined for each

time point. Cryostat sections (6 µm) were placed on

gelatin-coated slides, air dried overnight at 37°C,

rehydrated in phosphate-buffered saline (PBS) for

10 min, and blocked in 1% bovine serum albumin

(BSA) for 10 min at room temperature. The primary

antibody, anti-GLUT1, was applied at a 1:500

dilution and incubated for 1 hr at room temperature.

Then, fluorescein isothiocyanate (FITC)-conjugated

donkey anti-rabbit IgG (Jackson Immunoresearch,

West Grove, PA, U.S.A.), as a secondary antibody, was

applied and incubated for 1 hr at room temperature.

Negative controls were prepared by omitting the

primary antibody. The sections were photographed

with a Zeiss Axiophot microscope (Carl Zeiss, Thorn-

wood, NY, U.S.A.) equipped for epi-illumination.

652 H. TAKAHASHI ET AL.

Antibody binding was examined in the limbus, at the

leading edge of migrating epithelium, in the zone

between the limbus and the original debridement

(designated periphery), and following wound closure

in the epithelium that had migrated to cover the

debridement area.

3. Results

RT-PCR and Northern Blot Analysis

In an initial screening to determine if GLUT1 mRNA

levels were altered in response to an epithelial

debridement, a time-course analysis of the expression

of GLUT1 mRNA was performed using RT-PCR. The

PCR analysis revealed a rapid increase in the ex-

pression of GLUT1 mRNA after wounding. As seen in

Fig. 1, RT-PCR using primers specific for GLUT1

resulted in a PCR product of the expected size (292 bp).

To allow for semiquantitation of the relative GLUT1

mRNA levels, the GLUT1 PCR product levels were

compared to those of G3PDH. The GLUT1}G3PDH

ratios at 1, 2, 4 and 24 hr time points compared to

control were 3±7, 5±2, 9±2 and 9±8. In order to more

accurately quantify the GLUT1 mRNA levels, northern

blot analysis was performed. This analysis revealed a

2±8-kbp mRNA transcript in agreement with the size of

GLUT1 mRNA reported previously (Mueckler et al.,

1985) (Fig. 2). Northern blot analysis indicated that

GLUT1 mRNA levels increased following wounding,

showing highest levels at the four-hour time point.

When β-actin was used to standardize mRNA levels,

GLUT1 mRNA levels were found to be enhanced 1±2-,

1±7-, 3±7- and 3±1-fold compared to control (Fig. 3).

Western Blot Analysis and Glycosidase F Digestion

To determine if GLUT1 protein levels increased after

wounding, coinciding with the increased mRNA levels,

a time course analysis of the expression of GLUT1

protein was performed using western blotting (Fig. 4).

Although anti-Glut1 reacted primarily with a single

F. 1. RT-PCR analysis of RNA isolated from unwounded corneal epithelium (Con) and epithelium 1, 2, 4 and 24 hr afterdebridement. Arrows indicate GLUT1 and G3PDH PCR products. The PCR products are visualized on agarose gels stained withethidium bromide. Standard markers (Std) are shown in right lane.

Con 1 h 2 h 4 h 24 h

28 s

18 s

GLUT1

Actin

F. 2. Northern blot of control (Con) corneal epitheliumand epithelium harvested 1, 2, 4 and 24 hr after a 3-mmcentral corneal debridement. The cDNA probe for GLUT1detected a 2±8 kbp mRNA transcript. Molecular masses(kbp) are noted. Northern blot of β-actin is shown below.

band of approximately 55 kDa, a band of approxi-

mately 60 kDa was also present in some samples. The

higher molecular weight band was most obvious at

24 hr. Enhanced expression of GLUT1 protein was

seen first at 4 hr post wounding, and increased up to

24 hr, the time of epithelial wound closure. Inter-

estingly, the levels of GLUT1 remained elevated and

actually increased following wound closure, at two

and seven days post wounding. Levels decreased at 14

days but were still higher than control. Band densities

were 1.1-, 2±4-, 4±1-, 5±4-, 5±8-, 5±6- and 5±2-fold

higher than control at 2, 4, 6, and 24 hr, 2, 7, and 14

days post debridement, in a representative experiment

(Fig. 5).

