Investigative Ophthalmology & Visual Science, Vol. 30, No. 8, August 1989
Copyright © Association for Research in Vision and Ophthalmology
EGF Does Not Enhance Corneal Epithelial Cell Motility
H. Kaz Soong,* Brian McClenict James Varani,t Tarek Hassan,* Samuel C. M. Huang,* and Richard Brenz"
Although it is well known that epidermal growth factor (EGF) accelerates corneal epithelial wound
healing by stimulating mitosis, it is also believed that EGF may directly stimulate the motility of
individual corneal epithelial cells. We employed three different experimental methods to determine if
EGF does indeed enhance the motility of corneal epithelial cells (independent of mitotic effects). First,
the effects of EGF on the motility of tissue-cultured rat and rabbit corneal epithelial cells were
investigated by a Boyden chamber assay. In rat corneal epithelium, these effects were further investigated by a second method, the agarose drop assay. Both assay techniques demonstrated no increase in
corneal epithelial cell motility in the presence of EGF. These findings were corroborated by a third
method which consisted of measuring the closure rate of epithelial wounds in organ-cultured rat
corneas in the presence and absence of EGF, while concurrently arresting mitosis with colchicine. The
wound closure rate before addition of any drug was 0.46 ± 0.03 mm2/hr. The wound closure rate with
EGF (50 ng/ml) was 0.55 ± 0.03 mm2/hr, significantly {P < 0.005) more rapid than the drug-free
controls. However, when EGF (50 ng/ml) and colchicine (40 fig/m\) were used simultaneously, the
acceleration of wound closure by EGF was completely negated by the presence of colchicine, resulting
in a wound closure rate (0.46 ± 0.06 mm2/hr) that did not differ significantly (P > 0.50) from that of
the drug-free control. These results, together with the Boyden chamber and agarose drop assay
results, suggest that the acceleration of corneal epithelial wound healing by EGF is due primarily to
increased cell proliferation (with the increase in cell population causing the cells to spill over into the
epithelial defect). It appears that individual cell motility, independent of mitosis, is not primarily
responsible for accelerated wound healing in the presence of EGF. Invest Ophthalmol Vis Sci
30:1808-1812,1989
Almost immediately after corneal epithelial
wounding, the surrounding cells migrate into the
wound to cover the defect as rapidly as possible. This
first phase of wound healing is nonmitotic in nature
and consists entirely of cell migration. After a delay of
many hours, the second phase of healing, marked by
active mitosis, begins and serves to restore cellular
density.1"4 Failure of an epithelial defect to heal may
result in infection, enzymatic breakdown and permanent damage to the deeper structures.
Many drugs have been investigated for the purpose
of pharmacologically accelerating corneal epithelial
healing. Epidermal growth factor (EGF) has been reported to increase corneal epithelial wound closure
both in vivo and in vitro5"9 and to also stimulate
corneal epithelial cell proliferation and DNA synthesis.8 Although EGF is considered primarily a mitoFrom the Departments of "Ophthalmology and fPathology, the
University of Michigan Medical School, Ann Arbor, Michigan.
Supported by grants from the Michigan Eye Bank and Transplantation Center (HKS), Fight for Sight, Inc. (HKS), and the
American Cancer Society (grant PDT-324) (JV).
Submitted for publication: November 1, 1988; accepted February 23, 1989.
Reprint requests: H. Kaz Soong, MD, W. K. Kellogg Eye Center,
1000 Wall Street, Ann Arbor, MI 48105. •
genie agent, it also stimulates production of extracellular matrix components.10 In epidermal cells (keratinocytes), EGF not only increases mitosis, but also
appears to directly stimulate cell motility.'' Although
it has been commonly assumed that EGF may directly stimulate cell motility in corneal epithelial cells
as well,6'7'912 no studies have as yet specifically addressed and proven this.
