Restoration of corneal epithelial barrier function and wound healing by substance P and IGF-1 in rats with capsaicin-induced neurotrophic keratopathy

PURPOSE: To investigate the effects of topical application of the combination of substance P (SP) and insulin-like growth factor (IGF)-1 on corneal epithelial barrier function and epithelial wound closure in rats with capsaicin-induced neurotrophic keratopathy. METHODS: Neonatal rats were injected subcutaneously with a single dose of capsaicin to induce neurotrophic keratopathy. Corneal epithelial barrier function was evaluated with an anterior fluorophotometer. Tear fluid secretion was measured by the Schirmer test. Corneal epithelial wound healing was determined by measurement of the size of the epithelial defect after debridement of the entire epithelium. The combination of SP (1 mM) and IGF-1 (1 micro g/mL) in phosphate-buffered saline was administered in eye drops six times daily. RESULTS: Corneal epithelial barrier function was impaired and corneal epithelial wound healing was delayed in rats injected with capsaicin. The application of eye drops containing the combination of SP and IGF-1 to capsaicin-injected rats resulted in a significant improvement in corneal epithelial barrier function compared with that apparent in capsaicin-injected animals that received eye drops containing vehicle alone. Such treatment with SP and IGF-1 also significantly increased the rate of corneal epithelial wound closure in capsaicin-injected animals. CONCLUSIONS: Topical application of the combination of SP and IGF-1 improved both corneal epithelial barrier function and epithelial wound healing in an animal model of neurotrophic keratopathy.     

Regeneration of rabbit cornea following excimer laser photorefractive keratectomy: a study on gap junctions, epithelial junctions and epidermal growth factor receptor expression in correlation with cell proliferation

Corneal wound repair was investigated in rabbits following excimer laser ablation of a 6 mm diameter and 90 microm deep disc. In the healing process particular attention was focused on the epithelium where gap junction expression and the rearrangement of desmosomes and hemidesmosomes were correlated with cell proliferation and epidermal growth factor receptor expression. Immunofluorescence-based confocal laser scanning microscopy, semithin resin section morphology and electron microscopy were utilized. In resting cornea two isotypes of gap junctions, confined to different regions in the same basal epithelial cells, were detected. Particulate connexin43 (alpha1) immunostaining was concentrated on the apical while the connexin26 type (beta2) in the baso-lateral cell membranes. This is the first report of connexin26 in the cornea. Connexin43 was found also in corneal keratocytes and endothelial cell. Since the two connexins do not form functioning heteromeric channels and have selective permeabilities they may serve alternative pathways for direct cell-cell communication in the basal cell layer. During regeneration both connexins were expressed throughout the corneal epithelium including the migrating cells. They also showed transient up-regulation 24 hr after wounding in the form of overlapping relocation to the upper cell layers. At this time, basal epithelial cells at the limbal region, adjacent to the wound and those migrating over the wounded area all expressed membrane bound epidermal growth factor receptor and they were highly proliferating. In conclusion, like in other stratified epithelia connexin26 is also expressed in the cornea. Transient up-regulation and relocation of connexins within the regenerating epithelium may reflect the involvement of direct cell-cell communication in corneal wound healing. Mitotic activity in the migrating corneal epithelial cells is also a novel finding which is probably the sign of the excessive demand for new epithelial cells in larger wounds not met alone by the proliferating limbal stock.     

The contribution of blood flow by the anterior ciliary arteries to the anterior segment in the primate eye

The rate and mode of corneal wound healing in severely diabetic rats were studied by light microscopy and scanning electron microscopy. Diabetes mellitus was induced in 52 rats by alloxan injection, and 52 nondiabetic rats were used as controls. After 3 weeks, a nonpenetrating razor-blade wound was made in the central cornea of both eyes in 48 diabetic and 48 normal rats. The incision passed through the epithelium and into the stroma. The effects of diabetes on the unwounded cornea were observed by comparison with corneas from eight unwounded rats (four diabetic and four normal). Whole corneas from wounded diabetic and normal rats were studied at 0, 1, 3, 6, 12 and 24 hr and at 2–7 days after wounding. The rate and mode of healing were not found to differ between diabetics and normals. The surfaces of corneal wounds in both groups appeared to be completely healed and indistinguishable from the surrounding unwounded epithelium after 24 hr. The epithelial cells involved in the initial healing process were derived primarily from the layer of wing cells which progressed across the wound close to the connective-tissue base. Only in the final stages of healing, after the wound had been filled by the deeper epithelial cells, did superficial epithelial cells migrate. There appeared to be more exfoliating superficial epithelial cells over the entire cornea in diabetic rats than in normals. Because the healing of central corneal incisions occurs initially and primarily by sliding of the deeper epithelial cells, and because the diabetic condition appears to be associated with increased exfoliation of surface cells, the healing of central incisions may be less affected by diabetes than the healing of defects of the whole corneal surface, where the superficial epithelial cells have been reported to be the main migratory cells in the initial healing process and where healing in diabetics is delayed.     

