Downregulation of Reduced-Folate Transporter by Glucose in Cultured RPE Cells and in RPE of Diabetic Mice

Abstract

purpose. The polarized distribution of reduced-folate transporter (RFT)-1 to the apical retinal pigment epithelial (RPE) membrane was demonstrated recently. Nitric oxide (NO) significantly decreases the activity of RFT-1 in cultured RPE cells. NO is elevated in diabetes, and therefore in the present study the alteration of RFT-1 activity in RPE under conditions of high glucose was investigated.

methods. Human ARPE-19 cells were incubated in media containing 5 mM glucose plus 40 mM mannitol (control) or 45 mM glucose for varying periods and the activity of RFT-1 was assessed by determining the uptake of[ ³H]-N ⁵-methyltetrahydrofolate (MTF). The levels of mRNA encoding RFT-1 were determined by RT-PCR and protein levels by Western blot analysis. The activity of RFT-1 and expression of mRNA encoding RFT-1 were analyzed also in RPE of streptozotocin-induced diabetic mice.

results. Exposure of RPE cells to 45 mM glucose for as short an incubation time as 6 hours resulted in a 35% decrease in MTF uptake. Kinetic analysis showed that the hyperglycemia-induced attenuation was associated with a decrease in the maximal velocity of the transporter with no significant change in the substrate affinity. Semiquantitative RT-PCR demonstrated that the mRNA encoding RFT-1 was significantly decreased in cells exposed to high glucose, and Western blot analysis showed a significant decrease in protein levels. The uptake of [³H]-MTF in RPE of diabetic mice was reduced by approximately 20%, compared with that in nondiabetic, age-matched control animals. Semiquantitative RT-PCR demonstrated that the mRNA encoding RFT-1 was decreased significantly in RPE of diabetic mice.

conclusions. These findings demonstrate for the first time that hyperglycemic conditions reduce the expression and activity of RFT-1 and may have profound implications for the transport of folate by RPE in diabetes.

Diabetic retinopathy is the leading cause of blindness among working-aged adults in the United States, affecting approximately 10 to 12 million persons. ¹ Although retinal vasculature is particularly vulnerable to damage in diabetes, other retinal cells are at risk. Notable among these are retinal pigment epithelial (RPE) cells. ^{2 3} As described by Shiels et al., ³ after proliferation of new blood vessels from the neural retina, plasma leaks from these vessels and affects the posterior segment of the eye. Metabolic changes of diabetes mellitus cause vascular leakage, with alteration of the phenotype of RPE cells resulting in changes in cell function. The RPE is a monolayer of cuboidal cells that lies in close association with the highly metabolically active photoreceptor cells. Its many important functions include daily phagocytosis of rod and cone outer segment disks; uptake, processing, and the release of vitamin A; and mediation of the vectorial transport of nutrients from choroidal blood to photoreceptor cells. ^{4 5} How transport processes of the RPE are affected in diabetes, or other disease states, is only now beginning to be investigated due to the recent advances in our understanding of the transport mechanisms in mammalian cells. 

A nutrient whose transport has been investigated recently in RPE is folate. Folate, an essential vitamin required for DNA, RNA, and protein synthesis, ^{6 7} uses two different transport mechanisms to enter mammalian cells: folate receptor (FR)-α, and reduced-folate transporter (RFT)-1. ^{8 9} Recently, our laboratory demonstrated the polarized distribution of these two proteins in RPE. FRα is anchored to the basolateral RPE membrane ¹⁰ and RFT-1 is distributed in the apical membrane of the RPE. ^{11 12} FR, a glycosylated protein with a molecular mass of approximately 40 kDa, binds and internalizes folate through receptor-mediated endocytosis. ^{9 13} The entire FR protein is exposed to the exterior of the cell and is anchored to the plasma membrane through glycosylphosphatidylinositol. There are three isoforms of this receptor (α, β, and γ) among which only the α-isoform has been shown to participate in the cellular uptake of folates in normal cells. ¹³ Although the FRα has a very high affinity for nonreduced folate (k _d < 1 nM), it interacts also with reduced folates, although with much less affinity. The second transporter, RFT-1, has a molecular mass of approximately 60 kDa. ⁸ RFT-1 is a typical transporter protein with 12 membrane-spanning domains. It interacts with reduced folates, such as N ⁵-methyltetrahydrofolate (MTF), the predominant circulating form of folate, much more efficiently (k _m < 0.25 μM) than with nonreduced folate. This transporter has been cloned recently from mouse ¹⁴ and human tissues. ¹⁵  

Recently, the regulation of RFT-1 by nitric oxide (NO) was analyzed in human RPE. ¹⁶ NO is a molecule thought to be involved in the pathogenesis of diabetic retinopathy. ^{17 18 19} NO produces its biological effects by activating soluble guanylate cyclase, or by nitrosylation or oxidation of target proteins, either directly or through the formation of peroxynitrite. ²⁰ In the experiments analyzing RFT-1 activity, NO inhibited specifically and reversibly the uptake of N ⁵-MTF by a cGMP-independent mechanism. ¹⁶ These studies suggest that NO produced during retinal disease may affect the function of RFT-1 in adjacent RPE cells. There is evidence for increased production of NO in diabetes. The observation that NO attenuates the activity of RFT-1 prompted additional studies of the effects of hyperglycemia and diabetes on the activity and expression of this transporter. 

