Vitamin C Metabolomic Mapping in the Lens with 6-Deoxy-6-fluoro-ascorbic Acid and High-Resolution ¹⁹F-NMR Spectroscopy

Abstract

purpose. Metabolomics, or metabolic profiling, is an emerging discipline geared to providing information on a large number of metabolites, as a complement to genomics and proteomics. In the current study, a fluorine-labeled derivative of ascorbic acid (F-ASA), a major antioxidant- and UV-trapping molecule in the aqueous humor and the lens, was used to investigate the extent to which the lens accumulates potentially toxic degradation products of vitamin C.

methods. Human lens epithelial cells (HLE-B3) and rat lenses were exposed to hyperglycemic or oxidative stress in vitro or in vivo and probed for accumulation of F-ASA, fluoro-dehydroascorbate (F-DHA), fluoro-2,3-diketogulonate (F-DKG), and their degradation products in protein-free extracts, by proton-decoupled 750-MHz ¹⁹F-nuclear magnetic resonance (NMR) spectroscopy.

results. F-ASA and F-DHA were taken up into HLE B-3 cells by an Na⁺-dependent transporter. Their uptake was unexpectedly only slightly affected by hyperglycemia in vitro, unless glutathione was severely depleted. Glycemic stress catalyzed oxidation of F-ASA into a single novel F-compound at −212.4 ppm, whereas F-DHA and F-DKG were the major degradation products observed after GSH depletion. In contrast, F-ASA uptake was markedly suppressed in diabetic cataractous rat lenses, which accumulated both the F-DHA and the −212.4-ppm compound. In an unexpected finding, the latter formed only from F-ASA and not F-DHA or F-DKG, suggesting a novel pathway of in vivo F-ASA degradation. Both the cells and the intact rat and human lenses were permeable to several advanced F-ASA and F-DHA degradation products, except F-DKG. The unknown compound at −212.4 ppm was the only F-ASA degradation product that spontaneously formed in rabbit aqueous humor upon incubation with F-ASA.

conclusions. These studies suggest the existence of a novel ascorbic-acid–degradation pathway in the lens and aqueous humor that is influenced by the nature of the oxidant stress. Under similar culture conditions, intact lenses are more prone to hyperglycemia-mediated oxidant stress than are lens epithelial cells, but both are permeable to various F-ASA degradation products, the structure and biological roles of which remain to be established.

The aging human lens retains several components that formed at an embryonic stage. Therefore, these are exposed to physical and chemical stresses over a lifetime that result in irreversible modifications. Particular changes include protein aggregation, insolubilization, fragmentation, and coloration of lens crystallins. ¹ During cataractogenesis, both in diabetes and old age, these phenomena can be remarkably accelerated, leading to the formation of characteristic compounds that can act as covalent protein cross-links. Several mechanisms have been proposed to participate in the lenticular aging process, such as photo-oxidation, lipid peroxidation, transglutaminase activation, binding of tryptophan catabolites, and formation of advanced glycation end-products (AGEs)/Maillard reaction products derived from reducing sugars, oxoaldehydes, or ascorbic acid (ASA). ^{2 3 4 5 6 7 8} Specific cross-links and protein adducts that increase in the aging human lens have been described by others and ourselves. These include, for example, pentosidine, ⁹ vesperlysine A, ¹⁰ carboxymethyllysine (CML), ¹¹ and argpyrimidine/imidazo-lysine. ^{7 12} CML can be generated by glycoxidation, lipid peroxidation, and ascorbylation, ^{11 13} whereas argpyrimidine is a product of the condensation of methylglyoxal and arginine residues. ¹⁴  

The molecular origin of most of these compounds is unclear, because many of them, including argpyrimidine, can originate from ascorbic acid (ASA). ASA is a highly reactive glycating agent that is usually present at approximately 1 to 2 mM in the lens, compared with serum ASA levels (<100 μM). Therefore, ASA has been proposed to be a major source of AGE products in the lens. ¹⁵ When reacted with lens crystallins, ASA is 70-fold more reactive than glucose and 100-fold more reactive than fructose. ¹⁶ ASA oxidation products, which include dehydroascorbic acid (DHA), 2,3-diketoglulonic acid (DKG), and their degradation products, are important cross-linking precursors. ¹⁶  

So far, perhaps the most convincing evidence for the existence of ascorbic-acid–mediated modification in the lens is the discovery of oxalic acid monolysyl amide (OMA), ¹⁷ a modification of lysine that can be induced by incubation of proteins with ASA or its oxidation products. OMA, however, is not entirely specific for F-ASA, because ribose is also a precursor. ¹⁷ Thus, because both ASA and reducing sugars are sources of AGEs, it has not been possible to distinguish whether AGEs originate from ASA or reducing sugars under physiological conditions. For example, so far, it has not been possible to clarify whether CML originates from glycation, metabolites of glucose, ascorbate, or lipid peroxidation. 

These ambiguities imply that for ASA to participate in lens crystallin aging its degradation products should be detectable in the lens, especially under conditions that impair the redox state. To investigate these relationships, we have developed a metabolomic approach based on fluorine-labeled ASA (F-ASA) and high-resolution ¹⁹F-nuclear magnetic resonance (NMR) spectroscopy. Preliminary studies revealed that F-ASA and native ASA behave similarly in terms of oxidation kinetics, UV spectrum, and crystal structure, ¹⁸ and that five to seven degradation products may be observed during metal-catalyzed ascorbate oxidation. ¹⁸ Using this approach, we investigated the hypothesis that advanced degradation products of F-ASA accumulate in stressed lenses. We first report on the ability of cultured human HLE-B3 lens epithelial cells to take up F-ASA, F-DHA, and their degradation products. Next, we demonstrate the effects of stressing the cells with hyperglycemic conditions on uptake parameters and the degradation profile. Finally, we have investigated how intact rat and human lenses process F-ASA and how the F-ASA metabolome is altered by hyperglycemic stress in diabetic rats. 

