Healing of ocular surface wounds can be grossly hindered when there is depletion of corneal epithelial stem cells in the limbus. Persistent epithelial defects are the usual consequences and lead to ocular surface morbidity, such as persistent photophobia and overt conjunctivalization of the cornea. 1 Reestablishment of the normal epithelial surface requires replacement of the damaged corneal surface and defected limbal region.
Treatment involves transplantation of exogenous healthy limbal epithelium as an autograft obtained from normal contralateral eye if the damage is unilateral or as an allograft taken from relative or cadaveric eye. 2–6 The clinical outcome can be satisfactory, but there are limitations in this treatment mode: availability of donor tissue, risk of postoperative complications, and tissue rejection in allografting. There is a requirement for a considerable amount of limbal graft to be taken from the donor eye, thus increasing the risk of limbal stem cell exhaustion. 7 Another procedure uses ex vivo expansion of limbal stem cells for growth and differentiation to corneal epithelial cells, followed by transplantation of the confluent layers to the damaged ocular surface. 8 Using this, a small piece of healthy limbal tissue with an intact basal epithelium, 2 to 3 mm long, is adequate to generate sufficient cells for grafting. 9,10
Limbal cells are capable of attaining higher mitotic growth potential than corneal cells. They contain a subpopulation of stem cells, which are characterized by high capacity of self-renewal, slow cell cycle, and strong resistance toward cell differentiation. 11,12 After the first asymmetric cell division, a stem cell gives rise to two daughter cells; one retains in stem environment and the other commits to active proliferation to make up the population of transient amplifying cells. 13 Epidermal growth factor (EGF), acidic and basic fibroblast growth factors (FGF), and nerve growth factor stimulate the proliferation of transient amplifying and limbal stem cells, 14 which contain four- to five-fold higher levels of EGF receptors than corneal basal cells. 15 However, transforming growth factor (TGF) β inhibits the proliferation of limbal and corneal cells in culture. 16 It is also known that tissue inhibitors of metalloproteinases (TIMP) downregulate the enzymatic activity in the modulation of extracellular matrix during development. 17
Human amniotic membrane (HAM) as a grafting material has been used to encourage epithelialization in burns and skin ulcers. It is widely applied for conjunctival repair after chemical burns and for epithelial restoration in corneal ulcers, cicatricial eye diseases, and ptergyia. 18–21 Limbal grafting with HAM also helps ocular surface reconstruction after chemical or thermal burns and pterygia associated with symblepharon. 6,22,23 Recently, confluent cultures of rabbit and human corneal cells expanded from limbal tissue have been successfully generated on HAM. Subsequent transplantation to keratectomized and limbal-defected eyes resulted in satisfactory corneal reepithelialization. 10,24,25 This shows that HAM, which contains collagen types IV and V, fibronectin, and laminin, serves as a good substrate for epithelial cell growth and transfer to the damaged eye. 26 Meanwhile, the expression of growth factors in HAM epithelia or stroma may affect the reepithelialization of corneal cells on a damaged ocular surface. 27,28
In this study, the expression of growth factors, including EGF, TGFβ1, TGFβ2 and TIMP1, by human limbal epithelial cells cultivated on denuded HAM, was explored to show the effect of HAM on the growth and metabolism of human limbal epithelial cells.
MATERIALS AND METHODS
Cells, Media, and Culture Condition
Donor corneas from the Eye Bank of Hong Kong were obtained by conventional keratoplasty surgery. The donors were aged between 50 and 60 years. Human limbal epithelial cells were obtained from the corneoscleral rims by the following procedure. The limbal rim, approximately 1 mm in width, was trimmed, washed, and placed in Dulbecco's Modified Eagle's Medium (DMEM, Gibco, Rockville, MA, U.S.A.) containing 10% fetal bovine serum (Gibco). Tissue pieces were placed in culture dishes at 37°C and 5% CO 2 atm in SHEM (DMEM and Ham's F12 at a 1:1 vol/vol, supplemented with 5 mg/L transferrin, 5 mg/L insulin, 5 μg/L selenate, 0.1 mM ethanolamine, 0.1 mM o-phosphoethanolamine, 5 mg/L hydrocortisone, 0.1 mg/L cholera toxin α subunit, 10 μg/L EGF, 10 μg/L basic FGF, 0.5% dimethyl sulfoxide, 5% fetal bovine serum, and antibiotics).
