Glucose utilization by the retinal pigment epithelium: Evidence for rapid uptake and storage in glycogen, followed by glycogen utilization

Experimental Eye Research 83 (2006) 235e246 www.elsevier.com/locate/yexer

Glucose utilization by the retinal pigment epithelium: Evidence for rapid uptake and storage in glycogen, followed by glycogen utilization Preenie deS Senanayake a,*, Anthony Calabro b, Jane G. Hu c, Vera L. Bonilha a, Aniq Darr b, Dean Bok c, Joe G. Hollyfield a a

Department of Ophthalmic Research, The Cole Eye Institute, The Lerner Research Institute, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA b Department of Biomedical Engineering, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA c Jules Stein Eye Institute, Department of Neurobiology and Brain Research Institute, David Geffen School of Medicine‘, University of California, Los Angeles, CA 90095-7065, USA Received 7 July 2005; accepted in revised form 28 October 2005 Available online 11 May 2006

Abstract Glucose utilization and glycogen metabolism by human retinal pigment epithelium (RPE) cultures with high transepithelial resistance maintained on porous Millicell polycarbonate filters, were quantified by fluorophore-assisted carbohydrate electrophoresis (FACE). Glucose uptake was more efficient at the apical surface of the RPE. The utilization of glucose when restricted to either the apical or basal medium was also evaluated. Under both conditions, glucose was quickly transported to the opposite compartment and rapidly utilized. However, glucose from the apical compartment was depleted to a greater extent than from the basal compartment. The de novo synthesis and accumulation of glycogen accompanied glucose utilization. This was paralleled by a concomitant increase in lysosomal glycogen degradation measured as an increase in cell-associated maltodextrins. The highest levels of glucose in glycogen and maltodextrins occurred at 24 h, declining to basal levels at 72 h. Glucose transporter expression in the RPE cultures was evaluated with the reverse transcriptase-polymerase chain reaction. Glucose transporter-1 (GLUT 1) was the isoform expressed in these cells. GLUT 1 localization was determined by immunocytochemistry. GLUT 1 localizes to the apical and basolateral border of the RPE. The intensity of fluorescence was higher on the apical border. The rapid depletion of medium glucose suggests that RPE culture studies should replenish medium glucose more frequently than every 72 h to maintain physiologically relevant glucose concentrations. These studies are the first to demonstrate glucose, glycogen and maltodextrin metabolism by RPE cells, and their detection and quantitation by FACE. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: glucose; glucose transporter; glycogen; maltodextrins; retinal pigment epithelium

1. Introduction Glucose is the major energy source for retinal metabolism (Berman, 1991). In a variety of vertebrate species, retinal glucose utilization is several-fold higher than in other body tissues. In vivo glucose is delivered to the retina via two blood supplies. Blood vessels to the inner retina are derived from

* Corresponding author. Tel.: þ1 216 444 5920; fax: þ1 216 445 3670. E-mail address: [email protected] (P.deS Senanayake). 0014-4835/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.exer.2005.10.034

the central retinal artery that enters the eye with return venous drainage through the optic nerve head. The endothelial lining of this blood supply forms a barrier preventing passive diffusion of glucose and other molecules from the retinal capillaries into the surrounding retinal tissue (Cunz-Vaz et al., 1996; Shiose, 1970). The blood supply to the highly metabolically active photoreceptors in the outer retina is from the choroid, a richly vascularized tunic in the outer eye wall located between the sclera and the retinal pigment epithelium (RPE). The choroidal vessels terminate in a highly fenestrated, leaky capillary bed located proximal to the RPE, and is separated

236

P.deS Senanayake et al. / Experimental Eye Research 83 (2006) 235e246

from it only by Bru¨ch’s membrane, a thin acellular lamina. Solutes move passively from the choriocapillaris out into the surrounding connective tissue and through Bru¨ch’s membrane where they bathe the basolateral borders of the RPE. Passive movement of glucose and other solutes is restricted to the basolateral borders of the RPE by zonula occludens-type junctions that are located in junctional complexes near the apical surface (Hudspeth and Yee, 1973). Most of the glucose entering the outer retina appears to be supplied by the choriocapillaris (Foulds, 1990). This anatomical position of the RPE is at a highly critical interface, interposed between the choriocapillaris, the source of glucose (and oxygen), and the retinal neurons that utilize large amounts of glucose (and oxygen). A number of studies demonstrate that the barrier function and transport properties of the RPE are crucial to the maintenance of the outer retina (Bito and DeRoussrau, 1980; Steinberg and Miller, 1979), but the kinetics and patterns of glucose uptake and storage by the RPE have not been well documented. In addition, glucose metabolism has important functional and clinical significance. Studies of Badr et al. (2002) in experimental diabetes has demonstrated that the fraction of glucose entering the retina in diabetes is likely to be greater across the RPE than across the retinal vasculature. RPE is also the site of advanced glycosylation end product formation (Handa et al., 1999). Glucose transport into the retina is also a central component of the hypothesis of glucose toxicity in the pathogenesis of diabetic retinopathy (Diabetes Control and Complications Trial (DCCT) Research Group, 1993). In the studies reported here, we use confluent cultures of RPE cells with high transepithelial resistance to define the pattern of glucose utilization and the transporters involved in this process. Our studies show that glucose is rapidly removed from the culture medium by both the apical and basal RPE cell surfaces. Further, at the time extracellular glucose levels are high, a portion of the glucose recovered by the RPE cells is stored as glycogen with a portion of the glycogen depolymerized within lysosomes to maltodextrins such as maltose, maltotriose, maltotetraose, etc. As glucose in the culture medium is depleted, glycogen and maltodextrins also decrease. The GLUT 1 transporter appears to be the isoform involved in removal of glucose from the medium in these RPE cell cultures. 2. Materials and methods 2.1. Materials 2-Aminoacridone HCl (AMAC) was from Molecular Probes (Eugene, OR, USA). D-Glucose (Glc), D-mannose, glacial acetic acid (99.99%), dimethylsulfoxide (DMSO, 99.9%), sodium cyanoborohydride (95%), and glycerol (99.5%) were from AldricheSigma (Milwaukee, WI, USA). Proteinase K and phenol red (0.5% w/v) were from Gibco (Grand Island, NY, USA). MONOÔ composition gels (#60100) and MONO gel running buffer (#70100) were from Glyko Inc. (Novato, CA, USA). Dowex AG50W-X8 (200e400 mesh) was from Bio-Rad