Methods

Materials

Most reagents were from Sigma (St. Louis, MO) and were of the highest grade available. 6-Deoxy-6-fluoro-l-ascorbic acid (F-ASA) was synthesized as described by us. ¹⁸ 6-Deoxy-6-fluoro-l-dehydroascorbic acid (F-DHA) was freshly made from F-ASA by bromination on ice, as described earlier. ¹⁹ Briefly, 1 mL of 10 mM F-ASA was dissolved in resin (Chelex-100; Bio-Rad, Richmond, CA)-treated phosphate-buffered saline (PBS) and 10 μL bromine was immediately added under a fume hood. The bromine was purged by bubbling nitrogen gas over the solution until colorless. All PBS used in this study was treated with the resin overnight to remove heavy metal ions. Human lenses were obtained from the National Disease Research Interchange ([NDRI], Philadelphia, PA) and classified into types I to IV, based on degree of pigmentation, according to Pirie. ²⁰  

Cell Culture

Human lens epithelial cells (HLE-B3), the gift of Usha Andley (Washington University, St. Louis, MO), were cultured according to the protocol of Andley et al. ²¹ in Eagle’s minimum essential medium (EMEM; Mediatech, Herndon, VA) supplemented with 20% fetal bovine serum (FBS; Mediatech), 2 mM l-glutamine (Mediatech), 0.1 mM nonessential amino acids (Life Technologies, Rockville, MD), 100 U/mL penicillin (Mediatech), and 100 μg/mL streptomycin (Mediatech). The cultures were maintained at 37°C in humidified air containing 5% CO₂. Medium was changed twice a week. For subculture, cells were detached with 0.25% trypsin/EDTA (Mediatech). The experiments were performed with cells between passages 15 and 30. JAR cells (HTB-144; American Type Culture Collection [ATCC], Manassas, VA), a human choriocarcinoma cell line from placenta, were cultured in RPMI 1640 medium (Mediatech) supplemented with heat-inactivated 10% FBS, 2 mM l-glutamine, 10 mM HEPES, 1 mM sodium pyruvate, 100 U/mL penicillin, and 100 μg/mL streptomycin. The conditions of culture and subculture were the same as HLE B-3 cell culture. 17EM15 cells, a mouse lens epithelial cell line donated by John Reddan (Oakland University, Rochester, MI), were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Mediatech) supplemented with heat-inactivated 10% FBS, 2 mM l-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin. The conditions of culture and subculture were the same as for the HLE B-3 cell culture. 

Determination of F-ASA and F-DHA Cellular Uptake

For uptake experiments, HLE-B3 cells were cultured for 6 days after seeding at density of 1 × 10⁵/cm². The medium was changed every 2 days. For hyperglycemic conditions, cells were cultured with medium containing 50 mM d-glucose or 50 mM d-galactose for 6 days. F-ASA uptake was studied by exchanging the medium for 1 mM F-ASA containing 20% FBS/EMEM. At the appropriate time, the cells were collected by trypsinization and homogenized with a pellet pestle (Kontes Glass, Vineland, NJ). The lysate was mixed on ice with a final concentration of 4% metaphosphoric acid/PBS(−) containing 100 μM 6-deoxy-6-fluoro-d-glucose (F-Glc) as an internal standard for calibration and quantitation. The supernatants of the turbid solutions were obtained by centrifugation at 15,000 rpm for 20 minutes at 4°C. The supernatants were stored frozen at −80°C and transported on ice to the University of Akron for analysis. D₂O was added to the samples to a final concentration of 10%. An identical protocol was used to determine F-ASA and F-DHA uptakes in the JAR choriocarcinoma and 17EM15 mouse lens epithelial cells. 

Comparison of F-ASA and ASA Cellular Uptake

Kinetics of F-ASA uptake and processing were compared with those from native ASA. HLE-B3 cells were incubated with 1 mM F-ASA or native ASA for 24 hours. Levels of ASA and DHA in cell lysate and medium were measured by HPLC fluorescence. DHA was derivatized according to the method of Tessier et al. ²² Briefly, cell supernatant was treated with bromine to oxidize ASA to DHA for measuring total ascorbate. Forty microliters of the bromine-treated supernatant was derivatized by adding 30 μL of buffer (24.5 g NaH₂PO₄-H₂O and 10.4 mL of 4 normal NaOH adjusted to 100 mL [pH 5.4]) and 10 μL dimethyl-phenylenediamine (DMPD, 10 mg/mL 0.1 N HCl) with vigorous mixing and incubation for 5 to 120 minutes. The DMPD-DHA derivative was separated by an HPLC system (model 510) equipped with scanning fluorescence detector (model 470; both from Waters, Millipore Corp., Milford, MA) and a protein-peptide C18 reversed-phase column (diameter, 5 μm; 25 cm × 4.6 mm; Vydac, Hesperia, CA). The mobile phase was 25% methanol/phosphate buffer (0.08 M, pH 7.4; final pH 8.4). The column was eluted at 1.0 mL/min. The DMPD-DHA derivative was detected with excitation at 360 nm and emission at 440 nm. For measuring DHA content in cell supernatant, the supernatant was derivatized with DMPD without bromine oxidation. 

Inhibition of Glutathione Synthesis and Glutathione Contents in HLE-B3 Cells

To inhibit glutathione (GSH) synthesis, HLE-B3 cells were treated with 100 μM of l-buthionine-[S,R]-sulfoximine (BSO) overnight, i.e., at a concentration that did not have cytotoxicity in the MTT assay ((3-(4,5-cimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; Roche Molecular Biochemicals, Mannheim, Germany) (data not shown). Intracellular GSH levels were determined with a GSH assay kit (Calbiochem, San Diego, CA). Briefly, the cells (4 × 10⁶ cells) were collected by trypsinization and homogenized in 500 μL of 5% metaphosphoric acid with a pellet pestle (Kontes). The acid-precipitated protein was pelleted by centrifugation at 3000g for 10 minutes at 4°C. The supernatant was used for measurement of GSH content. 

Inhibition of F-ASA Transport with Cytochalasin B

Ascorbic acid can be taken up into cells by facilitated glucose transporters GLUT1 and GLUT3. ²³ To investigate whether F-ASA can be taken up by these transporters, an uptake experiment was performed in the presence of cytochalasin B, which is a known inhibitor of GLUT1 and GLUT3. HLE-B3 cells were cultured with 100 μM of F-ASA and 200 μM of cytochalasin B in 20% FBS/EMEM for 45 minutes at 37°C. The cells were collected by trypsinization and homogenized with a pellet pestle (Kontes) in 4% metaphosphoric acid/resin (Chelex-100; Bio-Rad)-treated PBS containing 100 μM of internal standard on ice. Supernatants were obtained from the lysate by centrifugation at 15,000 rpm for 20 minutes at 4°C and were analyzed by ¹⁹F-NMR spectroscopy after adding deuterium oxide to a final concentration of 10%. 