We initially used two different assay methods
(Boyden chamber1314 and agarose drop14"16) to study
the effects of EGF on the motility of individual corneal epithelial cells in tissue culture. The advantage of
both assay methods is that the motility of individual
cells could be measured independently of mitotic influences. In addition, we used a third experimental
method using organ-cultured corneas with epithelial
wounds. In this method, the effects of EGF on corneal epithelial wound closure rates were studied in
the presence and absence of colchicine. Colchicine,
applied to the organ culture system in order to inhibit
mitosis,1718 eliminates any acceleration of wound
closure related to the mitogenic effects of EGF. Thus,
in a multilayered cellular environment devoid of any
background cellular proliferation, it is possible to determine if EGF is capable of directly stimulating corneal epithelial cell motility. The organ culture model
1808
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933379/ on 06/16/2017
No. 8
EGF AND CORNEAL EPITHELIUM / Soong er ol
permits population migration behavior to be studied,
while the Boyden chamber and agarose drop models
permits individual cell migration properties to be
studied.
Materials and Methods
Cell Cultures
Adult male Sprague-Dawley rats and New Zealand
albino rabbits were killed with an overdose of intraperitoneal (rat) or intravenous (rabbit) ketamine.
Maintenance and handling of the animals were performed in accordance with NIH guidelines and the
ARVO Resolution on the Use of Animals in Research. Full-thickness sheets of corneal epithelium
were removed using neutral bacterial protease (Dispase II, Sigma Chemical Co., St. Louis, MO).19 The
sheets and fragments of epithelium were explanted
onto plastic tissue culture dishes and the cells were
allowed to grow for 1 week into a monolayer at 37°C
and 5% CO2 in denned minimum essential medium
(DMEM) (Eagle's minimum essential medium with
non-essential amino acids, L-glutamine, penicillin,
streptomycin and amphotericin B) with 10% fetal calf
serum and with or without 10 ng/ml EGF (Sigma).
Transformed cells from a human oral (epithelial)
squamous cell carcinoma line (SCC-1) in tissue culture were obtained from Dr. Thomas Carey (Department of Otolaryngology, University of Michigan
Medical School, Ann Arbor, MI).
Organ Cultures
Thirty adult male Sprague-Dawley rats (59 eyes)
were killed with an overdose of intraperitoneal ketamine. A 4 mm diameter, central trephine mark was
made on the surface of each cornea and the epithelium within was gently removed using a blunt,
rounded-edge razor blade (cutting edge smoothly
dulled on a fine whetstone). Wounded corneoscleral
buttons were excised and mounted endothelialrsidedown onto dome-shaped paraffin platforms inside
culture wells. The corneas were cultured for 21 hr at
37°C and 5% CO2 in DMEM without serum, and
with or without drugs (colchicine and/or EGF).
Drugs
For the Boyden chamber assay, EGF was added to
the tissue culture medium at concentrations of 10 or
50 ng/ml. For the agarose drop assay, EGF was added
to the medium at concentrations of 5, 50 or 100
ng/ml. The controls were free of EGF.
The organ culture study was organized into two
parts. The first part consisted of a comparison of epithelial wound healing rates between three popula-
1809
tions: Group 1) eleven corneas cultured in the presence of mouse submaxillary gland EGF (Sigma
Chemicals, St. Louis, MO), alone, at a concentration
of 50 ng/ml; Group 2) seventeen corneas cultured
with both EGF (50 ng/ml) and colchicine (40 Mg/ml)
(Sigma Chemicals); and Group 3) ten corneas (serving as controls for Groups 1 and 2) cultured in drugfree medium. The second part compared wound
healing rates between ten corneas cultured in drugfree medium and 11 corneas cultured in colchicine
(40 Mg/ml) alone.
Assay of Motility in Modified Boyden Chamber
Modified blind-well Boyden chambers were used
to study random cellular migration based on the
method described by Romualdez and Ward.13 The
chambers consist of two compartments partitioned
by 12 /tm pore nitrocellulose filters (Schleicher and
Schuell, Inc., Keene, NH). Top and bottom compartments each received 200 (x\ of serum-free
DMEM. In some chambers, EGF (either 10 or 50
ng/ml) was added only to the bottom compartment,
while in other chambers, EGF (50 ng/ml) was added
to both the upper and lower compartments. For the
controls, no EGF was added to. either compartment.
