A 76-bp deletion in the Mip gene causes autosomal dominant cataract in Hfi mice.

Hfi is a dominant cataract mutation where heterozygotes show hydropic lens fibers and homozygotes show total lens opacity. The Hfi locus was mapped to the distal part of mouse chromosome 10 close to the major intrinsic protein (Mip), which is expressed only in cell membranes of lens fibers. Molecular analysis of Mip revealed a 76-bp deletion that resulted in exon 2 skipping in Mip mRNA. In Hfi/Hfi this deletion resulted in a complete absence of the wildtype Mip. In contrast, Hfi/+ animals had the same amount of wildtype Mip as +/+. Results from pulse-chase expression studies excluded hetero-oligomerization of wildtype and mutant Mip as a possible mechanism for cataract formation in the Hfi/+. We propose that the cataract phenotype in the Hfi heterozygote mutant is due to a detrimental gain of function by the mutant Mip resulting in either cytotoxicity or disruption in processing of other proteins important for the lens. Cataract formation in the Hfi/Hfi mouse is probably a combined result of both the complete loss of wildtype Mip and a gain of function of the mutant Mip.

Abnormal eye development associated with Cat4a, a dominant mouse cataract mutation on chromosome 8.

PURPOSE: Cat4a, one of four mutant alleles at the mouse Cat4 locus, causes central corneal opacity and anterior polar cataract in heterozygotes and microphthalmia in homozygotes. The Cat4 locus has been mapped to chromosome 8, 31 cM from the centromere. In this study ocular development of Cat4a mutant mice was investigated to characterize the defects in eye morphogenesis. METHODS: Serial sections from eyes of wild-type, heterozygous, and homozygous littermates were examined by means of light microscopy at selected intervals from embryonic day 11 to postnatal day 1. Eyes of adult heterozygous and homozygous mice also were evaluated histologically. RESULTS: Failure of separation of the lens vesicle from the surface ectoderm was the earliest structural defect observed. In heterozygous embryos, the abnormality was limited to persistent connection of the anterior pole of the lens to the cornea. Adult heterozygotes had defects in the central corneal stroma and endothelium and anterior polar cataracts with or without keratolenticular adhesion. In homozygous embryos, the persistent connection of lens to surface ectoderm was associated with aborted lens development, failure of closure of the optic fissure, and impairment of growth of the eyecup. Microphthalmic eyes of adult homozygous mice had a poorly developed cornea, and the anterior chamber and vitreous compartment were absent. An extensively folded retina and remnants of a degenerated lens filled the interior of the globe. CONCLUSIONS: A developmental defect inhibits separation of the lens vesicle from surface ectoderm in mice heterozygous or homozygous for the Cat4a mutation. In homozygotes subsequent lens and eye morphogenesis are also severely affected. Cat4a shows phenotypical similarity to several other independent mouse mutations including Small eye, a mutation of the Pax6 gene. Cat4 may be one of several genes involved in a common developmental path and may be part of the Pax6-regulated gene cascade governing eye morphogenesis.

Neuropeptide Y-like immunoreactivity localizes to preganglionic axon terminals in the rhesus monkey ciliary ganglion.

PURPOSE: To characterize neuropeptide distribution in the ciliary ganglion of rhesus monkeys (Macaca mulatta). METHODS: Cryostat tissue sections of fixed rhesus monkey ciliary, pterygopalatine, superior cervical, and trigeminal ganglia were incubated with antisera to neuropeptide Y (NPY), calcitonin gene-related peptide (CGRP), substance P (SP), vasoactive intestinal peptide (VIP), tyrosine hydroxylase (TH), and dopamine-beta-hydroxylase (DBH). Antibody binding was visualized by indirect immunofluorescence. RESULTS: NPY-like immunoreactive (LI) nerve terminals surrounded 80% of ciliary ganglion cells, but ciliary ganglion cell somata were unstained. NPY-LI cells were present in the superior cervical ganglion, in which almost all cells were TH- and DBH-LI, and in the pterygopalatine ganglion, in which almost all cells were VIP-LI. Because neither TH, DBH, nor VIP immunoreactivity was detected in nerves contacting ciliary ganglion cells, the NPY-LI input to ciliary neurons does not likely derive from the autonomic ganglia. The trigeminal ganglion, another potential source, had no NPY-LI neurons. CGRP- and SP-LI axons from the nasociliary nerve traversed the ciliary ganglion; a small number of varicose axons were distributed among ganglion cells and rarely surrounded cell somata. Most ciliary ganglion cells were TH-LI, but only a few were DBH-LI. CONCLUSIONS: Based on these patterns of peptide immunoreactivities, the NPY-LI nerve fibers investing ciliary ganglion cells in the rhesus monkey are most likely preganglionic axon terminals of mesencephalic parasympathetic neurons. Although the origin and function of these NPY-LI nerves remains to be established, the present finding adds to the remarkable diversity of neuropeptide immunoreactivity so far identified in preganglionic and postganglionic cells of the ciliary ganglion in different species of birds and mammals, including primates.

