It is usually assumed that IRBP (interphotoreceptor retinoid-binding protein) is the only protein present in the interphotoreceptor matrix (IPM) capable of shuttling visual-cycle retinoids between photoreceptors and the retinal pigment epithelium. However, this laboratory previously presented qualitative evidence (Western blots) that serum albumin is present in human IPM. Furthermore, Ong and coworkers (1994) found that cultured RPE cells synthesize serum retinol-binding protein (RBP) and secrete it, mainly into the apical culture medium, which would correspond to the IPM in intact eyes. As both of these proteins can bind all- trans -retinol and 11- cis -retinal, it was of interest to quantify the amounts of albumin and RBP in human IPM. We used radial immunodiffusion to accomplish this. The average molar ratio of serum albumin to IRBP in these samples was 1.9; that of RBP to IRBP was 0.015. The presence of a high concentration of serum albumin in the IPM in situ was confirmed by the intense immunohistochemical staining seen in sections of fresh human eyes. The human case is not unique; various concentrations of albumin were found in the IPM of all vertebrate species examined (by gel electrophoresis). These results indicate that both serum albumin, because of its very high concentration in the IPM, and RBP, because of its comparatively tight binding to retinoids, need to be considered, along with IRBP, as proteins that may participate in visual-cycle transport. The accompanying paper addresses this concern.

Interphotoreceptor retinoid-binding protein (IRBP) greatly enhances the conversion of all-trans -retinol to 11- cis -retinal by the retinal pigment epithelium (RPE) and facilitates 11- cis -retinal release from the RPE. However, the mechanisms by which IRBP exerts these effects are not clear. Using a model system of purified bovine IRBP and isolated bovine RPE membranes, we investigated the possibility that IRBP may favor the delivery of all-trans -retinol to, or the release of 11- cis -retinal from, RPE membranes. As the interphotoreceptor space contains serum retinol-binding protein (RBP) and serum albumin in addition to IRBP, we similarly examined the exchange of retinoids between RPE membranes and human RBP or bovine serum albumin (BSA). Isolated RPE membranes were loaded with radioactive 11-cis -retinal and incubated with solutions of IRBP, RBP, BSA or with buffer alone. Membranes (pellet) and retinoid-binding protein or buffer (supernatant) were separated by centrifugation and analysed for radioactive 11- cis -retinal. Membranes incubated with buffer alone released only 4–5% of their 11- cis -retinal, while 25 μm IRBP removed 18–35%. More retinal was released as the membrane concentration was reduced. In contrast, RBP and BSA removed little retinal, even though both proteins are capable of binding this retinoid. Similar results were obtained with bovine liver membranes, consistent with the idea that the effects of IRBP do not depend on an RPE surface receptor for IRBP. IRBP was also markedly superior to RBP and BSA in removing all- trans -retinol from RPE membranes. In addition, IRBP efficiently delivered bound all- trans -retinol to membranes; however, in contrast to their differential removal of retinoids, all three binding proteins delivered comparable amounts of retinol to membranes. (This result supports the practice of using BSA as a retinoid carrier in in vitro experimental systems). We conclude that, whereas IRBP shares with other retinoid-binding proteins the ability to deliver retinol to membranes, IRBP is unique in its capacity to remove 11- cis -retinal from membranes. This may be the feature of IRBP that drives the vitamin A cycle to efficiently produce 11- cis -retinal.

We have investigated the synthesis and transport of apoE, the major apolipoprotein of the central nervous system, in the retina of the living rabbit. Four hours after the injection of [35S]methionine/cysteine into the vitreous, 44% of [35S]Met/Cys-labeled apoE is in soluble and membrane-enclosed retinal fractions, while 50% is in the vitreous. A significant amount of intact [35S]Met/Cys-labeled apoE is rapidly transported into the optic nerve and its terminals in the lateral geniculate and superior colliculus within 3-6 h in two distinguishable vesicular compartments. Müller glia in cell culture also synthesize and secrete apoE. Taken together, these results suggest that apoE is synthesized by Müller glia and secreted into the vitreous. ApoE is also internalized by retinal ganglion cells and/or synthesized by these cells and rapidly transported into the optic nerve and brain as an intact molecule. We discuss the possible roles of retinal apoE in neuronal dynamics.

The effect of VIP on the intracellular cyclic AMP of human retinal pigment epithelium cultures has been studied. Functional VIP receptor has been demonstrated in cultures from eyes given by five normal donors (age 16-64) (N-HRPE). But it has been found to be absent from high passage number cultures obtained from a retinitis pigmentosa eye of an 84-year-old patient (RP-HRPE). After 3 min of reaction with 1 x 10(-6) M VIP, the intracellular cyclic AMP level has increased to 5-15-fold over the basal level. The maximal effect of VIP (20-fold over the basal level) has been observed at 1 x 10(-7) M VIP. The half maximal activity of VIP is 3-5 x 10(-8) M. The present study also demonstrates the inducibility of the VIP responsiveness in RP-HRPE cultures after they have been treated with butyrate. Curr. Eye Res. 14: 1009-1014, 1995.