In the present study, we hypothesized that hyperglycemia and diabetes would impair the function of RFT-1. To test this hypothesis, an in vitro system was developed in which human RPE cells were exposed to varying concentrations of glucose and then assessed biochemically for their ability to take up folate. Kinetic parameters, levels of mRNA encoding RFT-1, and levels of RFT-1 protein were measured in control cells and in cells maintained under hyperglycemic conditions. Results of these experiments suggest that RFT-1 activity is attenuated in hyperglycemia and that this may occur at the mRNA and protein levels. The relevance of these observations to the in vivo diabetic condition was investigated by examining the transport activity of RFT-1 in RPE isolated from diabetic mice compared with normal mice. In addition, we analyzed the levels of mRNA encoding RFT-1 in RPE of diabetic versus control mice. In diabetic mice the transport activity was attenuated by approximately 20% compared with normal levels, and the mRNA levels were reduced markedly. This work represents the first report of attenuation of the functional activity of any folate transport protein under hyperglycemic conditions. 

Materials and Methods

Reagents

RPMI 1640 medium, TRIzol reagent, and penicillin-streptomycin were purchased from Gibco-Life Technologies (Rockville, MD); fetal bovine serum was from Sigma (St. Louis, MO);[ 3′,5′,7,9-³H]-N ⁵-methyltetrahydrofolate ([³H]-MTF; specific radioactivity, 30 Ci/mmol) was from Moravek Biochemicals (Brea, CA); l-[³H]-glutamine (specific radioactivity, 40 Ci/mmol) and l-[carboxyl-¹⁴C]-ascorbic acid (specific radioactivity 17.0 mCi/mmol) was from Amersham (Arlington Heights, IL). ARPE-19 cells were from American Type Culture Collection (Manassas, VA). Streptozotocin (STZ), d-(+)-glucose, monoclonal antibody to β-actin, and all other chemicals were purchased from Sigma. The urine strip test (Diascreen G) was from American Diagnostics (Pendleton, IN), and the glucose monitoring system (Prestige) was from Home Diagnostics (Ft. Lauderdale, FL). RNAWIZ reagent was from Ambion (Austin, TX); the ECL Western detection system was from Amersham Pharmacia Biotech (Piscataway, NJ); and the RNA PCR core kit was from Perkin-Elmer (Boston, MA). 

Cell Culture

Human ARPE-19 cells were maintained at 37°C in a humidified chamber of 5% CO₂ in RPMI 1640 medium containing 5 mM glucose, supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin. The culture medium was replaced with fresh medium every other day. At confluence, cultures were passaged by dissociation in 0.05% (wt/vol) trypsin in phosphate-buffered saline (PBS; pH 7.4). After trypsinization, the cells were seeded at a density of 1.9 × 10⁵ cells/well in 24-well culture plates and cultured in the presence of 2 mL/well of medium. When cultures reached 80% confluence, they were exposed to either high- or low-glucose medium, after which the uptake of radiolabeled compounds was measured, as described in a later section. In most experiments, the high-glucose RPMI 1640 medium contained 45 mM glucose and the low-glucose medium contained 5 mM glucose plus 40 mM mannitol. Mannitol was added to control for osmolar effects. In dose–response experiments, RPMI 1640 medium containing either 15, 25, 35, or 45 mM glucose was used, and the control medium contained 5 mM glucose plus 10, 20, 30, or 40 mM mannitol, respectively. In time-course studies, the exposure to high glucose was for 4, 6, 9, 12, or 24 hours. 

Animals

C57BL/6 mice (Harlan Sprague-Dawley, Indianapolis, IN) were used in these experiments. Type I (insulin-dependent diabetes mellitus) was induced chemically in 3- to 5-week-old mice, according to the method of Phelan et al. ²¹ Mice received an intraperitoneal injection of 75 mg/kg streptozotocin (STZ) dissolved in sodium citrate buffer (0.01 M, pH 4.5) on three successive days. Mice were screened for diabetes beginning 3 days after the first dose of STZ by testing for the presence of glucose in urine using the urine strip test. At the time of death, the diabetic state of the animal was confirmed by measuring blood glucose levels through a glucometer. Mice were maintained as described by Moore et al. ²² Care and use of the mice adhered to the principles set forth in the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Uptake Experiments in Cultured Cells

After exposure of cells to high glucose levels, the culture medium was removed and cells were washed once with warm uptake buffer (25 mM HEPES/Tris, 140 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl₂, 0.8 mM MgSO₄, 5 mM glucose [pH 7.5]). Uptake was initiated by adding 250 μL of uptake medium containing [³H]-MTF. The cells were incubated for 30 minutes at 37°C, which was in the linear range for uptake (data not shown). The medium was removed, and the cells were washed twice with ice-cold uptake buffer. The cells were solubilized with 0.5 mL of 1% sodium dodecyl sulfate and 0.2 N NaOH (SDS-NaOH) and used for determination of radioactivity by liquid scintillation spectrometry. 

