lable at ScienceDirect

Experimental Eye Research 87 (2008) 56–63

Contents lists avai

Experimental Eye Research

journal homepage: www.elsevier .com/locate/yexer

Corneal wound healing is modulated by topical application of amnioticfluid in an ex vivo organ culture model

Juan Castro-Combs a, Guillermo Noguera a, Marisol Cano a, Margaret Yew a, Peter L. Gehlbach a,Jonathan Palmer b, Ashley Behrens a,*

a The Wilmer Ophthalmological Institute, The Johns Hopkins University School of Medicine, Baltimore, MD, USAb The School of Veterinary Medicine, University of Pennsylvania, Kennett Square, PA, USA

a r t i c l e i n f o

Article history:Received 17 July 2007Accepted in revised form 22 April 2008Available online 30 April 2008

Keywords:corneawound healingreepithelializationamniotic fluid

* Corresponding author. The Wilmer OphthalmoHopkins University School of Medicine, 600 N Wolfe255, Baltimore, MD 21287, USA. Tel.: þ1 410 502 046

E-mail address: abehrens@jhmi.edu (A. Behrens).

0014-4835/$ – see front matter � 2008 Elsevier Ltd.doi:10.1016/j.exer.2008.04.010

a b s t r a c t

The purpose of this study was to evaluate the effects of topical human amniotic fluid (HAF) and equineamniotic fluid (EAF) on corneal reepithelialization and stromal wound healing.New Zealand white rabbit corneas (n¼ 52) were placed in an ex vivo air-interface organ culture. An8.5 mm-diameter mark in the center of the cornea was produced with a hand trephine to select the areafor epithelial scraping. A number 15 surgical blade was used to remove the epithelial layer within thedemarcated area in a standardized fashion. The corneas were assigned to one of four treatment groups(n¼ 8): fetal bovine serum (FBS), HAF, EAF, and a control group that was exposed to phosphate buffersolution (PBS). Corneal epithelial defects were imaged every 8 h for 72 h after the application of a 30 mldrop of 0.015% fluorescein. Five corneas of each treatment group were used for histology, proliferation,and apoptosis assay at 72 h after the epithelial defect was created.There was no significant difference in the mean rate of closure of the corneal epithelial defect betweenFBS treated corneas and controls (P> 0.06). The mean epithelial defect area (MEDA) was significantlysmaller in the EAF group as compared to control corneas at 24 h (P¼ 0.016), 40 h (P¼ 0.032), 64 h(P¼ 0.008) and 72 h (P¼ 0.007) following epithelial scrape. The MEDA in the HAF group was significantlysmaller at 16 h (P¼ 0.008), 64 h (P¼ 0.0072), and 72 h (P¼ 0.016) compared to the control group. TheMEDA in the HAF and EAF groups was smaller at all time points as compared to the FBS group, butthe difference was not significant. At histology, the mean keratocyte density was significantly higher inthe anterior stroma in the HAF (P< 0.001) and EAF groups (P¼ 0.001) as compared to control group. Thenumber of BrdU positive keratocytes was significantly higher in the superficial and deep stromal sub-areas in the HAF group as compared to control (P< 0.001 and P¼ 0.002, respectively). EAF and FBStreated corneas also showed a higher number of BrdU positive cells compared to control, but thisdifference was not significant. Finally, we did not observe any difference in the amount of TUNEL positivekeratocytes among the different groups.Our data indicates that the topical application of HAF and EAF is associated with acceleratedreepithelialization in this cornea organ culture model. Similarly, corneal keratocyte density appears to beless affected after epithelial injury using this treatment.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

Human amniotic membrane (HAM) was first used in ophthal-mology by de Rotth (1940) to repair conjunctival defects. Sincethen, HAM has been used to treat a wide variety of corneal disor-ders such as persistent epithelial defects (Shukla, 1962; Batmanovet al., 1990), neurotrophic corneas (Kruse et al., 1999; Lee and Tseng,

logical Institute, The JohnsStreet, Woods Building room1; fax: þ1 410 502 2461.

All rights reserved.