Identification of Human Limbal Epithelial Cells
Human limbal epithelial cells migrated out from limbal tissues were harvested and grown on glass coverslips for 2 days before fixation in methanol. After blocking the nonspecificities, the cells were incubated with antibodies against α-enolase (Santa Cruz, Santa Cruz, CA, U.S.A.), EGF receptor (Sigma, Saint Louis, MO, U.S.A.), cytokeratin type 3 (AE5, Sigma), and proliferating cell nuclear antigen (PCNA) (Santa Cruz), followed by fluorescein-conjugated secondary antibody.
Human Limbal Epithelial Cells Cultivated on HAM
Human amniotic membrane was obtained by elective cesarean section and was preserved sterile in 50% glycerin at −80°C. Before culture, HAM was firstly treated with 1.2 U/mL Dispase II (Boehringer Mann, Basel, Switzerland) in Ca ++ Mg ++ -free Hank's balanced salt solution (Gibco) for 30 minutes at 37°C. The loosened epithelium was removed, and the basement membrane was cut into 3 × 3 cm 2 sheets and placed in culture insert (Corning, Acton, MA, U.S.A.) with the epithelial side facing up in SHEM for an overnight at 37°C. Primary human limbal epithelial cells were seeded to culture plastic or on the epithelial side of HAM. The seeding density in both conditions was 5 × 10 3 cells/cm 2 . After cell attachment, both cultures were washed twice with Hank's balanced salt solution and replaced with serum and growth factor-free SHEM supplemented with 1% bovine serum albumin. The culture media were sampled at 24, 48, and 96 hours. After centrifugation, the supernatant was stored in aliquots at −80°C before assay.
Growth Factor Measurement by Enzyme-Linked Immunosorbent Assay
Culture media were completely thawed at 4°C, and growth factors were assayed by enzyme-linked immunosorbent assay kits (Quantikine, R&D Systems, Minneapolis, MN, U.S.A.). Parallel measurement of standards and samples was performed by applying them to antibody-precoated microplate, and the procedure was followed as indicated by the commercial kits. For total TGFβ measurement, the samples were acidified before assay. Color intensity after the reaction was measured by a microplate reader (3550-UV, BIORAD, CA, U.S.A.) at 450 nm with correction of interference measured at 540 nm. The samples and standards were assayed in triplicate, and each plate was scanned for three times to obtain the mean and standard deviation. Growth factor concentrations were determined from the best linear curve drawn with the absorbance of standards versus their concentrations. Protein concentration was determined by Bradford method. The concentration of growth factor was expressed in pg/mg protein.
Monoclonal antibodies against EGF and TIMP1 (R&D Systems) were separately incubated with human limbal epithelial cells for 48 hours in serum and growth factor-free SHEM supplemented with 1% bovine serum albumin. The cells were then fixed and stained for α-enolase, EGF receptor, cytokeratin, and PCNA.
Human Limbal Epithelial Cells Cultivated on HAM
The human limbal epithelial cells, at an early stage of explant (1 to 2 days), were small and polygonal. The primary cells propagated fast and duplicated in approximately 20 hours. Positive staining of PCNA was observed in all nuclei, whereas most had EGF receptor in the membranous region. α-Enolase was expressed perinuclearly and the signal was intense toward the outer nuclear membrane (Fig. 1A–C). In subpassages, human limbal epithelial cells readily attached on culture plastic and the epithelial side of denuded HAM. After 5 days, the cells formed a monolayer and showed an epithelioid morphology (Fig. 1D). In a prolonged culture of 12 days, the cells in the confluent cell layer showed large cytoplasmic-to-nuclear volume ratio (Fig. 1E). The expression of nuclear PCNA, cytoplasmic α-enolase, and membranous EGFR in our primary human limbal epithelial cells indicated that they attained some characteristics of stem cells. 15,29–31
ELISA for Growth Factors
On the plastic and denuded HAM, similar increasing levels of EGF with time of culture (24, 48, and 96 hours) were detected (Fig. 2A). At 96 hours, EGF expression was reduced in both cultures, with a lower concentration observed on plastic (185 ± 40 pg/mg protein) than on denuded HAM (220 ± 50 pg/mg protein), although the difference was insignificant.