During a bleach, all-trans-retinol passes from the photoreceptor outer segments to the retinal pigment epithelium (RPE), where retinol is found associated with cellular retinol-binding protein (CRBP). Interphotoreceptor retinoid-binding protein (IRBP) is thought to facilitate this exchange, but the transfer of retinol between IRBP and CRBP has not been explored. In this study we used a mixture of purified IRBP and CRBP as a model system to measure the amount and rate of retinol transfer between the two proteins. When retinol is transferred from IRBP to CRBP, its absorbance maximum shifts from 330 to 350 nm. By monitoring the increase in absorbance at 350 nm after mixing CRBP with IRBP-bound retinol, we measured the amount and time course of retinol transfer from IRBP to CRBP. To complement the absorbance measurements, the IRBP and CRBP in these mixtures were subsequently separated by size-exclusion HPLC and individually analysed for retinol content by scanning the effluent with a multiple-diode-array detector. As determined by measuring the change in absorbance at 350 nm, the mean percentage of IRBP-bound retinol transferred to CRBP was 103 +/- 11% (n = 9). The mean half-time of the transfer was 4.2 +/- 1.3 sec; time to reach equilibrium was 30-60 sec. IRBP that was separated from the mixture by HPLC contained little or no retinol, while the isolated CRBP was nearly saturated with retinol. No detectable transfer from CRBP to IRBP was observed. The distribution of retinol between these two proteins was consistent with the nearly 100-fold higher affinity of CRBP for retinol compared with IRBP. Both the degree and time-course of transfer support the idea that this difference in affinity contributes to the flow of retinol to the RPE during a bleach in vivo and, therefore, may play a role in the physiological regeneration of rhodopsin.

Previous observations have shown that Müller glial cells of the vertebrate retina contain cellular retinoid-binding proteins, that the retina contains retinoic acid, and that cellular retinoic acid-binding protein is present in amacrine neurons (and, in some species, Müller cells) within the retina. These findings led to the suggestion that Müller cells may synthesize retinoic acid and release it for use by other retinal cells. To test this possibility, we cultured Müller cells from adult rabbit retinas, incubated the cultures with radioactive retinol, and identified and quantified the resultant radioactive retinoids by HPLC. Retinaldehyde was rapidly synthesized from retinol, reaching a plateau of 1-2 pmol mg-1 cell protein by 30 min. Retinoic acid initially accumulated more slowly, but by 30 min constituted most of the synthesized retinoid. While the retinaldehyde remained within the cells, retinoic acid was rapidly released into the medium; extracellular retinoic acid exceeded the intracellular amount after 30 min of incubation. Smaller amounts of retinyl esters were also synthesized and retained by the cells. These results are consistent with the suggestion that Müller glia are a source of retinoic acid in the retina. The synthesis of retinoic acid by these cells, and the presence of retinal neurons that contain cellular retinoic acid-binding protein, raise the possibility that retinoic acid plays a role in the retina, although this role is not presently known. Furthermore, these results may have implications for other parts of the adult nervous system. Adult brain contains retinol- and retinoic acid-binding proteins, and, therefore, may also be a site of retinoic acid metabolism. Because of the relatively simple cellular organization of the retina and its demonstrated capacity to synthesize retinoic acid, the retina may be a system of choice for further studies of the synthesis and function of retinoic acid in adult neural tissue.

Previous work from these laboratories showed that the retention of retinyl ester synthetase activity by cultured human retinal pigment epithelium is up to tenfold greater with PM medium (Medium 199 plus insulin, other added defined components, 1% serum and 1% retina extract) than with conventional culture media. The present work shows that insulin is the component of PM medium required for maintenance of ester synthetase activity and that insulin-like growth factor type 1 (IGF-1) also is effective at maintaining ester synthesis. In addition, insulin can maintain ester synthetase activity in cultured rat RPE. Preliminary dose-response measurements provide additional support for these findings and strongly suggest that both insulin and IGF-1 are maximally effective at physiological concentrations (1-10 ng ml-1).

The phagocytosis of isolated rod outer segments by cultured rat retinal pigment epithelium (RPE) previously has been shown to be stimulated by serum in the culture medium. In vivo, serum is normally in contact with the basolateral surface of the RPE. Components of serum have also been detected in the interphotoreceptor matrix, raising the possibility that these components may be in contact with the apical RPE surface. However, with conventional culture techniques it is not clear whether the stimulation of phagocytosis by serum occurs at the basolateral or apical surface. To resolve this uncertainty, phagocytosis was studied using RPE cultured on microporous filters, permitting control of which RPE surfaces are in contact with serum. Serum was found to have no influence on phagocytosis when present at the RPE basolateral surface, regardless of the serum concentration (2 or 20%). In contrast, phagocytosis was elevated three- to fivefold when 2% serum was present at the apical RPE surface, irrespective of the presence of serum at the basolateral surface. It is concluded that serum stimulates the phagocytosis of ROS only when the serum is present at the RPE apical surface.

