Distribution of Ascorbate in the Anterior Bovine Eye

Amund Ringvold, Erlend Anderssen, and Inge Kjønniksen

PURPOSE. To analyze the ascorbate distribution in the anterior eye wall to better understand the functional signiﬁcance of this compound in the eye.

METHODS. Ascorbic acid was determined by high-performance liquid chromatography using an LC-10 system (Shimadzu, Kyoto, Japan). Bovine eye samples were used.

RESULTS. The highest ascorbate concentration was observed in the corneal epithelium, with signiﬁcantly higher values in the central (1.56 mg/g) than in the peripheral (1.39 mg/g) area. The ascorbate content was similar in the corneal stroma (0.22 mg/g), the Descemet’s membrane (DM)/endothelium (0.22 mg/g), and the aqueous humor (0.21 mg/ml). By comparison, the sclera (0.15 mg/g) and the conjunctiva (0.11 mg/g) showed lower values, as did the lacrimal gland (0.09 mg/g) and the serum (0.0008 mg/ml).

CONCLUSIONS. (1) Peak ascorbate concentration was observed in the central corneal epithelium covering the pupillary area. This is compatible with the idea that the ascorbate may act as an UV ﬁlter shielding internal eye structures from radiation damage. (2) The ascorbate concentration in the corneal stroma and DM/endothelium was as high as in the aqueous humor, and it is suggested that the aqueous humor plays a key role in the distribution of ascorbate to the anterior eye wall. (Invest Ophthalmol Vis Sci. 2000;41:20–23)

ince the first report of significant amounts of ascorbate in better understanding of the ascorbate distribution in the cor- the aqueous humor,1 high concentrations have been ob- nea and its surrounding tissues may provide some clues as to its served in many parts of the eye,2 with peak values in the functional aspects, so we decided to analyze the ascorbate S 3,4 corneal epithelium. It has turned out, however, that the concentration in different parts of the anterior eye. The bovine ascorbate content is higher in diurnal than in nocturnal mam- eye was chosen as the model because of its size, availability, mals, both in the aqueous humor5–7 and in the corneal epithe- and significant ascorbate content. lium.8 Indeed, the amount of ascorbate in different ocular compartments seems adjusted to the suggested ambient radiation dose at each particular level, and from these observations MATERIALS AND METHODS it has been deduced that the ascorbate acts as a UV filter protecting the eye from radiation damage. Fresh bovine eyes were collected at the abattoir. For practical During recent years, some experimental support has been reasons, the time span between killing and enucleation could presented for this hypothesis. Reddy et al.,9 who compared the not be reduced below 15 minutes. With subsequent time- effect of UV radiation on DNA strand breaks in the lens epi- consuming preparation procedures added, the total time from thelium of rat and guinea pig, concluded that high levels of death to frozen specimen was up to 2.5 hours in the most ascorbate in the aqueous humor of diurnal animals may protect detailed experiments. Two different pilot studies were there- the lens against UV radiation under physiological conditions. fore performed to test the stability of the cornea and its sur- Furthermore, it has been shown that the most efficient radia- rounding tissues/fluids during the preparation period. When tion in cell killing in cultured rabbit lens epithelium is the not otherwise specified, paired sampling was performed from 297-nm wavelength,10 i.e., just those rays most likely to be the single eye throughout. Eyes deliberately stored before absorbed by the aqueous ascorbate in vivo.5 However, Wil- preparation were kept in closed plastic bags at room temper- liams and Delamere11 have pointed out that the lack of anti- ature to reduce the loss of water and to prevent reduced ion oxidant protection due to low ascorbate in the nocturnal pump activity in the anterior eye. The bovine cornea is oval; aqueous might be compensated for by the high activity of a the longest and shortest diameters were 29.1 Ϯ 0.1 and 22.7 Ϯ peroxidase enzyme. 0.1 mm (10 eyes), respectively. Many questions remain unanswered concerning the sig- nificance of ascorbate for the cornea and its subjacent struc- Pilot Studies tures. Of particular interest is how it gets to the cornea. A (a) The amount of postmortem ascorbate loss was estimated by comparing the ascorbate content in en bloc specimens of cornea and conjunctiva/sclera at 15 minutes and 2.5 hours From The National Hospital Pharmacy, Oslo, Norway. after death, respectively. Such crude specimens were collected Submitted for publication April 8, 1999; revised August 16, 1999; readily within a few minutes of the desired time by opening the accepted September 3, 1999. cornea at the external limbal demarcation with a razor blade. Commercial relationships policy: N. Corresponding author: Amund Ringvold, Eye Department, Na- The specimen was cut free with a pair of scissors, and a 5-mm tional Hospital, University of Oslo, Pilestr. 32, N-0027 Oslo, Norway. broad rim of the adjacent conjunctival/scleral tissue was iso- [email protected] lated. The specimens were then immediately incubated en bloc

