Enhanced accumulation of A2E in individuals homozygous or heterozygous for mutations in BEST1 (VMD2).

Best vitelliform macular dystrophy (BMD) is an autosomal dominant inherited macular degenerative disease caused by mutations in the gene BEST1 (formerly VMD2). Prior reports indicate that BMD is characterized histopathologically by accumulation of lipofuscin in the retinal pigment epithelium (RPE). However, this accumulation has not been quantified and the chemical composition of lipofuscin in BMD has not been examined. In this study we characterize the histopathology of a donor eye from a rare individual homozygous for a mutation (W93C) in BEST1. We find that this individual's disease was not any more severe than has been described for heterozygotes. We then used this tissue to quantify lipofuscin accumulation by enriching intracellular granules from RPE cells on sucrose gradients and counting the granules in each density fraction. Granules from the homozygous donor eye as well as a donor eye from an individual heterozygous for the mutation T6R were compared with age-matched control eyes. Interestingly, the least dense fraction, representing classical lipofuscin granules was either not present or significantly diminished in the BMD donor eyes and the autoflourescence associated with lipofuscin had shifted to denser fractions. However, a substantial enrichment for granules in fractions of higher density was also noted in the BMD samples. Inspection of granules from the homozygous donor eye by electron microscopy revealed a complex abnormal multilobular structure. Analysis of granules by HPLC indicated a approximately 1.6- and approximately fourfold overall increase in A2E in the BMD eyes versus age-matched control eyes, with a shift of A2E to more dense granules in the BMD donor eyes. Despite the increase in A2E and total intracellular granules, the RPE in the homozygous donor eyes was relatively well preserved. Based on these data we conclude that the clinical and histopathologic consequences to the homozygous donor were not any more severe than has been reported previously for individuals who are established or presumptive heterozygotes. We find that A2E is a component of the lipofuscin accumulated in BMD and that it is more abundant than in control eyes suggesting that the etiology of BMD is similar to Stargardt's disease and Stargardt-like macular dystrophy. Finally, the changes we observe in the granules suggest that the histopathology and eventual vision loss associated with BMD may be due to defects in the ability of the RPE to fully degrade phagocytosed photoreceptor outer segments.

Myocilin-associated exosomes in human ocular samples.

Mutations in myocilin result in ocular hypertension, likely due to decreased drainage of aqueous humor through the trabecular meshwork. Since less myocilin is found in the aqueous humor of those with disease-causing mutations, understanding myocilin's role in the aqueous humor is of clinical importance. Recently, myocilin was shown to exit cultured trabecular meshwork cells in association with shed vesicles called exosomes. To examine relevance of this finding in a physiological setting, the present study examined three different types of ocular samples for the presence of myocilin-associated exosomes. Using differential centrifugation steps, we found myocilin associated with exosomes isolated from effluent collected from human anterior segments in organ culture and aqueous humor obtained from human cadaveric eyes or from patients undergoing excisional surgery. Similar to results with cultured cells, myocilin associated predominately with exosomes in fresh samples, appeared mostly soluble at later times, and had biochemical properties (density of 1.13-1.19 g/ml in linear sucrose gradient) similar to those characteristics of exosomes. These data indicate that exosomes are present and may facilitate the transport of myocilin into the extracellular space of human ocular cells.

Coiled-coil targeting of myocilin to intracellular membranes.

Mutations in myocilin (MYOC) associate with glaucoma and ocular hypertension. Unfortunately, the specific role of MYOC, a widely expressed protein of unknown function, in ocular hypertension is unknown. Since MYOC localizes both to intracellular membranes and to the cytosol, we tested the hypothesis that MYOC is a cytosolic protein that associates with cellular membranes via its coiled-coil domain. Using green fluorescent protein (GFP) chimeras in expression and metabolic labeling studies, we observed that MYOC's putative signal peptide failed to traffic GFP into the secretory machinery and out of transfected cells. Next, we tested which of MYOC's three folding domains were responsible for targeting. In cell fractionation and immunofluorescence microscopy studies, the coiled-coil, but not the helix-turn-helix or olfactomedin domains, was necessary and sufficient to target GFP chimeras to cell membranes. Interestingly, a vesicular phenotype required sequential addition of the helix-turn-helix and olfactomedin domains to the coiled-coil. Taken together, these data indicate that the coiled-coil domain, not the putative signal sequence, is responsible for the targeting of MYOC to the secretory machinery.