The culture of human limbal epithelial cells on plastic and on HAM showed a substantial and time-dependent increase in TIMP1 expression. Moreover, at 96 hours, there was more release of TIMP1 on HAM (1.98 ± 0.17 pg/mg protein) than on plastic (1.38 ± 0.42 pg/mg protein) (Fig. 2B). HAM alone under our culture conditions produced persistently less than 0.2 pg/mg protein of this growth factor at 24, 48, and 96 hours.
There was little secretion of TGFβ1 in human limbal epithelial cell culture. However, on HAM at 96 hours, the level of β1 isoform, approximately 35 ± 10 × 10 −3 pg/mg protein, was two-fold lower than 83 ± 22 × 10 −3 pg/mg protein on plastic (Fig. 2C). For TGFβ2, there was a slight decrease on HAM. The difference was less than one-fold at 96 hours (244 ± 23 × 10 −3 pg/mg protein on HAM compared with 363 ± 40 × 10 −3 pg/mg protein on plastic) (Fig. 2D). The experiments were performed in triplicate. Human limbal epithelial cells were obtained from corneoscleral rims of three different donors for separate limbal explant culture, cell cultivation on HAM, and ELISA of growth factors.
Growth Factor Effect on Human Limbal Epithelial Cells
With the depletion of EGF, the proliferation of human limbal epithelial cells was obviously slowed down, and the size of cell colonies was much smaller than in the control. The cells showed reduced staining of PCNA and α-enolase (Fig. 3B and D) when compared with the control (Fig. 3A and C). When TIMP1 was removed, the growth of human limbal epithelial cells was not much different from that observed in normal culture. However, the staining intensity of EGF receptor was much weaker (Fig. 3F) than in the control (Fig. 3E), showing that the presence of matrix metalloproteinases can reduce EGF receptor expression and may affect the downstream signaling cascade. Moreover, the staining of cytokeratin type 3 was different in TIMP1-depleted culture (Fig. 3H). The cytofilamentous protein was found dispersed in the cytoplasm, whereas in the control cells, it was filamentous and perinuclearly located (Fig. 3G).
Corneal epithelium is subjected to a constant process of cell renewal and differentiation. Cells in the uppermost layer are continuously sloughed off from the surface and replaced by basal cell proliferation and migration. Basal cells, with their conjuncture to the basement membrane, divide and are displaced into the suprabasal layers to become postmitotic and differentiated. Thus, the intrinsic environment in the basal region is determinative in regulating cell growth, proliferation, and postmitotic specialization. The epithelial side of amniotic basement membrane provides a substratum and facilitates adhesion, proliferation, and differentiation of cultured human limbal epithelial cells. 32–34 It inhibits protease activity and exhibits antifibrotic effect to the nearby tissues, providing a suitable substrate for epithelial cell growth. 35–37 Meanwhile, the membrane also produces growth factors, such as EGF, basic FGF, keratinocyte growth factor, and its receptor, and TGFα and TGFβ isoforms. 38 These cytokines stimulate proliferation of limbal and corneal epithelial cells by perpetuating the subpopulation of transient amplifying cells. 12 Without epithelium (i.e., in denuded HAM), their production is much reduced. 38 In this study, we showed the secretion of growth factors by human limbal epithelial cells cultured on plastic and on denuded HAM.