Role of Na⁺ for the Uptake of F-ASA and F-DHA by HLE-B3 Cells

The sodium-dependent vitamin transporter 2 (SVCT-2) is predominantly expressed on the membrane of HLE-B3. ²⁴ Sodium-dependent F-ASA or F-DHA uptake experiments were performed in Krebs solution containing 110 mM NaCl, 6 mM KCl, 1.1 mM MgCl₂, 5.5 mM glucose, 25 mM NaHCO₃, and 15 mM HEPES, adjusted to pH 7.4 with NaOH. For sodium-free conditions, all sodium was replaced with an equimolar amount of choline chloride, and the pH was adjusted to 7.4 with KOH. The cells were cultured to confluence, which was reached at 6 days. The culture medium was removed, and cells were washed three times with incubation buffer 30 minutes before the experiments. The uptake experiments were started by addition of buffer containing 1 mM F-ASA or 1 mM F-DHA. The incubations were continued for 2 hours with the buffer changed every 30 minutes. The incubations were terminated by removing the buffer and adding excess cold sodium-deficient buffer. After three washes with sodium-deficient buffer, the cells were prepared for ¹⁹F-NMR analysis, as mentioned earlier. 

Uptake of F-ASA Degradation Products by HLE-B3 Cells

Permeability to degradation compounds of F-ASA into HLE-B3 cells was tested. F-ASA (10 mM) was incubated for 3 days at 37°C in regular PBS to generate spontaneous degradation compounds of F-ASA. This mixture was diluted 10 times with 20% FBS/EMEM and added to the cells at the time they reached confluence. After 24 hours, the cells were collected by trypsinization and homogenized in 4% metaphosphoric acid as described earlier. The supernatants were obtained from the cell lysate by centrifugation at 15,000 rpm for 20 minutes at 4°C and analyzed by ¹⁹F-NMR spectroscopy, as described earlier. 

In Vivo Uptake of F-ASA into Rat Lenses

Animal experiments were performed according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Six-week-old Sprague-Dawley (SD) male rats were made diabetic by intraperitoneal injection of streptozotocin (65 mg/kg body weight). The rats were maintained for 6 months until cataracts were visible. F-ASA (4 mg/0.2 mL saline) was injected intraperitoneally every 12 hours for 5 days. Rats were killed with excess pentobarbital, eyeballs were removed, and lenses were carefully excised and immediately homogenized in 300 μL of metaphosphoric acid/resin (Chelex-100; Bio-Rad) buffer containing 100 μM internal standard. The clear supernatant obtained after centrifugation at 15,000 rpm for 20 minutes at 4°C was analyzed by ¹⁹F-NMR, as described earlier. 

F-ASA and F-DHA Uptake into Cultured Rat and Human Lenses

Rat lenses were cultured in Medium 199 (M199) containing Earle’s salts without phenol red (Life Technologies) and supplemented with 25 mM HEPES (pH 7.4), 100 mg/L glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin. The lenses were preincubated for 1 hour in culture medium at 37°C in a humidified atmosphere of 95% air and 5% CO₂ to check viability. If high protein outflow (more than 1 mg/mL) into the medium was observed, the lenses were discarded according to the recommendation of Tumminia et al. ²⁵ For F-ASA uptake experiments, 1 mM F-ASA was added to the medium, and the lenses were incubated as indicated for each experiment. For F-DHA uptake experiments, the medium containing 1 mM freshly prepared DHA was replaced every 2 hours for 12 hours, because it degrades rapidly. Human lenses were obtained within 24 hours of death (National Disease Research Interchange, Philadelphia, PA), in accordance with the provision of the Declaration of Helsinki for research involving human tissue, immediately placed into M199 containing Earle’s salts and cultured with 1 mM F-ASA or 1 mM F-DHA, as described earlier. Protein-free extract was prepared and analyzed by ¹⁹F-NMR, as described earlier. 

Degradation of F-ASA in Aqueous Humor from Rabbits

New Zealand White rabbit aqueous humor (∼0.2 mL) was collected from the anterior chamber of each eye of five animals. Local anesthesia was induced with topical application of 2% lidocaine gel (Xylocaine; AstraZeneca LP, Wilmington, DE). The cornea was punctured at the limbus with a 27-gauge needle attached to a 1-mL tuberculin-type syringe. Specimens were pooled (yield, ∼1.2 mL), purged with N₂ gas, and stored at −80°C. A 1-mM solution of F-ASA in aqueous humor was incubated in the dark for 24 hours in quartz NMR tubes. 

¹⁹F-NMR Spectroscopy

¹⁹F 705.5 MHz NMR spectra were obtained on a 750-MHz spectrometer (Unity Plus; Varian, Sunnyvale, CA) equipped with a 5 mm ¹H/¹³C/¹⁹F pulsed field gradient (PFG) triple-resonance probe (hardware configured for ¹⁹F detection and ¹H/¹³C decoupling). ¹⁹F spectra were acquired at 25°C using a 16.8-μs (∼90°C) ¹⁹F pulse width, 0.8-second acquisition time, 9.1-kHz spectral window, and gated ¹H decoupling (to suppress the nuclear Overhauser effect), using a 1.55-kHz decoupling field with Waltz modulation and 256 to 512 transients with a relaxation delay of 1 second. Data processing was performed on a computer workstation (Sun SPARC station-10; Sun Microsystems, Mountain View, CA; with VNMR software; Varian). The data were weighted with 2-Hz exponential line-broadening and zero filled before Fourier transformation. An internal standard consisting of F-d-glucose in D₂O was added immediately before the measurement. For calibration, the chemical shift of the downfield furanose conformer of F-glucose was set at −219 ppm, based on the shift of CCl₃F (δ_F = 76 ppm) as the external standard. The other F-glucose signal at −218.3 ppm probably corresponds to the pyranose conformer, as judged by the 60% to 40% ratio of these two signals. For quantitation, the sum of the two signals was assumed to represent 0.1 or 1.0 mM fluoro compound, as dictated by the experiments. 

Statistical Methods

The statistical significance was assessed on computer by analysis of variance (ANOVA; Excel; Microsoft, Redmond, WA). Data are expressed as the mean ± SD. 

Results

Processing of 6-Deoxy-6-fluoro-ascorbic Acid (F-ASA) by Cultured HLE-B3 Cells

We first investigated whether HLE-B3 cells can take up F-ASA from the culture medium and whether the latter contains degradation products that are detectable by ¹⁹F-NMR spectroscopy. A typical spectrum obtained with a 750-MHz instrument is shown in Figure 1 . Without proton decoupling, a multiplet was observed for F-ASA at −211.9 ppm with traces of F-DHA at −216.6 ppm (Fig. 1A) . Proton decoupling dramatically increased the sensitivity, as judged by the high signal-to-noise ratio in Figures 1A compared with Figure 1B . It also helped resolve peaks that were close to each other, because a single peak usually, but not necessarily, corresponds to a compound of distinct structure. In the case of the internal standard, F-Glu, two peaks corresponding to both F-Glu conformers at −218.3 and −219.0 ppm were obtained, the sum of which equaled either 1 mM or 0.1 mM of fluoro compound. 