Cell association increases RPE outgrowth from primary explant.

PURPOSE: Most studies of cell growth of the RPE employ cultures which have been previously passaged, but here we investigated freshly-explanted RPE cells, which may have been growth-quiescent for years, within the eye to determine whether co-culture affects initial outgrowth in primary culture. METHODS: Bovine or human RPE were co-cultured in primary culture with several cell types to test the effects of homologous or heterologous cell association. For bovine RPE, cell number was measured over 14 days in cultures of RPE alone, or RPE in co-culture with irradiated living cells, with fixed-killed cells, with cells separated from the RPE by a semipermeable membrane, or in medium conditioned by the cell types used for co-culture. For human RPE, isolates from all donors were randomized, over a 13-month period, to co-culture or to culture alone. The number of cultures attaining a cell number at one month that was sufficient for further propagation was compared. RESULTS: Co-culture with irradiated living cells increased the growth of primary cultures of both bovine and human RPE. Living cells were required; fixed-killed cells were ineffective. The outgrowth-promoting activity was not tissue or species specific, and it appeared to require close cell association between the RPE and the co-culture cell population. Conditioned media were ineffective and rather were slightly growth inhibitory. Primary RPE cells showed an earlier expression of vimentin (a marker of Gzero-G1 transition), more rapid cell spreading, and a greater increase in cell number between 7 and 14 days after explant when grown in co-culture than when cultured alone. CONCLUSIONS: The results indicate that cell association between non-mitotic RPE cells and previously cultured cells of many types increases the outgrowth of the RPE by accelerating the early stages of growth activation in vitro. Co-culture methods offer a practical means for increasing the likelihood of producing cultures from small RPE isolates. Further, should cells involved in proliferative pathologies in situ associate with non-mitotic RPE within the monolayer, the latter cells may also be activated, leading to an augmentation of the pathology.

Cell-cell adhesion molecules and the development of an epithelial phenotype in cultured human retinal pigment epithelial cells.

For most epithelial cells, the adherens junction protein E-cadherin is an epithelial morphogen, inducing the development of an epithelial phenotype in vitro after cell contact at confluency. Here retinal pigment epithelial cells (RPE), which lack E-cadherin but express a cadherin that is also found in many non-epithelial cells (N-cadherin), were examined for the ability to produce an epithelial phenotype in vitro. Subpopulations of grossly epithelioid or fusiform cells were selected for analysis from RPE cultures derived from adult human donors. After confluency, epithelioid RPE cells were observed to undergo time-dependent changes that were similar to those previously found in epithelial cells expressing E-cadherin: the cadherin gradually developed a zonular distribution of detergent-resistant protein that co-localized with forming circumferential actin bundles; Na/K ATPase accumulated at cell contact sites, then polarized to its tissue-specific domain (the apical membrane for RPE); the cells formed elevated domes on the impermeant culture substrate. In contrast to cells expressing E-cadherin, these events in RPE required weeks rater than days at confluency. Additional proteins were examined in epithelioid RPE cells revealing that cytokeratins reorganized after confluency producing a zonular array, and several other adhesion proteins (alpha5beta1 integrin, ICAM-1, PECAM-1, NCAM) became enriched at cell-cell contact sites, each developing a distinct pattern at a distinct postconfluency interval. In contrast to epithelioid RPE, in fusiform RPE the adhesion molecules did not develop discrete distribution patterns after confluency, although the same complement of adhesion proteins was expressed. In cells expressing E-cadherin, the absence of epithelial properties is often due to underexpression of the cadherin or of the catenins, adherens junction proteins that link the cadherin to actin. Fusiform RPE, however, were not deficient in these proteins, expressing amounts of N-cadherin, alpha-catenin, beta-catenin, plakoglobin, p120, alpha-actinin and vinculin that were equivalent to epithelioid cells. It appears, therefore, that a subset of epithelial cells that express N-cadherin can produce a highly-developed epithelial phenotype in vitro through a slow morphogenetic process. However, the expression alone of adhesion molecules, including those with a morphoregulatory function in other cells, is insufficient to produce an epithelial phenotype in all cells derived from the pigment epithelium.