Laboratories (Hercules, CA, USA). Glucoamylase, unsaturated hyaluronan (DDiHA), chondroitin (DDi0S), and chondroitin sulfate disaccharide standards (DDi2S, DDi4S, DDi6S, DDi2, 6S, DDi4, 6S), maltose (Glc-a1, 4-Glc), maltotriose (Glc-a1, 4-Glc-a1, 4-Glc), maltotetraose (Glc-a1, 4-[Glc -a1, 4]2Glc), maltopentose (Glc-a1, 4-[Glc-a1, 4]3-Glc), maltohexose (Glc-a1, 4-[Glc-a1, 4]4-Glc) and maltoheptose (Glc-a1, 4[Glc-a1, 4]5-Glc) were from Seikagaku America (Ijamsville, MD, USA). 2.2. Preparation of saccharide standards for fluorescent derivatization The AMAC derivatized HA, chondroitin and chondroitin sulfate Ddisaccharide standards, the AMAC derivatized 4SgalNAc and 6S-galNAc standards, and the AMAC derivatized mannose and glucose containing saccharide standards were prepared as previously described (Calabro et al., 2000a). AMAC derivatized glucose standards as previously described (Calabro et al., 2000a; Senanayake et al., 2000) were used to calibrate the AMAC derivatized DDi2S standards used for quantitation of the fluorescent bands in the gel images. 2.3. RPE cultures The use of all human tissues followed the tenets of the Declaration of Helsinki, and the donors or their guardians gave consent for donation of the tissues. Institutional Human Experimentation Committee approval was obtained for the use of human tissues. The RPE culture conditions have been described previously (Frambach et al., 1990; Senanayake et al., 2000) and are summarized as follows. Human RPE was collected and grown in Eagle’s minimum essential medium without calcium [EMEM (Joklik); Sigma Chemical, St. Louis, MO] with additives described previously until the resident cells proliferated, reached confluence and were released into the medium (Hu and Bok, 2001). These non-attached cells were collected and grown on Millicell chambers with polycarbonate filters (Millipore, Bedford, MA, USA) coated with mouse laminin (Collaborative Research, Bedford, MA). The Millicell chambers were maintained in multiwell plates, which allowed the separation of apical and basal medium compartments. The normal calcium medium used to maintain established cultures consisted of Eagle’s minimum essential medium (EMEM, Irvin Scientific, Santa Ana, CA) with supplements for optimal human RPE culture as described previously (Hu and Bok, 2001). Both the apical and basal compartments received the same medium volume. The transepithelial electrical resistance (TER) of each confluent monolayer was measured with an epithelial voltohmmeter (World Precision Instruments, New Haven, CT) as previously reported (Hu et al., 1994). Cell density measurements were performed on representative cultures as previously described (Bok et al., 1992). Briefly, cultures on polycarbonate supports were fixed with 4% formaldehyde in phosphate buffer, dehydrated and mounted with glass coverslips. One of the oculars was fitted with a calibrated

P.deS Senanayake et al. / Experimental Eye Research 83 (2006) 235e246

micrometer square, and the cells within the area were counted. Counts from 10 fields in four separate cultures were determined and the mean values were expressed as the cell density extrapolated to the total area occupied by the RPE on the Millicell plate. Culture media from apical and basal compartments were changed every three days. All of the studies were carried out as described below with RPE cultures at confluence (3.9
105 cells) with high resistance junctions (resistance range 800e1500 ohm-cm2). The RPE cells at confluence become heavily melanized and show a cobblestone appearance with most cells forming a hexagonal border (Senanayake et al., 2000). Each millicell filter was analyzed by counting the cells in an area that is 14.6
14.6 mm. From this the number of cells for the entire filter was calculated. Cell counts made from representative cultures reflect an average density of 395,000  28,800 cells/cm2 (mean  SD). Transmission electron microscopy revealed that the RPE exists as a single, low cuboidal layer, with apical microvillae and prominent melanin granules in the apical cytoplasm (Senanayake et al., 2000). 2.4. Processing of culture media and RPE cell layers The methods used for sample preparation and carbohydrate analysis have been described previously (Calabro et al., 2000a,b). RPE cell layers and their respective culture media were separately digested with proteinase K and boiled to inactivate the proteinase K (Calabro et al., 2000a,b; Senanayake et al., 2000). The saccharides in the proteinase K digests were double ethanol precipitated. The resulting pellet/precipitate, containing the macromolecular material such as glycogen, was separated from the supernatant solution containing the low molecular mass molecules including glucose and salts. The combined supernatant solutions were vacuum concentrated. The precipitate was dried for 5 min in a vacuum concentrator as well. The resulting fractions were resuspended in 0.0005% phenol red, 100 mM ammonium acetate, pH 7 and divided into equal aliquots and either vacuum concentrated for direct AMAC derivatization or enzyme digested followed by AMAC derivatization (Calabro et al., 2000a,b; Senanayake et al., 2000) as described below. The current study has extended the analytical capability of our previous study (Senanayake et al., 2000) by incorporating the following: (1) a second ethanol precipitation of the proteinase K pellet to maximize the partitioning of glucose into the supernatant solution; and (2) the utilization of the resultant combined supernatant solutions from (1) for the detection and quantitation of glucose and other molecules with low molecular mass. 2.5. Enzyme digestions and quantitation of glucose, glycogen and maltodextrins The respective aliquots of pellet and supernatant fractions were processed as follows: (1) direct AMAC derivatization to identify and quantitate those endogenous saccharide structures with a free reducing aldehyde group such as free glucose and its a1, 4-linked oligosaccharides such as maltose,

237

maltotriose, maltotetraose, etc. (GA); and (2) digestion with glucoamylase (1 h at 37  C with 100 mU/ml) followed by AMAC derivatization to release for identification and quantitation, the bound a1, 4- and a1, 6-linked glucose in glycogen and maltodextrins as glucose equivalents (þGA). Glucoamylase completely depolymerizes glycogen through hydrolysis of glucose from the non-reducing termini of glycogen, and is able to cleave both the a1, 4 linkages and a1, 6 branch points within glycogen. The net result is complete digestion of glycogen to glucose, which can be separated and quantified as glucose equivalents by fluorophore-assisted carbohydrate electrophoresis (FACE). We have compared the hexokinase/glucose-6-phosphate (glc-6-P) dehydrogenase enzymatic assay (Kunst et al., 1984) typically used for the quantitation of glucose with FACE analysis. The standard curves generated by both techniques were almost identical (correlation coefficient ¼ 0.998) indicating that the two techniques were comparable. The advantages of FACE over the hexokinase/glc-6-P dehydrogenase enzymatic assay and alternative glucose oxidase assay (Trinder, 1964) are: (1) the higher degree of sensitivity (pmole compared to nmole range); (2) identification of unknown saccharides; and (3) the simultaneous direct and separate measurement of glucose and other saccharides such as the maltodextrins. This is in contrast to the hexokinase/glc-6-P dehydrogenase assay, which measures both the glucose and glc6-P together. Hence to identify the contribution of glc-6-P to the measurement of glucose, treatment with hexokinase is required after the initial reading. Overall, FACE analysis enables rapid, simultaneous and parallel analyses of multiple samples with improved sensitivity and a wide range of data on glucose metabolism. 2.6. Fluorotagging procedures The undigested (GA) sample aliquot and glucoamylase (þGA) digestion products were derivatized by addition of 20 ml of 12.5 mM AMAC (250 nmol) in 85% DMSO/15% acetic acid, followed by incubation for 15 min at room temperature. Then 20 ml of 1.25 M sodium cyanoborohydride (25,000 nmol) in ultrapure water was added, followed by incubation for 16 h at 37  C. After derivatization, 10 ml of glycerol (20% final concentration) was added to each sample prior to electrophoresis. All derivatized samples were stored in the dark at 70  C (Calabro et al., 2000a; Senanayake et al., 2000). 2.7. Electrophoresis, imaging and quantitation Samples were run on MONO composition gels with MONO gel buffer. The samples were electrophoresed at 4  C for 80 min at a constant 500 V with a starting current of 25 mA/gel, and a final current of 10 mA/gel. The gels in their glass supports were illuminated with UV light (365 nm) from an Ultra Lum Transilluminator, and imaged with a Quantix cooled CCD camera (Roper Scientific/Photometrics). The images were analyzed using the Gel-Pro AnalyzerÔ program version 3.0 (Media Cybernetics).