After 24 hours of incubation of 1 mM F-ASA in culture medium (EMEM) at 37°C, three to five major products were observed at −211.5, −211.9 (F-ASA), −212.4, −213.8, and −218.2 ppm, with traces of degradation products at −213.15 and −216.6 ppm (F-DHA; Fig. 1C ). In contrast, one major product corresponding to F-ASA was present in the cell lysate (Fig. 1D) . Most of the degradation products, besides F-DHA and F-DKG at −213.8 ppm are not yet identified and are currently under investigation. 

To compare the degradation kinetics of F-ASA relative to ASA, both were added under identical experimental conditions at 1.0 and 1.5 mM concentrations to culture medium and assayed by ¹⁹F-NMR and HPLC, respectively, followed by incubation up to 24 hours at 37°C. Figure 2A shows that the half-life of F-ASA is 22.3 hours compared with 46.1 hours for ASA. This difference is most likely explained by the shorter half-life of F-DHA, which we found to be 2.5 minutes in 10% FBS (not shown), compared with DHA (t _1/2 = 6 minutes). ²⁶ The latter is more stable because of its ability to form a hemiketal ring involving the OH in position 6. ²³ Indeed, whereas DHA was detectable by HPLC, no F-DHA was observed by F-NMR in the culture medium (Fig. 2A) . F-DHA was observed only transiently in the initial phase of F-ASA oxidation (not shown). As shown in subsequent studies, this decreased stability of F-DHA turns into a major advantage for the study of F-ASA degradation products. In support, the absence of significant degradation products of F-ASA in the normal cultured cells shows that the redox state is stringently tilted toward reduction. 

To further investigate potential biological differences between F-ASA and native ASA, we determined cellular uptake kinetics of ASA by HPLC, compared with F-ASA by NMR spectroscopy. Figure 2B shows rapid and saturable uptake of native and fluorinated ASA. Plateau levels for F-ASA were 25% to 30% of native ASA. This lower plateau level could be explained by differences in V _max for uptake of F-ASA and ASA. Indeed, Rumsey et al. ²⁶ found 10 times higher uptake kinetics by fibroblasts for ascorbic acid than for 6-deoxy-6-iodo ascorbic acid. 

These experiments support the premise that the intracellular levels achieved by the fluorinated probe and the sensitivity of the method are sufficient to probe the effects of cellular stresses on F-ASA and its degradation products, despite differences in cellular uptake kinetics and steady state levels between F-ASA and native ASA. 

Transport Pathway of F-ASA into the Cells

F-ASA can be transported into cells in the form of dehydroascorbic acid by the facilitated hexose transporters GLUT1, GLUT3, and GLUT4, ²⁷ or in the form of ascorbic acid by sodium-dependent ascorbic acid transporters. ²⁸ To investigate the nature of the transporters potentially involved in F-ASA uptake, cytochalasin B, a known inhibitor of the hexose transporters, was applied with F-ASA. An F-ASA signal was detectable after 45 minutes of incubation with 100 μM F-ASA in control cells, but a minute signal for F-ASA was present in 200 μM cytochalasin B–treated cells (Fig. 3A 3B) . Quantitation revealed a decrease to 20.8% of control (Fig. 3C) , suggesting that F-ASA may be transported through GLUT1, GLUT3, or possibly GLUT4. However, because these transporters are specific for DHA, and because halogenated dehydroascorbic acids cannot form the hemiketal ring needed for uptake by these transporters, ²⁶ F-ASA must be transported by the SVCT2 transporter that has recently been described in this cell line. ²⁴ These investigators made the unexpected finding that cytochalasin B could inhibit up to 50% of ASA uptake by HLE-B3 cells. However, because cytochalasin B is not entirely specific for GLUT transporters, we tested the effects of Na⁺ depletion on the transport of F-ASA and F-DHA. Most surprising, the uptake of both molecules was totally suppressed (Fig. 3D) , suggesting that both F-ASA and F-DHA are taken up by the SVCT2 transporter. 

Effects of High Glucose and Galactose Levels on Processing of F-ASA by HLE-B3 Cells

Profound alterations of ascorbic acid homeostasis have been observed in diabetes and galactosemia. In the lens, exposure to high glucose and galactose levels leads to rapid depletion of GSH. ²⁹ ASA uptake by the epithelium, but not by the lens cortex, of diabetic rat is decreased. ³⁰ For this reason, we exposed subconfluent HLE-B3 cells to 50 mM glucose and galactose for 6 days. When the cells reached confluence, we investigated the impact of such treatment on F-ASA processing. Examples of spectra obtained 24 hours after addition of 1 mM F-ASA to control, galactose-, and glucose-treated cells are shown in Figures 4A 4B 4C . In all experiments, the major peak was F-ASA at −211.9 ppm. Small peaks for F-DHA and F-DKG were observed at −216.6 and −213.8 ppm, respectively. Quantitative analysis revealed that high-dose galactose and glucose treatment decreased F-ASA levels by only 16.8% and 15.7%, respectively (Figs. 4B 4C) . This decrease may be due to lower uptake or increased oxidation. In support of the latter possibility, there was a two-fold increase (P < 0.05) in the signal at −212.4 ppm in the galactose- and glucose-treated cells, respectively. F-DHA was also increased 1.5-fold (P < 0.05), but the magnitude of the increase was negligible, compared with the size of the F-ASA peak. Because the total recovery of the peaks was close to 100%, we concluded that high glucose and galactose levels caused relatively little change in the ascorbate homeostasis of cells. Most surprising, high glucose and galactose did not favor F-DHA and F-DKG as major degradation products of F-ASA, most likely because the effect on GSH levels were not sufficient to induce these changes. 

Given that intracellular GSH concentrations were much higher than the F-ASA concentrations—that is 19.32 and 15.64 nmol/10⁶ cells compared with approximately 0.390 nmol/10⁶ cells of F-ASA in galactose and glucose treatment, respectively—it is understandable that glucose and galactose had relatively little impact on the conversion of F-ASA to F-DHA and F-DKG. However, when the cells were treated with buthionine sulfoximine (BSO), an agent known to inhibit GSH synthesis, there was total loss of GSH (Table 1) that led to a 35.9% decrease in F-ASA levels and a 14-fold increase in F-DHA (P < 0.01). Most interesting, no other degradation products were observed besides the peak at −212.4 ppm. 