1997), chemical injuries (Sorsby and Symons, 1946; Sorsby et al.,1947), recurrent erosion syndrome and persistent epithelial defectsassociated with cicatricial conditions (Lee and Tseng, 1997). Positiveeffects on corneal reepithelialization, inflammatory response, andscar formation have been reported with the use of HAM (Kim andTseng, 1995; Meller and Tseng, 1998). In a randomized prospectivestudy of rabbit corneas undergoing photorefractive keratectomyexcimer laser Choi et al. (1998) demonstrated a reduction of haze inthe postoperative period after amniotic membrane was applied tothe wound area; they hypothesized that this may be due toa reduction in inflammatory cell infiltration in the earlypostoperative period. Moreover, Woo et al. (2001) created large

J. Castro-Combs et al. / Experimental Eye Research 87 (2008) 56–63 57

epithelial defects on rabbit corneas using excimer laser, reportingfaster rate of reepithelialization when the corneas were treatedwith amniotic membrane.

Most of the proteins present in HAM have also been found inHAF (Zhang et al., 2001). It has been reported that HAF is effective inthe reduction of corneal opacity and scar formation in mousecorneas after alkali corneal burn induction (Herretes et al., 2006). Inaddition, reports indicate that amniotic fluid accelerates recovery ofcorneal sensitivity through nerve regeneration and decreasescorneal scar formation after photorefractive keratectomy (Lee andKim, 1996). Potential therapeutic effects on corneal wound healingfor HAM and HAF are therefore expected to be similar.

Based on these observations, our objective with this study wasto evaluate the effects of topical application of HAF and EAF tomodulate corneal healing. To test this hypothesis, we utilized anestablished air-interface organ culture technique with New Zealandwhite rabbit corneas (Chuck et al., 2001).

2. Methods

2.1. Rabbit globe harvest

The eyes of 26 adult New Zealand white rabbits were enucleatedimmediately following euthanization and were placed in 50 mltubes (Becton Dickinson and Company, Franklin Lake, NJ) contain-ing Dulbecco Modified Eagle Medium (Gibco BRL, Life TechnologiesInc., Rockville, MD). The tubes were placed in a container with icefor preservation (4 �C). A circular central epithelial defect wascreated on each cornea 2 h after enucleation. The sclero-cornealring was excised from each globe with conjunctiva attached usingcurved scissors and leaving 3–5 mm of sclera.

2.2. Preparation of whole-organ rabbit cornea culture

With the use of a sterile blade, the ends of laboratory test tubes(2059; Falcon, Oxnard, CA) were cut at the indicator line nearest tothe bottom. The cut edges of the domes were smoothed, and thedomes were aseptically placed into six-well tissue culture plates(Becton and Dickinson, Franklin Lake, NJ). Each dome was gluedtightly concave-side down into each well using acrylic glue (GEWhite All Purpose Adhesive Caulk). Dulbecco Modified EagleMedium with antibiotic/antimycotic solution (1/200) (Mediatech,Inc., Herndon, VA) was then added to each well to just cover eachdome (12 ml per well). The sclero-corneal rings were placed in theculture plates to be incubated at 37 �C with 5% CO2. Culturemedium was changed once during the experiment (36 h).

2.3. Corneal epithelial defect

Using a dissecting microscope (Stemi 200-C, Carl Zeiss Inc.,Thornwood, NY) and a cold light source (Schott North America,Auburn, NY), an 8.5 mm-diameter circle located in the center of thecornea was produced with a hand trephine blade. A number 15surgical blade (Bard–Parker, Franklin Lakes, NJ) was used to scrapeand remove the epithelial layer including the basement membranewithin the demarcated area in a standardized fashion. The reepi-thelialization process was followed by imaging every 8 h usinga digital camera (Nikon Coolpix 990, Nikon Inc., Melville, NY) with a17�macro lens attached for 72 h. A 30 ml drop of 0.015% fluorescein(Pharmaceuticals, Inc., Aquebogue, NY) was applied to stain theepithelial defect just before every imaging time point. The excess offluorescein was rinsed away with 500 ml of phosphate-salinesolution (PBS) (Invitrogen Corporation, Grand Island, NY). Aseparate six-well tissue culture plate was made using black tubes.The sclero-corneal rings were placed in this special designed plateduring picture time to enhance the contrast during photography

and to avoid mixing of fluorescein and treatment drops withculture medium. The areas of corneal epithelial defect wereoutlined and measured using a digital imaging software (AxioVi-sion, Carl Zeiss Inc., Thornwood, NY). The wounded corneal surfacearea was calculated with a method described previously by Crossonet al. (1986).