238

P.deS Senanayake et al. / Experimental Eye Research 83 (2006) 235e246

The digital images shown in the results section depict an overestimated pixel value for the major derivatized structures in order to allow visualization of less abundant derivatized saccharides. Quantitation was performed on gel images having all pixels within a linear 12-bit intensity depth range. DDi2S standards at progressive dilutions (a range of 1.9 to 62.5 pmol was used, as determined by hexuronic acid analysis) were added to generate an internal concentration curve (R ¼ 0.991, P ¼ 0.001) for quantitation of resolved saccharides. 2.8. PCR analysis Total RNA was prepared from human RPE cultures by using an extraction reagent (TRIzol; Life Technologies Inc., Grand Island, NY) according to the manufacturer’s instructions. All the samples were treated with DNase to remove any contaminating genomic DNA. Expression of GLUT 1, 2, 3, 4 and 5 in the human RPE cultures was determined using RTePCR with isoform specific primers. First-strand DNA (cDNA) was prepared using 1 mg total RNA. The GLUT isoforms were amplified using the following primer sets GLUT 1: 50 Primer TGAACCTGCTGGC CTTC, 30 Primer GCAGCTTCTTTAGCACA, 399 bp; GLUT 2: 50 Primer CAACAGGTAATAATATC, 30 Primer CTCGCAC ACCAGGACAGG, 583 bp; GLUT 3: 50 Primer AAGGATAAC TATAATGG, 30 Primer GGTCTCCTTAGGAGGCT, 411 bp; GLUT 4: 50 Primer CAGAAGGTGATTGAACA, 30 Primer AGGTAGCACTGTGAGG, 492 bp; GLUT 5: 50 Primer GAATTCATGGAAGACTT, 30 Primer GCCATCTACGTTTG CAA, 398 bp (Takagi et al., 1994). PCR products were resolved by agarose gel electrophoresis and visualized with ethidium bromide staining.

for glucose utilization by the cultures during the 98 days in culture. Data between the two groups were compared using the Student’s t-test. In Fig. 1B, the bar graphs represent the mean  SEM, n ¼ 3. The data between the three groups was compared using analysis of variance (ANOVA). The criterion for statistical significance was P < 0.05. Each data point in experiments 2 and 3 represents the mean  SEM, n ¼ 3. 3. Results 3.1. Glucose utilization by RPE cultures 3.1.1. Experiment 1 To define the overall pattern of glucose utilization by RPE cells in culture (experiment 1A), medium fractions from both apical and basal compartments of a single culture were collected at 3-day intervals at the time of refeeding, beginning on day 14 through day 98 of the culture period (Fig. 1A). Fig. 1A shows a representative oversaturated FACE image focused on the AMAC-derivatized glucose bands from 59-, 65and 95-day time points that demonstrate the differences in band intensity in the apical (AM) and basal (BM) medium compartments. The average amount of apical and basal medium glucose remaining after each 3-day feeding interval was determined for the entire 98-day culture period. Quantitation was performed on a gel image similar to that in Fig. 1A,

A

2.9. Immunocytochemistry For confocal microscopy, RPE cultures on filters were fixed in 4% paraformaldehyde and processed as previously described (Bonilha et al., 2004). For ZO-1 staining, cells were permeabilized with methanol for 5 min at 20  C after fixation. Ezrin, b-catenin and GLUT 1 staining cells were permeabilized with 0.2% Triton X-100 for 10 min. Sections were analyzed using a Leica laser scanning confocal microscope (TCS-SP2, Leica, Exton, PA). A series of 1 mm xey (en face) sections were collected. Each individual xey image of the stained RPE cell cultures represents a three-dimensional projection of the entire optical section (sum of all images in the stack). Microscopic panels were composed using AdobePhotoshop 5.5. Bar ¼ 40 mm. Antibodies used included mouse anti-ezrin (NeoMarkers, Inc., Fremont, CA); rabbit anti-GLUT 1 (Chemicon International Inc., Temecula, CA); mouse anti-ZO-1 and mouse anti-b-catenin (Transduction Laboratories, Lexington, KY). Secondary antibodies included goat anti-mouse and rabbit Alexa 488 and Alexa 594 (Molecular Probes, Eugene, OR). 2.10. Statistics In experiment 1, the data represent the mean  SEM (standard error of the mean), n ¼ 22, for the time points monitored

B p< 0.002

p< 0.008

p< 0.02

Fig. 1. (A) FACE gel image showing the levels of AMAC-derivatized glucose recovered from apical (AM) and basal (BM) medium compartments of RPE cell cultures 72 h after refeeding. The image depicts oversaturated pixel values for the major derivatized components in order to allow visualization of less abundant derivatized components. (B) Quantitation by FACE of the glucose remaining in the apical and basal medium compartments of triplicate human RPE cell cultures at confluence with high resistance junctions. The samples were taken 72 h after refeeding at 14, 56 and 81 days of culture, respectively.

P.deS Senanayake et al. / Experimental Eye Research 83 (2006) 235e246

but having all pixels within a linear 12-bit intensity depth range. Clearly evident in Fig. 1A is the more intense glucose signal in each of the basal medium samples when compared to that in the corresponding apical medium sample. This trend was substantiated over all the time points monitored. The initial glucose concentration in the medium added to both apical and basal compartments was 5.8 mmol/ml, 0.5 ml of medium was added to each compartment. The mean concentration of glucose remaining at the end of three days averaged over this 98-day period was 0.508  0.075 mmol/ml (mean  SEM, n ¼ 22) in the basal medium and 0.058  0.021 mmol/ml (mean  SEM, n ¼ 22) in the apical medium. When compared to the starting medium glucose concentration of 5.8 mmol/ml, the amount of glucose remaining was only 1% in the apical medium and 9% in the basal medium. This suggests that the apical surface of the RPE is more efficient in the removal of glucose from the medium than the basolateral surface (data not shown). To provide quantitative data on additional cultures that would allow more definitive statistical comparison, the glucose remaining in separate RPE cultures was analyzed; apical and basal medium from three separate cultures at 14, 56 and 81 days (Fig. 1B).

3.1.2. Experiment 2 To establish the temporal pattern of glucose utilization by the RPE glucose remaining in the apical and basal compartments was evaluated at intervals of 0, 1, 2, 4, 6, 8, 16, 24, 48, and 72 h after the replacement of fresh medium beginning on day 84 of culture (Fig. 2). Three cultures were analyzed per time point. The rate of glucose utilization was slightly higher for basal compartment glucose at high glucose concentrations, but then ultimately more efficient for apical compartment glucose at low glucose concentrations. Glucose levels declined to approximately 50% of the original medium concentration within 6 h in both apical and basal compartments. By 24 h, only 10% of the original medium glucose remained. By 72 h, glucose remaining in the basal compartment (5%) was higher than that in the apical compartment (0.2%), but the total

239

amount of glucose remaining in both compartments represented less than 5% of that present at the time of refeeding (0 h). The above analysis indicates that medium glucose is utilized by the RPE rapidly over the first 24 h of culture with glucose levels at or below 10% of the starting medium glucose concentration for the last 48 h of culture. The RPE cell layers were processed to assess the synthesis of glycogen and accumulation of maltodextrins by the RPE in culture (Figs. 4e6).