Cells to be tested were seeded into the top compartment at concentrations of 3 X 105 (rat) and 5 X 105
(rabbit) cells per chamber. In order to provide a comparative positive control for migration, oral squamous carcinoma cells (SCC-1), which are known to
increase random migration in response to EGF, were
used in concurrent experiments with corneal epithelial cells. All chambers were incubated at 37°C in 5%
CO2 for 18 hr. The filters were removed and stained
with hematoxylin, dehydrated in propanol, cleared
with xylene and mounted onto glass slides. Cell migration was quantitatively read by light microscopy as
the number of cells per high-power field (X400) that
migrated into the pores of the filter. Readings were
performed in triplicate and compared statistically
with the student t-test.
Assay of Motility with Agarose Drop Explant
Method
The agarose drop explant method of assaying random cellular migration was a modification of the
methods described by Carpenter15 and by Harrington
and Stastny.16 Cells to be assayed were trypsined,
rinsed, and centrifuged into a pellet. The cells were
resuspended at a concentration of 1 X 107 cells/ml in
DMEM supplemented with 10% fetal calf serum and
0.2% agarose (Seaplaque Agarose, Marine Colloids,
Rockville, ME). Agarose-cell droplets of 5 fi\ size were
delivered by sterile micropipets and placed in the
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933379/ on 06/16/2017
1810
INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / August 1989
Table 1. Effects of EGF on cell migration in the
Boyden chamber assay
EGF concentration
in compartments
(ng/ml)
Cells
Rat
Rabbit
SCC-1
Number of migrated
cells per high-power
field (±SEM) at
filter depth of 10 nm
Top
Bottom
0
50
0
0
0
50
50
10
4± 1
2± 1
3± 1
0
50
0
0
0
50
50
10
2± 2
1± 1
3± 2
4± 2
0
50
10
0
50
10
38 ± 11
128 ±25
74 ± 15
5± 2
Boyden Chamber Studies
Wound Studies (Organ Culture System)
At the end of the 21 hr organ culture incubation
period, the cultured corneas were stained with RichTable 2. Effects of EGF on rat corneal epithelial cell
migration, compared with SCC-1 positive controls,
in agarose assay (data expressed as number of cells
migrated outside agarose droplet)
Number of migrated
cells (±SEM)
Concentration
ofEGF (ng/ml)
18 h
48h
Rat
0
100
50
5
<10
<10
<10
<10
<10
<10
<10
<10
SCC-1
0
100
50
5
13±
70 ±
107 ±
26 ±
5
13
10
2
ardson's stain (pH 7.0) to delineate the residual epithelial defects. Richardson's stain was used only at
the end of the incubation periods to avoid any toxic
effects on epithelial healing at the onset of wound
healing. The stained wounds were photographed
under a dissecting microscope and the photographically documented wound areas were measured with a
polar planimeter. The epithelial wound closure rates
were compared between controls, EGF alone and
EGF + colchicine, using the student t-test.
Results
center of the wells of a 96-well, microtiter tissue culture dish. The droplets were allowed to solidify at 4°C
for 10 min. The droplets were then covered with 200
n\ of DMEM supplemented with 200 /ig/ml of bovine
serum albumin with or without 5, 50 or 100 ng/ml
EGF. The microtiter plates were incubated at 37°C
with 5% CO2. Cell migration out of the agarose droplets was measured at 18 and 48 hr under an invertedphase microscope with a calibrated filar grid in the
eyepiece. The linear displacement of the migrating
cells from the agarose droplet was measured on four
sides of each droplet, as described previously.14 Measurements were performed in quadruplicate and
compared statistically with the student, t-test. SCC-1
cells were used as positive controls for migration.
Cells
Vol. 3 0
For both rat and rabbit corneal epithelial cells, no
significant migration (P > 0.05) was observed after 18
hr in EGF-stimulated cells over the EGF-free control
cells (Table 1). In contrast, SCC-1 cells migrated into
the filters in large numbers, even in the absence of
EGF (Table 1). The inherently higher motility is
characteristic of transformed cells. Use of a different
tissue culture medium (low-calcium, serum-free keratinocyte growth medium [KGM, Clonetics Corp.,
San Diego, CA]) did not increase the motility or
growth of the corneal epithelial cells.
