Organ culture preservation for corneal tissue. Technical and quality aspects.

INTRODUCTION: The technical and quality aspects of organ culture as a storage method for human donor corneas are described. MATERIALS AND METHODS: Data electronically stored since 1989 of > 41,000 corneas, processed in the Cornea Bank Amsterdam, are analysed. The technical information of eye banks collected in the Directory of the European Eye Bank Association (EEBA) is used as comparison. European Union (EU) directive for tissue banking and EEBA technical guidelines are references for the quality aspects. RESULTS: Organ culture allows the storage of donor corneas up to 4-5weeks. The storage phase is followed by a generally much shorter phase of 1-7 days, to reverse the corneal swelling occurring in the first phase and to transport the tissue to the clinic. Selection of the corneas based on inspection of the endothelium after storage as well as microbiological testing of the storage solution after a quarantine period are mandatory for this technique. General agreement exists about the outline of the method, but technical variations are applied to suit local circumstances and preferences of corneal surgeons. Agreement exists about a minimum endothelial cell count as selection criterion in case the donor endothelium is meant to be grafted. The use and cutoff points of other selection parameters for the cornea, e.g. the endothelial cell mosaic, are varying. According to EU regulations, a quality management system should be installed. This way each bank is able to issue a standardized product, while the production process is monitored with quality registrations. With the clinical outcome of the graft, the quality of the selection and storage procedures is verified. With the notification of adverse reactions such as primary graft failure and endophthalmitis, minimum risks will be assessed. CONCLUSION: The organ-cultured cornea is a well-documented product concerning microbiological safety and quality of the tissue. However, variations in performance and materials and no definite cut-off points for selection do not make an organ-cultured cornea a generally standardized product. The corneal surgeons have to ascertain themselves of the safety and quality of the followed procedure. It is up to an organization such as the EEBA to formulate tissue-specific additions to the EU regulations such as training opportunities, technical guidelines and criteria based on science.

Destruction of murine lymphoma cells by allogeneic immune peritoneal macrophages in vitro: an ultrastructure study.

SL2 lymphoma cells from inbred DBA/2 mice were added to monolayers of peritoneal macrophages isolated from immunized inbred C57BL/10ScCr mice. Tumor cells were removed from the supernatant most rapidly during the first few hours. Once tumor cells were bound to the macrophages, however, their destruction apparently proceeded at a constant rate. Most tumor cells initially had only very small areas of contact with macrophages. The first tumor cell changes were loss of microvilli and the formation of surface blebs that ultimately detached from the cell. Subsequently, the tumor cell rapidly rounded up while the nucleus became pyknotic and the cytoplasm vacuolated; finally, the plasma membrane lost its integrity. It was only in this stage that the macrophages actively phagocytized the tumor cells. The recorded extracellular killing of tumor cells, followed by phagocytosis of their remnants, was compared with the widely divergent descriptions of macrophage-tumor cell interaction in the literature.

Destruction of murine lymphoma cells by allogeneic peritoneal macrophages in vitro: influence of antiserum.

Peritoneal macrophages isolated from intact inbred C57BL/10ScCr mice and from inbred C57BL/10ScCr mice previously immunized with the allogeneic SL2 lymphoma were incubated with SL2 cells with and without antiserum. The mode of interaction of the cells during the first 4 hours was studied ultrastructurally. Normal macrophages interacted with the tumor cells only when antiserum was present. The main feature was a partial enveloping of the tumor cells by thin lamellipodia without any visible cell damage. Immune macrophages without antiserum killed the tumor cells extracellularly by a process of apoptosis and subsequently phagocytized the cell remnants, as described previously. In the presence of antiserum, immune macrophages phagocytized intact tumor cells with very long, slender lamellipodia. Subsequently, the tumor cells underwent intracellular necrosis that seemed to be initiated by the release by the macrophage of complete lysosomes into the phagosome containing the tumor cell. In addition to altering the mode of interaction, antiserum greatly increased the degree of cytotoxicity.

Demonstration of migration inhibitory factor (MIF) in a murine system using myelomonocytic leukemia cells (WEHI-3).

Cells of the myelomonocytic tumor cell line, WEHI-3, were used as indicator cells in the indirect capillary test for the detection of migration inhibitory factor (MIF). A migration inhibition of about 50% was found and the results were highly reproducible. The indicator cells can be obtained in large quantities, as the myelomonocytic cells grow as an ascitic tumor in the peritoneal cavity of BALB/c mice.