3.1.3. Experiment 3 In experiment 3, in two additional series, glucose was restricted to either the apical or basal culture medium of the RPE at the time of refeeding on day 84. Medium glucose content was then measured in both medium compartments at 0, 1, 2, 4, 6, 8, 16, 24, 48 and 72 h after refeeding (Fig. 3). Three separate cultures were analyzed per time point. Transepithelial resistance was measured in each culture with restricted glucose distribution and was not different from the resistance recorded when glucose was present in both compartments (800e 1100 ohm-cm2). Fig. 3A and B show the pattern of medium glucose utilization when glucose was present in only the apical or basal medium compartments, respectively. Under both experimental conditions glucose was quickly redistributed between apical and basal compartments. When glucose was provided only in the apical medium, glucose was present in the basal medium by 1 h, and achieved the approximate concentration remaining in the apical medium by 8 h. At 72 h, the levels of glucose detected were 0.1% and 0. 2% in the

A

B

Fig. 2. Quantitation of glucose recovered from apical and basal medium compartments from human RPE cells cultured for 0 to 72 h following refeeding on day 84. Each data point represents the mean  SEM of three separate cultures.

Fig. 3. Bi-directional transport of glucose by RPE cell cultures. (A) Time course for the utilization of glucose by RPE cell cultures when glucose was restricted to the apical medium. (B) Time course for the utilization of glucose by RPE cell cultures when glucose was restricted to the basal medium. Each data point represents the mean  SEM of three separate cultures.

240

P.deS Senanayake et al. / Experimental Eye Research 83 (2006) 235e246

apical and basal medium compartments, respectively. When glucose was provided only in the basal compartment, glucose was present in the apical medium by 2 h, and achieved the approximate concentration remaining in the basal medium by 4 h. At 72 h, the levels of glucose detected were 0.4% and 6% in the apical and basal medium compartments, respectively. Only when the basal medium contains high glucose concentrations does it retain 5e10% glucose. 3.2. Glucose storage by the RPE To determine the intracellular fate of glucose removed from the culture medium, we used enzymatic digestion coupled with FACE analysis of the RPE cell layers that were used to analyze medium glucose depletion presented in Fig. 2. Two sets of cultured cells were collected at 0, 1, 2, 4, 6, 8, 16, 24, 48 and 72 h following addition of fresh medium on day 84. Fig. 4A shows a representative oversaturated FACE image (6-h time point) that demonstrates the typical pattern for AMAC-derivatized glucose present in the ethanol precipitate and ethanol supernatant fractions of proteinase K digested cell layers before (GA) and after (þGA) glucoamylase digestion. As expected, unbound glucose is not present in the precipitate fraction (GA), but is completely soluble and partitions into the supernatant fraction (GA). This small amount of unbound glucose (containing a free reducing aldehyde) present in the cell layer compartment is derived from glucose

A

B

Fig. 4. (A) FACE gel image showing the levels of AMAC-derivatized glucose recovered from an 84 day RPE cell culture 6 h after refeeding. (B) Quantitation of glycogen-derived glucose from RPE cell layer compartments of duplicate cultures collected from 0 to 72 h following refeeding on day 84. Each data point represents the mean of two separate cultures. Extracellular matrix associated glucose (B), glycogen (C), maltodextrins associated with glycogen (D) or free in solution (E) as described in Fig. 9. The image depicts oversaturated pixel value for the major derivatized components in order to allow visualization of less abundant derivatized components.

trapped within the volume of cell-associated extracellular matrix and in residual medium that remains associated with the cell layer after removal of the medium fraction. As expected, after glucoamylase digestion (þGA), glucose appears in the precipitate fraction having been liberated from intracellular glycogen pools isolated after the proteinase K digestion and subsequent ethanol precipitation. Surprisingly, when compared to the untreated sample (GA), the fluorescence intensity of the glucose band in the supernatant fraction also increased upon glucoamylase digestion (þGA), indicating the presence of molecular species containing glucose linked by a1, 4-/a1, 6-linkages that are small enough to partition into the supernatant fraction (maltodextrins). For both supernatant and precipitate fractions, the fluorescence intensity of the glucose after glucoamylase digestion (þGA) minus that without digestion (GA) represents the glycogen-derived glucose in each fraction. Here glycogen-derived glucose is defined as glucose covalently bound with other glucose molecules through a1, 4-/a1, 6-linkages. To determine the amount of glycogen-derived glucose in each culture fraction, the fluorescence intensities of the glucose bands were quantified using gel images similar to that in Fig. 4A, but with all pixel values within a linear 12-bit intensity range. Fig. 4B shows the time course for glycogen-derived glucose in the precipitate and the supernatant fractions of RPE cultures from 0 to 72 h. The results show that in both fractions the amount of glycogen-derived glucose remained at low initial levels for the first 8 h of culture and increased rapidly to maximal levels between 16 and 24 h, after which the levels decreased to the initial levels present by 72 h. We also compared the glycogen levels in RPE cultures fed daily for 3 days with RPE cultures fed every three days as seen in Fig. 4. In the cultures fed daily, the levels of glycogen were 10e12 fold higher than in the cultures fed every 3 days (data not shown). This indicates that excess glucose can be stored as glycogen by the RPE. 3.3. Identification of glycogen and maltodextrins in the RPE The ethanol-soluble, molecular species responsible for the unexpected presence of a1, 4-/a1, 6-linked glucoamylasegenerated glucose in the supernatant fractions of the RPE cell layers shown in Fig. 4 were identified in Figs. 5 and 6. The oversaturated FACE gel image in Fig. 5 is an expanded view of the same gel image shown in Fig. 4 with the AMAC-derivatized glucose bands of Fig. 4 clearly evident in the center of the image. In Fig. 5, there are several slower migrating bands in the undigested (GA) precipitate (macromolecular-associated) and supernatant (free-soluble) fractions. These bands disappear upon glucoamylase digestion (þGA) with a concomitant increase in the corresponding AMACderivatized glucose bands. These bands have been identified as AMAC-derivatized maltose, maltotriose, maltotetraose, etc. based on: (1) their direct fluorescence labeling due to the presence of a free-reducing aldehyde; (2) their susceptibility to glucoamylase digestion; and (3) their co-localization