Agarose Drop Explant Studies
For rat corneal epithelial cells, no significant migration (P > 0.05) was observed at 18 hr and 48 hr in
EGF-stimulated cells over EGF-free controls (Table
2). SCC-1 cells, in contrast, showed strikingly higher
numbers of migrated cells (Table 2). As in the case of
the Boyden chamber observations, SCC-1 cells had
inherently higher migration than the corneal epithelial cells, once again implying the more aggressive
nature of the transformed cells.
Organ Culture Wound Studies
In the first part of the organ culture studies (38
eyes), there was no significant difference (P > 0.50)
between wound healing rates of drug-free controls
(Group 3) and simultaneous EGF + colchicinetreated cases (Group 2) (Table 3). The wound healing
rate in the presence of EGF alone (Group 1) was
significantly faster (P < 0.005) than both the control
and EGF + colchicine cases (Groups 2 and 3).
In the second part of the organ culture studies (21
eyes), there was no significant difference (P > 0.50)
between wound healing rates of colchicine alone
versus drug-free medium.
Discussion
EGF was first isolated from mouse submaxillary
glands20 in 1962 and numerous subsequent studies
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933379/ on 06/16/2017
No. 8
1811
EGF AND CORNEAL EPITHELIUM / Soong er ol
have shown the ability of EGF to accelerate reepithelialization of the cornea.5"9'2122 Although the mechanism of acceleration of epithelial wound healing is
thought to be mainly by stimulation of cellular mitosis, it was widely believed that perhaps there was
also direct stimulation by EGF of cellular migration.6'7'912 In other cell types (eg, keratinocytes), EGF
has been found by both Boyden chamber and agarose
drop assays to enhance cell motility." A possible
mechanism for this may be that EGF stimulates synthesis of extracellular matrix components, such as
fibronectin, which has been reported to increase cellular migration and cell-substrate adhesion.610'23"25
For corneal epithelial cells, in contrast, we found that:
(1) EGF does not promote motility of the individual
cells; and (2) pharmacological blockage of mitosis
completely inhibits the accelerating effects of EGF on
epithelial wound closure. Recently, Watanabe et al
also reported an absence of chemotactic activity by
EGF on rabbit corneal epithelial cells,26 using a Boyden chamber assay modified according to the method
of Gauss-Muller. Thus, unlike the case of other cell
types, such as keratinocytes, it appears that acceleration of corneal epithelial wound healing by EGF is
entirely due to a stimulation of cellular mitosis and
not due to increased individual cell motility. The
ability of EGF to accelerate corneal epithelial wound
closure may, therefore, be due to more cells being
pushed or crowded into the defect area by a sheer
increase in cell numbers. It is possible that the net
increase in the matrix proteins due to stimulation by
EGF may not be of sufficient quantity to significantly
influence cell motility in the corneal epithelial cell.
The three migration assay methods used in this
study measure corneal epithelial cell migration under
several different conditions and systems. By combining the results of these related but distinct assay
methods, the reliability of the observations is enhanced. In the modified Boyden chamber and agarose drop assays, individual cell movement is measured, while in the organ culture assay, population
cell movement is measured. In the modified Boyden
chamber assay, cells migrate into the interstices of a
micropore filter and cellular migratory behavior depends, to a certain degree, on the ability of the cells to
deform and squeeze through the pores. The length of
the incubation period is relatively short (18 hr) and
the effects of cell division are negligible. As the cells
are fixed before the migrating cells are counted, this
method is not a dynamic assay in the strictest sense.