Beneficial effect of HLA-DR matching on the survival of corneal allografts.

BACKGROUND: Although HLA typing and matching have been used for 3 decades, that practice has been poorly implemented in corneal transplantation, mainly because of inconclusive or contradictory analytical results. Consequently, we studied the immune response of corneal transplant recipients to HLA histoincompatibilities in a large homogeneous study. METHODS: All corneal transplantations performed by a single surgeon between 1976 and 1996 were studied. HLA-AB matching was used for recipient selection. All HLA typings were performed by a single experienced laboratory. Population genetic techniques were used to assess the validity of the HLA typings. Mono- and multivariate analyses were performed to identify the factors which significantly influence the survival of corneal allografts. Simulation studies were carried out to demonstrate the effects of mis-typed donor and recipient HLA-DR typings on analytical results. RESULTS: Retransplantation, degree of vascularization, HLA-AB and DR matching, endothelial cell count, graft size, recipient gender, and storage method were identified as significant factors by our monovariate analyses. A Cox proportional hazards survival analysis model identified degree of vascularization and HLA-AB and DR matching as significant prognostic factors when all immunological rejection episodes were used, P=0.000001. When only irreversible immunological rejection episodes were used, panel reactive antibodies, retransplantation, and number of rejection events were also identified, P=0.000001. Simulation studies showed that the effects of HLA-DR matching are abrogated by poor HLA-DR typings. CONCLUSIONS: Corneal allograft recipients have a normal alloimmune response to histoincompatibilities. Demonstration of that fact requires accurate HLA typings.

Effects of platelet-derived growth factor on endothelial wound healing of human corneas.

PURPOSE: To investigate the effects of platelet-derived growth factor (PDGF) on endothelial wound healing of organ-cultured human corneas. METHODS: The endothelia of paired human donor corneas (age, 71 +/- 11 years; total 84 pairs) were mechanically wounded (area, 5.6 +/- 0.8 mm2). Of each pair, one cornea was treated with 10 ng/ml human recombinant PDGF-BB while its mate served as control. The endothelial wound closure time was assessed by daily staining of the corneas with trypan blue. Morphometric data (endothelial cell density, shape, coefficient of variations) were obtained in the wound area after alizarin red staining. DNA synthesis was assessed using 3H-thymidine autoradiography. RESULTS: Although significant, the time of complete wound closure shortened only marginally on addition of PDGF to the culture medium. In the closed wound center (between 4 and 9 days), all corneas exposed to PDGF had significantly higher endothelial cell densities (737 +/- 126 cells/mm2) than the control corneas (515 +/- 89 cells/mm2). Fifteen days after wounding, the mean endothelial cell density averaged 526 +/- 93 and 708 +/- 135 cells/mm2 in the control and PDGF-treated groups, respectively. PDGF did not affect the final cell shape within the closed wounds. DNA synthesis was significantly but only marginally enhanced in PDGF-treated corneas. CONCLUSION: In organ-cultured human corneas, PDGF-BB promotes endothelial wound healing predominantly by cell migration, at least in corneas from senior donors.

Basic fibroblast growth factor stimulates corneal endothelial cell growth and endothelial wound healing of human corneas.

PURPOSE: To determine the dose response of human recombinant basic fibroblast growth factor (bFGF) on mitogenic activity, and the supplementary role of serum in cultured bovine and human corneal endothelial cells (BCECs, HCECs). To investigate the effect of bFGF on endothelial wound healing of human corneas in vitro. METHODS: In cell culture, DNA synthesis was assessed by 3H-thymidine incorporation. Wound healing was studied using paired human corneas after mechanical damaging of the endothelium. One cornea was treated with bFGF, and the mate served as control. Wound closure was determined after staining with trypan blue. Endothelial cell density (ECD) was assessed in the closed wound area after alizarin red staining. DNA synthesis was assessed using 3H-thymidine autoradiography. RESULTS: In cell culture, bFGF induced a dose-dependent mitogenic response on BCECs and HCECs. Addition of serum to the culture medium shifted the dose-response curve to considerably lower bFGF concentrations. In organ culture, the time of complete wound closure shortened only marginally (0.5 day) after bFGF treatment (P < 0.01). In the closed wound center, ECD was significantly higher in 1 ng/ml bFGF-treated corneas (686 +/- 134 cells/mm2) than in controls (554 +/- 117 cells/mm2), an increase of +25%. Doses of 0.1 and 10 ng/ml also were effective, but less so than with 1 ng/ml (+11% and +15%, respectively), whereas a dose of 100 ng/ml even had a negative effect (-11%). DNA synthesis was marginally enhanced in bFGF-treated (1 ng/ml) corneas. CONCLUSIONS: The maximal effective dose of bFGF producing a BCEC mitogenic response is dependent on serum. In human senior donor corneas, bFGF promotes endothelial wound healing predominantly by stimulation of cell migration.