Downloaded from iovs.arvojournals.org on 09/26/2021 IOVS, January 2000, Vol. 41, No. 1 Ascorbate and Anterior Bovine Eye 21

(see below), there being no further time-consuming proce- (0.5–1.5 ml) of metaphosphoric acid of 100 g/l water were dures. used, depending on the size of the specimens. The epithelial, (b) The aqueous humor, tear fluid, and limbal blood ves- conjunctival, DM/endothelial, and lacrimal gland specimens sels are putative sources of ascorbate to the cornea, and their were homogenized as far as possible with a Teflon pestle in the ascorbate concentrations were analyzed accordingly. The original vials (10 ml disposable glass containers without any aqueous humor was tested 15 minutes, 1, 2, and 3 hours after coating) for 1 minute, before being frozen for storage in the death. Blood was collected from living animals and lacrimal same containers. Because the corneal stroma and sclera were gland from 30 to 60 minutes postmortem. assumed to resist this treatment and because homogenization Subsequently, the ascorbate distribution in the various with a metal pestle would induce ascorbate oxidation, speci- parts of the anterior eye wall was analyzed in four different mens containing these elements were frozen slowly experiments, each of which was run twice. Separately isolated, (Ϫ5–10°C) and thawed three times in the refrigerator to create fresh tissues were used for each run. a spongious tissue accessible to the incubation solution. Minc- ing such specimens by cryosectioning also was tested in sep- Experiment I arate experiments. However, a ceramic knife was too blunt, The ascorbate contents in the central versus the peripheral and a metal knife reduced the amount of ascorbate by some corneal epithelium were compared. The central area was de- 50% (not shown). Specimens were stored at Ϫ35°C and ana- fined by a corneal trephine 9 mm in diameter placed center to lyzed within 2 weeks. center with the pupil, the peripheral area outside reaching to The high-performance liquid chromagraphy (HPLC) sys- the limbus. The respective specimens were then rubbed off tem consisted of a Shimadzu LC-10 system (Kyoto, Japan) using with a diamond knife or a microscope slide. a SPD-M10AVP detector, a SIL-10A autoinjector, an LC-10AS pump, degasser, and LC-10 software. The chromatographic Experiment II conditions were adapted from Rodriguez et al.12 The column used was a Supelcosil LC-18, 250 ϫ 4.6 mm, 5-␮m particle size The central epithelium was further analyzed separately. In this (Supelco, Bellafonte, PA), with a small guard column contain- experiment, the central epithelium was defined as the circular ing the same material. The mobile phase consisted of HPLC- area covering the total pupil. A plastic template imitating the grade water acidified to pH 2.2 with sulfuric acid (isocratic pupil’s form, which is oval in the bovine eye, was created. This method). The flow rate was 1.0 ml/min, with the detection template, formed by two parallel sides 6 mm apart closed by performed at 243 nm. The column was washed regularly in two half-circles with the top points 12 mm apart, was placed grade water. on the