Since anti-GLUT1 showed a positive reaction with

two bands in western blotting experiments, the

samples were treated with glycosidase F to determine

if the bands represented alternate glycosylation forms

of GLUT1. As seen in Fig. 6, anti-GLUT1 reacted with

two bands of 55 kDa and 60 kDa in epithelium

harvested 24 hr after wounding. Following digestion

EXPRESSION OF GLUCOSE TRANSPORTER 1 653

4

024 hr

Time post-debridement

Rel

ativ

e G

LU

T1

mR

NA

leve

ls 3

2

1

4 hr2 hr1 hrControl

F. 3. Quantitation of GLUT1 mRNA levels shown in Fig.2. GLUT1 levels were determined, using image analysis asdescribed in Materials and Methods, and compared to thelevels of β-actin. The results are expressed as -fold en-hancement compared to control, which was given anarbitrary value of one.

Con 6 h2 h 4 h 24 h 2 d 7 d 14 d

112

84

53

35

F. 4. Western blot of control (Con) corneal epitheliumand epithelium harvested 2, 4, 6, and 24 hr; 2, 7, and 14days after a 3 mm central corneal debridement. The blot wasreacted with anti-GLUT1 as described in Materials andMethods. Molecular masses (in kilodaltons) determined fromstandard proteins are noted. The figure shows a rep-resentative experiment. Note : two bands shown at 24 hrtime point.

of the epithelial proteins with glycosidase F, only a

single band of 40 kDa was observed. The results

suggest that the 60-kDa band is an alternate glyco-

sylation form of GLUT1.

Immunohistochemistry

To determine the cellular localization of GLUT1,

indirect immunofluorescence, using anti-GLUT1, was

6

0

Con

trol

Time post-debridement

Rel

ativ

e G

LU

T1

prot

ein

leve

ls

1

5

4

3

2

2 h

r

4 h

r

6 h

r

24 h

r

2 da

ys

7 da

ys

14 d

ays

F. 5. Quantitation of relative GLUT1 concentrations incontrol and regenerating corneal epithelium (shown in Fig.4) determined as described in Materials and Methods. Valuesare compared to control, which was given an arbitrary valueof one.

+ –

112

84

53

35

F. 6. Western blot of epithelium harvested 24 hr afterdebridement, incubated with glycosidase F (­) or incubatedunder the same conditions without the enzyme (®) asdescribed in Materials and Methods. The blot was reactedwith anti-GLUT1. Molecular masses (in kilodaltons) de-termined from standard proteins are noted. The faint band ofapproximately 40 kDa in the (®) lane may indicate either abreakdown product or deglycosylated GLUT1. Proteaseinhibitors were not added to the sample, as they interferedwith the action of glycosidase F.

carried out on unwounded corneas as well as corneas

2, 4, 6, and 24 hr; 2, 7, 14, and 21 days post

debridement. The characteristic pattern of anti-GLUT1

binding to unwounded corneal epithelium showed

654 H. TAKAHASHI ET AL.

weak binding in the limbal basal cells [Fig. 7(A)], little

or no binding in central corneal epithelium [Fig. 8(A)],

and an occasional weakly binding cell in the peripheral

cornea [Fig. 9(A)]. At 2 hr after debridement, a slight

increase in GLUT1 was detected in cells at the leading

edge of migrating epithelium [Fig. 8(B)], while no

change in binding was seen in the other areas of the

corneal epithelium or in the limbus (not shown). At

4 hr [Fig. 8(C)] and 6 hr [Fig. 8(D)], the binding

intensity increased at the leading edge. In addition, the

expression of GLUT1 was greatly enhanced in the

limbal region [Fig. 7(B)] and peripheral cornea [Fig.