Effects of human epidermal growth factor on endothelial wound healing of human corneas.

Paired human donor corneas (age, 73 +/- 12 yr), preserved in organ culture medium, were used to evaluate the effect of human epidermal growth factor (hEGF) on endothelial wound closure rate (WCR), on morphometric parameters (cell size, shape, and density), and on cell division in the wound area. The endothelium of the corneas was mechanically wounded (area, 4.9 +/- 0.9 mm2). For each pair, one cornea was treated with 10 ng/ml hEGF, while the mate served as control. WCR was assessed by daily staining of the corneas with trypan blue. Morphometric data were obtained after alizarin staining. Mitotic activity was assessed using 3H-thymidine autoradiography. Addition of hEGF significantly increased the WCR compared to the control group. In the closed wound (between 4-9 d), the mean cell size in the center averaged 1940 microns2 in the control group and 1287 microns2 in the hEGF-treated group (P less than 0.01). Fifteen days after wounding, the mean cell sizes averaged 1910 microns2 and 1427 microns2 in the control and hEGF-treated group, respectively (P less than 0.01). All corneas exposed to hEGF had higher endothelial cell densities than the control corneas. In the early stages of wound closure, the cells in the transitional zone in hEGF-treated corneas had a somewhat more elongated shape. However, hEGF did not affect the final cell shape within the closed wound. Autoradiographic results revealed that hEGF accelerated DNA-synthesis, although only to a limited extent. The results indicate that, in human corneas, hEGF promotes endothelial wound healing predominantly by cell migration, at least in corneas from senior donors.

Platelet-derived growth factor: receptor expression in corneas and effects on corneal cells.

PURPOSE: Platelet-derived growth factor (PDGF), a major mitogen and chemoattractant, is a dimeric molecule of disulfide-bonded A and/or B polypeptide chains (PDGF-AA/AB/BB). Two PDGF receptors (PDGFR) exist, alpha and beta, which dimerize after ligand exposure. The alpha-receptor binds both A- and B-chains, whereas the beta-receptor preferentially binds the B-chain. Whether PDGFR are present on, and whether PDGF is mitogenic for, corneal cells was investigated. METHODS: For receptor determination, a two-step immunoperoxidase technique with monoclonal antibodies against both alpha- and beta-receptors was applied on frozen sections of human and bovine corneas. To test the mitogenic activity of PDGF-BB, two proliferation assays, the DNA synthesis assay (3H-thymidine incorporation) and the colorimetric MTT assay, were used for cultured bovine corneal endothelial cells (BCEC) and human corneal fibroblast (HCF). RESULTS: Both receptors were present on epithelial cells, stromal fibroblasts, and endothelial cells, the beta-receptor being most abundant. In BCEC, minimal and maximal effects on DNA synthesis occurred at 10 ng/ml and 50-100 ng/ml PDGF, respectively. For HCF, the minimal and maximal effective doses were 1 ng/ml and 25-100 ng/ml of PDGF, respectively. The MTT assay, carried out in BCEC only, showed a minimal and maximal cell activity at 1 ng/ml and 10-100 ng/ml of PDGF, respectively. CONCLUSIONS: The presence of PDGFR in human corneal epithelium, fibroblasts, and endothelium and the mitogenic effects of PDGF on corneal cells indicate that PDGF may play a role in corneal wound healing.

Lymphocyte induced macrophage cytotoxicity: characterization of the macrophage cytotoxicity-inducing lymphocyte.