Mapping of the autosomal dominant cataract mutation (Coc) on mouse chromosome 16.

PURPOSE: To characterize the mouse cataract mutation Coc. METHODS: Coc is an X-radiation-induced autosomal dominant cataract mutation maintained on a murine C3H inbred strain. The affected heterozygotes were outcrossed to C57BL/6, and (C3H Coc/+ x C57BL/6) mice that were Coc/+ were then backcrossed to C57BL/6 to generate a panel of 103 progeny for mapping. For linkage analysis, microsatellites from each autosome were selected. The maximum distance between markers was 30 centimorgans (cM). RESULTS: The initial genome-wide screen of 14 backcrossed progeny indicated that the Coc locus resides on chromosome 16. Further mapping with additional markers from chromosome 16 for all 103 backcrossed progeny positioned Coc between markers D16Mit134 and D16Mit63. This region is syntenic to human chromosome 3. CONCLUSIONS: Mapping of the Coc locus to mouse chromosome 16 provides the positional information necessary to identify the candidate gene responsible for the Coc phenotype. The molecular characterization of the gene disrupted in the Coc mutation will provide insight into the mechanisms involved in cataract formation.

Nitric oxide synthase in autonomic innervation of the cat carotid body.

In the cat carotid body, nitric oxide synthase (NOS) immunoreactivity and NADPH diaphorase activity localize in nerve fibers mainly associated with blood vessels and occasionally lying close to glomus cells. The NOS-positive innervation originates in part from multipolar ganglion cells scattered in and around the carotid body and in the glossopharyngeal nerve. In the superior cervical ganglion, NOS and diaphorase staining localizes to many preganglionic axons and also to a small population of vasoactive intestinal peptide-positive, presumably cholinergic, ganglion cells. Positively stained ganglion cells are absent in the petrosal ganglion and very rare in the nodose ganglion, although both sensory ganglia display characteristic distributions of cells immunoreactive for calcitonin gene-related peptide, substance P and tyrosine hydroxylase. The NOS-positive innervation of the carotid body thus appears to be autonomic, originating mainly from a population of dispersed ganglion cells, and probably parasympathetic in nature. The superior cervical ganglion also may supply some pre- or postganglionic NOS-positive axons. Nitric oxide released from these nerves could affect glomus cell activity directly or indirectly by vasoregulation.

Helospectin-like immunoreactivity in the rat eye and pterygopalatine ganglion.

Helospectin I and II, nearly identical peptides isolated from lizard venom, show close sequence homology with vasoactive intestinal peptide (VIP). In mammals, the helospectins have been localized immunohistochemically to neurons of the brain, digestive tract and respiratory system, commonly co-existing with VIP. Using an antiserum that recognizes both forms of the peptide, we localized helospectin-like immunoreactivity in the rat eye and pterygopalatine ganglion. In the eye, helospectin-positive nerve fibers were evident mostly in the choroid, associated with small arterial vessels or distributed diffusely in the stroma; they were only occasionally seen in the anterior uvea. All of the helospectin nerve fibers also appeared to be immunoreactive for VIP. In the pterygopalatine ganglion, the principal source of VIP-containing innervation of the posterior uvea, approximately 25% of ganglion cells were helospectin-positive; all helospectin-reactive cells were also strongly positive for VIP. This immunohistochemical localization of helospectin indicates possible involvement in ocular autonomic functions, particularly regulation of blood flow.

Galanin immunoreactivity in autonomic innervation of the cat eye.