Effects of High Glucose and Galactose Levels on Processing of F-DHA by HLE-3B Cells

Depending on the type of cells, ascorbic acid is taken up into cells either as ASA or DHA that is then rapidly reduced by GSH to ASA. To investigate whether F-DHA can also be taken up, the F-ASA experiment was repeated with F-DHA instead. Because of the low stability of the former at physiological pH, any degradation product stemming from DHA that would accumulate in the culture medium might be taken up by the cell if the latter is permeable to them. Indeed several degradation products were present in the culture medium as little as 2 hours after incubation with F-DHA (Fig. 5B) and no F-DHA was detected compared with the signal in the medium at the start of the experiment. After 24 hours of incubation with repetitive changes of the medium to replenish F-DHA, a single major intracellular peak corresponding to F-ASA was detected, with less than 1% comprising F-DHA and F-DKG. No signal whatsoever was detected at −212.4 ppm (Fig. 5C) . It should be noted that, because of repetitive replenishment, the intracellular F-ASA concentration was 2.5 times higher than when the cells were incubated with F-ASA (Fig. 5D versus Fig. 4E ), thereby confirming the unexpected observation that F-DHA can be taken up by the cultured cells. 

Another surprise was the finding that prior treatment of the cells with 50 mM glucose or galactose not only failed to affect the intracellular F-ASA formation from F-DHA, but also had no impact on cellular F-DHA concentration, which remained very low (Fig. 5D) . Thus, whereas high glucose and galactose levels slightly decreased F-ASA uptake, they did not affect the uptake of F-DHA and its conversion to F-ASA. Because the level of intracellular GSH is important for maintenance of homeostasis of ascorbic acid, we also performed the F-DHA uptake experiment with cells treated with BSO. In this instance, F-DHA and F-DKG contents were significantly increased with diminution of F-ASA content (69.5% versus control; Fig. 5D ). Overall these results are very similar to those obtained upon treatment of HLE-B3 cells with glucose or galactose in the presence of F-ASA, and support the proposition that GSH is a key mediator of ascorbic acid homeostasis. Again, it is quite surprising that there were no significant fluorinated degradation products besides F-DHA, nor was the −212.4-ppm compound observed, suggesting the latter can form only from F-ASA. 

Comparative Processing of F-ASA by Human Choriocarcinoma and Mouse Lens Epithelial Cells

To investigate the specificity of the foregoing findings, identical experiments were performed with JAR cells (i.e., human choriocarcinoma cells of placental origin) as well as with the 17EM15 mouse lens epithelial cell line. The former were chosen based on a report describing the presence of SVCT2 in the cells. ³¹ Very similar results were obtained with the JAR cells, in that they accumulated mainly F-ASA from F-ASA or F-DHA on treatment with high glucose or galactose levels (not shown). However, F-ASA and F-DHA uptake was 10 times lower in the mouse lens cells (not shown), a finding that strongly suggests presence of low levels of SVCT2 transporter in these cells. 

HLE-B3 Cell Permeability to F-ASA Degradation Compounds

The remarkable resistance of HLE-B3 cells to forming significant amounts of F-ASA degradation products raised the question of whether alternative mechanisms such as uptake of reactive degradation products from, for example, the aqueous humor might be a source of intracellular carbonyl stress. Indeed, previous experiments in our laboratory showed that cataractous, but not normal, lenses are permeable to (DKG). ¹⁹ To investigate this possibility, F-ASA degradation products were first generated by incubation of F-ASA in non–resin (Chelex-100; Bio-Rad)-treated PBS for 72 hours (see the Methods section). Uptake of F-ASA degradation products by HLE-B3 cells was determined after a 24-hour incubation period and the results were compared with degradation products present in the medium. No F-ASA or F-DHA were present in the medium (Fig. 6A ; Table 2 ). However, prominent peaks were observed at −211.6, −212.4, −213.3, and −214.1 ppm, and an additional small peak for F-DKG at −213.8 ppm. Except for F-DKG and the peak at −214.1 ppm, three of the five peaks present in the medium were detected intracellularly (Fig. 6B , Table 2 ). In other words, these results show that the cell membrane is permeable to F-ASA degradation products. Of importance, the major signal at −212.4 ppm in Figures 6A and 6B , was also observed in 24-hour incubation of cells with 1 mM F-ASA in the presence of glucose or galactose or BSO (Fig. 4) . 

F-ASA and F-DHA Uptake by Cultured Normal and Diabetic Cataract Rat Lenses

The ability of intact lenses from normal or streptozotocin-diabetic rats to process F-ASA and F-DHA was investigated first by incubating the cultured lenses (n = 6) with 1 mM F-ASA for 24 hours or with 1 mM F-DHA with change of the medium every 2 hours. In the normal and diabetic lenses, 26% and 35% of the F-compounds were recovered as F-ASA, respectively on incubation with F-ASA (Fig. 7A) . As described earlier for ascorbic acid, ³² there was a 35% decrease in F-ASA uptake by diabetic, compared with normal, lenses. Of note, the major degradation product in both diabetic and normal lenses was the unknown compound at −212.4 ppm. The normal lenses accumulated more F-ASA degradation products than the diabetic lenses (Fig. 7B) , possibly reflecting a better ability to detoxify these products through oxidoreduction. 

In contrast to the HLE-B3 cells, no F-ASA was recovered on incubation with F-DHA in either normal or diabetic lenses. None of the unknown compound at −212.4 ppm was observed. Instead, F-DKG was the major product in the normal lens, which also tended to be more permeable to F-DHA degradation compounds. 

Effect of Cataractogenesis on Processing of F-ASA and F-DHA In Vivo

To reflect the in vivo situation accurately, F-ASA was administered intraperitoneally at a dose of 4 mg twice per day for five consecutive days to six control rats and rats with 6 months of diabetes. All diabetic rats had cataracts. After 5 days of injection, lens extracts were rapidly prepared and analyzed. F-ASA was detected in both normal and cataractous rat lenses (Fig. 8A) . As in the in vitro model, mean levels were 50% lower in the diabetic lenses. F-DHA was close to undetectable in normal lenses. Consistent with increased oxidant stress, F-DHA was dramatically elevated in the cataractous lenses (Fig. 8B) . Thus, these results are similar to those of the lenses incubated in vitro, except for the fact that no F-DHA was detected in lenses from normal rats injected with F-ASA. The unknown compound at −212.4 ppm was present in approximately half of the diabetic lenses tested (Fig. 8C) . Recovery analysis showed that the sum of the fluorinated degradation products equaled approximately 70% of F-ASA concentration in the normal lenses (i.e., approximately 30% of total fluoro compounds were unaccounted for as a result of diabetes and cataractogenesis). These results indicate that cataractous lenses could not take up or maintain ascorbic acid in reduced form and had more catabolites than normal rat lenses. In addition, DKG was not detected, suggesting it was either not formed or taken up. Alternatively, it was rapidly degraded or reacted with lens crystallins. 