We found no statistical difference in EGF expression, irrespective of the culture substrata. However, more EGF was detected in prolonged culture under both conditions (Fig 2A). This indicates that with cell proliferation, either on plastic or on denuded HAM, the limbal cells can produce a substantial amount of its own EGF. We have shown that stripping off EGF in the culture medium resulted in much slower growth of limbal cells. The administration of extra EGF is helpful in maintaining human limbal epithelial cell growth in culture. 14 Although we showed production of EGF by human limbal epithelial cells with and without HAM in our culture system, its level (approximately 190 pg/mL on plastic and 270 pg/mL on HAM) would not be sufficient for healthy maintenance of human limbal epithelial cells. In fact, when we supplemented our human limbal epithelial cell culture on denuded HAM with EGF to ng/mL level, better growth was attained.
In contrast, variations in TGFβ and TIMP1 levels showed the influence of amniotic basement membrane on human limbal epithelial cell growth. In prolonged culture of 96 hours, release of TGFβ1and TGFβ2 decreased more rapidly in the presence of denuded HAM than without. However, this pattern was reversed for TIMP1. Because TIMPs are antagonists for matrix metalloproteinases, their presence may downregulate the catalytic activity of matrix metalloproteinases in the modulation of extracellular matrix during development, tissue remodeling, and wound healing. 17 It has been shown that the expression of TIMP1 and TIMP3 are paralleled to gelatinases, in which they were intensively stained in corneal epithelial cells in pterygial specimens, containing proliferative and inflammatory growth of limbal epithelial stem cells. 39 Furthermore, when we incubated our human limbal epithelial cells with anti-TIMP1 antibody, the expression of EGF receptor was reduced (Fig. 3F). This suggests that matrix metalloproteinases may affect human limbal epithelial cell growth through an EGF receptor-signaling pathway. In head and neck squamous cancer, overexpression of EGF receptor is a marker for carcinogenesis, and it has a positive correlation with matrix metalloproteinase-9 and an inverse correlation with TIMP1. 40 The slightly increased level of TIMP1 secretion by human limbal epithelial cells on denuded HAM indicates a possible role of amniotic basement membrane on HLE growth. TIMP1 expression can lead to sustained EGF effects through the enhanced expression of EGF receptor.
The decreased expression of TGFβ1 and TGFβ2 owing to the presence of HAM (Fig. 2C and D) persisted for 96 hours. The repression was more dramatic in TGFβ1, almost a two-fold decrease. We found that such a decrease coincided with human limbal epithelial cell proliferation and colony expansion on HAM in our culture. A similar finding was reported in a study by Tseng et al., in which the levels of TGFβ isoforms and their receptor type II in cultured human corneal and limbal fibroblasts were suppressed dramatically in 24 hours after being placed on amniotic stromal matrix. It was suggested to relate with the antiscarring effect exerted by HAM. 28 TGFβ signaling system can modulate cell proliferation. Furthermore, TGFβ isoforms are potent inhibitors in the synthesis of matrix degrading enzymes. 41 The downregulation in this signaling system may facilitate the regulation of TIMPs and related enzymes on matrix degradation. However, the actual association of these cytokines is yet to be determined.
In human limbal stem cell allografting, the donor cells have been shown to be progressively replaced by host cells. 42 This suggests that the donor graft, with the provision of suitable cytokine and growth environment, is likely to influence tissue repair in the healing mechanism and its temporal sequence. In this study, we showed changes in growth factor release by human limbal epithelial cells because of the presence of HAM. These variations would affect the growth, metabolism, and differentiation of cells by triggering various cascades of signal transduction and gene expression. The presence of amniotic membrane therefore may favor corneal epithelialization and thereby benefit ocular surface reconstruction.
1. Dua HS, Saini JS, Azuara-Blanco A, et al. Limbal stem cell deficiency: concept, etiology, clinical presentation, diagnosis and management. Ind J Ophthalmol 2000; 48: 83–92.
2. Rao SK, Rajagopal R, Sitalakshmi G, et al. Limbal allografting from related live donors for corneal surface reconstruction. Ophthalmology 1999; 106: 822–8.