In the agarose droplet explant assay, cells migrate out
from a concentrated cell suspension in agarose onto
the surface of a microtiter tissue culture dish and
cellular migratory activity is much less dependent on
cell deformability. The length of the incubation pe-
Table 3. Comparative corneal epithelial wound
healing rates in rat organ culture
Treatment
condition
Epithelial wound healing
rate (mm2/hr) ± SEM
Drug-free
EGF + colchicine
EGF alone
0.461 ±0.030
0.462 ± 0.055
0.550 ± 0.026
riod is relatively long (18 and 48 hr) and cellular
motility is assessed daily. With longer incubation
times, significant cell division may occur and it may
be difficult sometimes to determine whether cell displacement is primarily due to individual motility or
to increased numbers of cells. One advantage of a
longer incubation time with the agarose droplet
method is that it allows fastidious cells more time to
adapt, adjust and grow in the system before the cells
are counted. Therefore, the condition of the cells at
the onset of the incubation becomes less crucial. Also,
cellular motility can be measured noninvasively and
in situ without having to fix the cells first.
The disadvantage of the first two assay methods
(Boyden chamber and agarose drop) is that they both
use tissue-cultured corneal epithelial cells separated
from their stroma or native substrata. In contrast, the
third method (organ culture) may better approximate
the true clinical situation, as it involves a multicellular environment and native substrata. The disadvantage of the organ culture method, however, is that it
involves a complex, multilayered mass of confluent
cells; pharmacological effects on single-cell motility
are, therefore, difficult to isolate. Colchicine was used
at a dose of 40 /xg/ml in the organ culture experiments to eliminate mitosis. This drug level is more
than sufficient to assure complete abolishment of mitosis; for instance, treatment of epithelial cell lines in
tissue culture with as little as 4 iig/m\ of colchicine
has been shown to completely disrupt the normal
interphase microtubular network17 and concentrations in the order of 10 Mg/ml are used to arrest mitosis in karyotypical studies.18 Previous studies by
Gipson et al27 have demonstrated that although mitosis is completely abolished at 40 ng/m\, the migration of corneal epithelial cells at this level of colchicine remains unaffected. The second part of our
organ culture studies also showed no significant difference in epithelial wound healing rate between colchicine (40 jug/ml) alone and drug-free medium.
These findings are corroborated by the fact that the
first phase of corneal epithelial wound healing involves only migration and no mitosis. In our organ
culture assay, wound closure rates were equal between the drug-free control corneas and the corneas
cultured with colchicine + EGF, implying that EGF-
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933379/ on 06/16/2017
1812
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Augusr 1989
mediated acceleration of epithelialization is due primarily to mitogenesis.
With the inability of EGF alone to promote corneal
epithelial cell motility, the addition of fibronectin
eyedrops to the EGF treatment regimen for persistent
corneal epithelial defects appears to be a rational
therapeutic approach, with EGF providing mitogenic
stimulation and fibronectin enhancing cell-to-substratum adhesion and providing chemotactic and
haptotactic stimulation of cell migration. Particularly
in severely diseased or injured corneas with an abnormal substratum for the epithelium, treatment
with topical EGF alone is probably not optimal.
11.
12.
13.
14.
15.
Key words: corneal epithelium, epidermal growth factor,
mitosis, cell motility
16.
Acknowledgments
The authors wish to thank Brian Biesman, MD and Ms.
Patricia Perone for their technical assistance.
17.
18.
References
19.
1. Arey LB and Cavode WM: The method of repair in epithelial
wounds of the cornea. Anat Rec 86:75, 1943.
2. Cintron C, Kublin CL, and Covington HC: Quantitative studies of corneal epithelial wound healing in rabbits. Curr Eye Res
1:507, 1982.
3. Buck RC: Cell migration in repair of mouse corneal epithelium. Invest Ophthalmol Vis Sci 18:767, 1979.
4. Kuwabara T, Perkins DG, and Cogan DG: Sliding of the epithelium in experimental corneal wounds. Invest Ophthalmol
15:4, 1976.
5. Savage CR Jr and Cohen S: Proliferation of corneal epithelium
induced by epidermal growth factor. Exp Eye Res 15:361,
1973.
6. Watanabe K, Nakagawa S, and Nishida T: Stimulatory effects
offibronectinand EGF on migration of corneal epithelial cells.
Invest Ophthalmol Vis Sci 28:205, 1987.
7. Brightwell JR, Riddle SL, Eiferman RA, Valenzuela P, Barr
PJ, Merryweather JP, and Schultz GS: Biosynthetic human
EGF accelerates healing of neodecadron-treated primate corneas. Invest Ophthalmol Vis Sci 26:105, 1985.