Propagation and phenotypic preservation of rabbit limbal epithelial cells on amniotic membrane

PURPOSE: To describe the phenotypic characteristics of a limbal epithelial cell sheet outgrowth from a limbal explant cultured on amniotic membrane. METHOD: Immunofluorescent staining and confocal microscopy were used to examine the expressions of p63, Ki-67, keratins 3 and 14, connexin 43, and the integrin alpha6/beta4 and alpha3/beta1 subunits in corneal and limbal tissues in a limbal explant and epithelial outgrowth cultured for 2 weeks on amniotic membrane. RESULTS: The expression patterns of p63, Ki-67, keratins, integrins, and connexin 43 in a limbal explant with an epithelial outgrowth cultured for 2 weeks on amniotic membrane resembled those in freshly prepared limbus. Moreover, the distribution of integrin subunits in positive cells of the limbal explant and its epithelial outgrowth was similar to that of the corneal epithelial cells during wound repair. CONCLUSIONS: The epithelial cell sheet grown from a limbal explant on amniotic membrane exhibited a phenotype similar to that of the limbus, suggesting that amniotic membrane is a substrate capable of supporting the propagation and preservation of p63-positive limbal epithelial cells.     

Loss of fenamate-activated K+ current from epithelial cells during corneal wound healing.

PURPOSE: The corneal epithelium provides a barrier between the external environment and the cornea. It also serves as an ion transporting epithelium. Because of its proximity with the external environment, the corneal epithelium is frequently injured through physical or chemical insult. The purpose of this study was to determine whether corneal epithelial cell whole-cell currents change during corneal wound healing as the author of the present study has previously reported for corneal keratocytes and endothelial cells. METHODS: Rabbit corneal epithelial cells were injured by scraping, heptanol exposure, or freezing. The epithelium was allowed to heal for 12 to 74 hours. Cells were dissociated from corneas, and whole-cell currents were examined using the amphotericin-perforated-patch technique. RESULTS: Cells from the wounded corneal groups had significantly increased capacitance values, indicating increased surface area compared with that of control cells. As previously reported, the primary control whole-cell current was a fenamate-activated K+ current. An inwardly rectifying K+ current and a Cl- current were also observed. In epithelial cells from heptanol-wounded corneas, these conductances were generally unchanged. In cells from scrape- and freeze-wounded corneas, however, the fenamate-activated current was absent or significantly attenuated. CONCLUSIONS: As they do in corneal keratocytes and endothelial cells, K+ channels disappear during some models of corneal epithelial wound healing. In addition, cell capacitance, a measurement of cell surface area, increases. These results suggest that substantial K+ channel activity is not required for in vivo epithelial cell proliferation during corneal wound healing.     

20%乙醇处理兔角膜后上皮增生和细胞凋亡的研究

目的探讨准分子激光角膜上皮瓣下磨镶术(LASEK)中采用20%乙醇浸润兔角膜40s后角膜上皮增生和角膜细胞凋亡情况与机械刮除角膜上皮后的异同。方法实验组42只新西兰大白兔,用直径为8mm的LASEK专用角膜上皮刀切割角膜上皮,20%的乙醇浸润单眼40s,机械刮除对侧眼中央8mm直径的角膜上皮,随机分7组,于术后0、4h,1、3、5、8、30d取材;6只兔眼为空白对照。角膜冰冻切片,行Ki67免疫组化检查和TUNEL检测,计数角膜中央前基质细胞。结果乙醇浸润后5d中央角膜上皮增生达峰值,术后1d周边角膜上皮增生达峰值;术后4h上皮刀口下方局限的角膜基质细胞TUNEL染色阳性,数量最多;各组角膜中央前基质细胞计数和空白对照比差异无统计学意义(P=0.68)。机械刮除角膜上皮后3d周边角膜上皮增生达峰值,其高于乙醇浸润后角膜上皮的增生峰值;术后4h可见大量中央前基质细胞TUNEL阳性;术后1d中央前基质细胞数量最少(P<0.05)。结论与机械刮除角膜上皮相比,20%乙醇浸润40s对角膜损伤轻,恢复快,乙醇浸润后的角膜上皮对基质细胞有保护作用。     