P.deS Senanayake et al. / Experimental Eye Research 83 (2006) 235e246

241

A

Fig. 5. Identification of glycogen and maltodextrins in human RPE cell cultures. FACE gel image shows the separation of saccharides recovered from an RPE cell layer (3.9
105 cells) 6 h after the change in medium. Saccharide profiles for the precipitate (lanes 3 and 4) and supernatant (lanes 5 and 6) fraction from proteinase K digested cell layers before (lane GA) and after glucoamylase digestion (lane þ GA). Lane 7 (GA) shows the separation of a series of synthetic maltodextrins. Lane 8 (þGA) shows that the series of synthetic maltodextrins in lane 7 was digested by glucoamylase to produce glucose. Lane 1 shows the separation of a mixture of 13 AMAC derivatized synthetic saccharides standard (STD 1). N-Acetylgalactosamine (galNAc), mannose, glucose, HA disaccharide (DDiHA), unsulfated (DDiOS), monosulfated (DDi2S, DDi4S, DDi6S), di-sulfated (DDi4, 6S and DDi2, 6S), trisulfated (DDi2, 4, 6S) CS disaccharides 6S-galNAc and 4S-galNAc. The image depicts an oversaturated pixel value for the major derivatized components in order to allow visualization of less abundant derivatized components, BeE as in Fig. 4.

with a1, 4-linked glucose oligosaccharide standards. These bands are collectively referred to here as maltodextrins, and are presumed to have originated via the lysosomal processing of glycogen within the RPE cells, which is consistent with other reports (Winchester, 1996). The contribution that the molecular species in the maltodextrin bands make to the measurements of glycogen-derived glucose shown in Fig. 4 can be calculated using gel images similar to that in Fig. 5, but having all pixels within a linear 12-bit intensity depth range. The contribution from the maltodextrins measured as glucose equivalents was calculated based on the fluorescence intensity of each of the individual maltodextrin bands, and the known number of glucose equivalents in each oligosaccharide. Only the reducing ends of the oligosaccharides are fluorescently tagged, thus the intensity of the fluorescence measured in each of these bands reflects the molar amount of each oligosaccharide structure. While the values calculated for the maltodextrins account for all of the glycogen-derived glucose in the supernatant fractions, there are insufficient amounts to account for all of the glycogen-derived glucose in the precipitate fractions as reported in Fig. 4B. The remaining glycogen-derived glucose in the precipitate fractions is derived from glycogen, which is detectable by FACE analysis as glucose equivalents only after glucoamylase digestion. The undigested glycogen is initially undetectable, as it has no intrinsic free-reducing groups for fluorescence derivatization. The glucose equivalents derived from glycogen and maltodextrins in the supernatant, precipitate and total cell layer fractions of RPE cultures over the 72 h after feeding with fresh medium are plotted versus time in Fig. 6. Note

B

C

Fig. 6. Time course for total glycogen-derived glucose, maltodextrins and glycogen in RPE cell cultures. (A) Macromolecular associated saccharides in RPE cell cultures. (B) Free-soluble saccharides in the RPE cell cultures. (C) Total saccharides in RPE cell cultures.

that the curves for the glycogen-derived glucose are the same as in Fig. 4. The results presented in Fig. 6 show the same temporal pattern for both the glycogen and maltodextrins in the total cell layer compartment (Fig. 6C) with low initial levels for the first 8 h of culture. There is then an increase to maximal levels between 16 and 24 h, followed by a return to levels below the initial values at 72 h. Interestingly, the total amount of glucose in both glycogen and maltodextrins is approximately the same at all time points, but with more maltodextrins present during the initial 8-h period, and more glycogen during the peak period between 16 and 24 h (Fig. 6C). The finding that maltodextrins were present in the ethanol precipitate fractions was surprising in that the maltose,

242

P.deS Senanayake et al. / Experimental Eye Research 83 (2006) 235e246

maltotriose and maltotetraose standards completely partition into ethanol supernatant fractions (data not shown). The explanation may be that a certain portion of the maltodextrins remain associated with glycogen, perhaps through intermediary molecules such as lysosomal amylases, which accounts for the partitioning of some of the maltodextrins into the macromolecular associated fraction after ethanol precipitation (see Fig. 9 and Section 4). Interestingly, all of the maltodextrins detected during the initial 8-h period are in the free-soluble supernatant fraction (Fig. 6B), which may represent residual lysosomal degradation products of glycogen. As expected, no glycogen was found in the supernatant fractions (Fig. 6B). The maltodextrins in the supernatant fraction show a similar temporal pattern to the glycogen and maltodextrins in the total cell layer fractions (Fig. 6C). The glycogen and maltodextrins in the precipitate fractions also follow the same temporal pattern (Fig. 6A). A larger proportion of the maltodextrins is present in the free soluble fraction compared to the macromolecular associated fraction. Dramatically evident in comparing data from Figs. 2 and 6 is the inverse relationship between medium glucose loss and the accumulation of glucose in the form of glycogen and maltodextrins. At later recovery intervals, as glucose in the medium is depleted, glucose present in glycogen and maltodextrins disappears, suggesting the mobilization of glucose recently stored in glycogen by the RPE. 3.4. Glucose transporter expression in RPE cells: Experiment 4 Glucose uptake requires the presence of specific glucose transporters in the RPE cells. RPE cell layers from cultures 72 h after refeeding on day 84 and 150 were processed to determine the expression of GLUT 1. In addition to GLUT 1, on day 150 expression of GLUT 2, 3, 4 and 5 was also evaluated. To determine the specific isoforms present we performed RTePCR using appropriate primers on 150-day cultures (Fig. 7). GLUT 1 expression was detected in day 84 RPE cultures. The expression of GLUT 1 was confirmed in day 150 cultures. Although we used primers specific for GLUT 1, 2, 3, 4, and 5 on day 150 cultures, we observed the amplification signal only for GLUT 1 (Fig. 7). These results indicate that GLUT 1 is the major isoform expressed by the RPE. 3.5. Glucose transporter localization in RPE cells: Experiment 5 RPE cell layers from cultures 72 h after refeeding on days 84 and 150 were fixed in 4% paraformaldehyde for evaluating the distribution of GLUT 1 by immunocytochemistry and confocal microscopy. In addition to GLUT 1, on day 150 localization of ZO-1, b-catenin and ezrin was also evaluated (Fig. 8). Confocal imaging showed a highly intense distribution of GLUT 1 in 84-day and 150-day RPE cultures. To define the distribution of GLUT 1 in the RPE on day-150 cultures, we used confocal imaging of GLUT 1 immunoreactivity for

Fig. 7. Expression of GLUT transporters in RPE cells. RTePCR analysis of GLUT 1, 2, 3, 4 and 5 in confluent human RPE cell cultures that have established high resistance junctions. RTePCR was performed with isoformspecific primer sets. Lanes 1 and 8, DNA size markers (100 base-pair DNA ladder; Life Technologies, Inc); lane 2, negative control; lane 3, GLUT 1 (399 bp); lane 4, GLUT 2 (583 bp); lane 5, GLUT 3 (411 bp); lane 6, GLUT 4 (492 bp); lane 7, GLUT 5 (398 bp).