Uptake Experiments in Intact RPE Tissue

With a procedure adapted from Vilchis and Salceda, ²³ the uptake of radiolabeled MTF was measured in the RPE-eyecups of six mice that had been diabetic 12 weeks and in six age-matched control mice. At the time of death, blood glucose levels were measured and were 419 ± 30 mg/dL and 131 ± 12 mg/dL in diabetic and control mice, respectively. To obtain RPE-eyecups, the corneas were slit with a sharp scalpel and the lens and vitreous humors extruded. The retinas were lifted from the RPE-eyecups allowing the apical RPE surface to be exposed to the incubation medium. The dissected RPE-eyecups were immediately placed in folate-free RPMI 1640 medium containing [³H]-MTF (3 nM) and were incubated for 30 minutes at 37°C in a humidified chamber with a gas flow of 0.1 L/min 95% oxygen and 5% carbon dioxide. This organ culture system has been described. ^{23 24} After incubation, the tissues were washed five times with ice-cold uptake buffer and subsequently weighed using an analytical balance. Each RPE-eyecup was sonicated (10 pulses; Virsonic 50 sonicator; Virtus, Gardiner, NY) in 0.5 mL 1% SDS-NaOH. Tissues were incubated in the SDS-NaOH solution for 90 minutes, and radioactivity was determined by liquid scintillation spectrometry. 

Semiquantitative RT-PCR Analysis of RFT-1 mRNA in Human ARPE-19 Cells Cultured in High or Low Glucose

Subconfluent ARPE-19 cells were cultured in RPMI medium in the presence of 5 mM glucose plus 40 mM mannitol or 45 mM glucose for 6, 12, and 24 hours. Total RNA was prepared using TRIzol. RT-PCR was performed with primer pairs specific for human RFT-1 and human glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The primers for RFT-1 were 5′-CAGCGTGTGGCCGGCTACTC-3′ (sense) and 5′-TCTGCCGCGGGCCGTTGTAGA-3′ (antisense), corresponding to nucleotide positions 542-561 and 1005-1023, respectively, of the human RFT-1 cDNA. ¹⁵ RT-PCR was performed in 9 to 32 cycles, with a denaturing phase of 30 seconds at 94°C, an annealing phase of 30 seconds at 63.9°C, and an extension of 2 minutes at 72°C. The PCR products (10 μL) were gel electrophoresed and subjected to Southern hybridization with a ³²P-cDNA probe specific for human RFT-1. Human GAPDH was used as an internal control. The upstream primer 5′-AAGGCTGAGAACGGGAAGCTTGTCATCAAT-3′ (sense), and the downstream primer 5′-TTCCCGTTCAGCTCAGGGATGACCTTGCCC-3′ (antisense) correspond to nucleotide positions 241-270 and 711-740, respectively, in human GAPDH cDNA. ²⁵ The hybridization signals were quantified with a phosphorescence imaging system (STORM; Molecular Dynamics, Sunnyvale, CA) and processed on computer (ImageQuaNT software; ver. 4.2a; Molecular Dynamics). Intensities were analyzed using the linear range of the PCR cycle. 

Semiquantitative RT-PCR Analysis of RFT-1 mRNA in RPE of Control and STZ-Induced Diabetic Mice

The RPE-eyecup was isolated from mice that had been diabetic for 3, 6, or 12 weeks. Average blood glucose levels were 436 ± 22, 383 ± 20, and 351 ± 34 mg/dL in animals that had been diabetic 3, 6, and 12 weeks, respectively. Age-matched, nondiabetic mice were used as control subjects; blood glucose levels were 172 ± 14 mg/dL. Six eyes per group per experiment were pooled for analysis. Total RNA was isolated using the RNAWIZ. RT-PCR was performed using primer pairs specific for mouse RFT-1. The upstream primer 5′-GCGTCTTCCCTGTCTAAA-3′ and the downstream primer 5′-GTCTCCCCTGTCGTCCTC-3′ correspond to nucleotide positions 1182-1199 and 1539-1557, respectively, in the cloned mouse RFT-1 cDNA. ¹⁴ For semiquantitative RT-PCR, PCR after reverse transcription was performed over a range of 9 to 32 cycles. Similar experiments were performed with primer pairs specific for mouse GAPDH (upstream primer 5′-ACCGGATTTGGCCGTATT-3′, downstream primer 5′-TCTGGGATGGAAATTGTGAG-3′ correspond to positions 65-82 and 1132-1151, respectively). The products were size fractionated on agarose gels and subjected to Southern hybridization with probes specific for each of the products. These probes were generated by labeling the respective subcloned RT-PCR products with [³²P]dCTP. The intensity of the hybridization signals for RFT-1 versus GAPDH in diabetic and control mice was quantified using a phosphorescence imaging system (STORM; Molecular Dynamics) and processed using the accompanying software (ImageQuaNT; Molecular Dynamics). The relationship between the intensity of the signal and the PCR cycle number was analyzed to determine the linear range for the PCR product formation. The intensities of the signals within the linear range were used for data analysis. 