3. Dua HS, Azuara-Blanco A. Allo-limbal transplantation in patients with limbal stem cell deficiency. Br J Ophthalmol 1999; 83: 414–9.
4. Dua HS, Azuara-Blanco A. Autologous limbal transplantation in patients with unilateral corneal stem cell deficiency. Br J Ophthalmol 2000; 84: 273–8.
5. Tsubota K, Satake, Y, Kaido M, et al. Treatment of severe ocular surface disorders with corneal epithelial stem cell transplantation. N Engl J Med 1999; 340: 1697–703.
6. Basti S, Rao SK. Current status of limbal conjunctival autograft. Curr Opin Ophthalmol 2000; 11: 224–32.
7. Tseng SCG. Concept and application of limbal stem cells. Eye 1989; 3: 141–57.
8. Shimazaki J. The evolution of lamellar keratoplasty. Curr Opin Ophthalmol 2000; 11: 217–23.
9. He YG, Alizadeh H, Kinoshita K, et al. Experimental transplantation of cultured human limbal and amniotic epithelial cells onto the corneal surface. Cornea 1999; 18: 570–9.
10. Tsai RJF, Li LM, Chen JK. Reconstruction of damaged corneas by transplantation of autologous limbal epithelial cells
. New Engl J Med 2000; 343: 86–93.
11. Kruse FE. Stem cells and corneal epithelial regeneration. Eye 1994; 8: 170–83.
12. Zieske JD. Perpetuation of stem cells in the eye. Eye 1994; 8: 163–9.
13. Hall PA, Watt FM. Stem cells: the generation and maintenance of cellular diversity. Development 1989; 106: 619–33.
14. Kruse FE, Tseng SCG. Growth factors modulate clonal growth and differentiation of cultured rabbit limbal and corneal epithelium. Invest Ophthalmol Vis Sci 1993; 34: 1963–76.
15. Zieske JD, Wasson M. Regional variation in distribution of EGF receptor in developing and adult corneal epithelium. J Cell Sci 1993; 106: 145–52.
16. Kruse FE, Tseng SCG. Inhibitory effect of TGFβ on corneal epithelial proliferation can be reversed by fibroblast coculture [abstract]. Invest Ophthalmol Vis Sci 1993; 34: 1011.
17. Massova I, Kotra LP, Fridman R, et al. Matrix metalloproteinases: structures, evolution and diversification. FASEB J 1998; 12: 1075–95.
18. Tsubota K, Satake Y, Ohyama M. Surgical reconstruction of the ocular surface in advanced ocular cicatricial pemphigoid and Stevens-Johnson syndrome. Am J Ophthalmol 1996; 122: 38–52.
19. Prabhasawat P, Barton K, Burkett G, et al. Comparison of conjunctival autografts, amniotic membrane
grafts and primary closure for pterygium excision. Ophthalmology 1997; 104: 974–85.
20. Kruse FE, Rohrschneider K, Volcker HE. Multilayer amniotic membrane
transplantation for reconstruction of deep corneal ulcers. Ophthalmology 1999; 106: 1504–11.
21. Meller D, Pires RT, Mack RJ, et al. Amniotic membrane
transplantation for acute chemical or thermal burns. Ophthalmology 2000; 107: 980–9.
22. Shimazaki J, Shinozaki N, Tsubota K. Transplantation of amniotic membrane
and limbal autograft for patients with recurrent pterygium associated with symblepharon. Br J Ophthalmol 1998; 82: 235–40.
23. Tseng SC, Prabhasawat P, Barton K, et al. Amniotic membrane
transplantation with or without limbal allografts for corneal surface reconstruction in patients with limbal stem cell deficiency. Arch Ophthalmol 1998; 116: 431–41.
24. Koizumi N, Inatomi T, Quantock AJ, et al. Amniotic membrane
as a substrate for cultivating limbal corneal epithelial cells for autologous transplantation in rabbits. Cornea 2000; 19: 65–71.