epithelium to indicate the area of study. The template The specimens were thawed to room temperature and was subsequently circumscribed with a diamond knife, and the centrifuged at 12,000g for 10 minutes, and the supernatants marked area was rubbed off. Because paired sampling was were injected (20 ␮l) in triplicate into the HPLC apparatus impossible, the number of eyes was doubled. Two separate from injector glass vials without further dilution. The analyst series were run, one with the template corresponding to the was not informed about the type and order of the specimens. pupil and one with the template rotated 90° to it. In the Standard curves were obtained after triplicate injection of 0.01, following, these specimens are referred to as “horizontal” and 0.1, 0.25, and 0.5 mg/ml of analytical grade ascorbic acid “vertical,” respectively. dissolved in 10% metaphosphoric acid. The correlation coeffi- Experiment III cient of the standard curves was always greater than 0.999, and linearity was shown to be between 0.005 and 0.5 mg/ml. Anterior eye wall tissue was analyzed. The total epithelium was Minimum detection limit of the standard was found to be 10 rubbed off, and a limited aqueous sample aspirated before the ng/ml. The variation coefficient of the injection repeatability cornea was cut out along the external limbus with a pair of was less than 1%, while interday variation coefficient of iden- scissors. Finally, the corneal stroma and the Descemet’s mem- tical bovine corneal samples was 13%. brane (DM)/endothelium were collected. Subsequently, the adjacent 5 mm broad rim of conjunctival and scleral tissues was Statistics isolated separately. Ascorbate concentrations are presented as means Ϯ SD. Stu- Experiment IV dent’s t-test (two-sided) was used. To analyze the cornea in more detail, central and peripheral specimens were prepared. The epithelium was again subdivided into a central and a peripheral area using the 9 mm RESULTS trephine and collected as described above, before the stroma was arbitrarily subdivided into two separate sheets with a Pilot Studies spatula from a limbal incision. A limited sample of aqueous was (a) The ascorbate contents of the en bloc cornea and conjunc- then aspirated, the cornea cut out along the external limbus, tiva/sclera specimens collected 15 minutes after death were and a central corneal tissue button (9 mm across) punched out 0.33 Ϯ 0.03 and 0.107 Ϯ 0.004 mg/g, respectively. The con- from behind. Each of these corneal specimens was separated centrations were reduced by 15.2% and 11.2%, respectively, in into anterior, posterior stroma, and DM/endothelium. These tissues analyzed 2.5 hours after death. Six eyes were used. were the most time-consuming experiments. (b) An ascorbate content of roughly 0.2 mg/ml in the Aliquots of aqueous humor were mixed 50/50 with pre- aqueous humor remained unchanged from 15 minutes to 1, 2, cooled metaphosphoric acid (50 g/l deionized water) before and 3 hours after death (6 eyes in each group). The lacrimal storage. For serum and tissue specimens, different volumes gland (30–60 minutes postmortem) and serum (from living