9(B)]. At 24 hr, the wound was closed. Binding to the

basal cells in the limbus remained elevated but

appeared to decrease compared to binding at 4 and

6 hr [Fig. 7(C)]. In central cornea, the binding

remained intense and began to concentration to the

basal cell layer [Fig. 8(E)]. At 2 days, binding to the

central cornea was much more intense than at

previous time points and was significantly concen-

trated to the basal cell layer [Fig. 8(F)]. In contrast,

binding in peripheral corneal epithelium was present

in most cell layers [Fig. 9(C)]. At the limbus, binding

decreased remarkably compared to that at 4 and 24 hr

[Fig. 7(D)]. At seven days, binding to central cornea

was similar to that at two days; binding to limbal basal

cells [Fig. 7(E)], and peripheral cells [Fig. 9(D)]

returned to the unwounded level. At 14 days, the

expression of GLUT1 was still elevated in central

cornea [Fig. 8(G)], but the binding was no longer

concentrated to the basal cell layer. At 21 days, the

expression of GLUT1 returned to levels present in

unwounded corneas [Fig. 8(H)]. Little or no binding

was observed when the primary antibody was omitted

[Fig. 7(F)].

4. Discussion

Corneal wound healing involves a number of

metabolically demanding processes including migra-

tion of the epithelium to close the wound, cell

proliferation allowing repopulation of the wound area,

synthesis of extracellular matrix, and reassembly of

adhesion structures (Khodadoust et al., 1968; Pfister,

1975; Buck, 1979; Gipson and Kiorpes, 1982; Jester,

Rodrigues and Sun, 1985; Zieske et al., 1986;

Cotsarelis et al., 1989; Zieske, Bukusoglu and Gipson,

1989). The primary source of energy for these

processes appears to be glycolysis, based on exper-

imental data showing that inhibition of glycolysis

blocks migration (Kuwabara et al., 1976). Interest-

ingly, Kuwabara et al. (1976) found that epithelial

glycogen stores were dissipated in the migrating

epithelium within 3 hr after wounding, suggesting

that an alternate source of glucose was present. In the

current study, it was investigated (1) whether levels of

GLUT1 protein were enhanced during the wound

repair process, which potentially would allow the

enhanced transport of glucose into the cell to meet the

metabolic demand, and (2) whether the expression of

GLUT1 protein was associated with the increased

expression of GLUT1 mRNA. The results of the studies

clearly demonstrate that GLUT1 protein levels are

dramatically increased in association with mRNA

levels, after wounding. Enhanced expression of GLUT1

mRNA was observed at 1 hr—by RT-PCR and at

2 hr—by northern blot, followed by upregulated

expression of GLUT1 protein at 4 hr—by western blot.

Immunohistochemistry suggested that GLUT1 protein

levels increase at the leading edge of migrating

epithelium as early as two hours after wounding. Cell

migration is believed to begin 2–4 hr after wounding

(Pfister, 1975). Thus, the initiation of migration along

with its metabolic demands correlates well with the

increase in GLUT1 mRNA and protein levels. This

rapid induction of GLUT1 mRNA after wounding may

be stimulated by growth factors, as GLUT1 mRNA

expression is known to be induced rapidly by several

growth factors including epidermal growth factor,

fibroblast growth factor, and platelet-derived growth

factor (PDGF) (Hiraki et al., 1988). Interestingly,

PDGF has been shown not only to upregulate GLUT1

mRNA but also to stabilize the mRNA (Rollins et al.,

1988). In 3T3 fibroblasts, the addition of PDGF

increased the half-life of GLUT1 mRNA by up to 4-fold.

One of the unexpected results of the study was that

GLUT1 protein levels continued to increase after

epithelial wound closure (24 hr in the rat debridement

model), did not begin to decrease until 14 days after

wounding, and did not return to control levels until 21

days post wounding. One interpretation of this data is

that GLUT1 protein provides the glucose necessary for

epithelial migration and then is involved in a separate

process following wound closure. It has previously

been observed that α-enolase levels remain elevated

for as long as 4 weeks after wounding (Chung et al.,

1995), suggesting that glycolytic rates may be

elevated for a considerable time after the initial wound

closure. An alternate possibility is that GLUT1 protein

is acting as a water channel and that the prolonged

expression may be necessary for the cornea to regain

its normal homeostatic volume. This possibility is

supported by the finding that tight junction refor-

mation requires several days following corneal debride-

ments (McCartney and Cantu-Crouch, 1992). This

transport of water has been suggested to occur

independently of glucose transport (Fischbarg et al.,

1990).