Kinetically distinct classes of cyclic AMP (cAMP) and cyclic GMP (cGMP) phosphodiesterase activities (PDEs) were detected in homogenates of cultured pigment epithelium (PE) from both normal and Royal College of Surgeons (RCS) rats. PDE activities with apparent low Michaelis constants (Low Km cAMP- and cGMP-PDEs) were associated with the supernatant, while PDE activities with apparent high Michaelis constants (high Km cAMP- and cGMP-PDEs) were slightly higher in the pellet than in the supernatant after ultracentrifugation (100,000 g). Activity of the low Km PDEs was significantly reduced while that of high Km PDEs was not affected by known inhibitors of PDE. In both normal and RCS rat PE low Km PDEs required calcium and magnesium ions for optimal activity while the high Km PDEs required neither. In homogenates of cultured RCS rat pigment epithelium (PE), the kinetic parameters for cAMP- and cGMP-PDEs were comparable to normal, with the exception of the Km value of the low Km cGMP-PDE. This Km value was two-fold higher in the RCS compared with the normal (indicative of a reduced affinity for cGMP). It remains to be determined if the reduced affinity for cGMP in the RCS PE is related to the genetic defect which is expressed as a deficiency in the phagocytosis of outer segments by these cells.

Retinal pigment epithelium (RPE) cultured on microporous filter supports was compared to RPE cultured on plastic and evaluated for features characteristic of RPE in vivo. RPE cells grown on filters were cuboidal, formed junctional complex structures between cells, and had elaborate microvilli and basal infoldings similar to RPE in vivo, while RPE grown on plastic also formed intercellular junctions but appeared squamous and had few microvilli and basal infoldings. RPE grown on filters or plastic secreted an extracellular matrix at the basal surface and ingested isolated rat rod outer segments at the apical surface. RPE grown on filters coated with laminin or fibronectin became confluent more rapidly than RPE grown on uncoated filters, while RPE grown at the same density on filters coated with collagen type I did not become confluent. The laminin and fibronectin coatings did not alter the RPE cell morphology; however, cells seeded on collagen-coated filters grew in large disorganized clusters. RPE grown on laminin-coated filters formed functional tight junctions as evidenced by the capacity of RPE monolayers to prevent the bulk flow of medium and the passage of trypan blue across the filter. Radiolabeled sucrose and inulin were used to measure the paracellular flux through the tight junctions between cells. The passage of these tracers was linear over time, with the lower molecular weight tracer, sucrose, passing through the monolayer more readily than inulin. Values for the flux of radiolabeled bovine serum albumin across RPE monolayers fell between values for sucrose and inulin. The results from these studies show that RPE monolayers cultured on laminin-coated filters maintain a morphology similar to that of RPE in vivo, are capable of ingesting rod outer segments, and form a selectively permeable barrier to various tracers. This culture system should be useful for studies of transepithelial transport, secretion, endocytosis and exocytosis that require independent control of the extracellular environment at the apical and basolateral cell surfaces.

The site and substrate for all-trans to 11-cis isomerization in the visual cycle have remained obscure for several decades. Only recently studies on a subcellular level have begun to shed some light on these phenomena. We have addressed this system on a cellular level by utilizing intact isolated bovine retinal pigment epithelial cells, maintained during short-term incubation in vitro. Supplementation with labeled all-trans retinol incorporated in a lipid vesicle carrier, in a range of 1–6 nmol per 106 cells, resulted in a rapid uptake of retinol. The majority of the internalized retinol was processed prior to mixing with endogenous retinoid pools and most of it was converted into all-trans retinylester. Up to 10% of the incorporated label was isomerized yielding 11-cis retinol, 11-cis retinaldehyde and 11-cis retinylester. The kinetics of the 11-cis retinoid formation indicated that 11-cis retinol is the first isomerization product. The level of 11-cis retinol apparently ‘triggered’ further processing into other 11-cis retinoids. An updated model with discussion topics is presented for the retinoid pathway relevant to the visual cycle.

A method of preparing monolayer cultures of retinal pigment epithelium from normal pigmented neonatal rats is described. Critical features include incubating the eyes in balanced salt solution and treating with trypsin before dissecting the eyes. The tissue also has been cultured from RCS rats with inherited retinal degeneration. Since the pigment epithelium has been shown to be the primary site of action of the gene for retinal dystrophy in the RCS rat, the method should be useful in studying the defect(s) associated with this mutation.