2.4. Treatment and experimental design

After the epithelial defect was created, the sclero-corneal ringswere assigned to four different groups according to the treatmentadministered: human amniotic fluid (HAF) (n¼ 8), equine amnioticfluid (EAF) (n¼ 8), fetal bovine serum (FBS) (n¼ 8) (Benchmarktriple filtered, Gemini Bio-products, West Sacramento, CA), andcontrol group that was exposed to PBS (n¼ 8). A 20 ml drop of therespective treatment was applied every 8 h.

2.5. Amniotic fluid

After approval of the Institutional Review Board of the JohnsHopkins University, HAF was obtained from patients with low riskpregnancies at the Department of Gynecology and Obstetrics at theJohns Hopkins Hospital. A 16–21 week amniotic fluid pool wasmade using 12 samples to be discarded after routine amniocentesisfor karyotyping. EAF was obtained from the School of VeterinaryMedicine at the University of Pennsylvania. The samples wereaseptically recollected by direct amniocentesis at the moment ofdelivery. Both HAF and EAF samples were centrifuged at 1800 rpmfor 10 min and the supernatant separated and preserved at �70 �C.At the moment of utilization, aliquots of fluid were thawed andthen stored at 4 �C during treatment to minimize bacterialproliferation (Herretes et al., 2006).

2.6. Corneal histology

Twenty corneas (n¼ 5 from each group) were fixed in 10%buffered formaldehyde at 72 h after epithelial defect creation. Thecorneas were then immersed and oriented in paraffin. The sampleswere cut, mounted and stained with hematoxylin and eosinaccording to standard methods. Epithelial and stromal character-istics were analyzed using an inverted microscope with a 40�objective (Axiovert 200M, Carl Zeiss Inc., Thornwood, NY). Digitalimages of the central cornea were captured using a 10� objective.A 300 mm wide cross-section of the corneas was created on thedigital images using AxioVision software. This cross-area wasdivided into three horizontal sub-areas (superficial, medial anddeep stroma) with the same dimensions each (Fig. 1). Cell densityin these areas was calculated by counting the number of cells ineach section.

2.7. Cell proliferation assay

5-Bromo-2-deoxyuridine (100 mg/ml) (BrdU, Sigma–Aldrich)was added to culture wells of 20 corneas (n¼ 5 each group) 2 hbefore fixation. Tissue sections were processed as described byLevin and Verkman (2006). The slides were incubated with a sheeppolyclonal antibody to BrdU (50 mg/ml; Abcam, Cambridge, MA).Bound antibody was detected using the Vectastain Elite ABC Kit(Vectastain; Vector Laboratories, Burlingame, CA). Digital images ofthe central cornea were captured using a 10� objective. BrdUpositive keratocytes were counted in the superficial, medial anddeep stromal sub-areas of a central corneal section divided aspreviously explained.

Fig. 1. Corneal cross-section for keratocyte quantification. A 300 mm wide cross-sec-tion of the corneal stroma was created on the digital images. The field was divided intothree equal regions designated superficial, medial and deep corneal stroma (10�, H–E).

Fig. 2. Reepithelialization rate. (A) There was no significant difference in the cornealepithelium closure between FBS and control. (B) The rate of reepithelialization wasstatistically higher in the EAF compared to control at 24, 40, 64, and 72 h. (C) Cornealepithelium healing was significantly faster in the HAF compared to control at 16, 64and 72 h.

J. Castro-Combs et al. / Experimental Eye Research 87 (2008) 56–6358

2.8. TUNEL assay

The terminal deoxynucleotidyl transferase (TdT)-mediateddUTP nickendlabeling (TUNEL) technique was performed to detectkeratocyte apoptosis (ApopTag Plus In Situ Apoptosis FluoresceinDetection Kit, Chemicon International). Twenty corneas keptin organ cultures for 72 h were fixed with 10% buffered para-formaldehyde for 24 h and embedded in paraffin. The slides wereincubated with TUNEL reaction and counterstained with DAPI(1 mg/mL; Vector Laboratories, Burlingame, CA). The samples wereobserved with a reverse phase contrast microscope. Digital imagesof the central cornea were captured using a 10� objective. TUNELpositive keratocytes were counted in the superficial, medial anddeep stromal sub-areas of a central corneal section divided aspreviously explained.