Molecular requirements for assembly and function of a minimized human integrin alphaIIbbeta3.

Integrin subunit compatibility within and between species plays a major role in heterodimer assembly and ligand specificity. As an example, human alphaIIb pairs only with human beta3 and does not assemble a heterodimer with beta3 from other species. We use interspecies subunit chimeras to identify molecular requirements for subunit compatibility and show that species-restricted heterodimer assembly depends on a unique hexapeptide VGSDNH in an extended loop of the hypothetical human beta3 MIDAS domain. This allows us to express alphaIIb(1-233) and beta3(111-318) as a soluble, mini-integrin that retains RGD-dependent ligand recognition. Thus, in the case of one integrin, alphaIIbbeta3, the molecular requirements for integrin subunit compatibility and ligand recognition are intimately related.

Phenotypic heterogeneity of retinal pigment epithelial cells in vitro and in situ.

We have previously developed methods based upon differential cell adhesion to select for cells of two different phenotypes, epithelioid and fusiform, from cultures of human RPE. Here we considered whether the differences in cell shape correlated with differences in protein tyrosine phosphorylation since it is known that elevated phosphorylation perturbs the stability of the adherens junction, a major determinant of epithelial phenotype. In cultures of both phenotypes we found low tyrosine phosphorylation levels in postconfluent cultures, and the same complement of tyrosine phosphoproteins after treatment with a phosphatase inhibitor. However, in contrast to epithelioid cells, fusiform RPE failed to localize the phosphoproteins to junctional sites on the cell periphery. We also re-evaluated primary and passaged RPE cultures for additional cell shape variants. Several discrete phenotypes were identified within the same cultures. They required several weeks at confluency to develop in primary as well as in passaged cultures, but after developing they remained stable for months. Since explanted RPE cells manifest several shape variants in an identical culture environment we examined bovine RPE cells in situ to determine whether the cells were heterogenous with regards to some cell surface and cytoskeletal proteins that might contribute to cell shape. Circumferential actin microfilament bundles and the occluding junction protein ZO-1 had fairly uniform distributions among cells in the monolayer, but the intermediate filament protein vimentin and the pericellular expression of phosphotyrosine varied among individual cells. Therefore, despite its grossly homogeneous appearance, the RPE monolayer in situ might be considered a mosaic of structurally heterogeneous cells which can give rise to phenotypically-distinct subpopulations when propagated in vitro.

Separation of phenotypically distinct subpopulations of cultured human retinal pigment epithelial cells.

Like most epithelial cells that are isolated from tissues and placed in culture, the phenotype of human retinal pigment epithelial cells (RPE) in vitro is more variable than for the same cells in situ. The phenotypic heterogeneity of the cultures has been attributed to varying stages of differentiation of the cells induced by the culture environment. In this study we show that within the same cultures RPE cells exist as phenotypic variants which are distinct and stable. Two phenotypically distinct subpopulations were identified, one epithelioid and one fusiform, that were present from the first spreading event in primary culture through multiple serial passages and when maintained for extended periods at confluency. Cell aggregation studies indicated differences in cellular adhesions, known determinants of cell shape, between the subpopulations. Methods to separate the subpopulations were developed which are based on the differential trypsin sensitivity of adhesions. The separated subpopulations had the same RPE cytokeratins by immunoblotting, but cytokeratin filaments (and actin filaments) had different organizations. The study indicates that RPE cell cultures contain at least two subpopulations of phenotypically distinct cells under identical culture conditions that can be separated and propagated independently. The phenotypic variants offer a model system for investigating determinants of epithelial cell shape in RPE. Further, the separation methods might be applied to test for phenotypic variants in other types of cultured epithelia.

The response of cultured human retinal pigment epithelium to hypoxia: a comparison to other cell types.