Western Blot Analysis of RFT-1 Levels in ARPE-19 Cells Cultured in High or Low Glucose

ARPE-19 cells were cultured in medium containing 45 mM glucose or 5 mM glucose plus 40 mM mannitol and incubated for 6 or 24 hours. Cells were washed with 0.01 M PBS and lysed in cold lysis buffer (50 mM Tris-HCl [pH 7.4] containing 1% Triton X-100, 10 mM EDTA, 2 mM Na₃VO₄, 0.5% deoxycholate, 10 mM sodium pyrophosphate and 50 mM NaF). Cells were scraped off the flask, passed through a 26-gauge needle 15 times to create a homogenous mixture, and sonicated for 15 to 20 seconds. The cell debris was removed by centrifugation at 10,000g for 10 minutes at 4°C. Protein concentrations in the supernatant were determined according to the method of Lowry et al. ²⁶ Equivalent amounts of protein (10, 20, and 30 μg) from the total cell lysates were boiled in Laemmli’s buffer ²⁷ for 5 minutes and analyzed by 7.5% SDS-PAGE. After the separated proteins were transferred onto nitrocellulose membranes, the membranes were blocked for 1.5 hours at room temperature with Tris-buffered saline-0.1% Tween-20 containing 5% nonfat milk. The membranes were incubated for 2 hours with a polyclonal antibody that was raised against the peptide sequence RPKRSLFFNRDDRGRC, which corresponds to residues 205-220 of human RFT-1. ¹² The specificity of the antibody has been determined using Western blot analysis. The antibody identified a major protein in ARPE-19 cell membranes with a molecular size that corresponded to that of RFT-1. In addition, there were two faint immunopositive bands detected. Preincubation of the antibody with the antigenic peptide before Western blot analysis no longer recognized the major RFT-1 band, although the two faint bands were still observed. These data suggest that the major band corresponds to RFT-1, and this was the band that was used in densitometric quantification. The membranes were probed with a secondary horseradish peroxidase (HRP)–conjugated goat anti-rabbit IgG antibody (1:3000) for 1.5 hours and washed, and the proteins were visualized using the ECL Western detection system (Amersham Pharmacia Biotech). The membranes were washed three times and reprobed with an antibody against β-actin. After immunoblots were transferred and visualized, films were placed on a white light box vertically and image capture was performed (AlphaImager 2200 digital imaging system; Alpha Innotech Corp., San Leandro, CA). A drag-and-drop rectangular grid of identical size was placed on the band of interest and clicked for processing. Background density was automatically subtracted. Data represent an average of all the pixel values enclosed in the grid. 

Data Analysis

Data were analyzed on computer (NCSS 97 statistical package; NCSS, Kaysville, UT). In cases of multiple comparisons, ANOVA was used followed by the Tukey-Kramer paired comparison test. P < 0.05 was considered significant. 

Results

Time Course of Attenuation of MTF Uptake under Hyperglycemic Conditions

The uptake of [³H]-MTF (3 nM) by ARPE-19 cells exposed to 45 mM glucose for 4, 6, 9, 12, and 24 hours was compared with control cells cultured in 5 mM glucose plus 40 mM mannitol. Cells exposed to 45 mM glucose for 4 hours showed a 10% stimulation of MTF uptake compared with control levels (Fig. 1) . Within 6 hours, however, the uptake of MTF in cells exposed to 45 mM glucose decreased by approximately 35% compared with control levels. Similar results were obtained in cells exposed to high glucose for 9 hours and longer. 

Specificity of the Glucose-Induced Attenuation of RFT-1

Specificity of the glucose-induced attenuation of[ ³H]-MTF (3 nM) uptake in ARPE-19 cells was not nonspecific. It is not likely that the attenuation of uptake was due to cell damage, because the uptake of other nutrients such as[ ³H]-glutamine and[ ¹⁴C]-ascorbic acid was not reduced under identical experimental conditions (Fig. 2) . The uptake of glutamine and ascorbic acid was enhanced somewhat by exposure of cells to high glucose. 

Dose–Response Relationship with the Effect of High Glucose on the Uptake of MTF

The effects of increasing concentrations of glucose on the uptake of [³H]-MTF (3 nM) by ARPE-19 cells are shown in Figure 3 . ARPE-19 cells were incubated for 6 hours with 15, 25, 35, or 45 mM glucose. The uptake of MTF was compared with that in cells that had been incubated with 5 mM glucose and the appropriate concentration of mannitol. In cells incubated with 15 mM glucose, the uptake of[ ³H]-MTF was not significantly different from that of the osmolar control (5 mM glucose plus 10 mM mannitol). Cells incubated for 6 hours with 25 mM glucose demonstrated a slight attenuation of MTF uptake, but this attenuation did not differ significantly from control values. When cells were exposed to 35 mM glucose for 6 hours, however, the uptake of MTF was reduced by nearly 30%, compared with the osmolar control (5 mM glucose plus 30 mM mannitol). These data were similar to the attenuation of MTF uptake observed in the cells incubated 6 hours with 45 mM glucose. 