comparison with the localization of ZO-1 (marker for the location of the apical junctional complex, Stevenson et al., 1986), b-catenin (present on the lateral plasma membranes) and ezrin (present in the apical and basal plasma membranes, Bonilha et al., 1999) (Fig. 8). GLUT 1 immunoreactivity is evident on the apical surface of the RPE and extends below the ZO-1 signal onto the lateral plasma membrane as well (Fig. 8AeF). b-Catenin labels the lateral plasma membrane and shows some colocalization with GLUT 1 (Fig. 8GeL). Ezrin immunoreactivity is principally localized to the apical plasma membrane and shows extensive overlap with GLUT 1 in this location (Fig. 8MeR). Clearly evident in these comparisons is the presence of GLUT 1 on both the apical and basolateral surface of the RPE, but the intensity of the GLUT 1 signal is much higher on the apical surface than it is below the junctional complex on the basolateral surface. 4. Discussion The present studies demonstrate that the apical and basal surfaces of the RPE cells rapidly remove glucose from the culture medium. Glucose is normally presented to the RPE from the basal side and would not be expected to be present at high levels on the apical side. The RPE is located between the rich choroidal blood supply and the metabolically highly active photoreceptors. In vivo the RPE transports glucose down a concentration gradient. Thus the transporters at the apical membrane would be for exporting rather than importing glucose into the RPE. Unlike in the in vivo condition the GLUT 1 transporters in both apical and basolateral membranes of the RPE cells in culture are exposed to the same glucose concentrations. The RPE in cells in culture utilize the medium glucose transported into the RPE cells by the

P.deS Senanayake et al. / Experimental Eye Research 83 (2006) 235e246

A

C

E

B

D

F

G

I

K

H

J

L

M

O

Q

N

P

R

243

Fig. 8. Confocal images of GLUT 1 localization in RPE cells. In panels A and B, cells were probed with mouse anti-ZO-1 (green), cell-cell junctions are indicated by arrowheads in the xey/‘‘en face’’ image (panel A) and xez/cross-section (panel B). In panels C and D cells were probed with rabbit anti-GLUT 1 (red). GLUT 1 was detected on the apical and basolateral borders of the RPE cells. Staining was more intense on the apical border. Panel E shows the merge between A and C, and panel F shows the merge between B and D. GLUT 1 staining below the tight junctions as revealed by ZO1 staining in the Z section (arrow, panel F) confirms the lateral localization of GLUT 1. In Panels G and H, cells were probed with a mouse anti-b-catenin (green, double arrows) a marker for the lateral border of the RPE cells. In panels I and J, cells were probed with rabbit anti-GLUT 1 (red). GLUT 1 was detected on the apical and basolateral borders of the RPE cells. Staining was more intense on the apical border. The intensity of staining in panels J and I was stronger compared to C and D. This may be a reflection of differential permeability between MeOH and Triton X-100. Panel K shows the merge between G and I, and panel L shows the merge between H and J. In panels M and N, cells were probed with a mouse anti-ezrin (green) a marker of apical and basolateral borders of the RPE cells. In panels O and P the cells were probed with rabbit anti-GLUT 1 (red). Basolateral localization of GLUT 1 was confirmed in panels P and R. Goat anti-mouse and rabbit Alexa 488 and Alexa 594 were the secondary antibodies used. The images show a highly dense distribution of GLUT 1 in the RPE cultures. The transporters were localized preferentially on the apical surface of the RPE cells. Scale bar: 20 mm.

GLUT 1 transporter. The faster depletion of glucose from the apical medium compartment observed in this study may reflect the higher abundance of GLUT 1 transporters on the apical border of the RPE cells.

In this analysis, when glucose was restricted to either the apical or basal compartment, it was rapidly removed from the glucose enriched compartment and then appeared in the initially glucose free compartment. This redistribution

244

P.deS Senanayake et al. / Experimental Eye Research 83 (2006) 235e246

continued until medium glucose was equilibrated on either side of the RPE cells before its eventual depletion from both compartments. The maintenance of transepithelial resistance throughout these studies suggests that glucose was transported across the RPE cell layer (transepithelial transport), and was not simply passively equilibrated by moving between the individual cells. Furthermore, since glucose concentration was equilibrated when restricted to either the basal or apical compartment, this suggests that glucose transport by the RPE cells can be bi-directional. Glucose disposal and glycogen metabolism begin with the transport of glucose across the cell membrane, enabled by members of the family of facilitated glucose transporters. The primary cellular storage form of glucose in mammals is glycogen. Evaluation of RPE glycogen has been made accessible in this study by coupling specific enzymatic digestions with FACE. A special feature of this technology is that glucose and glycogen in a sample can be quantified simultaneously: a clear advantage over the traditional methods for glucose (Kunst et al., 1984; Trinder, 1964) and glycogen (Bueding and Hawkins, 1964; Nahorski and Rogers, 1972; Passonneau et al., 1967; Templeton, 1961) evaluation. In humans and other vertebrates studied, glucose utilization by the retina is three-fold higher than in any other tissue in the body. Therefore, an obvious role for RPE glycogen may be as an emergency source of glucose during glycemic stress. In the cytosol, glucose in glycogen is classically mobilized as glucose-1-phosphate via the action of glycogen phosphorylase-A. Glycogen can also be observed normally in the lysosomes of cells (Winchester, 1996). It is presumed that glycogen enters the lysosome by the process of autophagocytosis of the cytoplasm (Winchester, 1996). Glycogen in lysosomes can be depolymerized via the action of amylase- and maltase-like enzyme activities (Winchester, 1996). In vivo, when low glucose levels approach, the RPE may mobilize the lysosomal pathway as a source of unphosphorylated glucose for transport to the metabolically demanding photoreceptors. Thus, in the RPE cells, the lysosomal pathway could serve as a source of glucose as well as a transport function for glucose. The potential fate of medium glucose that enters the glycogen pathway in RPE cell cultures is shown schematically in Fig. 9. Pathway P1 shows a simplified version of the conversion of Glc-6-P to glycogen. Pathway P2 shows the degradation of glycogen in the lysosome via maltodextrins. The discovery of the lysosomal pathway for the degradation of glycogen in the RPE cells was a direct consequence of the application of the FACE technology. It is important to note that the lysosomal pathway (P2, Fig. 9) has the potential to generate unphosphorylated glucose. In mammalian cells, studies have revealed the existence of an unexpected large amount of regulated exocytosis in cells such as fibroblasts and epithelial cells, previously believed to exhibit only constitutive exocytosis. Regulated extracellular release of lysosomal contents has been described for several cell types, such as hepatocytes (LeSage et al., 1993), RPE cells (Hoppe et al., 2004) and macrophages (Tapper and Sundler, 1990). The findings of Rodrigues et al. (1997) in epithelial cells and fibroblasts cells