Cellular proliferation and leukocyte infiltration in the rabbit cornea after photorefractive keratectomy

PURPOSE: To map the proliferative activity of corneal cells during wound healing following photorefractive keratectomy (PRK) and to compare two markers for proliferation. METHODS: PRK, 5- mm in diameter with a -6 D setting, was performed in one eye of 28 New Zealand White Rabbits. The rabbits were sacrificed at time points between 12 hours and three months after surgery. The treated and fellow corneas were fixed in 10% formaldehyde, paraffin embedded, and immunohistochemically stained for proliferate cell nuclear antigen (PCNA) and at one time point, 1 week, also for Ki-67. RESULTS: Following initial sliding of the epithelial cells, the proliferative activity in the wound area starts in the leading edge (24 hours) and is spread towards the periphery. The proliferative activity peaks after one week and subsides during the following two weeks. Early (24 hours) proliferative activity is also seen in the limbal epithelium which peaks after three days. The keratocytes express PCNA in the peripheral stroma 48 hours after injury. They then also migrate to repopulate the stroma under the wound area. The expression period lasts 1 week and subsides the following week. Leukocytes are found in the wound as early as 12 hours after injury. The cells disappear around the time of epithelial wound closure, i.e. after 3 days. The two proliferative markers PCNA and KI 67 show a similar distribution after surgery. CONCLUSION: Epithelial proliferative activity starts earlier after injury, and is preceded by leukocyte presence in the wound. The PCNA expression starts later in the keratocytes but lasts somewhat longer (3 weeks). PCNA expression appears more efficient than Ki-67 to show proliferative activity of slow cycling cells in the cornea     

Expression of HGF, KGF, EGF and receptor messenger RNAs following corneal epithelial wounding

Hepatocyte growth factor (HGF), keratinocyte growth factor (KGF), epidermal growth factor (EGF), and their receptors have been associated with homeostasis and wound healing in the cornea. The purpose of this study was to examine the expression of the messenger RNAs for these growth factors and receptors in a wounded series of mouse corneas using in situ hybridization. In situ hybridization was performed with 3H-labeled riboprobes on unwounded corneas and corneas at 30 minutes, 4, 12, 24, 48 and 72 hr, and 7 days after epithelial scrape wounds in Balb/C mice. Qualitative and semi-quantitative analyses were performed. Expression of HGF, KGF and EGF mRNAs in keratocytes in the unwounded cornea was low. EGF mRNA was also expressed in unwounded corneal epithelium. Following wounding, however, these growth factor mRNAs were markedly upregulated in keratocytes. EGF mRNA expression in the epithelium appeared unaffected by wounding. At seven days after wounding and several days following closure of the epithelial defect, HGF mRNA and KGF mRNA were still expressed at higher levels in keratocytes compared with unwounded corneas. No difference in expression of HGF or KGF mRNAs between limbal, peripheral corneal, or central corneal keratocytes was noted in the unwounded cornea, KGF receptor mRNA was prominently expressed throughout the unwounded corneal epithelium. HGF receptor mRNA and EGF receptor mRNAs were expressed at low levels in unwounded cornea epithelium. Following scrape injury, expression of HGF receptor mRNA and KGF receptor mRNA were markedly upregulated in the corneal epithelium, while no significant increase in EGF receptor mRNA expression was noted. These studies suggest a prominent role for HGF and KGF in modulating corneal epithelial wound healing following injury. Less prominent changes in EGF mRNA and EGF receptor mRNA in the corneal epithelium following wounding may suggest that EGF has more of a role in homeostasis in the mouse corneal epithelium.     