PURPOSE: To characterize the response of cultured human retinal pigment epithelial (RPE) cells to lowered environmental oxygen. METHODS: The response of cultured RPE cells to lowered oxygen environments was compared to that of cell types of presumed high (Madin-Darby canine kidney [MDCK] cells, an epithelial cell line) and low (CSF, corneal stromal fibroblasts) aerobic requirements. Cultures in a range of densities were exposed for 7 days to 3%, 8%, or 20% O2 with measurements of adenosine triphosphate (ATP), cell number, and cytochrome oxidase (CO) activity (an enzyme marker of aerobic metabolism). RESULTS: RPE cells had levels of CO activity and total cellular ATP intermediate between those of CSF (low) and MDCK (high) in all oxygen environments. Hypoxia led to modestly lowered ATP pools and CO activity for RPE cells over a wide range of culture densities. Hypoxia induced a greater cell loss in MDCK cells than in RPE cells, and the effects of hypoxia were greater in dense cultures of both epithelial cell types. Hypoxia had little effect on cell number for CSF. CONCLUSIONS: The data suggest that cultured RPE cells, though aerobically active, are not as dependent upon oxidative phosphorylation and are more resistant to hypoxia than MDCK cells, a cell type derived from another well-perfused tissue. The authors conclude that RPE cells are unlikely to suffer from hypoxic injury in situ because of a moderate aerobic demand and an abundant oxygen supply.

In vitro aging of bovine and human retinal pigment epithelium: number and activity of the Na/K ATPase pump.

We have previously observed that the density of Na/K ATPase pumps is lower in RPE cells in the posterior pole of both bovine and human eyes. The posterior pole in human eyes includes the macula, a region which is predisposed to aging pathology. In this study we examined the effect of age on the sodium pump using cultures of bovine (bRPE) and human (hRPE) RPE cells that were aged in vitro by repeated passage. The cultures were assayed for cell number, protein, pump density (specific binding of [3H]ouabain) and pump activity (specific uptake of 86Rb) at confluency at each passage. In culture, bRPE had more pumps per cell (3.2 x 10(6)) than hRPE (1.2 x 10(6)), but the bRPE pumps were less active so the pumping capacity per cell was nearly equal. Bovine RPE aged more rapidly in vitro (survived fewer passages) than hRPE. With aging, RPE cells from both species showed declines in cell number at confluency. Pump number and pump activity per cell remained constant. Because cell number declined, the pumping capacity per unit area of confluent epithelium was diminished with culture aging. RPE cell number is known to decline with age in situ, especially in the macula. If Na/K ATPase pump number and activity per RPE cell remain constant with aging in vivo as shown here in vitro, the effective pumping capacity of the RPE per unit area of 'monolayer' would decline in aged eyes.(ABSTRACT TRUNCATED AT 250 WORDS)

Modulation of potassium transport in cultured retinal pigment epithelium and retinal glial cells by serum and epidermal growth factor.

The ionic environment of retinal photoreceptors is partially controlled by potassium transporters on retinal glial and retinal pigment epithelial cells (RPE). In this study, serum and epidermal growth factor (EGF) were examined as modulators of potassium transport in confluent cultures of human RPE and rabbit retinal glia. EGF is a known mitogen for confluent RPE cultures and was shown here to also stimulate [3H]thymidine incorporation in cultures of retinal glia. For potassium transport studies 86Rb was used as a tracer during a 17-min incubation. For both retinal cell types the mean total 86Rb uptake in 10% serum was approximately 60% above basal, serum-free controls. For EGF, tested in several experiments in a concentration range from 1 to 100 ng/ml, maximal total uptake was 33 and 24% above controls for RPE and glia, respectively. Inhibitor studies suggested that basal and serum-stimulated uptake for both cell types occurred by the ouabain-sensitive Na-K ATPase pump and by the furosemide- or bumetanide-sensitive Na-K-Cl cotransporter. EGF-stimulated uptake appeared to be due predominantly to the cotransporter. The data suggest that serum components and EGF, which may be available to retina-derived cells under pathologic conditions, may not only stimulate proliferation but may also act as short-term modulators of potassium ion movement and thus affect physiologic processes that are sensitive to ion homeostasis.

Collagen shield delivery of tissue plasminogen activator: functional and pharmacokinetic studies of anterior segment delivery.