Kinetic Analysis of Glucose-Induced Attenuation of RFT-1 Activity

The kinetics of RFT-1 activity were analyzed in ARPE-19 cells exposed for 6 hours to 45 mM glucose versus those incubated in 5 mM glucose plus 40 mM mannitol. Uptake of [³H]-MTF (3 nM) was measured in the presence of cold MTF ranging from 0.05 to 1μ M (Fig. 4) . The analysis showed that the decrease in transport activity of RFT-1 observed under hyperglycemic conditions compared with control was associated with a decrease in the maximal velocity (50.4 ± 3.1 vs. 77.9 ± 4.7 pmol/mg protein/30 minutes for 45 mM glucose versus 5 mM glucose plus 40 mM mannitol, respectively) with no significant change in the substrate affinity (k _m = 0.28 ± 0.05 and 0.22 ± 0.04 μM for 45 mM glucose and 5 mM glucose plus 40 mM mannitol, respectively). The marked change in V _max with no change in k _m suggests a change in protein density of RFT-1. The Eadie-Hofstee plot (uptake/concentration versus uptake) was linear for both 45 mM glucose and control conditions (r ² = 0.992 and 0.990, respectively). It is unlikely that at higher MTF concentrations, FRα contributed significantly to the observed uptake of MTF. This was investigated recently by Huang et al. ¹¹ FRα binding activity (in which membrane binding of [³H]-folic acid in the absence or presence of unlabeled folic acid is performed at 4°C, a temperature at which transport is minimized) showed virtually no specific binding of folate on the apical RPE surface. ¹¹ It is also unlikely that the uptake was mediated by simple diffusion, because folate is a hydrophilic (lipophobic) molecule and requires an active transport mechanism. 

Semiquantitative RT-PCR Analysis of Hyperglycemia-Induced Attenuation of RFT-1 in ARPE-19 Cells

The influence of high glucose on the steady state levels of mRNA transcripts specific for RFT-1 was investigated using semiquantitative RT-PCR. mRNA encoding RFT-1 was normalized against GAPDH in treated and control cells. In cells treated with 45 mM glucose for 6 hours, the steady state levels of RFT-1 mRNA were significantly less than in control cells (Fig. 5A) . As the exposure times to high glucose increased (12 and 24 hours), the mRNA levels encoding RFT-1 decreased. The phosphorescence imaging analysis showed that the ratio of RFT-1 mRNA bands to GAPDH bands decreased as exposure to high glucose increased (Fig. 5B) . These results suggest that the hyperglycemia-induced decrease in the transport activity of RFT-1 probably is due to decreased de novo synthesis of the transporter protein resulting from the reduced steady state levels of the transporter mRNA. 

Western Blot Analysis of RFT-1 in ARPE-19 Cells Exposed to High Glucose

After determining that the mRNA levels encoding RFT-1 were reduced in ARPE-19 cells exposed to 45 mM glucose compared with the osmolar control, we asked whether there was any change in the level of RFT-1 protein in these cells. ARPE-19 cells were exposed to 45 mM glucose for 6 or 24 hours, and control cells were incubated with 5 mM glucose plus 40 mM mannitol for the same periods. The cells were lysed, and the lysate proteins were subjected to SDS-PAGE followed by immunoblot analysis, using an antibody against RFT-1. Figure 6A shows scans of gels loaded with 10 and 20 μg protein for treated and control cells. Densitometric scans of gels showed that in cells incubated 6 hours with 45 mM glucose, there was a 16% decrease in detectable RFT-1 protein compared with control cells. The decrease in protein levels was similar for both loading concentrations. After 24 hours in high glucose, there was a 25% decrease in detectable RFT-1 in the cells exposed to 45 mM glucose compared with control cells (Fig. 6B) . Levels of β-actin were not altered under hyperglycemic conditions. 

MTF Uptake in RPE of Diabetic Mice

To determine whether the activity of RFT-1 was affected under diabetic conditions, the uptake of MTF was analyzed in isolated RPE. The RPE was isolated from 12-week diabetic or age-matched control mice, as described in the Methods section. Because in the eye RFT-1 is present in RPE ¹² and not in the choroid or sclera, the presence of these latter two tissues does not affect uptake studies. The uptake of radiolabeled MTF was expressed as the amount of[ ³H]-MTF taken up per unit weight of the tissue. As shown in Figure 7 , the uptake of [³H]-MTF by the RPE of diabetic mice was reduced by approximately 20% compared with that in nondiabetic control mice. 