Processing of F-ASA and F-DHA in Cultured Human Lenses

Central to our hypothesis is that the human lens should be susceptible to the accumulation or formation of F-degradation products that may act as precursors of lenticular chromophores, fluorophores, and cross-links. Therefore, we investigated the degradation pattern of F-ASA or F-DHA in human lenses obtained within 24 hours of death (NDRI) and shipped on ice in DMEM. On receipt, the lenses were preincubated for 60 minutes in ascorbic acid, and protein content was assayed as an indicator of viability according to Tumminia et al. ²⁵ Incubations with 1 mM F-ASA or F-DHA were performed for 12 hours. 

A typical profile of F-ASA degradation products in the medium and an extract from a 61-year-old human lens is shown in Figures 9A and 9B . The two profiles differed little, except for the new peak at −212.2 ppm (Fig. 9B , arrow) that appears to have originated from a conversion of the peak at −212.4 ppm (Fig. 9A , arrow). A total of 12 lenses from donors 19 to 74 years of age were analyzed. The data are shown in table format (Table 3) to better reflect the different behavior of individual lenses. Mean values for each signal were associated with large SDs, in part due to the “all or none” presence of certain peaks in some lenses. The results (Table 3) show that F-ASA (at −211.9 ppm) is the major peak detected, with a mean concentration of 29.3 ± 17.1 nmol/lens. A second important peak that was almost ubiquitously present at −212.4 ppm corresponded to a concentration of 8.4 ± 6.6 nmol/lens. This peak was the same as that previously observed in HLE-B3 cells and selected diabetic rat lenses. Other peaks were present in concentrations below 1.6 nmol per lens, and there did not appear to be a relationship with age, except perhaps for the unexpected finding that the F-ASA concentration was highest in the older lenses. 

When the contralateral lenses (n = 14) were incubated with F-DHA, using change of medium every 2 hours to compensate for the rapid degradation of F-DHA, little or no F-ASA was recovered with only one exception. However, repeat experiments with short-term incubation showed the lenses were indeed able to form F-ASA (not shown). The lenses accumulated a large number of degradation products. The data in Table 4 summarize the pattern of degradation products seen in individual lenses. Although the pattern was similar in many lenses, none of the observed signals corresponded to F-DHA, and the intralenticular pattern resembled the composition of the medium. 

Discussion

We have used a metabolomics approach to assess, without chromatographic separation, a large number of metabolites and catabolites of vitamin C in the lens. Depending on the experimental conditions, up to 20 signals derived from F-ASA, each corresponding to a distinct structure, was distinguished by high-resolution NMR spectroscopy with a dedicated fluorine probe and proton decoupling. Using a similar approach, others have recently performed metabolomic studies for fast initial screening of biodegradation pathways in newly isolated microorganisms. ³³ A number of unexpected findings came to light in our study. 

First, uptake studies revealed that both F-DHA and F-ASA are taken up into HLE-B3 cells, probably by the same transporter (i.e., the SVCT2 rather than the SVCT1 transporter). ²⁴ This interpretation is compatible with the fact that both transports are suppressed by removal of Na⁺ from the medium, and that F-DHA, being unable to form the cyclic hemiketal structure that is recognized by GLUT transporters, is structurally more closely related to native ASA than native DHA. Thus, one can conclude that the intracellular F-ASA concentration that is recorded in cells incubated with F-DHA is a reflection of the sum of the activity of SVCT2 plus the ability of the cellular redox system to reduce F-DHA to F-ASA. Similarly, because no F-DHA is taken up by GLUT transporters, ²⁶ the F-DHA to F-ASA ratio should be a good measure of the integrity of the GSH dependent redox state of the cell. This was most apparent in the experiment in which treatment with BSO to deplete GSH led to an increase in F-DHA. 

The uptake data obtained with intact rat and human lenses confirm that whole lenses are able to take up F-ASA and F-DHA. Compared with the cultured cells, however, the cultured lenses were not as efficient at keeping F-ASA in reduced form. Typically, F-ASA formation from F-DHA was observed only in the initial incubation period up to approximately 3 hours (not shown). Instead, the lenses avidly accumulated F-degradation products. Considerable efforts were made to minimize the elapsed time between death of the donor and the beginning of the experiment and to exclude lenses judged unsuitable based on protein leakage. ²⁵ However, we cannot exclude significant stress due to this type of processing artifact or the harshness of the culture conditions themselves. This may explain the paradoxical finding of absence of conversion of F-DHA into F-ASA by intact rat or human lenses after 12 and 24 hours of incubation, compared with the apparently more resistant HLE B-3 cells. 

Another remarkable finding is that the F-ASA homeostasis in HLE-B3 cells is highly resistant toward stress from high glucose and galactose levels. Profound depletion of GSH levels with agents such as BSO was needed to obtain a significant increase in F-DHA when cells were exposed to either F-ASA or F-DHA. This suggests that lens epithelial cells are highly efficient at keeping ascorbic acid in reduced form. Nevertheless, the sugar stress was associated with a twofold increase in the unknown compound at −212.4 ppm when exposed to F-ASA (Fig. 5D) . Although it is possible that this degradation product was formed intralenticularly, the experiments in which the cells were exposed to 24-hour degradation products of F-ASA (Fig. 6B) clearly show that it can be taken up passively, along with a few other products, except for the negatively charged F-DKG. It is noteworthy that cellular levels of the −212.4-ppm compound, in contrast to F-DHA, were not further increased on BSO depletion (Fig. 4E) , suggesting that its intracellular presence is not related to oxidant stress. 

The exposure of rat and human lenses to advanced degradation products of F-ASA and F-DHA has revealed that these lenses were permeable to several products. Most interesting, the pattern of degradation products from F-ASA and F-DHA was markedly different, revealing complete absence of the compound at −212.4 ppm in the F-DHA spectrum. This strongly suggests that ascorbic acid can be degraded without first being oxidized to DHA. A pathway involving triplet oxygen attack of ASA with threonic acid as a potential end product has been described by Miyake et al. ³⁴ Threonic acid, however, would also be expected to form from DKG, which is the immediate degradation product of DHA. Thus, threonic acid appears to be an unlikely end product. Of interest, the data in Fig. 9 suggest that the −212.4-ppm compound is converted by the lens into another product at −212.2 ppm, suggesting the former might be a substrate for oxidoreductases. 