2.9. Statistical analysis

Comparisons between groups were made using statisticalsoftware for Windows (StatsDirect 2.4.1, Cheshire, UK). Compari-sons between epithelial defect areas were made using the non-parametric Mann–Whitney test. Comparisons between groupswere made with Kruskal–Wallis test, and all pairwise multiplecomparison procedures were made with Bonferroni test.

3. Results

3.1. Corneal reepithelialization

All epithelial wounds healed in a concentric fashion. Thedecrease in epithelial defect area over time (mm2/h) in all groupswas not a linear process. There was a significant reduction in thewound area during the first 48 h, after which epithelial healingslowed.

The mean decrease in epithelial wound area in the FBS groupoccurred at a faster rate than the decrease evident in the controlgroup but this difference was statistically significant only at onetime point (48 h) (P¼ 0.031) (Fig. 2A). The epithelial defects weresignificantly smaller in the EAF group as compared to controlcorneas at 24 h (P¼ 0.016), 40 h (P¼ 0.0317), 64 h (P¼ 0.008) and72 h (P¼ 0.007) following epithelial defect creation (Figs. 2B and 3).Moreover, the corneal epithelium defect area in the HAF group wasalso significantly smaller at 16 h (P¼ 0.008), 64 h (P¼ 0.0072), and

72 h (P¼ 0.016) after corneal wound creation when compared withthe control group (Figs. 2C and 3). The mean epithelial defect areain the HAF group was smaller at all time points as compared to theFBS group, but this difference was statistically significant only at 8 h(P¼ 0.016) and 16 h (P¼ 0.02). Epithelial healing was faster in EAFtreated corneas when compared to the FBS group but this differ-ence was statistically significant only at 24 h (P¼ 0.008). At the last

Fig. 3. Corneal reepithelialization. Representative corneas from control, EAF and HAF groups, showing the epithelial defect area (green staining) from the moment the defect wascreated to the last recorded time point. Six of the 10 time points were included in this figure and are representative.

J. Castro-Combs et al. / Experimental Eye Research 87 (2008) 56–63 59

time point of observation (72 h), none of the corneas of the controlgroup was completely healed and only one from the FBS group wascompletely closed. Conversely, the epithelial defect of 2 and 3corneas were completely closed by 72 h in the HAF and EAF groups,respectively.

3.2. Histology

Corneal epithelium maintained its normal appearance in theperipheral cornea in all groups. Peripheral corneal epitheliumcould be described as a nonkeratinized, stratified, squamousepithelium consisting of basal cells, wing cells, and superficialcells distributed in 5–7 layers. On the other hand, the regen-erated epithelium in the central cornea was composed by

a single layer of basal cells with 1–3 more layers of superficialcells. We did not observe morphologic difference in the regen-erated epithelial layer when control corneas were compared totreated groups by light microscopy. The overall thickness ofcentral epithelium was decreased in all groups when comparedto peripheral epithelium.

A decrease in corneal stromal cell density (keratocytes) in thesuperficial corneal stroma beneath the area of epithelial scrapingwas present. This decrease in keratocyte density was evident inall groups but the mean decrease in cell count varied among them(Fig. 4). Mean keratocyte cell density and P values of the differ-ences in cell density in the three different stroma sub-areasamong groups have been presented in Tables 1A and 1B,respectively.

Table 1ACorneal keratocyte density in superficial, medial and deep stromal layers (mean andstandard deviation have been included)

Mean [SD] cell density

Superficial stroma Medial stroma Deep stroma

Control 9.7 [7.6] 29 [7.4] 33 [8.6]FBS 11.5 [1.9] 28.6 [7.4] 32 [8.1]HAF 38 [7.3] 34.4 [5.3] 43.8 [9.8]EAF 30 [5.3] 40.5 [8.7] 26.6 [6.1]

Table 1BComparisons between mean values of keratocyte cell density in the three stromalsub-areas

Table of mean differences

Superficial stroma(P< 0.001)a

Medial stroma(P> 0.05)a

Deep stroma(P¼ 0.04)a

Control vs. FBS 1.000 >0.05 >0.05Control vs. HAF <0.001 >0.05 >0.05Control vs. EAF 0.001 >0.05 >0.05HAF vs. FBS <0.001 >0.05 >0.05EAF vs. FBS 0.003 >0.05 >0.05EAF vs. HAF 0.328 >0.05 0.033

a Kruskal–Wallis.