Immunolocalization of GLUT1 protein showed in-

tense binding at the leading edge of migrating

epithelium from 4 hr post wounding to 24 hr, the time

of epithelial wound closure. The binding gradually

diminished in the areas peripheral to the original

debridement, while limbal basal cells, where the

corneal epithelial stem cells are believed to exist

(Zieske, 1994), showed enhanced binding up to 24 hr

after wounding. This may indicate the increased

EXPRESSION OF GLUCOSE TRANSPORTER 1 655

F. 7. Immunolocalization of GLUT1 in the limbus of unwounded eyes (A), and 4 hr, (B), 24 hr (C), 2 days (D), and 7 days(E) following a 3-mm central corneal epithelial debridement. Note, intense binding to the endothelial cell layer can be seen in(B) and (D). (F) Specificity control showing absence of binding when primary antibody is omitted. Central cornea extends tothe right in all micrographs. Bar¯50 µm.

energy demand for cell proliferation in limbal stem

cells. The binding to limbal basal cells, however,

decreased remarkably at 2 days and returned to

control levels at 7 days post wounding. One possible

interpretation of this result is that the stem cells

undergo mitosis in order to supply daughter cells

repopulating the wound area, but then exit the cell

cycle after one round of proliferation. In contrast,

central cornea showed very intense binding at 2 and

7 days after wounding, in agreement with the western

blot data, which showed the highest levels at these

time points. The area of intense localization corre-

sponded to the original 3-mm debridement zone,

suggesting a vigorous energy demand for cell pro-

liferation and remodeling of the tissue. Immuno-

localization also revealed that GLUT1 protein was

concentrated in the basal cells at these time points, in

contrast to earlier time points, where GLUT1 protein

was present in several cell layers. Studies by others

(Gherzi et al., 1992) have shown that GLUT1 protein

is concentrated in the basal cells of epidermal and

conjunctival epithelium. The change of GLUT1 protein

localization from all cell layers to primarily the basal

cell layer may represent a change in energy demands

needed for migration, which affects multiple cell layers,

to energy demands for cell proliferation, which occurs

almost exclusively in the basal cell layer. In support of

this hypothesis, GLUT1 protein localization in the

limbus was observed exclusively in the basal cells,

which would be expected to proliferate but not migrate

extensively.

Finally, it is intriguing that two GLUT1 protein

bands were detected, using western blotting, at the

24 hr time point. GLUT1 is known to show a different

656 H. TAKAHASHI ET AL.

F. 8. Immunolocalization of GLUT1 in the central cornea of unwounded eyes (A), at the leading edge of migratingepithelium 2 hr (B), 4 hr (C), 6 hr (D) post wounding and in the debridement zone 24 hr (E), 2 days (F), 14 days (G), and 21days (H) post debridement. Bar¯50 µm.

molecular weight depending on its N-linked glyco-

sylation. Although the regulation of this variable

glycosylation is poorly understood, the suggestion has

been made that N-glycosylation is essential for its

glucose transport activity (Feugeas et al., 1990) and

increases an affinity for glucose (Asano et al., 1991,

1993), indicating that GLUT1-mediated glucose up-

take is partly regulated by its glycosylation. Interest-

ingly, Masumi et al. (1993) reported that in Swiss 3T3

cells, 60–65 kDa GLUT1 protein was induced by

transforming growth factor-β1) (TGF-β1). Treatment

of membrane proteins with glycosidase F generated a

EXPRESSION OF GLUCOSE TRANSPORTER 1 657

F. 9. Immunolocalization of GLUT1 in the peripheral cornea of unwounded eyes (A), and 6 hr (B), 2 days (C), and 7 days(D) post debridement. Bar¯50 µm.

single band of 40-kDa GLUT1 protein. It was observed

that digestion of GLUT1 protein from 24 hr samples

with glycosidase F clearly produced a single band of

40 kDa agreeing with their findings. Northern blot

analysis showing a single band for GLUT1 mRNA

(2±8 kbp) at 24 hr time point also supports the

conclusion that this 60-kDa protein is GLUT1. The

possibility that the 60-kDa GLUT1 protein was induced

by TGF-β1 around the time of wound closure is

currently being investigated.