The ability of the intact lens to take up potentially damaging degradation products of F-ASA and F-DHA from the medium suggests that it is of paramount importance to explore the stability of ascorbic and dehydroascorbic acid in the aqueous humor. In that regard, it should be pointed out that all experiments described in this work were performed at a physiological concentration (1 mM)—that is, that of ascorbic acid in the aqueous humor. 

Diabetes was associated with marked suppression of F-ASA uptake into the lens. Similar observations were made by DiMattio, ³⁰ who found a suppression of uptake of ¹⁴C-ASA into the lens epithelium from diabetic rats. One possible reason for the low F-ASA levels in the lenses of diabetic rats injected intraperitoneally with F-ASA may be the low plasma levels that were decreased by 65%. However, the decrease was also present when the cells were directly exposed to F-ASA (Fig. 7) . This suggests that SVCT2 expression may be impaired in diabetes. Whereas the in vitro uptake of several degradation products was decreased in diabetic lenses, their overall pattern was similar to that in normal lenses, except for the total absence of F-DHA by diabetic lenses (Fig. 7B) . Paradoxically, however, F-DHA and the −212.4-ppm compound were present in cataractous lenses from diabetic rats, suggesting that the cultured lens model does not replace in vivo experimentation. The latter data thus strongly suggest the −212.4-ppm compound originated intralenticularly, rather than from passive uptake from the aqueous humor. Nevertheless, experiments with rabbit aqueous humor also show this compound can form through spontaneous degradation of F-ASA in this medium (not shown). The tentative relationship between this compound, the culture medium, the aqueous humor, and the lens is summarized in Figure 10 . 

The chemical identity of most F-compounds derived from F-ASA and F-DHA remains to be established. Recently, while this study was in progress, Simpson and Ortwerth ³⁵ investigated the degradation of U-¹³C- ascorbic acid and came to the conclusion that erythrulose was a major degradation product from the “nonoxidative” degradation of ASA, DHA, and DKG. The discovery raised questions concerning the true existence of l-threose and l-xylosone as legitimate degradation products of ascorbic acid. ^{36 37} It also showed that l-threonic acid is obtained if oxidizing conditions are present. These findings raise the question of whether the −212.4-ppm compound is either erythrulose or threonic acid. F-erythrulose would appear unlikely, because the −212.4-ppm compound was also obtained under oxidizing conditions that would preclude formation of erythrulose. F-threonic acid is another possibility. However, the −212.4-ppm compound could not be obtained from F-DHA, a necessary precursor of F-DKG. Clearly, future experiments are needed to elucidate unequivocally the structure of this important degradation product. 

F-DHA, when incubated with lens crystallins, was as efficient at cross-linking as native DHA (not shown). However, attempts to detect fluorine incorporation into proteins incubated with F-DHA failed. Instead, free fluoride at −135 ppm was recovered (not shown), suggesting that enolization and fluoride elimination had occurred during the ascorbylation process and that some of the cross-links might involve carbon-6 of ascorbic acid. 

In summary, our studies have shown that F-ASA and F-DHA are taken up into HLE-B3 cells, most likely by the SVCT2 transporter, and that the cells are highly efficient at maintaining F-ASA in reduced form, even in the presence of substantial hyperglycemic stress. Both these cells and intact lenses are permeable to F-ASA and F-DHA degradation products, the composition of which was surprisingly markedly different. A major unknown compound at −212.4 ppm is present in cataractous lenses from diabetic rats that is obtained only from F-ASA. 

²  Present affiliation: Department of Chemistry, University of Gdansk, Gdansk, Poland.

 Supported by National Eye Institute Grant EY07099 and P30-EY11373 from the National Eye Institute and a Mentorship Grant from the American Diabetes Association. The Kresge Foundation provided funds from the Kresge Challenge fund at the University of Akron to purchase the nuclear magnetic resonance equipment used in this work.

 Submitted for publication June 12, 2002; revised November 13, 2002; accepted December 5, 2002.

 Disclosure: M. Satake, None; B. Dmochowska, None; Y. Nishikawa, None; J. Madaj, None; J. Xue, None; Z. Guo, None; D.V. Reddy, None; P.L. Rinaldi, None; V.M. Monnier, None

 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.

 

A 750-MHz ¹⁹F-NMR spectrum of F-ASA and its degradation products. (A) F-ASA (1 mM) in fresh culture medium, (B) F-ASA proton decoupled spectrum, (C) F-ASA (1 mM) incubated for 24 hours in culture medium, (D) NMR spectrum of cell lysate after a 24-hour incubation with 1 mM F-ASA.

Figure 2.

 

Kinetics of ASA and F-ASA degradation in culture medium (A) and uptake into HLE-B3 cells (B). The culture medium (A) and HLE-B3 cells (B) were periodically assayed for ASA and DHA by HPLC with a o-phenylenediamine derivative and F-ASA and F-DHA by ¹⁹F-NMR. Data represent means ± SD (n = 3).

Figure 3.

 

Effect of cytochalasin B and sodium depletion on F-ASA uptake by HLE-B3 cells. Cells treated without (A) and with (B) 200 μM cytochalasin B for 30 minutes before F-ASA incubation. (C) Quantitative analysis summarizing (A) and (B). Data represent the mean ± SD (n = 3). Effect of Na⁺ depletion on the uptake of F-ASA and F-DHA by HLE-B3 cells (D). Note: cells in panel A–C received 0.1 instead of 1 mM F-ASA.

Figure 4.

 

Effects of d-galactose and d-glucose on F-ASA uptake into HLE-B3. F-ASA was taken up into HLE-B3 cells for 24 hours after treatment with hyperglycemic conditions or BSO. ¹⁹F-NMR signals of control cells (A), cells treated with 50 mM d-galactose (B) or d-glucose for 6 days (C), or cells treated with 100 mM BSO for 24 hours (D). (E) Quantitative analysis of the major signals from (A–D) . Data represent the mean ± SD (n = 3).

Table 1.

 

Intra- and Extracellular Glutathione Levels under Hyperglycemic Conditions and after Treatment with BSO

	Intracellular (nmol/106 cells)	Medium (μM)
Control	20.83 ± 0.08	90.66 ± 6.6
50 mM d-galactose	19.32 ± 0.09	82.50 ± 3.1
50 mM d-glucose	15.64 ± 0.46	103.50 ± 3.7
100 μM BSO	1.30 ± 0.01*	29.50 ± 0.9*

Figure 5.