highlight a novel role for lysosomes in cellular membrane traffic and suggest that lysosomes fuse with the plasma membrane. The lysosomes may therefore be multiple function organelles. Cathepsin D, the most abundant lysosomal protease in RPE and frequently used as a marker for lysosomes, was present in the interphotoreceptor matrix (IPM) of normal and retinitis pigmentosa eyes (Schmidt et al., 1988). Thus, the potential exists for glucose generated in the lysosome to be transported to the IPM by the exocytic pathway in the RPE. Our results indicate that glucose entering the RPE cells is rapidly converted to glycogen with substantial amounts appearing as maltodextrins. In fact, our data represents the first evidence for the temporal generation of maltodextrins in eukaryotic cells. The maltodextrins are present in the precipitate and supernatant fractions of the RPE cell layers, suggesting that a portion of the maltodextrins may be associated with macromolecular ethanol insoluble molecular species such as glycogen. Maltodextrins are thought to represent degradation products liberated when glycogen is depolymerized and are not intermediates in glycogen synthesis (Winchester, 1996). It is interesting that the time course of glycogen accumulation and loss directly parallels the appearance, accumulation and loss of the maltodextrins. This suggests near simultaneous synthesis and degradation of glycogen by the RPE cells and may be a mechanism to restrict glycogen accumulation and storage within RPE cells as limiting glucose levels approach. When glucose levels are low it is necessary to restrict glycogen accumulation and storage within RPE cells, which might retard or prevent glucose from reaching the retinal cell layer. Also evident in the quantitative data is that the total glucose equivalents present in the maltodextrins is similar to the glucose equivalents present in glycogen, suggesting a significant role for the lysosomal pathway for glycogen processing in RPE cells. The physiological relevance of the lysosomal processing of glycogen is not understood. It is not known whether maltodextrins have any biological actions in eukaryotic cells or simply reflect the state of glycogen catabolism. The GLUT transporters are a key component in the facilitative transport of glucose in RPE and other cells. The more efficient depletion of glucose from the apical compartment in cultures monitored at three-day intervals over a 98-day period suggests that the apical surface of the RPE cells is more efficient than the basal surface in glucose transport. These functional data correlate directly with the observation of higher levels of GLUT 1 transporter immunoreactivity associated with the apical surface compared to the basolateral surfaces. The expression of GLUT 1 in the current study is in agreement with the findings of Takagi et al. (1994) in isolated human RPE cell cultures showing that GLUT 1 is the most abundantly expressed isoform. This is also consistent with GLUT 1 being the transporter in cells that form interepithelial occludens-type junctions that function to establish bloodtissue barriers, like the one formed by the RPE (Harik et al., 1990; Takata et al., 1990, 1992). GLUT 1 has been identified in proteomic analyses of the human RPE (West et al., 2003) and isolated intact mouse RPE microvilli (Bonilha et al., 2004). In addition Kumagai et al. (1994) have reported the

P.deS Senanayake et al. / Experimental Eye Research 83 (2006) 235e246

245

Glucose (A)

ECM

Apical Medium

Glucose (B)

T Glycolysis

Glc-6-P

(D) (E)

P1

Glucose Maltose

Glc-1-P

Cytoplasm

UDP-Glc

P2 GP

Ly s

oso m e

Maltotriose Maltotetraose

Glycogen (C) T

ECM Membrane

Glucose (A)

Basal Medium

Fig. 9. A schematic model of the glucose pathways and events involved in the synthesis and degradation of glycogen observed in this RPE cell culture system. The RPE rests on a porous membrane allowing culture medium to bathe both apical and basolateral surface. GLUT 1 transporters (T) are located on both apical and basolateral surfaces of the cells. Glucose depletion data presented in Figs. 1e3, and confocal microscopy presented in Fig. 8 indicate that GLUT 1 transporters are more numerous on the apical than on the basolateral border. Medium glucose (A) is present in the apical and basolateral compartments with a portion of this glucose (B) associated with the extracellular matrix (ECM). Glucose enters the cell via glucose transporters (GLUT 1) along both the apical and basolateral surfaces, and is immediately phosphorylated to glucose-6-phosphate (Glc-6-P). Glc-6-P is an important branch point in carbohydrate metabolism, serving as substrate for several enzymes that initiate different pathways. Our analysis is limited to following glucose as it moves into the glycogen pathways designated as P1and P2. P1 shows a simplified version of the conversion of Glc-6-P to glycogen, and consists of four types of reactions: (1) conversion of Glc-6-P to glucose-1-P (Glc-1-P); (2) activation of Glc-1-P via the formation of uridine diphosphate glucose (UDP-Glc); (3) formation and elongation of glycogen chains (a1, 4 linked); and (4) creation of branch points (a1, 6 linked) that can be further elongated. Glycogen breakdown may occur in the lysosome via pathway P2. Glycogen (C) is depolymerized to produce glucose and maltodextrins such as maltose, maltotriose and maltotetraose (E). These breakdown products of glycogen are demonstrated in Fig. 5. A portion of the maltodextrins may remain associated with the depolymerization enzymes (D) and glycogen during the transition state. When RPE cell cultures are digested with proteinase K and double ethanol precipitated, glycogen (C) and associated maltodextrins (D) partition into the precipitate fraction. The supernatant fraction contains the free maltodextrins (E) and free extracellular glucose (B) associated with the extracellular matrix or perhaps intracellular unphosphorylated glucose.

apical and basolateral distribution of GLUT 1 in diabetic and non-diabetic human eye. The expression of minor levels of GLUT 3 and GLUT 5 were reported by Takagi et al. (1994), but in our study these two isoforms were either absent or below detection. In summary, this study is the first to: (1) directly follow and quantify the depletion of glucose from the culture medium by

RPE cells; (2) demonstrate the synthesis of glycogen by the RPE; (3) identify maltodextrins and their metabolism by RPE cells; (4) simultaneously quantify glucose and glycogen metabolism by the RPE cells, and (5) demonstrate that in culture glucose is more efficiently taken up by the apical surface of the RPE cells than by the basolateral surface. Although possibly not reflective of the RPE in vivo, the rapid depletion of glucose