In an immunohistochemical study, we find that galanin is much more widely distributed in the peripheral innervation of the cat eye than in other animals so far examined. Previous studies of rat and pig eyes have revealed sparse galanin-positive nerves that presumably originate in the trigeminal ganglion. In contrast, the cat has a rich supply of galanin-containing nerve fibers throughout the uvea. Galanin-positive varicose nerves concentrate densely in iris muscles and distribute more sparsely in the ciliary muscle. The ciliary processes have a plexus of galanin-positive nerves underlying the ciliary epithelium at their base and positive nerve fibers coursing within their stroma. The ciliary artery and its branch vessels in the uvea are invested with a dense plexus of galanin-positive nerves. All autonomic ganglia supplying the eye contain cells that express galanin. It is present in 97% of superior cervical ganglion cells, coexisting with both tyrosine hydroxylase and neuropeptide Y; in 80% of pterygopalatine ganglion cells, most of which also contain vasoactive intestinal peptide; and in approximately 25% of ciliary ganglion cells. After unilateral superior cervical ganglionectomy, galanin-positive nerves almost totally disappear from the iris muscles, demonstrating that they are predominantly of sympathetic origin. Galanin-positive nerves investing the ciliary artery and choroidal blood vessels are not detectably reduced by sympathectomy, indicating that perivascular parasympathetic nerves from the pterygopalatine ganglion also express galanin. Other galanin-containing nerves in the eye can originate from the trigeminal and ciliary ganglia. The prominence of galanin in the ocular autonomic innervation of the cat provides an opportunity to explore the physiological effects of this neuropeptide in the eye.

Lens development in a dominant X-linked congenital cataract of the mouse.

Xcat is a recently identified mouse mutation causing X-linked dominant congenital cataract. The mutation is of particular interest as a possible animal model for the human X-linked cataract syndrome. Using light microscopy, we examined the histological changes of mutant lenses at selected intervals between embryonic (E) day 14 and postnatal (P) day 21. At E14, primary fiber formation completely fills the former lens vesicle in both normal and mutant mice, but in affected animals the primary fibers are irregularly arranged and show small foci of cellular disintegration. Progressive degeneration of primary fibers occurs from E15 to E18 and, during late gestation, secondary lens fibers also begin to degenerate. The lens epithelium and newly differentiated fibers, however, show no evident abnormality. Postnatally, most of the lens substance becomes amorphous; the cataractous process terminates in rupture of the posterior lens capsule by P21. Analysis of crystallin and cytoskeletal proteins of postnatal cataractous lenses revealed no significant abnormalities when compared to normal lenses. The observed sequence of histological changes indicates that the Xcat mutation affects the differentiation of lens fiber cells at some point after their initial elongation.

Expansion of the retinal pigment epithelium in experimental myopia.

Expansion of the retinal pigment epithelium was studied in neonatal chicks after one or two weeks of unilateral form vision deprivation to investigate altered ocular growth mechanisms in this experimental model of myopia. The area of individual retinal pigment epithelial (RPE) cells, measured in tangential sections, was greater in myopic eyes than in contralateral control eyes at both times. The mean RPE cell area in myopic eyes increased to the same extent as the area of the retinal pigment epithelium as a whole. In control eyes between one and two weeks, RPE cell expansion occurred predominantly in the periphery; in myopic eyes, it occurred more generally across the epithelium but was less pronounced in the temporal region. Given the absence of detectable mitotic figures in control and myopic eyes, expansion of the epithelial layer is attributable to passive stretch or growth of existing cells. Whether scleral growth or stretch occurs selectively beneath the areas of more pronounced RPE cell expansion is unknown.

An immunohistochemically distinct population of cat ciliary ganglion cells.

In the cat ciliary ganglion, 16% and 23% respectively of the neurons are surrounded by nerve fibers immunoreactive to calcitonin gene-related peptide (CGRP) and somatostatin (SOM). One-third of the ganglion cell perikarya are immunoreactive to neuropeptide Y (NPY). Analysis of the coincidence of immunoreactivities shows a striking correlation. Practically all of the ganglion cells surrounded by CGRP-like immunoreactive (LI) nerve fibers also are surrounded by SOM-LI nerve processes and demonstrate NPY-LI perikarya. These observations define a subset of NPY-LI ciliary ganglion cells with a particular peptidergic input and perhaps a specific physiologic function.

Peptide immunoreactivity of the ciliary ganglion and its accessory cells in the rat.