 

Effect of high glucose and galactose stress on the processing of F-DHA and its degradation products by HLE-B3 cells. ¹⁹F-NMR spectra of freshly made F-DHA through bromine oxidation of F-ASA (A), of F-DHA (1 mM) incubated for 2 hours in culture medium (B), and of HLE-B3 cell lysate on exposure to F-DHA (1 mM) which was pulsed every 2 hours for 24 hours (C). (D) Quantitative analysis of the major F-compounds recovered in HLE-B3 cells exposed to 50 mM d-glucose or d-galactose for 6 days or BSO for 24 hours. Data represent the mean ± SD (n = 3).

Figure 6.

 

Uptake of F-ASA degradation products by HLE-B3 cells. ¹⁹F-NMR spectra of F-ASA (1 mM) which was allowed to decompose spontaneously into the culture medium for 24 hours at 37°C (A) and from lysate of cells exposed to the medium for 24 hours (B).

Table 2.

 

Quantitative Analysis of the Fluoro Compounds Recovered Inside HLE-B3 Cells Exposed to F-ASA Degradation Products for 24 Hours

Unit	Signal									Total
		*	*	*	F-DKG	*
ppm	−211.6	−212.4	−213.3	−213.8	−214.1
μM	249.4	513.5	106.5	26.6	112.4	1008.3
pmol/106 cells	40.5 ± 2.4	111.3 ± 7.9	32.5 ± 6.5	ND	ND	184.2 ± 6.8

Figure 7.

 

Representative ¹⁹F-NMR spectra of lens extract from normal and diabetic rats lenses that were incubated with 1.0 mM F-ASA or F-DHA for 24 hours (A). Quantitative analysis shows 40% to 60% suppression of F-ASA and F-degradation products uptake (P < 0.05) and marked increase in F-DHA (−216.6 ppm) (P < 0.05) in diabetic lenses (B). The unknown compound at −212.4 was present in all lenses treated with F-ASA. F-ASA is absent in F-DHA incubated lenses for 24 hours, but detectable if shorter incubation times are used (not shown). Diabetes tended to suppress uptake of many of the DHA-derived peaks. Data represent the mean ± SD of results in three pools of each of two lenses.

Figure 8.

 

Representative ¹⁹F-NMR spectra of lens extract from normal (A) and diabetic rats (B) that were injected with 4 mg/kg F-ASA for 5 days before lens extraction. Quantitative analysis shows 50% suppression of F-ASA uptake (P < 0.05) and presence of F-DHA in diabetic lenses (C). The unknown compound at −212.4 was present in some but not all lenses. Data represent the mean ± SD from three pools of two lenses each. n.d., not detected.

Figure 9.

 

Representative ¹⁹F-NMR spectra of lens extract from 61-year-old donor lenses exposed to F-ASA or F-DHA degradation products. Signals observed after 12 hours of incubation of 1 mM F-ASA in medium (A) and lens (B), and after a 2-hour incubation of 1 mM F-DHA in medium (C) and a lens exposed for 12 hours to medium that was replaced every 2 hours (D).

Table 3.

 

Intralenticular Concentration of F-Compounds in Cultured Human Lenses Incubated in F-ASA

Age (y)	* −211.6, †	F-ASA −211.9, †	* −212.4, †	F-DKG −213.8, †	F-DHA −216.6, †	* −218.0, †	Total
19	0.0	34.9	6.2	0.0	0.0	0.0	41.1
44	0.0	10.9	3.2	0.0	0.0	0.0	14.1
45	1.6	13.7	13.2	2.4	0.0	2.9	33.8
50	1.5	24.6	4.8	0.0	0.0	2.5	33.4
58	0.0	26.5	1.4	0.0	0.0	0.0	27.9
59	0.0	28.0	2.3	1.3	0.0	0.0	31.6
61	4.0	12.8	17.7	3.0	0.0	2.7	40.3
63	2.1	10.8	16.0	2.7	0.0	2.7	34.3
64	0.0	25.6	0.0	0.0	12.6	0.0	38.3
66	2.5	52.0	17.2	3.2	2.4	3.3	80.7
72	2.2	57.9	13.0	3.1	0.0	5.1	81.4
74	1.3	53.8	5.1	1.7	0.0	0.0	61.9

Table 4.

 

Intralenticular Concentration of F-Compounds in Human Lenses Incubated in F-DHA

Age (y)	* −211.6, †	F-ASA −211.9, †	* −213.2, †	F-DKG −213.8, †	* −214.7, †	* −214.8, †	* −216.2, †	* −217.8, †	Total
19	0.0	4.6	0.0	0.0	0.0	0.0	0.0	0.0	4.6
44	0.0	78.8	3.8	1.8	3.2	8.5	0.0	0.0	96.1
45	1.3	0.0	1.2	2.0	1.3	1.5	0.0	0.0	7.3
50	2.9	0.0	6.1	10.2	15.8	9.4	4.7	2.3	51.3
52	3.6	0.0	3.6	2.7	5.3	5.3	0.0	0.0	20.6
58	3.9	2.2	8.8	9.2	20.8	12.7	4.3	6.7	68.6
59	4.0	3.6	3.6	8.1	10.0	8.1	2.0	3.6	43.2
61	10.4	2.3	11.8	23.7	17.5	13.8	6.2	10.5	96.1
63	12.5	1.4	11.6	19.6	16.0	14.4	5.7	6.2	87.3
66	3.0	0.0	4.4	6.1	8.9	5.9	3.0	2.5	33.7
66	6.2	0.0	6.9	5.0	6.5	4.8	2.8	5.3	37.5
70	4.0	0.0	0.0	5.0	0.0	0.0	0.0	0.0	9.0
74	4.2	0.0	5.0	4.6	8.4	5.8	2.5	3.3	33.8

Figure 10.

 

Conceptual relationship between F-ASA degradation in the extra- and intralenticular space. The data in bold mostly represent degradation/uptake pathways that have been observed in cultured lens epithelial cells or lenses from diabetic rats injected with F-ASA, except for F-DKG that was not found in diabetic lenses. The shaded products/pathways, including the −212.4-ppm compound, were observed in culture medium. Only the −212.4-ppm compound was present in rabbit aqueous humor at 24 hours. Although F-DKG was not passively taken up by cultured cells and lenses, several F-compounds readily accumulated intracellularly and intralenticularly in vitro. Finally, in all experiments no evidence for F-ASA generation from F-DHA in culture medium was observed.