246

P.deS Senanayake et al. / Experimental Eye Research 83 (2006) 235e246

from the culture medium by these cells suggests that daily additions of glucose may be needed for optimum maintenance of RPE cultures. Indeed, it has been demonstrated that RPE cultured in a relatively low concentration of glucose does not properly polarize its Na-K ATPase to the apical plasma membrane (Hu and Bok, 2001). Acknowledgements This work was supported by grants from The Foundation Fighting Blindness, Owings Mills, MD. (J.G.H and D.B), NIH, The National Eye Institute, Bethesda, MD (D.B, EY00331 and EY00444; P.deS.S., EY013752; J.G.H., EY014240 and EY015638) and the Dolly Green Endowed Chair at UCLA (D.B.). References Badr, G.A., Tang, J., Ismail-Beigi, F., Kern, T.S., 2002. Diabetes downregulates GLUT 1 expression in the retina and microvessels but not in the cerebral cortex or its microvessels. Diabetes 49, 1016e1021. Berman, E.R., 1991. Biochemistry of the Eye. Plenum Press, New York. 309e467. Bito, L.Z., DeRoussrau, C.J., 1980. Transport functions of the blood-retinal barrier systems and the micro-environment of the retina. In: CunhaVaz, J.G. (Ed.), The Blood Retinal Barriers. Plenum Press, New York, pp. 133e136. Bok, D., O’Day, W., Rodriguez-Boulan, E., 1992. Polarized budding of vesicular stomatis and influenza virus from cultured human and bovine retinal pigment epithelium. Exp. Eye. Res. 55, 853e860. Bonilha, V.L., Finnemann, S.C., Rodriguez-Boulan, E., 1999. Ezrin promotes morphogenesis of apical microvilli and basal infoldings in retinal pigment epithelium. J. Cell Biol. 147, 1533e1547. Bonilha, V.L., Bhattacharya, S.K., West, K.A., Sun, J., Crabb, J.W., Rayborn, M.E., Hollyfield, J.G., 2004. Proteomic characterization of isolated retinal pigment epithelium microvilli. Molecular and Cellular Proteomics 3, 1119e1127. Bueding, E., Hawkins, J.T., 1964. An enzymatic method for determination of glycogen. Anal. Biochem. 7, 26e36. Calabro, A., Benavides, M., Tammi, M., Hascall, V., Midura, R., 2000a. Microanalysis of enzyme digests of hyaluronan and chondroitin/dermatan sulfate by fluorophore-assisted carbohydrate electrophoresis (FACE). Glycobiology 10, 273e281. Calabro, A., Hascall, V., Midura, R., 2000b. Adaptation of FACE methodology for microanalysis of total hyaluronan and chondroitin sulfate composition from cartilage. Glycobiology 10, 283e293. Cunh-Vaz, J., Shalkib, M., Ashton, N., 1996. Studies on the permeability of the blood-retinal barrier. Br. J. Ophthalmol. 50, 441e453. Diabetes Control and Complications Trial (DCCT) Research Group, 1993. The efficacy of intensive treatment of diabetes on the development and progression of long-term complications in insulin dependent diabetes. N. Eng. J. Med. 329, 977e986. Foulds, W.S., 1990. The choroidal circulation and retina metabolism: Part 2. An overview. Eye 4, 243e248. Frambach, D.A., Fain, G.L., Faber, D.B., Bok, D., 1990. Beta adrenergic receptors on cultured human retinal pigment epithelium. Invest. Ophthalmol. Vis. Sci. 31, 1767e1772. Handa, J., Verzijl, N., Matsunaga, F.M., 1999. Increase in the advanced glycation end product pentosidine in Bruch’s membrane with age. Invest. Ophthalmol. Vis. Sci. 40, 775e779. Harik, S., Kalaria, R.N., Whitney, P.M., Andersson, L., Lundahl, P., Ledbetter, S.R., Perry, G., 1990. Glucose transporters are abundant in cells with ‘‘occluding’’ junctions at the blood-eye barriers. Proc. Natl. Acad. Sci. USA 87, 4261e4264.

Hoppe, G., O’Neil, J., Hoff, H.F., Sears, J., 2004. Products of lipid peroxidation induce missorting of the principal lysosomal protease in retinal pigment epithelium. Biochem. Biophys. Acta 1689, 33e41. Hu, J.G., Bok, D., 2001. A cell culture medium that supports differentiation of human retinal pigment epithelium into highly polarized monolayers. Mol. Vision 7, 14e19. Hu, J., Gallemore, R., Bok, D., Lee, A., Frambach, D., 1994. Localization of NaKATPase on cultured human retinal pigment epithelium. Invest. Ophthalmol. Vis. Sci. 35, 3582e3588. Hudspeth, A., Yee, A., 1973. The intercellular junctional complexes of the retinal pigment epithelium. Invest. Ophthalmol. Vis. Sci. 12, 354e365. Kumagai, A.K., Glasgow, B.J., Pardridge, W.M., 1994. GLUT1 glucose transporter expression in the diabetic and nondiabetic human eye. Invest. Ophthalmol. Vis. Sci. 35, 2887e2894. Kunst, A., Draeger, B., Ziegenhorn, J., 1984. Methods of Enzymatic Analysis. In: Bergmeier, H.U. (Ed.), third ed. Vol. 2, Academic Press, New York, pp. 163e172. LeSage, S., Robertson, W.E., Baumgart, M.A., 1993. Bile acid-dependent vesicular transport of lysosomal enzymes into bile in the rat. Gastroenterology 105, 889e900. Nahorski, S.R., Rogers, K.J., 1972. An enzymic fluorometric micro method for determination of glycogen. Anal. Biochem. 49, 492e497. Passonneau, J.V., Gatfield, P.D., Schulz, D.W., Lowry, O.H., 1967. An enzymic method for measurement of glycogen. Anal. Biochem. 19, 315e326. Rodrigues, A., Webster, P., Ortego, J., Andrews, W., 1997. Lysosomes behave as Ca2þ-regulated exocytic vesicles in fibroblasts and epithelial cells. J. Cell Biol. 137, 93e104. Schmidt, S.Y., Heth, C.A., Edwards, R.B., Brandt, J.T., Adler, A.J., Spiegel, A., Shichi, H., Berson, E.L., 1988. Identification of proteins in retinas and IPM from eyes with retinitis pigmentosa. Invest. Ophthalmol. Vis. Sci. 11, 1585e1593. Senanayake, P.deS., Calabro, A., Hu, J.G., Bok, D., Hollyfield, J.G., 2000. Glycosamionglycan synthesis and secretion by the retinal pigment epithelium: polarized delivery of hyaluronan from the apical surface. J. Cell Sci. 114, 199e205. Shiose, Y., 1970. Electron microscopic studies on blood-retinal and bloodaqueous barriers. Jpn. J. Ophthalmol. 14, 78e87. Steinberg, R.H., Miller, S.S., 1979. Transport and membrane properties of the retinal pigment epithelium. In: Zinns, K.M., Marmor, M.F. (Eds.), The Retinal Pigment Epithelium. Harvard University Press, Cambridge, MA, pp. 205e255. Stevenson, B.R., Siliciano, J.D., Mooseker, M.S., Goodenough, D.A., 1986. Identification of ZO-1: a high molecular weight polypeptide associated with the tight junction (zonula occludens) in a variety of epithelia. J. Cell Biol. 103, 755e766. Takagi, H., Tanihara, H., Seino, Y., Yosimura, N., 1994. Characterization of glucose transporter in cultured human retinal pigment epithelium cells: Gene expression and effects of growth factors. Invest. Ophthalmol. Vis. Sci. 35, 170e177. Takata, K., Kasahara, T., Kasahara, M., Ezaki, O., Hirano, H., 1990. Erythrocyte /hepG2-type glucose transporter is concentrated in cells of bloodtissue barriers. Biochem. Biophys. Res. Commun. 173, 67e73. Takata, K., Kasahara, T., Kasahara, M., Ezaki, O., Hirano, H., 1992. Ultracytochemical localization of the erythrocyte/hepG2-type glucose transporter (GLUT 1) in cells of the blood-retinal barrier in the rat. Invest. Ophthalmol. Vis. Sci. 33, 377e383. Tapper, H., Sundler, R., 1990. Role of lysosomal and cystolic pH in the regulation of macrophage lysosomal enzyme secretion. Biochem. J. 272, 407e414. Templeton, M., 1961. Microdetermination of glycogen with anthrone reagent. J. Histochem. Cytochem. 9, 670e672. Trinder, P., 1964. Determination of blood glucose using glucose oxidase with an alternative oxygen acceptor. Ann. Clin. Biochem. 6, 24e27. West, W.A., Yan, L., Shadrach, K., Sun, J., Hassan, A., Miyagi, M., Crabb, J.S., Hollyfield, J.G., Marmorstein, A.D., Crabb, J.W., 2003. Protein database, human retinal pigment epithelium. Mol. Cell. Proteom. 2, 37e49. Winchester, B.G., 1996. Lysosomal metabolism of glycoconjugates. In: Lloyd, J.B., Mason, R.W. (Eds.), Subcellular Biochemistry: Biology of the Lysosome, Vol. 27. Plenum Press, New York, pp. 191e238.