By means of immunohistochemistry, calcitonin gene-related peptide, Leu-enkephalin and neuropeptide Y localize to rat ciliary and accessory ganglion cells. The proportion of ciliary and accessory neurons immunoreactive to each peptide is provided and compared to previous data for vasoactive intestinal peptide. These findings indicate considerable neurochemical complexity for a parasympathetic ganglion with a small cell population.

Carboxyfluorescein distribution in ocular tissues of normal and diabetic rats.

Quantitative fluorescence microscopy was used to study carboxyfluorescein distribution across the blood-ocular barrier of control and streptozotocin diabetic rats at 2 min, 1 and 2 hr after dye injection (125 mg/kg iv). Measurement of dye concentrations in plasma, urine and feces demonstrated increased plasma clearance and increased urinary clearance of carboxyfluorescein in diabetic rats. Unbound plasma dye concentration in the diabetic animals fell to 34% of the control level at 1 hr after injection; the corresponding plasma concentration vs. time integral was reduced to only 74% of the control value. The presence of a less fluorescent glucuronide conjugate of carboxyfluorescein was not detected in plasma, urine or feces. Fluorescence intensity in the ocular tissues measured, including choriocapillaris, pigment epithelium, retina, ciliary epithelium, iris, and cornea, was not higher for diabetic than for control rats. In addition, there was no indication of localized dye leakage into retina through defects in the pigment epithelial and vascular endothelial barriers or of increased dye entry at the optic disc, a site of blood-retinal barrier discontinuity. Normalization of tissue fluorescence intensity measurements at the different time intervals to compensate for disparity in concurrent plasma dye concentrations resulted in significantly higher levels in diabetic retinas at 1 hr. However, because this difference was not manifest when the plasma dye concentration vs. time integral was used to normalize the data, it is concluded that no greater accumulation of carboxyfluorescein occurs in the retina of diabetic rats over the time period studied.

Nerve pathways between the pterygopalatine ganglion and eye in cats.

By dissection of thiocholine-stained orbital preparations, it has been determined that three different nerve pathways link the pterygopalatine ganglion and the eye in cats. 1) Nerves from the proximal half of the ganglion join a plexus of nerves and ganglion cells in the rete mirabile of the maxillary artery. Branches of the internal carotid nerve also supply this plexus. Fine nerves from the plexus travel to the optic nerve and then to the eye, accompanying both the nasociliary nerve that passes through the rete and the ciliary arteries that arise from the rete. 2) One or more nerves from the nerve of the pterygoid canal and from a prominent accessory ganglion near the orbital apex course to the inferior optic nerve surface at the optic foramen; these then run distally along the optic nerve to fuse with ciliary nerves or to accompany ciliary arteries entering the eye. 3) Other nerves from the pterygopalatine ganglion travel medially around the extraocular muscle cone to join the ethmoidal and infratrochlear branches of the nasociliary nerve; some nerves from the ganglion then take a retrograde course to the optic nerve, where they join ciliary nerves or arteries to the eye. All three pathways may transmit sympathetic, parasympathetic and somatic sensory nerve fibers.

Carboxyfluorescein transfer across the blood-retinal barrier evaluated by quantitative fluorescence microscopy: comparison with fluorescein.