The phenotype of lymphocytes, obtained from mice immunized with allogeneic tumor cells, with the capacity to induce macrophage cytotoxicity was determined. Macrophage cytotoxicity was induced, either by incubating the macrophages with Macrophage Arming Factor (MAF) containing supernatants of cultures of sensitized lymphocytes and tumor cells (arming) or by incubating the macrophages directly with sensitized lymphocytes and tumor cells (activation). The MAF producing or activating capacity of the lymphocytes was not only "triggered" by the sensitizing tumor cells but also by normal cells and other tumor cells bearing the H-2 determinants of the sensitizing tumor cell. The capacity to render macrophages cytotoxic was not reduced after treatment of the lymphocytes with mitomycin-C or treatment with anti-murine Ig and complement. This capacity of the lymphocytes was abrogated after treatment with anti-T-cell serum or anti-Thy 1.2 serum and complement. After treatment with anti-Lyt 1 or anti-Lyt 2 serum and complement, the activating capacity was significantly reduced and the MAF producing capacity of the lymphocytes abrogated. Mixing the Lyt 1 depleted and Lyt 2 depleted lymphocytes or addition of normal lymphocytes to the Lyt 1 depleted or Lyt 2 depleted populations did not restore the MAF producing and activating capacities. This indicated that the lymphocytes inducing macrophage cytotoxicity in this allogeneic system are Lyt-1+2+ T-lymphocytes, which do not need to divide prior to perform their action.

Biodegradable three-dimensional networks of poly(dimethylamino ethyl methacrylate). Synthesis, characterization and in vitro studies of structural degradation and cytotoxicity.

In ophthalmology, there is a need for novel degradable biomaterials for e.g. controlled drug release in the vitreous body. These degradable materials should feature both excellent biocompatibility, and well-defined kinetics of degradation. In most cases, poly(D,L-lactic acid), or poly(lactic-co-glycolic acid) are used. These materials, however, suffer from some serious drawbacks, since the degradation kinetics are difficult to control, especially since the so-called 'burst-degradation' occurs. Here, we describe a set of novel polymeric networks which largely consist of poly(dimethylamino ethyl methacrylate) (poly(DMAEMA)); these materials are crosslinked via a dimethacrylate molecule that contains two carbonate groups. This system is susceptible to hydrolytic scission. The degradation products do not exert a catalytic effect on the ongoing degradation reaction (i.e. there is no burst effect). We describe the synthesis of three of these materials, which differ merely with regard to the crosslinker content. These materials were characterized through DMTA, 1H NMR and FT-IR spectroscopy, and scanning electron microscopy. The reaction DMAEMA + 2-hydroxyethyl methacrylate (HEMA) was studied in detail, using 1H NMR spectroscopy, and these experiments revealed that the reaction of DMAEMA and HEMA produces a random (Bernouillian-type) copolymer. From this, we contend that the new materials have more or less uniform distribution of the crosslinks throughout their volume. Structural degradation of the three materials was studied in vitro, at pH 7.4, 9.1 and 12.0. It is found that the materials exhibit smooth hydrolysis, which can be controlled via the crosslink density and the pH, as was expected a priori. It should be noted that degradation of these materials produces non-hydrolysable, but water-soluble, oligo(DMAEMA) and poly(DMAEMA) molecules. We subsequently performed in vitro studies on the biocompatibility of these materials. The MTT cytotoxicity assay revealed that the materials were cytotoxic to chondrosarcoma cells. This is most probably due to local increase of the pH due to the basic character of the pending dimethylamino groups. Cytotoxicity remained virtually unchanged after extended washing with water. This indicates that the cytotoxicity is an intrinsic property of the material and was not caused by remnants of free monomer. Cytotoxicity was also seen in cell cultures (human fibroblasts isolated from donor corneas) which were grown in contact with the materials. It is concluded that the new materials have attractive degradation characteristics, but their cytotoxicity makes them unsuitable for applications in ophthalmology.

Toxic effects of detergents on the corneal endothelium.

Eighteen patients developed a toxic endothelial cell destruction syndrome following normal intraocular surgery, caused by a detergent residue originating from irrigating cannulas. The residue occurred after the concentration of a detergent solution has been increased from 0.4% to 4%, in combination with insufficient cleaning of the cannulas. Mass spectrometric analysis revealed the detergent to contain a nonionic ethoxylated fatty alcohol (6% vol/vol). Quantitative endothelial vital staining and in vitro corneal endothelial perfusion demonstrated endothelial toxic effects at the 1% and 0.06% level for the detergent and the pure ethoxylated fatty alcohol, respectively. Permeability studies showed that the toxic effects occurred as a result of endothelial barrier breakdown.

Clinical comparison of grafts stored in McCarey-Kaufman medium at 4 degrees C and in corneal organ culture at 31 degrees C.