Analysis of Steady State Levels of mRNA for RFT-1 in the RPE of Diabetic and Control Mice

The RPE of control mice and mice that had been diabetic for 3, 6, or 12 weeks was used for semiquantitative RT-PCR for the determination of levels of mRNA transcripts encoding RFT-1. As an internal control, the steady state levels of GAPDH mRNA in the samples were determined in parallel. The results show that the RPE of 3-week diabetic mice had approximately 15% to 20% less RFT-1 mRNA compared with that of control animals, whereas the RPE of 6-week diabetic mice had approximately 40% less RFT-1 mRNA compared with control mice (Fig. 8A) . Levels of the housekeeping gene, GAPDH, did not change in the diabetic mice, suggesting that the effect of diabetes on mRNA encoding RFT-1 was specific. RPE of mice that had been diabetic for 12 weeks demonstrated a 90% decrease in steady state levels of mRNA encoding RFT-1. The phosphorescence analysis (Fig. 8B) confirmed that RFT-1 mRNA levels decreased as diabetes progressed in these animals. These results suggest that the decrease in the uptake of radiolabeled MTF observed in the RPE isolated from diabetic mice (shown in Fig. 7 ) may be due to decreased expression of the gene coding for RFT-1 or to decreased stability of the RFT-1 mRNA. 

Discussion

RFT-1 is hypothesized to play a key role in delivery of folate to adjacent photoreceptor cells. ¹² The present study assessed the effects of high glucose on the transport activity of RFT-1 in the RPE. Earlier studies showed that exposure of RPE cells to NO, which has been implicated in the pathogenesis of diabetic retinopathy, ^{17 18 19 28 29 30 31} attenuated the activity of RFT-1. ¹⁶  

Most of our experiments were performed using the ARPE-19 cell line. These cells retain features characteristic of RPE cells, including defined cell borders, a cobblestone appearance, noticeable pigmentation, ^{32 33} and the capacity to phagocytose outer segment disks. ³⁴ The usefulness of these cells in studying the transport of folate was established recently. ¹² Owing to the distribution of RFT-1 to their apical membrane, ARPE-19 cells are an ideal model in which to study the effects of high glucose on the transport function of this protein. It should be noted that RFT-1 is a bidirectional transporter ⁸ —that is, it can transport MTF from within the RPE cells outward to the adjacent subretinal space, and it can transport in the opposite direction. For the studies reported herein, the function of RFT-1 was assessed by analyzing its ability to take up MTF (inward transport). Additional studies demonstrate, however, that RFT-1 can transport MTF from within the RPE outward through the apical membrane (Bridges C, Ganapathy V, and Smith S, unpublished observations, 2001). 

In the present experiments using ARPE-19 cells, we observed an approximate 35% decrease in MTF uptake in the presence of high glucose levels. This decrease in uptake activity occurred within 6 hours of exposure to 45 mM glucose. Cells exposed to 35 mM glucose had a similar attenuation of RFT-1 activity. The effects of high glucose do not affect all transport systems, however, because the transport systems for glutamine and ascorbic acid were not reduced when cells were grown in 45 mM glucose. After determining that MTF uptake was attenuated under the high-glucose condition, we examined the kinetics of MTF uptake. The data showed a marked difference in the V _max of uptake under high-glucose conditions with no significant change in substrate affinity. These data suggest a possible change in the density of RFT-1 as a function of exposure of cells to high glucose. To explore this possibility, we cultured ARPE-19 cells in the presence of 45 mM glucose for 6, 12, or 24 hours and examined the level of mRNA encoding the protein using semiquantitative RT-PCR. In addition, we analyzed the level of RFT-1 protein in ARPE-19 cells exposed to high glucose for 6 or 24 hours using Western blot analysis. The semiquantitative RT-PCR analysis showed a 30% and 40% decrease in mRNA levels encoding RFT-1 as the exposure time to high glucose increased. The Western blot analysis showed a 16% and 25% decrease in protein density in cells exposed to high glucose for 6 and 24 hours, respectively. 

The levels of glucose chosen for the in vitro experiments were high and were selected to represent a significant elevation above the level of glucose preferred by ARPE-19 cells. These cells thrive and demonstrate their polarized characteristics in medium containing 17 mM glucose. ³³ Thus, the 35- and 45-mM glucose levels used in these studies are approximately three times higher than normal for cell culture conditions. Although these conditions simulate a hyperglycemic state, they cannot be extrapolated directly to the diabetic condition. It has been shown by others that primary cultures of RPE cells demonstrate a greater glucose utilization when cultured under higher than normal glucose concentrations. ³⁵  