J. Castro-Combs et al. / Experimental Eye Research 87 (2008) 56–6360

3.3. TUNEL and cell proliferation assay

Seventy-two hours after epithelial scraping, TUNEL assaydetected DNA fragmentation on the transition area between theanterior hypocellular stroma and the medial stroma, adjacent to theepithelial scraping site. The amount of TUNEL positive cells was notsignificantly different when treatment groups were compared tocontrol (Fig. 5). Moreover, the amount of epithelial BrdU positivecells was not different when treatment groups and control werecompared. BrdU positive keratocytes were present in all groups,and these proliferating cells were localized in the three areas inwhich the corneal stroma was subdivided (superficial, medial anddeep stroma) (Fig. 6). Mean and standard deviation of BrdU positivekeratocytes and P values of the comparisons among groups in thethree different stroma sub-areas have been presented in Tables 2Aand 2B, respectively.

4. Discussion

4.1. Reepithelialization

The ex vivo air-interface organ culture technique was firstdescribed by Richard et al. (1991) and is considered to be an

Fig. 4. Corneal keratocyte density. Representative figure indicating keratocyte density differscale bars 100 mm). Control (A), FBS (B), HAF (C), and EAF (D).

appropriate method for the maintenance of the cornea. It permitsa short-term culture of ex vivo corneas with an adequate preser-vation of epithelial integrity. This approach seems to decrease theintracellular edema observed in submerged models improvingepithelial cell morphology and preserving stromal keratocytes forat least 4 weeks. On the other hand, some weaknesses of the systemmay include the use of acrylic glue in the culture (which may betoxic), the absence of tear secretion, the isolation of nerve regula-tion and the lack of immune system cells closely involved in thecorneal healing process (Wilson et al., 2001).

Continuous epithelial renewal is a process that depends on cellproliferation, migration, and shedding of the epithelium from thecorneal surface (Thoft and Friend, 1983). This process is essential topreserve corneal integrity and to maintain corneal transparencyduring baseline and wound healing conditions. Corneal reepithe-lialization is a complex process involving the flattening andelongation of epithelial cells during migration (Zieske et al., 2000).Important variations in proliferative and migration-related activityoccur across the surface of the epithelium with some cells

ences in the anterior and medial stroma among treatment groups (arrows) (10�, H–E.

Fig. 5. TUNEL assay. Representative figure showing TUNEL positive keratocytes (Green [FITC]) (arrows). Sections are counterstained with DAPI for nuclei (blue). Control (A) and (B),FBS (C) and (D), EAF (F), HAF (E), negative control (G), control time 0 (H) (10�, scale bars 100 mm, 20�, scale bars 50 mm).

J. Castro-Combs et al. / Experimental Eye Research 87 (2008) 56–63 61

migrating to close the defect and others proliferating to provideadditional cells (Gipson et al., 1993).

The in vitro culture of animal cells commonly requiresa medium rich in nutrients. These nutrients are usually supplied by

supplementing a formulated medium with animal serum. FBS isone of the key ingredients used in corneal epithelial and stromalcell cultures (Bonfiglio et al., 2006; Kawakita et al., 2004), becauseit offers a complex of proteins and amino acids, lipids and

Fig. 6. Corneal keratocyte proliferation assay. Representative figure showing BrdU positive keratocytes (dark brown nuclei, [arrows]) in cross-sections of the central cornea. Control(A) and (B), FBS (C), EAF (D), HAF (E), negative control (F) (10�, scale bars 100 mm, 20� scale bars 50 mm). Corneal stroma sub-areas: superficial (1), medial (2), and deep (3) cornealstroma.

Table 2BComparisons between mean values of BrdU positive keratocytes in the three stromal

J. Castro-Combs et al. / Experimental Eye Research 87 (2008) 56–6362

triglycerides, vitamins, inorganic minerals and salts, and growthfactors that promote culture growth. However, the application ofFBS to the corneas in our ex vivo model did not produce a signifi-cant change in epithelial wound closure when compared to control.The epithelial wound area was smaller in the FBS treated corneasfrom 48 to 72 h, but this difference was only statistically significantat one time point (48 h) (P< 0.05). The epithelial defect area in theHAF group was smaller than the defect in the control group corneasfrom 16 to 72 h. This difference was statically significant at 16, 64and 72 h (P< 0.01), which corresponds well with the observationsmade by Herretes et al. (2006), suggesting a beneficial effect of HAFon reepithelialization following alkali burns of cornea in an in vivoanimal model. The healing rate in the EAF treated group was alsofaster than control from 8 to 72 h. This difference was found to besignificant at 24, 40, 64 and 72 h (P< 0.01) (Fig. 3).