In summary, GLUT1 mRNA and protein levels were

dramatically enhanced in response to corneal wound-

ing. Surprisingly, the protein expression remained

elevated up to 2 weeks after epithelial wound closure.

Expression of GLUT1 protein was increased at times

and in locations that may reflect its involvement in the

migratory and proliferative phases of corneal wound

repair. These findings lead to the speculation that

there may be a link between alterations in enhanced

glucose transport and impaired wound healing seen in

clinical pathologies. This concept is supported by the

observation that wound healing is impaired in patients

with diabetes mellitus. Several investigations have

shown that GLUT1 expression is altered in this disease

(Mueckler, 1990; Pardridge, Triguero and Farrell,

1990; Estrada et al., 1994; Kumagai et al., 1994).

In the future, we plan to examine the effect of

growth factors (TGF-β1 in particular) on GLUT1 gene

and protein expression in corneal wound healing and

whether there is a relationships between GLUT1 levels

and healing rates in diabetic animals.

Acknowledgements

Supported by research grant R01-EY05665 from theNational Institutes of Health to JDZ and an ARVO}JapanNational Society for the Prevention of Blindness ResearchFellowship to HT.

References

Aguayo, J. B., McLennan, I. J., Graham, C. J. and Cheng,H. M. (1988). Dynamic monitoring of corneal carbo-hydrate metabolism using high-resolution deuteriumNMR spectroscopy. Exp. Eye Res. 47, 337–43.

Asano, T., Katagiri, H., Takata, K., Lin, J.-L., Ishihara, H.,Inukai, K., Tsukuda, K., Kikuchi, M., Hirano, H., Yazaki,Y. and Oka, Y. (1991). The role of N-glycosylation ofGLUT1 for glucose transport activity. J. Biol. Chem. 266,24632–6.

Asano, T., Takata, K., Katagiri, H., Ishihara, H., Inukai, K.,Anai, M., Hirano, H., Yazaki, Y. and Oka, Y. (1993).The role of N-glycosylation in the targeting and stabilityof GLUT1 glucose transporter. FEBS Lett. 324, 258–61.

Bell, G. I., Burant, C. F., Takeda, J. and Gould, G. W. (1993).Structure and function of mammalian facilitative sugartransporters. J. Biol. Chem. 268, 19161–4.

Birnbaum, M. J., Haspel, H. C. and Rosen, O. M. (1987).Transformation of rat fibroblasts by FSV rapidlyincreases glucose transporter gene transcription. Science235, 1495–8.

Buck, R. C. (1979). Cell migration in repair of mouse cornealepithelium. Invest. Ophthalmol. Vis Sci. 18, 767–84.

Chomczynski, P. and Sacchi, N. (1987). Single-step methodof RNA isolation by acid-guanidinium thiocyanate-phenol-chloroform extraction. Analyt. Biochem. 162,156–9.

Chung, E. H., Bukusoglu, G. and Zieske, J. D. (1992).

658 H. TAKAHASHI ET AL.

Localization of corneal epithelial stem cells in thedeveloping rat. Invest. Ophthalmol. Vis. Sci. 33, 2199–206.

Chung, E. H., DeGregorio, P. G., Wasson, M. and Zieske, J. D.(1995). Epithelial regeneration after limbus-to-limbusdebridement : Expression of alpha-enolase in stem andtransient amplifying cells. Invest. Ophthalmol. Vis. Sci.36, 1336–43.

Cotsarelis, G., Cheng, S, Z., Dong, G., Sun, T. T. and Lavker,R. M. (1989). Existence of slow-cycling limbal epithelialbasal cells that can be preferentially stimulated toproliferate : implications on epithelial stem cells. Cell 57,201–9.

Estrada, D. E., Elliot, E., Zinman, B., Poon, I., Liu, Z., Klip, A.and Daheman, D. (1994). Regulation of glucosetransport and expression of GLUT3 transporters inhuman circulating mononuclear cells : studies in cellsfrom insulin-dependent diabetic and nondiabetic indi-viduals. Metabol. Clin. Exper. 43, 591–8.