To confirm the relevance of our in vitro findings, we examined the uptake of radiolabeled MTF in RPE-eyecups obtained from mice that had been diabetic 12 weeks. To do this, we adapted an organ culture system that has been used widely to assess function in tissues from whole animals. ^{23 24} The uptake studies showed that hyperglycemia caused an approximate 20% decrease in RFT-1–mediated activity. These functional studies were followed by molecular analysis of RFT-1 in the RPE of diabetic mice. Semiquantitative RT-PCR demonstrated a marked decrease (90%) in the mRNA levels encoding RFT-1, particularly in mice that had been diabetic 12 weeks, whereas levels of the housekeeping gene GAPDH did not change in the diabetic mice. There are several explanations for the discrepancy in the 20% decrease in functional activity compared with a 90% decrease in mRNA levels under diabetic conditions. One is that due to relatively slow protein turnover, protein levels (and hence activity) remained elevated longer than RNA levels. In experiments with the cultured RPE cells, however, protein and RNA profiles seemed to decrease in tandem (Figs. 5 6) . An alternative explanation for the discrepancy between activity and mRNA levels is that non–RFT-1–mediated uptake of MTF occurs in the RPE-eyecups of diabetic mice, perhaps by FRα, known to be present in the choroid and sclera. ³⁶ Because RFT-1 is present only in the RPE, ¹² measurements of RNA are likely to reflect more accurately the impact of hyperglycemia on RFT-1 than are the uptake experiments in which FRα may play a role. To date, there have been no analyses of FRα activity or expression under hyperglycemic conditions. 

Taken together, our data suggest that RFT-1 activity in the RPE is altered under hyperglycemic conditions. It is recognized that alterations of the RPE are not early events in diabetic retinopathy; however, Shiels et al. ³ have described that plasma leaking from damaged retinal blood vessels can have a significant impact on disease in the posterior segment of the eye. They state that metabolic changes caused by diabetes mellitus result in vascular leakage with alterations in the phenotype of RPE. 

Our findings that folate transport by the RPE may be affected in diabetes are significant. The apically placed RFT-1 presumably serves to transport folate from RPE to the adjacent, highly metabolically active photoreceptor cells. ¹² Thus, a decrease in RFT-1 activity would likely have an impact on the amount of folate delivered to adjacent photoreceptor cells, subsequently compromising their function. Decreased levels of folate in these cells would impact primarily on the RNA and protein synthetic capacity of these cells, because photoreceptors are terminally differentiated and are not likely to be involved in DNA synthesis. There is evidence that photoreceptor cell function is compromised in diabetic retinopathy. ^{37 38 39} Analysis of b-wave and oscillatory potential parameters showed rod and cone abnormalities. ³⁷ Cho et al. ³⁸ have reported a selective loss of short-wavelength (S)-cones in diabetic retinopathy that is thought to account for the acquired tritan-like color confusion found in this disease. 

In summary, the present study represents the first report of attenuation of the activity and expression of a folate transport protein under hyperglycemic conditions. Functional studies in cultured RPE cells demonstrated that incubation in 45 mM glucose for as little as 6 hours led to an attenuation of RFT-1 activity. Molecular studies indicated a decrease in the mRNA transcripts encoding RFT-1, and Western blot analysis revealed a decrease in the RFT-1 protein level in cells exposed to high levels of glucose. The relevance of these observations to the diabetic condition was demonstrated by a functional attenuation of RFT-1 activity in organ culture experiments with RPE- eyecups isolated from diabetic mice, compared with those of control mice. A significant decrease in the mRNA transcripts encoding RFT-1 was observed in diabetic mice compared with control subjects. The findings of these studies will form the basis of future experiments to understand alterations in the transport of folate in diabetic retinopathy. 

 Supported by National Institutes of Health Grants EY12830 and EY13089, Fight for Sight-Prevent Blindness America, and an unrestricted award from Research to Prevent Blindness, Inc. to the Department of Ophthalmology, Medical College of Georgia and the Medical College of Georgia Research Institute.

 Submitted for publication June 7, 2001; revised September 21, 2001; accepted October 18, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Time course of attenuation of MTF uptake in the presence of high glucose. ARPE-19 cells were exposed to 45 mM glucose for various lengths of time, and the uptake of [³H]-MTF (3 nM) was determined. Parallel experiments were performed with cells cultured in the presence of 5 mM glucose plus 40 mM mannitol (osmolar control). Results are expressed as the percentage of MTF uptake measured in corresponding control cells not treated with high glucose. Data are the mean ± SEM of four determinations from two independent experiments. *Significantly different from control (P < 0.05).

Figure 2.

 

Specificity of glucose-induced attenuation of MTF uptake. ARPE-19 cells were exposed to 45 mM glucose for 6 hours before measuring the uptake of [³H]-MTF (3 nM), [³H]-glutamine (25 nM), or [¹⁴C]-ascorbic acid (3 μM). Parallel experiments were performed with cells cultured for 6 hours in the presence of 5 mM glucose plus 40 mM mannitol (osmolar control). Results are expressed as the percentage of MTF uptake measured in corresponding control cells not treated with high glucose. Data are the mean ± SEM of four determinations from two independent experiments. *Significantly different from control (P < 0.05).

Figure 3.