Twenty-eight paired human corneas were preserved in minimal essential medium at 31 degrees C and in McCarey-Kaufman medium at 4 degrees C. These grafts were then transplanted in pairs of patients with keratoconus who were age matched as closely as possible. These pairs received donor corneas from the same donor, so for each pair the donor age and time from death to preservation were the same. Visual acuity, central corneal thickness, and endothelial cell counts were compared. During the 1- to 2-year study period, no statistically significant difference in visual acuity, corneal thickness, or endothelial cell density was found between grafts stored in minimal essential medium and those stored in McCarey-Kaufman medium.

Electron microscopy of cultured human corneas. Osmotic hydration and the use of a dextran fraction (dextran T 500) in organ culture.

Human corneas were studied by means of electron microscopy after culture at 31 degrees C for two to 20 days in a medium containing 8% dextran T 500. Dextran T 500, a strong osmotic agent, was included in the culture medium to prevent excessive swelling of the cornea. In order to exclude the possibility that the observed effects were the result of osmotic changes during fixation, in each experiment, fixatives with different osmolalities were used (430, 574, 727, and 812 mOsm). After 8, 16, and 20 days of culture, vacuoles appeared that were filled with dextran; the cytoplasm was completely filled with these vacuoles. The vacuolization was not limited to the endothelium, but was also observed in stromal keratocytes and to a limited extent in the epithelium. It was concluded that the monolayer of endothelium was still intact after 20 days of culture in medium containing 8% dextran T 500, and, secondly, that the dextran might have been taken up by endocytosis. Whether or not the uptake of dextran has a long-term toxic effect on the corneal cells remains to be elucidated.

Toxic endothelial cell destruction of the cornea after routine extracapsular cataract surgery.

Eighteen patients developed an acute corneal decompensation following normal intraocular surgery (cataract extraction in 17 patients), characterized by star-shaped endothelial folds, a twofold increase in corneal thickness, and a visual acuity of counting fingers during several postoperative days. In some cases, there was an additional iritis and transient hypotony. There was no effect of topical and/or subconjunctival corticosteroids on the course of the decompensation. Endothelial morphometric analysis showed a mean endothelial cell loss of 72%. Endothelial wound healing, as determined by coefficient of variation and percentage hexagonals, stabilized 6 months postoperatively. We coined the term toxic endothelial cell destruction for this syndrome. Epidemiological evaluation revealed the toxic endothelial cell destruction syndrome to be linked with the 10-fold increase of a detergent solution in the ultrasonic bath for cleaning the surgical instruments.

Corneal endothelium and growth factors.

Human corneal endothelium is characterized by a low regenerative capacity, mainly because of a low mitotic activity, and therefore complete regeneration of the endothelial layer after injury is precluded. A decrease in endothelial cell density and an abnormal cell mosaic, which may occur after many conditions, can compromise the integrity of the endothelial monolayer, resulting in corneal decompensation with reduced vision and eventual need for penetrating keratoplasty. It would be beneficial to have growth factors that can help to maintain or restore the integrity of this delicate endothelial monolayer by maintaining or increasing the endothelial cell density or to stimulate the regeneration during wound healing. Growth factors represent a group of signalling peptides which influence diverse cellular processes, including cell proliferation, differentiation, migration and survival. One of the areas that has received great interest is its enhancement of wound healing. In this review the effects of three most effective growth factors (EGF, PDGF, FGF) on corneal endothelium, especially on wound healing in human corneal buttons, will be discussed.

Histopathologic effects of mitomycin C after trabeculectomy in human glaucomatous eyes with persistent hypotony.

We evaluated the histopathologic findings in seven patients who underwent surgical revision of the filtration site after trabeculectomy with mitomycin C because of persistent hypotonous maculopathy. Light microscopic examination of subconjunctival tissue and sclera demonstrated hypocellularity of fibroblasts and disruption of the normal architecture. Tissue fragments at the margin of the bleb wall demonstrated scarring and contained multiple fibroblasts. Additionally, we investigated the histopathologic changes in an eye obtained from a patient who died one week after a trabeculectomy with mitomycin C. Transmission electron microscopy showed myelin figures, increased melanolipofuscin granules, vacuolated cytoplasm, and disrupted mitochondria of the ciliary body epithelium underlying the site of mitomycin C application. On the basis of these findings, both overfiltration because of tissue disorganization of the filtering bleb and aqueous hyposecretion because of ciliary body toxicity might be involved in the causes of persistent hypotony after mitomycin C trabeculectomy.

Histocompatibility and corneal transplantation.