Effects of substance P and IGF-1 in corneal epithelial barrier function and wound healing in a rat model of neurotrophic keratopathy

PURPOSE: To establish a rat model of neurotrophic keratopathy and to examine the effects of the combination of substance P (SP) and insulin-like growth factor (IGF)-1 on corneal epithelial barrier function and wound healing in this model. METHODS: Corneal denervation was achieved by thermocoagulation of the ophthalmic branch of the trigeminal nerve. A modified Schirmer test was performed without topical anesthesia. Corneal epithelial barrier function was assessed by measurement of fluorescein permeability with an anterior fluorophotometer. Epithelial wound healing was evaluated by measurement of the area of the defect at various times after removal of the entire epithelium. Eye drops containing both 1 mM SP and IGF-1 (1 micro g/mL) were administered six times daily. RESULTS: The Schirmer test result in eyes subjected to trigeminal denervation was lower than that in control eyes. The fluorescein permeability of the corneal epithelium of denervated eyes was increased relative to that of control eyes. Furthermore, trigeminal denervation induced a delay in corneal epithelial wound healing. Application of eye drops containing SP and IGF-1 to denervated corneas restored the fluorescein permeability of the corneal epithelium to control levels and abolished the delay in epithelial wound healing. CONCLUSIONS: A rat model of neurotrophic keratopathy, characterized by reduced tear secretion, loss of corneal sensation, impaired epithelial barrier function, and delayed epithelial wound healing, was established by trigeminal denervation. Treatment with both SP and IGF-1 improved corneal epithelial barrier function and stimulated corneal epithelial wound healing in this model.     

Development of a three-dimensional organ culture model for corneal wound healing and corneal transplantation

PURPOSE: To develop and evaluate a three-dimensional organ culture system of the cornea anterior chamber that could replicate the in vivo processes occurring during corneal wound healing and corneal transplantation. METHODS: Bovine corneoscleral buttons were clamped in a specially designed chamber through the sclera outside the limbus. The epithelium was exposed to air, and its anterior surface was automatically irrigated. The endothelial layer was perfused separately with media under normal intraocular pressure. Wound healing and corneal transplantation were observed using light, scanning, and transmission electron microscopy. RESULTS: The organ culture system maintained the epithelium, the putative epithelial stem cells in the limbus, the stroma, and the endothelium in good condition for the 10-day period during which the system was evaluated. The authors observed that the processes of wound healing and corneal transplantation in the model appeared similar to those occurring in vivo. CONCLUSIONS: In vitro model closely replicated the in vivo processes of wound healing and corneal transplantation. The authors believe this model will be useful for basic investigations into the cornea, such as study of the response of the cornea to surgery, wound healing, toxins, and therapeutic agents.     

A new in vitro corneal preparation to study epithelial wound healing.

Corneal epithelial wound healing is an important process necessary for maintenance of visual integrity. Corneal epithelial wound healing occurs by cellular migration and proliferation. However, the molecular basis of reepithelialization is not known. To investigate individual molecular contributions to the wound healing process, an in vitro corneal preparation comparable to the in vivo condition is needed. This investigation developed a new whole mount in vitro rabbit cornea preparation and studied epithelial wound healing rates for epithelial and subepithelial wounds. The wound closure rates obtained in this study for epithelial and subepithelial wound healing (52 +/- 14 microns/hr and 38 +/- 7 microns/hr, respectively) are comparable to in vivo rates of wound healing determined by other laboratories for rabbits. This preparation, achieved by functionally separating the epithelial and endothelial sides of the cornea, allows application of agents to the cornea in a manner that approximates the in vivo condition. This in vitro system is promising for future studies designed to investigate corneal wound healing while reducing potential ocular discomfort associated with in vivo corneal wounding.     

Platelet response to corneal abrasion is necessary for acute inflammation and efficient re-epithelialization