BACKGROUND: Postoperative fibrin formation remains a major complication associated with intraocular surgery, especially after vitreoretinal surgery for proliferative vitreoretinopathy, proliferative diabetic retinopathy, trauma, or endophthalmitis. Tissue plasminogen activator (tPA) has been shown, both in experimental studies and clinical trials, to specifically dissolve formed intraocular fibrin after intracameral or intravitreal injection. We studied collagen shield delivery of tPA to the anterior segment and vitreous of rabbit eyes to evaluate a noninvasive delivery modality. METHODS: Anterior segment fibrin clots were formed in rabbit eyes by injecting citrated rabbit plasma. The tPA hydrated collagen shields, or control shields, were then placed on the rabbit corneas and the extent of fibrin clot was followed. In other rabbit eyes, tPA hydrated collagen shields were placed on the rabbit corneas and an enzyme-linked immunosorbent assay (ELISA) was utilized to determine aqueous, vitreous, and blood levels of tPA over time. RESULTS: Collagen shield tPA delivery shortened the time to fibrin clot lysis by 50% (mean clearance time = 49 +/- 23 hours; P less than .05). ELISA for tPA levels noted measurable vitreous levels by 2 hours after tPA hydrated collagen shield application with a peak at 24 hours. Aqueous tPA levels were not measurable until 18 hours after tPA collagen shield application and peaked at 36 hours. Vitreous tPA levels were greater than aqueous tPA levels at all time points (P less than .05). No evidence of corneal edema or opacification, hemorrhage, or cataract was seen. CONCLUSIONS: These results document the efficacy and safety of tPA delivery to the aqueous and vitreous via a hydrated collagen shield in this animal model.

Retinal pigment epithelial cells of the posterior pole have fewer Na/K adenosine triphosphatase pumps than peripheral cells.

The density of Na/K adenosine triphosphatase (ATPase) pumps in retinal pigment epithelial (RPE) cells in different retinal regions was quantified by measuring the binding of 3H-ouabain to RPE in cow and human eyecups. In bovine eyes, pump density was estimated in RPE samples isolated from three retinal regions outlined with a 7-mm trephine: one from the posterior pole in the area centralis and two from the superior, equatorial retina representing unpigmented (in the tapetum) and pigmented zones. In human eyes, RPE samples were isolated from a posterior region centered around the macula and one superior region. Ouabain binding to RPE of the posterior pole of both species was approximately 40-60% lower than binding to RPE of more peripheral regions in the same eyes. For bovine eyes, ouabain binding did not differ between pigmented and unpigmented cells of the superior retina, suggesting that reduced binding in the relatively amelanotic posterior cells was not related to levels of pigmentation. For human RPE, binding to posterior cells was lower in eyes from donors of all ages (range, 17-90 yr). The data suggest that Na/K ATPase pump site density is lower in posterior RPE cells of both bovine and human eyes, perhaps due to a regional difference in requirements for ionic regulation.

Intravitreal clearance of tissue plasminogen activator in the rabbit.

The clearance of tissue plasminogen activator (t-PA) injected into the midvitreous cavity was studied in the phakic, vitrectomized rabbit eye with and without intravitreal fibrin clots. The quantity and activity of t-PA in the vitreous, serum, and aqueous were determined at ten minutes and at 3, 6, 15, 24, and 48 hours after initial injection by an enzyme-linked immunosorbent assay (ELISA) and a spectrophotometric solid-phase fibrin assay (SOFIA). In eyes without an intravitreal fibrin clot, the estimated half-life for t-PA was 4.3 hours by SOFIA and 5.8 hours by ELISA. In eyes containing a vitreal fibrin clot, the half-life increased to 9.8 hours by SOFIA and 11.9 hours by ELISA. Both of these half-lives were significantly greater than the half-life for eyes without fibrin. Regardless of the presence of fibrin, intravitreal t-PA activity was significantly less than t-PA quantity, suggesting the presence of a t-PA inhibitor. A peak in aqueous t-PA occurred before six hours, indicating that t-PA was cleared in part through the anterior chamber. There was no measurable serum t-PA at any of the sampling times.