Downloaded from iovs.arvojournals.org on 09/26/2021 22 Ringvold et al. IOVS, January 2000, Vol. 41, No. 1

FIGURE 1. Ascorbate concentrations in various regions of the anterior bovine eye. The data are expressed as milligrams per gram tissue, except for the aqueous humor and serum, which are milligrams per milliliters.

animals) values were considerably lower [0.09 Ϯ 0.03 mg/g (3 together the following values were obtained: central cornea: samples) and 0.0008 Ϯ 0.001 mg/ml (3 samples), respectively]. epithelium (1.56 Ϯ 0.18 mg/g), anterior (0.21 Ϯ 0.04 mg/g) and posterior (0.18 Ϯ 0.03 mg/g) stroma, and DM/endothelium Experiment I (0.19 Ϯ 0.04 mg/g); peripheral cornea: epithelium (1.39 Ϯ The ascorbate concentration in the central (9 mm diameter 0.11 mg/g), anterior (0.24 Ϯ 0.02 mg/g) and posterior (0.20 Ϯ trephine) versus peripheral epithelium was tested. Two sepa- 0.02 mg/g) stroma, and DM/endothelium (0.23 Ϯ 0.02 mg/g). rate runs, each with six paired specimens, were performed on The aqueous humor showed 0.21 Ϯ 0.02 mg/ml. Again there different days and showed essentially the same results. The was a significant difference in ascorbate concentration be- mean ascorbate concentrations were 1.58 Ϯ 0.22 and 1.28 Ϯ tween the central and peripheral epithelium (P ϭ 0.01), 0.09 mg/g in the central and peripheral corneal epithelium, whereas central and peripheral concentrations were essentially respectively. This difference is statistically significant (P Ͻ the same in the stroma and in the DM/endothelium. The 0.001). The central and peripheral epithelium taken together ascorbate contents of the epithelium, stroma, and DM/endo- showed 1.31 Ϯ 0.08 mg/g. thelium as separate layers were calculated at 1.41 Ϯ 0.11, 0.22 Ϯ 0.02, and 0.22 Ϯ 0.02 mg/g, respectively, and taken Experiment II together the pooled value of the total cornea was 0.36 Ϯ 0.04 The ascorbate distribution in the central epithelial area was mg/g. tested in more detail by comparing horizontal and vertical A short version of the ascorbate content in various ante- specimens. Two runs, each with 12 eyes, showed mean ascor- rior eye structures is highlighted in Figure 1. bate concentrations of 1.22 Ϯ 0.13 and 1.23 Ϯ 0.17 mg/g, respectively, i.e., the values from the horizontal and vertical areas were not significantly different (P ϭ 0.9). DISCUSSION Three methodological aspects of this study deserve special Experiment III attention: the HPLC technique, the possibility of postmortem The ascorbate distributions in the anterior eye segment were ascorbate loss from the cornea, and the tissue extraction pro- analyzed in two separate runs (6 ϩ 6 eyes) showing essentially cedure of the ascorbate. the same results: corneal epithelium 1.33 Ϯ 0.15 mg/g, corneal The HPLC technique was performed according to well- stroma 0.19 Ϯ 0.03 mg/g, DM/endothelium 0.22 Ϯ 0.05 mg/g, established strategies12 and has proved reliable in a previous aqueous humor 0.19 Ϯ 0.04 mg/ml, conjunctiva 0.11 Ϯ 0.02 work.8 In addition, all aqueous humor samples, i.e., specimens mg/g, and sclera 0.15 Ϯ 0.02 mg/g. As seen, the stroma, without extraction bias, came out with predictable results. DM/endothelium, and aqueous humor were similar and signif- Together, these factors confirm the reliability of the HPLC icantly lower than the epithelium (P Ͻ 0.001). In addition, the procedure. scleral concentration was lower than the stromal, and the The level of ascorbate is remarkably stable in the aqueous conjunctival lower than the scleral (P Ͻ 0.008). By weight, the humor from stored eyes. A previous report13 revealed that the epithelium, stroma, and DM/endothelium made up 14%, 75%, ascorbate decline in bicarbonate–Ringer medium with EDTA is and 11% of the total cornea, respectively. For purposes of negligible after 90 minutes even when the solution is well comparison with the pilot results, pooled values of the total aerated, and we found that the ascorbate level in aqueous cornea and conjunctiva/sclera were calculated at 0.34 Ϯ 0.03 stored intraocularly was stable for 3 hours. The ascorbate loss and 0.14 Ϯ 0.02 mg/g, respectively. from the tissues during preparation was estimated at 15% for the most time-consuming experiments (experiment IV). This is Experiment IV a moderate reduction, and it is noteworthy that the results The cornea was separated into a central (9 mm diameter from experiment IV match the comparable figures obtained by trephine) and a peripheral part and then analyzed with regard shorter preparation (experiments I and III). We therefore con- to ascorbate content in the epithelium, anterior stroma, poste- clude that the postmortem ascorbate loss during this study was rior stroma, and DM/endothelium in each of these pieces. The a minor problem that does not discredit our results. aqueous humor also was tested. As in experiment II, the two As to the tissue extraction procedure of ascorbate, differ- separate runs (6 ϩ 6 eyes) showed similar results, and taken ent approaches were used. The epithelium, a purely cellular