Carboxyfluorescein levels in ocular tissues of normal rats were measured using quantitative fluorescence microscopy and compared with fluorescein levels to determine the extent to which blood-retinal barrier permeability is affected by the difference in lipid solubility of these two dyes. Retinal fluorescence intensity measurements at 2 min after i.v. dye injection were very much lower for carboxyfluorescein than for fluorescein despite similar plasma free dye concentrations. Marked leakage of dye from the optic disc into peripapillary retina was identified. At 1- and 2 hr, retinal levels of the two dyes became more similar, because fluorescein was removed from retina faster than carboxyfluorescein. After sodium-iodate-induced damage of the pigment epithelium, high levels of both dyes were evident in retina, but carboxyfluorescein was localized chiefly within extracellular space whereas fluorescein also densely stained cell somata. The fluorescence intensity levels recorded, which are proportional to the total mass of dye in the tissue, were correspondingly lower for carboxyfluorescein than for fluorescein, indicating that they were markedly affected by the different distribution of the two dyes. To relate tissue fluorescence intensity directly to dye concentration in the extracellular fluid, measurements were obtained from isolated retinas incubated in dye solutions of known concentration. Log-log plots demonstrated a linear relation between fluorescence intensity and medium concentration for both dyes, but retinal fluorescence of carboxyfluorescein, in correspondence with its limited distribution in the tissue space, was consistently less than that of fluorescein. The ratio of carboxyfluorescein to fluorescein fluorescence varied with the retinal layers but was constant for each layer over the concentration range tested. These fluorescence intensity ratios then were used to adjust the in vivo data so that comparison between the two dyes more closely reflected their extracellular dye concentration. With this correction the amount of carboxyfluorescein present in outer retina shortly after dye injection was approx. 1/10 that of fluorescein, indicating that carboxyfluorescein penetrates the pigment epithelium less readily than fluorescein, as expected from the difference in lipid solubility of the two dyes. However, fluorescence of both dyes in retina and presumably in vitreous humor eventually reached similar levels. This is attributed to entry of the dyes at sites of barrier discontinuity, as at the optic disc, and by a difference in their rates of removal from the intraocular compartment.

Autonomic neurons supplying the rat eye and the intraorbital distribution of vasoactive intestinal polypeptide (VIP)-like immunoreactivity.

We traced the origin and path of autonomic nerves to the rat eye using, as an aid to dissection, a modified thiocholine method for the histochemical demonstration of cholinesterase. When applied to whole nerves and ganglia supplying the rat eye, this procedure is not specific for cholinergic neurons; instead it stains both sympathetic and parasympathetic nerves, many of which are otherwise too fine to identify in dissection. We found that nerves from the superior cervical and pterygopalatine ganglia form a plexus at the orbital apex corresponding to the retro-orbital plexus found in rabbit, monkey and man. In the rat, nerves from the retro-orbital plexus travel peripherally to the superior surface of the optic-nerve sheath. Here, they fuse with long ciliary nerves and the post-ganglionic nerves from the ciliary ganglion to form another dense nerve-fiber plexus that ultimately supplies the eye. Importantly, the plexus on the optic nerve contains many isolated or aggregated ganglion cells. These are comparable in number to those in the ciliary ganglion itself and are assumed to be accessory ciliary neurons. Using immunohistochemistry, we also sought evidence for vasoactive intestinal polypeptide in these ganglia and nerves. As previously known, many pterygopalatine ganglion cells are immunoreactive for this peptide. Vasoactive intestinal polypeptide (VIP)-like immunoreactive nerve fibers are present in nerves from the retro-orbital plexus to the optic-nerve sheath plexus, in most nerves of the latter plexus, and in most nerves entering the eye. Furthermore, a small proportion of nerve cells in the main and accessory ciliary ganglia also are immunoreactive for VIP. We conclude that in addition to the pterygopalatine ganglion, the ciliary ganglion and its accessory ganglia are sources of VIP-like immunoreactive nerves in the rat eye.

Fluorescein distribution in retinas of normal and diabetic rats.

Quantitative fluorescence microscopy was used to study fluorescein distribution across the blood-retinal barrier in control and streptozotocin diabetic rats at 2 min, 1 and 2 hr after dye injection (12.5 and 125 mg kg-1 i.v.). Fluorescence intensities of choriocapillaris and retina were compared with plasma fluorescein levels. Diabetic rats had only one-half the total plasma fluorescein concentration of controls by 1 hr after injection, but the proportion of free fluorescein was greater in diabetic animals than in controls, providing the total plasma dye levels did not exceed the binding capacity of plasma proteins. Diabetic rats also had more unbound fluorescein glucuronide in plasma. With fluorescence microscopy no focal fluorescein leakage from retinal capillaries or pigment epithelium was seen in any diabetic or control eye. However, the dye was detected in retinas of diabetic and control animals at all intervals except 2 hr after injection of the lower dose when its fluorescence was too low to measure. Initial fluorescein entry occurred from the choroid by diffusion through pigment epithelial cells creating a steep intensity gradient decreasing from outer to inner retinal layers. With time this gradient flattened and then completely reversed, suggesting removal of dye from outer retinal layers and concomitant equilibration of inner layers with a pool of fluorescein in vitreous humor. Although the pattern of transretinal fluorescence distribution was similar in all rats, in specific instances, retinal fluorescence intensity differed significantly between control and diabetic animals. Fluorescence intensity was higher in diabetic than in control rats at 2 min after the 12.5 mg kg-1 dose and lower at 1 hr after the 125 mg kg-1 dose (P less than 0.05). These differences were directly related to transient differences in plasma free fluorescein concentrations, and in both cases, retinal fluorescence in diabetic rats returned to control values in conjunction with return to control levels of their plasma free fluorescein concentrations. The amount of dye detected in diabetic retinas was not in excess of normal levels at any interval after injection when related to concurrent plasma free fluorescein concentrations. These data do not indicate blood-retinal barrier dysfunction in the diabetic rat, but interpretation of the results is limited by the experimental conditions. With the high dose there was probably initial saturation of active transport mechanisms for dye removal, and with the low dose, dye distribution could not be followed to 2 hr when evidence of abnormal accumulation in retina may be more apparent.