 

Dose–response relationship showing the effect of high glucose on the uptake of MTF. ARPE-19 cells were exposed for 6 hours to 15, 25, 35, or 45 mM glucose, after which the uptake of [³H]-MTF (3 nM) was measured. Parallel experiments were performed with cells cultured 6 hours in the presence of 5 mM glucose plus 10, 20, 30, or 40 mM mannitol (osmolar controls). Results are expressed as the percentage of MTF uptake measured in corresponding control cells not treated with high glucose. Data are the mean ± SEM of four determinations from two independent experiments. *Significantly different from control (P < 0.05).

Figure 4.

 

Kinetic analysis of MTF uptake in ARPE-19 cells treated with high levels of glucose. ARPE-19 cells were treated with 45 mM glucose for 6 hours. Parallel experiments were performed with cells cultured 6 hours in the presence of 5 mM glucose plus 40 mM mannitol (osmolar control). Uptake of MTF was measured in these cells for 30 minutes over an MTF concentration range of 0.05 to 1 μM. Data are expressed as the mean ± SEM of three determinations from one independent experiment. Results are presented as plots describing the relationship between MTF concentration and MTF uptake rate and also as Eadie-Hofstee plots (inset; V/S versus V) V, MTF uptake in picomoles per milligram of protein per 30 minutes; S, MTF micromolar concentration.

Figure 5.

 

Analysis of steady state levels of mRNA for RFT-1 and GAPDH in ARPE-19 cells exposed to high levels of glucose. ARPE-19 cells were treated with 45 or 5 mM glucose plus 40 mM mannitol (control) for 6, 12, and 24 hours at 37°C. Total RNA was then isolated from these cells and used for semiquantitative RT-PCR. Primer pairs specific for human RFT-1 and GAPDH mRNA were used. RT-PCR was performed with a wide range of PCR cycles (n = 9–32). The resultant products were run on a gel and then subjected to Southern hybridization with ³²P-labeled cDNA probes specific for RFT-1 and GAPDH. The hybridization signals were quantified by phosphorescence imaging, and the intensities that were in the linear range with the PCR cycle number were used for analysis. (A) Representative Southern hybridization signal of bands from the 18th cycle for RFT-1 and the 15th cycle for GAPDH. (B) Relative band density (RFT-1/GAPDH) in cells treated with high glucose relative to that in control cells. The RFT-1-to-GAPDH ratio in control cells was taken as 1. Data are the mean ± SEM of three determinations from two independent experiments. *Significantly different from control (P < 0.05).

Figure 6.

 

Western blot analysis of RFT-1 in ARPE-19 cells exposed to high glucose for 6 and 24 hours. ARPE-19 cells were exposed to 45 mM glucose for 6 or 24 hours and lysed and the lysate subjected to SDS-PAGE followed by immunoblot analysis with a polyclonal antibody recognizing RFT-1. Membranes were washed and reprobed with an antibody against β-actin. Parallel experiments were performed with cells cultured in the presence of 5 mM glucose plus 40 mM mannitol (osmolar control). The density of the bands was quantified with a phosphorescence imaging system. (A) Immunoblot from a representative experiment showing two loading concentrations: 10 and 20 μg. (B) Band intensity from densitometric scans (RFT-1/β-actin) in cells treated with high glucose relative to that in control cells for both loading concentrations. The RFT-1-to-β-actin ratio in control cells was taken as 1. Data are the mean ± SEM of two determinations from four independent experiments. *Significantly different from control (P < 0.05).

Figure 7.

 

Uptake of MTF by RPE of normal and diabetic mice. C57BL/6 mice were made diabetic using three consecutive 75 mg/kg doses of STZ. At 12 weeks after onset of diabetes, the RPE was dissected from remaining ocular tissues. RPE from age-matched C57BL/6 mice was used as the control. Incubation of RPE was performed at 37°C for 30 minutes in folate-free RPMI 1640 medium supplemented with [³H]-MTF (3 nM). Results are the mean ± SEM (n = 12). Data are the mean ± SEM of six determinations from two independent experiments. *Significantly different from control (P < 0.05).

Figure 8.

 

Analysis of steady state levels of mRNA for RFT-1 and GAPDH in RPE of control (C) and diabetic (D) mice. C57BL/6 mice were made diabetic using three consecutive 75-mg/kg doses of STZ. At 3, 6, and 12 weeks after onset of diabetes, mice were killed, and the RPE was dissected from the adjacent neural retina. Age-matched control mice were used in parallel. Total RNA was isolated from the RPE and used for semiquantitative RT-PCR. Primer pairs specific for mouse RFT-1 and mouse GAPDH mRNA were used, and RT-PCR performed as described in Figure 5 . (A) Representative Southern hybridization signal showing bands from the 21st cycle of RT-PCR for RFT-1 and 15th cycle for GAPDH. (B) Band intensity (RFT-1/GAPDH) in diabetic RPE relative to that in control RPE. The RFT-1-to-GAPDH ratio in control cells was taken as 1. Data are the mean ± SEM of three determinations from two independent experiments. *Significantly different from control (P < 0.05).

The authors thank Amira El-Sherbeny for growing the cells used in the RT-PCR experiments and Ramesh Kekuda and Puttur Prasad for their advice in performing the RT-PCR experiments.