BACKGROUND: HLA typing and matching have been poorly implemented in corneal transplantation, mainly because of inconclusive or contradictory analytical results. Consequently, we studied the immune response of corneal transplant recipients to HLA histoincompatibilities in a large homogeneous study. METHODS: All corneal transplantations were performed by a single surgeon in a single center between 1976 and 1996. Population genetic and other statistical analyses were performed. Simulation studies assessed the effects of HLA-DR mistypings on analytical results. RESULTS: Mono- and multivariate analyses identified retransplantation, degree of vascularization, HLA-AB and -DR match grades, endothelial cell count, graft size, recipient gender, storage method and panel-reactive antibodies as significantly influencing the survival of corneal transplants. Simulation studies showed that the beneficial effect of HLA-DR matching is abrogated by HLA-DR mistypings. CONCLUSIONS: Corneal transplant recipients have a normal immune response to HLA incompatibilities. Demonstration of that fact requires accurate HLA typings.

Microbial decontamination of human donor eyes with povidone-iodine: penetration, toxicity, and effectiveness.

BACKGROUND/AIMS: Povidone-iodine (PVP -I) is applied for microbial decontamination of human eyes donated for transplantation. Concentrations and immersion times vary greatly. The effectiveness and toxicity of PVP-I were assessed for different decontamination protocols. METHODS: Human donor eyes and corneas were immersed in different concentrations (5-100 mg/ml) of PVP-I for different times (2-30 minutes). The penetration of iodine into the corneal tissue was assessed by x ray microanalysis. Microbial contamination was determined by taking cultures of the limbal areas and storage solutions and by incubation of the corneoscleral buttons in antibiotic-free culture medium. Cytotoxicity of PVP-I for corneal fibroblasts in culture was assessed using the MTT assay. RESULTS: Depending on concentration and immersion time iodine was found to penetrate into the epithelium, Bowman's layer, and stroma in amounts equivalent to 2-40 mg/ml PVP-I. The MTT assay demonstrated that 2.5 mg/ml PVP-I caused total damage to fibroblasts in vitro. Rinsing eyes with tap water and subsequent immersion in PVP-I reduced the rate of contamination from 82 out of 106 to 69 out of 106 and 37 out of 106, respectively. Antibiotics in the storage medium further reduced contamination from about 40% to 3%. Microbial contamination was not reduced by increasing the concentration and immersion times beyond 5 mg/ml PVP-I for 2 minutes. CONCLUSION: Immersion of human donor eyes in 5 mg/ml PVP-I solution for 2 minutes significantly reduces microbial contamination of donor corneas without relevant penetration of iodine into the corneal layers. Higher PVP-I concentrations and longer immersion times do not further reduce contamination, whereas the amount of iodine penetrating the corneal layers is elevated above the level cytotoxic for corneal fibroblasts. In view of this, concentrations above 5 mg/ml of PVP-I and immersion periods over 2 minutes are not recommended for reduction of the contamination rate of donor eyes.

The specific architecture of the anterior stroma accounts for maintenance of corneal curvature.

AIM: To analyse the human corneal stroma in extreme hydration to discover if its structure is responsible for corneal stability. METHODS: Corneas in several hydration states were used: postmortem control corneas (PM; n=3), corneas left for 1 day in phosphate buffered saline (PBS; n=4), and corneas left for 1 day (n=4), 2 days (n=4), 3 days (n=2), and 4 days (n=4) in deionised water. All corneas were fixed under standardised conditions and processed for light and electron microscopy. In addition, two fresh corneas from the operating theatre were studied which were processed 6 months after storage in sodium cacodylate buffer. RESULTS: After 1 day in deionised water maximal stromal swelling was reached which did not change up to 4 days. The stroma of deionised water corneas (1400 microm) was much thicker than that of PBS corneas (650 microm) and PM corneas (450 microm). Deionised water treatment led to disappearance of all keratocytes leaving only remnants of nuclei and large interlamellar spaces. In these specimens the distance between the collagen fibres had increased significantly, but the diameter of the collagen fibres did not seem to be affected. A remarkable observation was that the most anterior part of the stroma (100-120 microm) in all deionised water specimens and those stored for 6 months in buffer was not swollen, indicating that the tightly interwoven anterior lamellae are resistant to extreme non-physiological hydration states. CONCLUSIONS: The rigidity of the most anterior part of the corneal stroma in extreme hydration states points to an important role in maintenance of corneal curvature. Since a large part of this rigid anterior part of the stroma is either removed (PRK) or intersected (LASIK), it is possible that in the long run patients who underwent refractive surgery may be confronted with optical problems.