PURPOSE: Adhesion molecules play a critical role in leukocyte emigration to wound sites, but differences are evident in different vascular beds. In this study, the contributions of P-selectin to neutrophil emigration into the cornea after central epithelial abrasion were investigated. METHODS: Re-epithelialization, neutrophil influx, and platelet accumulation were assessed in C57BL/6 mice after removal of a 2-mm diameter area of central corneal epithelium that did not directly injure the limbal vessels or the avascular stroma of the cornea. Comparisons were made between wild-type (WT) mice and mice with targeted deletions of genes for P-selectin, CD18, or CD54, or mice with antibody-induced neutropenia or thrombocytopenia. RESULTS: After central corneal epithelial abrasion, platelets localized in the limbal vessels and neutrophils emigrated from the limbal vessels to the region of the epithelial wound. There was temporal correspondence of platelet and neutrophil localization, peaking within 12 hours of wounding. Platelet accumulation, neutrophil emigration and corneal epithelial healing as measured by wound closure, basal epithelial cell density, and epithelial cell division were significantly reduced in P-selectin-deficient mice (P-sel(-/-)). Anti-GP1balpha antibody-induced thrombocytopenia in WT mice significantly reduced platelet and neutrophil accumulation and wound healing. Passive transfer of wild-type platelets into P-sel(-/-) mice significantly restored platelet localization in limbal vessels, neutrophil emigration, epithelial cell division, and epithelial cell migration into the abraded region of the cornea. CONCLUSIONS: Platelet localization in the limbus of abraded corneas contributes to re-epithelialization, and P-selectin provides a necessary step in this process.     

Hyaluronan Facilitates Corneal Epithelial Wound Healing in Diabetic Rats

We investigated the effect of hyaluronan on corneal epithelial wound healing in rats affected by diabetes. Furthermore, because hyaluronan is thought to affect corneal epithelial wound healing through the mechanism of binding of hyaluronan to provisional fibronectin in the wounded area, we compared the localization of fibronectin immunohistochemically during corneal epithelial wound healing in diabetic and non-diabetic rats. Streptozotocin was used to induce diabetes in half the rats. Two weeks after treatment, the whole corneal epithelium of diabetic and untreated rats was debrided. The rats were divided into groups (seven or eight rats per group), and hyaluronan eye drops at concentrations of 0.03, 0.1, or 0.3%, chondroitin sulfate (3%), or phosphate buffered saline (PBS) was given in eye drops 6 times a day for 4 days, starting immediately after debridement. The area of the corneal epithelial wound was measured immediately after debridement and at 12, 18, 24, 30, 48, 72, and 96 hours afterwards. Although the healing process was similar in non-diabetic and diabetic rats, the healing rate in diabetic rats was slower than that in normal controls. In both diabetic and non-diabetic rats, hyaluronan increased the healing rate in a dose-dependent manner; the difference was significant compared with the PBS-treated group, at hyaluronan doses of 0.1% and 0.3%. However, chondroitin sulfate did not affect corneal epithelial wound closure, regardless of whether the rats were diabetic or not; the healing rates were identical to those of PBS-treated diabetic and non-diabetic controls. In both diabetic and non-diabetic corneas, fibronectin was localized in the corneal subepithelial region, and in streaks between collagen fibers of the stroma. One day after debridement, a layer of fibronectin immunofluorescence was clearly visible on the surface of the denuded stroma. As healing progressed staining of fibronectin diminished at the interface between the new epithelium and the stroma. These changes in localization of fibronectin during corneal epithelial wound healing were similar in both diabetic and non-diabetic rats. Our results demonstrate that hyaluronan facilitates corneal epithelial wound healing in diabetic rats, and suggest that one possible mechanism of its stimulatory effect lies in its binding to a provisional fibronectin matrix, in both diabetic and non-diabetic rats.     

In-situ porcine corneal matrix hydrogel as ocular surface bandage

PurposeBioactive substrates can be used therapeutically to enhance wound healing. Here, we evaluated the effect of an in-situ thermoresponsive hydrogel from decellularized porcine cornea ECM, COMatrix (COrnea Matrix), for application as an ocular surface bandage for corneal epithelial defects.MethodsCOMatrix hydrogel was fabricated from decellularized porcine corneas. The effects of COMatrix hydrogel on attachment and proliferation of human corneal epithelial cells (HCECs) were evaluated in vitro. The effect of COMatrix on the expressions of the inflammatory genes, IL-1β, TNF-α, and IL-6 was assessed by RT-PCR. The in-situ application and also repairing effects of COMatrix hydrogel as an ocular bandage was studied in a murine model of corneal epithelial wound. The eyes were examined by optical coherence tomography (OCT) and slit-lamp microscopy in vivo and by histology and immunofluorescence post-mortem.ResultsIn vitro, COMatrix hydrogel significantly enhanced the attachment and proliferation of HCECs relative to control. HCECs exposed to COMatrix had less induced expression of TNF-α (P < 0.05). In vivo, COMatrix formed a uniform hydrogel that adhered to the murine ocular surface after in-situ curing. Corneal epithelial wound closure was significantly accelerated by COMatrix hydrogel compared to control (P < 0.01). There was significant increase in the expression of proliferation marker Ki-67 in wounded corneal epithelium by COMatrix hydrogel compared to control (P < 0.05).ConclusionsCOMatrix hydrogel is a naturally derived bioactive material with potential application as an ocular surface bandage to enhance epithelial wound healing.     