In this study we have evaluated corneal reepithelialization inthe absence of the intact immune system, tear secretion andcorneal innervation, all known to be important participants in thewound healing process. However, the use of this ex vivo model hasallowed us to observe and study the relationships between accel-erated reepithelialization and changes in corneal stromal celldensity. Topical HAF and EAF were effective promoters of cornealreepithelialization when compared to topical PBS. This stimulatoryeffect most likely results from a group of factors present in HAF and

Table 2AMean corneal keratocyte proliferation in superficial, medial and deep stromal Layers(mean and standard deviation have been included)

Mean [SD] BrdU positive cell density

Superficial stroma Medial stroma Deep stroma

Control 6.2 [6.5] 5.7[1.7] 6.2 [4.7]FBS 7 [2.1] 10.7[5.1] 17 [5.5]HAF 25.2 [4.5] 31.5 [9.1] 27.7 [7.8]EAF 5.2 [3.8] 19.2 [4.6] 14.5 [6.5]

EAF. HAF is rich in growth factors known to be critical for fetaldevelopment (Mulvihill et al., 1986). The use of equine amnioticfluid to treat corneal epithelial healing is, to our knowledge,completely novel and with this study we have shown that amnioticfluid from other species may have beneficial effects similar to HAFin corneal epithelial wound healing. The use of EAF representsa reliable source of amniotic fluid since the volumes of fluid obtainfrom mares are larger than humans. In addition, mares have beenused to safely isolate medical formulations such as conjugated es-trogens and serpent antivenom from their natural fluids (Chippauxet al., 2007; Zhao and Diaz-Brinton, 2006).

4.2. Stromal cell density

Injuries such as corneal scraping, mechanical pressure and viralinfection of the epithelium trigger keratocyte apoptosis in thesuperficial stroma (Wilson et al., 1994, 2001). Prior work suggeststhat this regulated cell death is mediated by cytokines releasedfrom the injured epithelium such as IL-1 and TNF alpha (Masuret al., 1999; Mohan et al., 1996; Wilson et al., 1996). Keratocyte

sub-areas

Table of mean differences

Superficial stroma(P< 0.001)a

Medial stroma(P> 0.05) a

Deep stroma(P¼ 0.003) a

Control vs. FBS 1.000 >0.05 >0.05Control vs. HAF <0.001 >0.05 0.002Control vs. EAF 0.097 >0.05 >0.05HAF vs. FBS <0.001 >0.05 >0.05EAF vs. FBS 0.149 >0.05 >0.05EAF vs. HAF 0.054 >0.05 >0.05

a Kruskal–Wallis.

J. Castro-Combs et al. / Experimental Eye Research 87 (2008) 56–63 63

apoptosis begins soon after epithelial injury and continues for atleast 1 week following injuries such as epithelial scraping, PRK ora microkeratome cut into the cornea (Gao et al., 1997; Mohan andWilson, 2001).

In this study we observed an increase in keratocyte celldensity during the early stages of corneal wound healing (3days) following HAF and EAF treatment. Our data indicates thatthe topical application of amniotic fluid is associated with theaccelerated reepithelialization observed in the treated groupsand protection of corneal keratocytes. The higher keratocytedensity observed in the anterior stroma 72 h after epithelialscraping in the corneas treated with HAF and EAF seems tobe, by the light of our results, a product of cell proliferationrather than differences in keratocyte apoptosis. It has beenreported that activated keratocytes and myofibroblasts regulateproliferation, motility, and differentiation of epithelial cells withthe production of HGF and KGF (Weng et al., 1997). It is possiblethat a higher keratocyte cell density in the anterior stroma couldincrease HGF and KGF production which may accelerate cornealreepithelialization.

Given the results obtained in this study, we postulate thattopical HAF and EAF seem to have a positive effect in cornealwound healing in this model. Further in vivo studies arerequired to demonstrate the effectiveness of this treatment inanimals, to confirm the obtained results in a live model ofcorneal injury.