8. Frati L, Daniele S, Delogu A, and Covelli I: Selective binding
of the epidermal growth factor and its specific effects on the
epithelial cells of the cornea. Exp Eye Res 14:135, 1972.
9. Petroutsos G, Courty J, Guimaraes R, Pouliquen Y, Barritault
D, Plouet J, and Courtois Y: Comparison of the effects of EGF,
pEGF, and EDGF on corneal epithelium wound healing. Curr
Eye Res 3:593, 1984.
10. Castor CW and Cabral AR: Growth factors in human disease:
20.
21.
22.
23.
24.
25.
26.
27.
Vol. 30
The realities, pitfalls, and promise. Sem Arthritis Rheum
15:33, 1985.
Nickoloff BJ, Mitra RS, Riser BL, Dixit VM, and Varani J:
Modulation of keratinocyte motility. Am J Pathol 132:543,
1988.
Thompson P, Desbordes JM, Giraud J,-Pouliquen Y, Barritault D, and Courtois Y: The effect of an eye derived growth
factor (EDGF) on corneal epithelial regeneration. Exp Eye Res
34:191, 1982.
Romualdez AG Jr and Ward PA: A unique complement derived chemotactic factor for tumor cells. Proc Natl Acad Sci
USA 72:4128, 1975.
Varani J, Orr W, and Ward PA: A comparison of the migration
patterns of normal and malignant cells in two assay systems.
Am J Pathol 90:159, 1978.
Carpenter RR: In vitro studies of cellular hypersensitivity: I.
Specific inhibition of migration of cells from adjuvant-immunized animals by purified protein derivative and other protein
antigens. J Immunol 91:803, 1963.
Harrington JT Jr and Stastny P: Macrophage migration from
an agarose droplet: Development of a micromethod for assay
of delayed hypersensitivity. J Immunol 110:752, 1973.
Wolf KM and Mullins JM: Cytochalasin B-induced redistribution of cytokerain filaments in PtK| cells. Cell Motility
7:347, 1987.
Dustin P: Historical background. In Microtubules, Dustin P,
editor. Berlin, Springer-Verlag, 1984, pp. 8-18.
Gipson IK and Grill SM: A technique for obtaining sheet of
intact corneal epithelium. Invest Ophthalmol Vis Sci 23:269,
1982.
Cohen S: Isolation of a mouse sub-maxillary gland protein
accelerating incision eruption and eyelid opening in the newborn animal. J Biol Chem 237:1555, 1962.
Ho PC, Davis WH, Elliot JH, and Cohen S: Kinetics of corneal
epithelial regeneration and epidermal growth factor. Invest
Ophthalmol 13:804, 1974.
Daniele S, Frati L, Fiore C, and Santoni G: The effect of
epidermal growth factor (EGF) on the corneal epithelium in
human. Graefes Arch Klin Exp Ophthalmol 210:159, 1979.
Nishida T, Tanaka H, Nakagawa S, Sasabe T, Awata T, and
Manabe R: Fibronectin synthesis by the rabbit cornea: Effects
of mouse epidermal growth factor and cyclic AMP analogs.
Jpn J Ophthalmol 28:196, 1984.
Phan TM, Foster CS, Boruchoff SA, Zagachin LM, and Colvin
RB: Topical fibronectin in the treatment of persistent corneal
epithelial defects and trophic ulcers. Am J Ophthalmol
104:494, 1987.
Nishida T, Nakagawa S, Awata T, Ohashi Y, Watanabe K, and
Manabe R: Fibronectin promotes epithelial migration of cultured rabbit cornea in situ. J Cell Biol 97:1653, 1983.
Watanabe K, Nakagawa S, and Nishida T: Chemotactic and
haptotactic activities of fibronectin for cultured rabbit corneal
epithelial cells. Invest Ophthalmol Vis Sci 29:572, 1988.
Gipson IK, Westcott MJ, and Brooksby NG: Effects of cytochalasins B and D and colchicine on migration of the corneal
epithelium. Invest Ophthalmol Vis Sci 22:633, 1982,
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933379/ on 06/16/2017