Nerve growth factor and corneal wound healing in dogs

Nerve growth factor in the tear film and corneal epithelium is hypothesized to play an important role in ocular surface maintenance and corneal wound healing. The purpose of this study was to determine the expression of nerve growth factor and its high affinity (trkA) receptor in tears, cornea, and lacrimal glands of normal dogs, the modulation of nerve growth factor and its trkA receptor during corneal wound healing, and the effect of topical nerve growth factor application on canine corneal epithelial wound healing. In the first of three experiments, the nerve growth factor content of tears, corneal epithelium, lacrimal gland, and 3rd eyelid gland was determined in normal dogs by enzyme-linked immunosorbent assay and the expression of nerve growth factor and its trkA receptor were evaluated in the cornea and lacrimal glands by immunohistochemistry. In a second experiment, unilateral corneal epithelial defects were created, and tissues were evaluated for changes in nerve growth factor or trkA expression for 1 week. In a third experiment, bilateral corneal epithelial defects were created and the right eyes in each animal were treated 4 times daily with either recombinant human nerve growth factor, murine nerve growth factor, or nerve growth factor-blocking antibody. The results of this study showed that nerve growth factor levels in normal dog tears, corneal epithelium, third eyelid gland and lacrimal gland were 15.4+/-4.6 ng ml(-1), 33.5+/-12.3, 52.4+/-17.4 and 48.8+/-9.4 ng g(-1), respectively. NGF and trkA receptors were identified by immunohistochemistry in all tissues examined. After unilateral corneal wounding, nerve growth factor concentration increased in the tears bilaterally for 3 days, especially in the wounded eye, and then returned to pre-wounding values. Nerve growth factor content, and immunohistochemical staining for nerve growth factor and trkA, increased significantly in the ipsilateral cornea epithelium following unilateral wounding. Nerve growth factor concentrations in lacrimal and third eyelid glands also increased bilaterally (p<0.01) after unilateral wounding. Time to wound closure and rate of epithelial migration did not differ significantly between nerve growth factor-treated, nerve growth factor antibody-treated, and control eyes. In conclusion, nerve growth factor is present under resting physiologic conditions in normal canine tears, and nerve growth factor and its trkA receptor are present under resting conditions in normal canine corneal epithelium, lacrimal gland and third eyelid gland. Nerve growth factor is elevated in the tears, cornea, and lacrimal glands after corneal epithelial wounding; however, topical application of nerve growth factor, or its blocking antibody does not modulate corneal wound healing in the normal dog eye.     

Saxitoxin: an anesthetic of the deepithelialized rabbit cornea

PURPOSE: The aim of the current study was to define the efficacy of saxitoxin as a corneal anesthetic in rabbits after mechanical corneal abrasion and photorefractive keratectomy (PRK). METHODS: Twelve Dutch belted rabbits were given a single 1.2-microg dose of saxitoxin or vehicle after mechanical abrasion of the cornea. Corneal sensation was evaluated hourly for 6 hours. A second group of 12 Dutch belted rabbits was given a 1.2-microg dose of saxitoxin or vehicle every 5 hours for 30 hours after PRK. Corneal sensation was evaluated after 5, 10, 15, 20, 25, and 30 hours. Pachometry was performed before PRK and again after the epithelial defects had healed. The rate of epithelial defect closure was assessed by measuring the epithelial defect size 25, 42, 65, 88, and 113 hours after PRK. RESULTS: A dose of 1.2 microg of saxitoxin given every 5 hours produced continuous corneal anesthesia after PRK. There was no difference in the rate of wound healing between eyes treated with saxitoxin and vehicle. There was no difference in the degree of wound healing, as measured by pachometry, between eyes treated with saxitoxin and vehicle. There were no apparent ocular or systemic toxic effects from saxitoxin administration. CONCLUSION: At a dose of 1.2 microg, saxitoxin is a safe, effective, long-acting corneal anesthetic in rabbits after PRK.