Increase of basal cell membrane area of the retinal pigment epithelium in experimental diabetes.

Stereological analysis of electron micrographs of the pigment epithelium of rats with drug-induced diabetes demonstrated an increase of plasma membrane surface area at the basal aspect of the cells. In none of the diabetic animals examined was there any evidence of breakdown of the blood-retinal barrier to the protein tracer, horseradish peroxidase. Statistically significant increases in basal plasma membrane length and surface density (surface area per unit cell volume) were measured in both streptozotocin and alloxan-injected rats after four weeks of diabetes. When hyperglycemia in streptozotocin-injected rats was promptly reversed by transplantation of normal pancreatic islets, the increase of membrane surface area did not occur. We conclude, therefore, that increased basal surface area of pigment epithelial cells is related to the diabetic condition rather than to a toxic action of the diabetogenic agents. Furthermore, increased membrane surface area was present in streptozotocin-diabetic rats killed after six months of diabetes indicating that the structural change is relatively stable. Relation of basal membrane alteration in the pigment epithelium to any functional disturbance of the barrier cell layer or of the retina in diabetes remains to be established.

Acidosis alters fluorescein permeability. Differential tracer penetration through pigment epithelium.

The permeability of the blood-retinal barrier to infusion of fluorescent tracers was examined in tissue sections of frozen eyes from conscious rats that had breathed either air (control) or 25% carbon dioxide in air for one hour. For all animals the blood-retinal barrier remained impermeable to either carboxyfluorescein or Evans blue, indicating a functionally intact tight junctional barrier during hypercapnia. However, in hypercapnic animals, fluorescein penetrated into the neural retina, mainly via the pigment epithelium. Fluorescein also passed through the pigment epithelium in rats with metabolic acidosis induced by intravenous infusion of ammonium chloride, which lowered arterial blood pH without raising the Paco2 or BP. The results indicate an increased permeability of fluorescein through the cell membranes of the pigment epithelium during respiratory and metabolic acidosis. The effect on fluorescein may be due to a pH-dependent increase in the plasma concentration of the associated, more lipid-soluble form of this weak acid.

Carboxyfluorescein. A probe of the blood-ocular barriers with lower membrane permeability than fluorescein.

We have examined the permeability of the blood-ocular barriers to carboxyfluorescein, a dye similar in spectral properties but more polar than fluorescein. Octanol-buffer partition ratios of carboxyfluorescein, measured as an indication of lipid solubility, were approximately 1,000 times lower than those of fluorescein at pH values between 6.40 and 8.03. The partition ratios of both dyes show pronounced pH dependence. We also evaluated intraocular dye distribution by fluorescence microscopy after intravenous (IV) injection in rats. Carboxyfluorescein does not penetrate ciliary or iris epithelial cells, whereas fluorescein prominently stains these cells. Quantitative measurement of fluorescence intensity demonstrates that carboxyfluorescein does not enter the retina even when high doses are administered. Fluorescein, in contrast, can be detected throughout the retina with fluorescence intensity levels proportional to the IV dose administered. The relative inability of carboxyfluorescein to penetrate the blood-ocular barriers is not caused by greater binding to plasma proteins, since the plasma concentration of free carboxyfluorescein is greater than that of fluorescein. We conclude that carboxyfluorescein has potential experimental and clinical use as a probe of the blood-ocular barriers. Because of its low membrane permeability, it may yield a better definition of the nature of barrier abnormalities than is now possible with fluorescein.