Acknowledgements

The authors would like to thank the Department of PrenatalDiagnosis and Treatment Center of the Johns Hopkins Hospital, andThe School of Veterinary Medicine at the University of Pennsylvaniafor their continuous support in providing the amniotic fluidsamples necessary to perform this study.

Financial Support: Research to Prevent Blindness Inc., NewYork, NY.

References

Batmanov, Iu.E., Egorova, K.S., Kolesnikova, L.N., 1990. Use of fresh amnion in thetreatment of corneal diseases. Vestn. Oftalmol. 106, 17–19.

Bonfiglio, V., Camillieri, G., Avitabile, T., Leggio, G.M., Drago, F., 2006. Effects of theCOOH-terminal tripeptide alpha-MSH (11–13) on corneal epithelial woundhealing: role of nitric oxide. Exp. Eye Res. 83, 1366–1372.

Chippaux, J.P., Massougbodji, A., Stock, R.P., Alagon, A., 2007. Clinical trial of anF(ab0)2 polyvalent equine antivenom for African snake bites in Benin. Am. J.Trop. Med. Hyg. 77, 538–546.

Choi, Y.S., Kim, J.Y., Wee, W.R., Lee, J.H., 1998. Effect of the application of humanamniotic membrane on rabbit corneal wound healing after excimer laserphotorefractive keratectomy. Cornea 17, 389–395.

Chuck, R.S., Behrens, A., Wellik, S., Liaw, L.L., Dolorico, A.M., Sweet, P., Chao, L.C.,Osann, K.E., McDonnell, P.J., Berns, M.W., 2001. Re-epithelialization in corneaorgan culture after chemical burns and excimer laser treatment. Arch.Ophthalmol. 119, 1637–1642.

Crosson, C.E., Klyce, S.D., Beuerman, R.W., 1986. Epithelial wound closure in therabbit cornea. Investig. Ophthalmol. Vis. Sci. 27, 464–473.

Gao, J., Gelber-Schwalb, T.A., Addeo, J.V., Stern, M.E., 1997. Apoptosis in the rabbitcornea after photorefractive keratectomy. Cornea 16, 200–208.

Gipson, I.K., Watanabe, H., Zieske, J.D., 1993. Corneal wound healing and fibronectin.Int. Ophthalmol. Clin. 4, 149–163.

Herretes, S., Suwan-Apichon, O., Pirouzmanesh, A., Reyes, J.M., Broman, A.T.,Cano, M., Gehlbach, P.L., Gurewitsch, E.D., Duh, E.J., Behrens, A., 2006. Use of

topical human amniotic fluid in the treatment of acute ocular alkali injuries inmice. Am. J. Ophthalmol. 142, 271–278.

Kawakita, T., Espana, E.M., He, H., Yeh, L.K., Liu, C.Y., Tseng, S.C., 2004. Calcium-induced abnormal epidermal-like differentiation in cultures of mouse corneal–limbal epithelial cells. Investig. Ophthalmol. Vis. Sci. 45, 3507–3512.

Kim, J.C., Tseng, S.C., 1995. Transplantation of preserved human amniotic membranefor surface reconstruction in severely damaged rabbit corneas. Cornea 14,473–484.

Kruse, F.E., Rohrschneider, K., Volcker, H.E., 1999. Multilayer amniotic membranetransplantation for reconstruction of deep corneal ulcers. Ophthalmology 106,1504–1510.

Lee, H.S., Kim, J.C., 1996. Effect of amniotic fluid in corneal sensitivity and nerveregeneration after excimer laser ablation. Cornea 15, 517–524.

Lee, S.H., Tseng, S.C., 1997. Amniotic membrane transplantation for persistentepithelial defects with ulceration. Am. J. Ophthalmol. 123, 303–312.

Levin, M.H., Verkman, A.S., 2006. Aquaporin-3-dependent cell migration andproliferation during corneal re-epithelialization. Investig. Ophthalmol. Vis. Sci.47, 4365–4372.

Masur, S.K., Dewal, H.S., Dinh, T.T., Erenburg, I., Petridou, S., 1999. Myofibroblastsdifferentiate from fibroblasts when plated at low density. Proc. Natl. Acad. Sci.U.S.A. 93, 4219–4223.

Meller, D., Tseng, S.C., 1998. Reconstruction of the conjunctival and corneal surface.Transplantation of amnionic membrane. Ophthalmologe 95, 805–813.