Feugeas, J. P., Neel, D., Pavia, A. A., Laham, A., Goussault,Y. and Derappe, C. (1990). Glycosylation of the humanerythrocyte glucose transporter is essential for glucosetransport activity. Biochim. Biophys. Acta. 1030, 60–4.

Fischbarg, J., Kuang, K. Y., Vera, J. C., Arant, S., Silverstein,S. C., Loike, J. and Rosen, O. M. (1990). Glucosetransporters serve as water channels. Proc. Natl. Acad.Sci. U.S.A. 87, 3244–7.

Flier, J. S., Mueckler, M. M., Usher, P. and Lodish, H. F.(1987). Elevated levels of glucose transport and trans-port messenger RNA are induced by ras or srconcogenes. Science 235, 1492–5.

Friend, J. (1979). Biochemistry of ocular surface epithelium.Int. Ophthalmol. Clin. 19, 73–91.

Gherzi, R., Melioli, G., De, L. M., D’Agostino, A., Guastella,M., Traverso, C. E., D’Anna, F., Franzi, A. T. andCancedda, R. (1991). High expression levels of the‘erythroid}brain’ type glucose transporter (GLUT1) inthe basal cells of human eye conjunctiva and oralmucosa reconstituted in culture. Exp. Cell Res. 195,230–6.

Gherzi, R., Melioli, G., de L. M., D’Agostino, A., Distefano, G.,Guastella, M., D’Anna, F., Franzi, A. T. and Cancedda,R. (1992). ‘HepG2}erythroid}brain’ type glucose trans-porter (GLUT1) is highly expressed in human epidermis :keratinocyte differentiation affects GLUT1 levels inreconstituted epidermis. J. Cell. Physiol. 150, 463–74.

Gipson, I. K. and Kiorpes, T. C. (1982). Epithelial sheetmovement : protein and glycoprotein synthesis. Dev.Biol. 92, 259–62.

Gottsch, J. D., Chen, C. H., Aguayo, J. B., Cousins, J. P.,Strahlman, E. R. and Stark, W. J. (1986). Glycolyticactivity in the human cornea monitored with nuclearmagnetic resonance spectroscopy. Arch. Opththalmol.104, 886–9.

Gould, G. W. and Holman, G. D. (1993). The glucosetransporter family : structure, function and tissue-specific expression. [Review]. Biochem. J. 295, 329–41.

Harik, S. I., Kalaria, R. N., Whitney, P. M., Andersson, L.,Lundahl, P., Ledbetter, S. R. and Perry, G. (1990).Glucose transporters are abundant in cells with ‘oc-cluding’ junctions at the blood-eye barriers. Proc. Natl.Acad. Sci. U.S.A. 87, 4261–4.

Haspel, H. C., Wilk, E. W., Birnbaum, M. J., Cushman, S. W.and Rosen, O. M. (1986). Glucose deprivation andhexose transporter polypeptides of murine fibroblasts. J.Biol. Chem. 261, 6778–89.

Hiraki, Y., Rosen, O. M. and Birnbaum, M. J. (1988). Growthfactors rapidly induce expression of the glucose trans-porter gene. J. Biol. Chem. 263, 13655–62.

Jester, J. V., Rodrigues, M. M. and Sun, T. T. (1985). Changein epithelial keratin expression during healing of rabbit

corneal wounds. Invest. Ophthalmol. Vis. Sci. 26,828–37.

Johnson, J. A. and Fusaro, R. M. (1972). The role of the skinin carbohydrate metabolism. [Review]. Adv. Metabol.Disord. 60, 1–55.

Kaulen, P., Kahle, G., Keller, K. and Wollensak, J. (1991).Autoradiographic mapping of the glucose transporterwith cytochalasin B in the mammalian eye. Invest.Ophthalmol. Vis. Sci. 32, 1903–11.

Khodadoust, A. A., Silverstein, A. M., Kenyon, D. R. andDowling, J. E. (1968). Adhesion of regenerating cornealepithelium. The role of basement membrane. Am. J.Ophthalmol. 65, 339–48.