25. Schwab IR. Cultured corneal epithelia for ocular surface disease. Trans Ann Ophthalmol Soc 1999; 97: 891–986.
26. van Herendael BJ, Oberti C, Brosens I. Microanatomy of the human amniotic membrane
. A light microscopic, transmission and scanning electron microscopic study. Am J Obstet Gynecol 1978; 131: 872–80.
27. Sato H, Shimazaki N, Shimazaki J. Role of growth factors for ocular surface reconstruction after amniotic membrane
transplantation. Invest Ophthalmol Vis Sci 1998; 39: S428.
28. Tseng SCG, Li DQ, Ma X. Suppression of transforming growth factor
-beta isoforms, TGF-beta receptor type II and myofibroblast differentiation in cultured human corneal and limbal fibroblasts by amniotic membrane
matrix. J Cell Physiol 1999; 179: 325–35.
29. Zieske JD, Bukusoglu G, Yankauckas MA. Characterization of a potential marker of corneal epithelial stem cells. Invest Ophthalmol Vis Sci 1992; 33: 143–52.
30. Chung EH, DeGregorio PG, Wasson M, et al. Epithelial regeneration after limbus-to-limbus debridement. Expression of α-enolase in stem and transient amplifying cells. Invest Ophthalmol Vis Sci 1995; 36: 1336–43.
31. Gan L, van Setten G, Seregard S, et al. Proliferating cell nuclear antigen colocalization with corneal epithelial stem cells and involvement in physiological cell turnover. Acta Ophthalmol Scand 1995; 73: 491–5.
32. Khodadoust AA, Silverstein AM, Kenyon KR. Adhesion of regenerating corneal epithelium. The role of basement membrane. Am J Ophthalmol 1968; 65: 339–48.
33. Talmi YP, Sigler L, Inge E. Antibacterial properties of human amniotic membranes. Placenta 1991; 12: 285–8.
34. Boudreas N, Sympson CJ, Werb Z. Suppression of ICE and apoptosis in mammary epithelial cells by extracellular matrix. Science 1995; 267: 891–3.
35. Kim JS, Park SW, Kim JH. Temporary amniotic membrane
draft promotes healing and inhibits protease activity in corneal wound induced by alkali burns in rabbit. Invest Ophthalmol Vis Sci 1998; 39: S428.
36. Na BK, Hwang JH, Shin EJ. Analysis of human amniotic membrane
components as proteinase inhibitors for development of therapeutic agent of recalcitrant keratitis. Invest Ophthalmol Vis Sci 1998; 39: S90.
37. Tseng SCG, Li DQ, Ma X. Downregulation of TGFβ1, β2, β3 and TGFβ receptor II expression in human corneal fibroblasts by amniotic membrane
. Invest Ophthalmol Vis Sci 1998; 39: S428.
38. Koizumi N, Inatomi T, Sotozono C, et al. Growth factor
mRNA and protein in preserved human amniotic membrane
. Curr Eye Res 2000; 20: 173–7.
39. Di Girolamo N, Wakefield D, Coroneo MT. Differential expression of matrix metalloproteinases and their tissue inhibitors at the advancing pterygium head. Invest Ophthalmol Vis Sci 2000; 41: 4142–9.
40. O-Charoenrat P, Ryhs-Evans P, Modjtahedi H, et al. Overexpression of epidermal growth factor
receptor in human head and neck squamous carcinoma cell lines correlates with matrix metalloproteinase-9 expression and in vitro invasion. Int J Cancer 2000; 86: 307–17.
41. Hagedorn H, Sauer U, Schleicher E, et al. Expression of TGF-β1 protein and mRNA and the effect on the tissue remodeling in laryngeal carcinomas. Anticancer Res 1999; 19: 4265–72.
42. Williams KA, Brereton HM, Aggarwal R, et al. Use of DNA polymorphisms and the polymerase chain reaction to examine the survival of a human stem cell allograft. Am J Ophthalmol 1995; 120: 342–50.