Mohan, R.R., Kim, W.J., Mohan, R.R., Chen, L., Wilson, S.E., 1996. Bone morphogenicproteins 2 and 4 and their receptors in the adult human cornea. Investig.Ophthalmol. Vis. Sci. 39, 2626–2636.

Mohan, R.R., Wilson, S.E., 2001. Ex vivo human corneal epithelial cells expressmembrane-bound precursor and mature soluble epidermal growth factor(EGF) and transforming growth factor (TGF) alpha proteins. Exp. Eye Res. 68,129–131.

Mulvihill, S.J., Stone, M.M., Fonkalsrud, E.W., Debas, H.T., 1986. Trophic effect ofamniotic fluid on fetal gastrointestinal development. J. Surg. Res. 40,291–296.

de Rotth, A., 1940. Plastic repair of conjunctival defects with fetal mebranes. Arch.Ophthalmol. 23, 522–525.

Richard, N.R., Anderson, J.A., Weiss, J.L., Binder, P.S., 1991. Air/liquid corneal organculture: a light microscopic study. Curr. Eye Res. 10, 739–749.

Shukla, I.M., 1962. Amniotic membrane grafts in corneal ulcer. J. All IndiaOphthalmol. Soc. 10, 55–60.

Sorsby, A., Symons, H.M., 1946. Amniotic membrane grafts in caustic burns of theeye (burns of the second degree). Br. J. Ophthalmol. 30, 337–345.

Sorsby, A., Haythorne, J., Reed, H., 1947. Further experience with amnioticmembrane grafts in caustic burns of the eye. Br. J. Ophthalmol. 31, 409–418.

Thoft, R.A., Friend, J., 1983. The X, Y, Z hypothesis of corneal epithelial maintenance.Investig. Ophthalmol. Vis. Sci. 24, 1442–1443.

Weng, J., Mohan, R.R., Li, Q., Wilson, S.E., 1997. IL-1 upregulates keratinocyte growthfactor and hepatocyte growth factor mRNA and protein production by culturedstromal fibroblast cells: interleukin-1 beta expression in the cornea. Cornea 16,465–471.

Wilson, S.E., He, Y.G., Weng, J., Zieske, J.D., Jester, J.V., Schultz, G.S., 1994. Effect ofepidermal growth factor, hepatocyte growth factor, and keratinocyte growthfactor, on proliferation, motility and differentiation of human corneal epithelialcells. Exp. Eye Res. 59, 665–678.

Wilson, S.E., He, Y.G., Weng, J., Li, Q., McDowall, A.W., Vital, M., Chwang, E.L., 1996.Epithelial injury induces keratocyte apoptosis: hypothesized role for theinterleukin-1 system in the modulation of corneal tissue organization andwound healing. Exp. Eye Res. 62, 325–338.

Wilson, S.E., Mohan, R.R., Mohan, R.R., Ambrosio Jr., R., Hong, J., Lee, J., 2001.The corneal wound healing response: cytokine-mediated interaction ofthe epithelium, stroma, and inflammatory cells. Prog. Retin. Eye Res. 20,625–637.

Woo, H.M., Kim, M.S., Kweon, O.K., 2001. Effects of amniotic membrane onepithelial wound healing and stromal remodeling after excimer laser keratec-tomy in rabbit cornea. Br. J. Ophthalmol. 85, 345–349.

Zhang, Q., Shimoya, K., Moriyama, A., Yamanaka, K., Nakajima, A., Nobunaga, T.,Koyama, M., Azuma, C., Murata, Y., 2001. Production of secretory leukocyteprotease inhibitor by human amniotic membranes and regulation of its con-centration in amniotic fluid. Mol. Hum. Reprod. 7, 573–579.

Zhao, L., Diaz-Brinton, R., 2006. Select estrogens within the complex formulation ofconjugated equine estrogens (Premarin) are protective against neurodegener-ative insults: implications for a composition of estrogen therapy to promoteneuronal function and prevent Alzheimer’s disease. BMC Neurosci. 7, 24.

Zieske, J.D., Takahashi, H., Hutcheon, A.E., Dalbone, A.C., 2000. Activation ofepidermal growth factor receptor during corneal epithelial migration. Investig.Ophthalmol. Vis. Sci. 41, 1346–1355.