Kitagawa, T., Masumi, A. and Akamatsu,, Y. (1991).Transforming growth factor-beta 1 stimulates glucoseuptake and the expression of glucose transporter mRNAin quiescent Swiss mouse 3T3 cells. J. Biol. Chem. 266,18066–71.

Kumagai, A. K., Glasgow, B. J. and Pardridge, W. M. (1994).GLUT1 glucose transporter expression in the diabeticand nondiabetic human eye. Invest. Ophthalmol. Vis. Sci.35, 2887–94.

Kuwabara, T., Perkins, D. G. and Cogan, D. G. (1976).Sliding of the epithelium in experimental cornealwounds. Invest. Ophthalmol. Vis. Sci. 15, 4–14.

Mantych, G. J., Hageman, G. S. and Devaskar, S. U. (1993).Characterization of glucose transporter isoforms in theadult and developing human eye. Endocrinology 133,600–7.

Masumi, A., Akamatsu, Y. and Kitagawa, T. (1993).Modulation of the synthesis and glycosylation of theglucose transporter protein by transforming growthfactor-beta 1 in Swiss 3T3 fibroblasts. Biochim. Biophys.Acta. 1145, 227–34.

McCartney, M. D. and Cantu-Crouch, D. (1992). Rabbitcorneal epithelial wound repair : tight junction re-formation. Curr. Eye. Res. 11, 15–24.

Mueckler, M., Caruso, C., Baldwin, S. A., Panico, M., Blench,I., Morris, H. R., Allard, W. J., Lienhard, G. E. andLodish, H. F. (1985). Sequence and structure of ahuman glucose transporter. Science 229, 941–5.

Mueckler, M. (1990). Family of glucose-transporter genes.Implications for glucose homeostasis and diabetes.Diabetes 39, 6–11.

Pardridge, W. M., Triguero, D., and Farrell, C. R. (1990).Downregulation of blood-brain barrier glucose trans-porter in experimental diabetes. Diabetes 39, 1040–4.

Pfister, R. R. (1975). The healing of corneal epithelialabrasions in the rabbit : a scanning electron microscopestudy. Invest. Ophthalmol. Vis. Sci. 14, 648–61.

Rollins, B. J., Morrison, E. D., Usher, P. and Flier, J. S.(1988). Platelet-derived growth factor regulates glucosetransporter expression. J. Biol. Chem. 263, 16523–6.

Takagi, H., Tanihara, H., Seino, Y. and Yoshimura, N.(1994). Characterization of glucose transporter incultured human retinal pigment epithelial cells : geneexpression and effect of growth factors. Invest.Ophthalmol. Vis. Sci. 35, 170–7.

Takata, K., Kasahara, T., Kasahara, M., Ezaki, O. andHirano, H. (1991). Ultracytochemical localization of theerythrocyte}HepG2-type glucose transporter (GLUT1)in the ciliary body and iris of the rat eye. Invest.Ophthalmol. Vis. Sci. 32, 1659–66.

Thoft, R. A. and Friend, J. (1983). The X, Y, Z hypothesis ofcorneal epithelial maintenance [letter]. Invest. Ophthal-mol. Vis. Sci. 24, 1442–3.

Wadhwani, K. C., Fukuyama, R., Giordano, T., Rapoport,S. I. and Chandrasekaran, K. (1993). Quantitativereverse transcriptase-polymerase chain reaction ofglucose transporter 1 mRNA levels in rat brain micro-vessels. Analyt. Biochem. 215, 134–41.

EXPRESSION OF GLUCOSE TRANSPORTER 1 659

Zieske, J. D. and Gipson, I. K. (1986). Protein synthesisduring corneal epithelial wound healing. Invest. Ophthal-mol. Vis. Sci. 27, 1–7.

Zieske, J. D., Bukusoglu, G. and Gipson, I. K. (1989).Enhancement of vinculin synthesis by migrating strati-fied squamous epithelium. J. Cell Biol. 109, 571–6.

Zieske, J. D., Bukusoglu, G. and Yankauckas, M. A. (1992).Characterization of a potential marker of cornealepithelial stem cells. Invest. Ophthalmol. Vis. Sci. 33,143–52.

Zieske, J. D. (1994). Perpetuation of stem cells in the eye[Review]. Eye 8, 163–9.