Downloaded from iovs.arvojournals.org on 09/26/2021 IOVS, January 2000, Vol. 41, No. 1 Ascorbate and Anterior Bovine Eye 23

tissue, was easily homogenized, and so there is unlikely to be range of aqueous humor values.16 The possibility that the any extraction bias for these specimens. The corneal stroma corneal epithelium may receive some ascorbate from the tears and sclera, on the other hand, contain for the most part colla- therefore cannot be excluded. However, the main concentra- gen, which made homogenization with nonmetallic equipment tion mechanism in the epithelium is not confined to the super- impossible. This is why the freeze–thaw procedure was cho- ficial cells, which are preparing for desquamation. The basal sen. The fact that similar concentrations were obtained from layer is a more reasonable entrance to the epithelium, and stromal tissue extracted in one piece (experiment III) com- these cells are reliant on supplies from the stroma and aqueous pared to stromal tissue subdivided before extraction (experi- humor. ment IV) indicates that the results are reliable. The figures are useful at least as minimum values. Acknowledgments The corneal epithelium contained significantly more The authors thank Diana de Besche, Eli Gulliksen, and Astrid Østerud ascorbate than the stroma in both experiment III and experi- for technical assistance. ment IV, and so our study supports previous reports showing an ascorbate concentration mechanism in the corneal epithe- References 3,4,8 lium. However, the main observation in the present study 1. Harris LJ. Chemical test for vitamin C, and the reducing substances is that the level of ascorbate in the central region of the present in tumour and other tissues. Nature. 1933;132:27–28. epithelium is 12% to 23% higher than in the periphery (exper- 2. Heath H. The distribution and possible functions of ascorbic acid iments I and IV). This is not entirely unexpected, since the in the eye. Exp Eye Res. 1962;1:362–367. 3. Pirie A. Ascorbic acid content of cornea. Biochem J. 1946;40:96– corneal epithelial distribution of all small solutes concentrated 100. from the aqueous would be higher centrally because of diffu- 4. Reim M, Seidl M, Brucker K. Accumulation of ascorbic acid in the sion toward the limbus and loss to the blood. However, two corneal epithelium. Ophthalmic Res. 1978;10:135–139. aspects do not fit in with this view: (1) The bovine cornea is 5. Ringvold A. Aqueous humour and ultraviolet radiation. Acta Oph- oval, and diffusion toward the limbus should tend to create an thalmol (Copenh). 1980;58:69–82. oval form of the central high-loaded field. According to exper- 6. Reiss GR, Werness PG, Zollman PE, Brubaker RF. Ascorbic acid levels in the aqueous humor of nocturnal and diurnal mammals. iment II, the field with the highest concentration is circular. (2) Arch Ophthalmol. 1986;104:753–755. Diffusion toward the limbus also should be traceable within 7. Koskela TK, Reiss GR, Brubaker RF, Ellefson RD. Is the high the stroma. Experiment IV did not indicate any difference in concentration of ascorbic acid in the eye an adaptation to intense ascorbate concentration in central versus peripheral stroma. solar irradiation? Invest Ophthalmol Vis Sci. 1989;30:2265–2267. However, regardless of which mechanism is responsible, the 8. Ringvold A, Anderssen E, Kjo¨nniksen I. Ascorbate in the corneal fact remains that the highest epithelial ascorbate concentration epithelium of diurnal and nocturnal species. Invest Ophthalmol Vis Sci. 1998;39:2774–2777. covers the pupil. One cannot exclude that the UV-absorbing 9. Reddy VN, Giblin FJ, Lin L–R, Chakrapani B. The effect of aqueous ability of ascorbate may help minimize radiation damage in humor ascorbate on ultraviolet-B-induced DNA damage in lens subjacent eye structures. epithelium. Invest Ophthalmol Vis Sci. 1998;39:344–350. How does the ascorbate get to the cornea? Principally, 10. Andley UP, Lewis RM, Reddan JR, Kochevar IE. Action spectrum three sources are possible: the aqueous humor, the tears, and for cytotoxicity in the UVA- and UVB-wavelength region in cultured lens epithelial cells. Invest Ophthalmol Vis Sci. 1994;35: the conjunctival blood vessels. 367–373. The high ascorbate content in the aqueous humor, its 11. Williams RN, Delamere NA. A comparative study on the peroxi- proximity to the cornea, and the suggested ascorbate pump in dase activity within the aqueous humor of the rabbit, guinea-pig the corneal endothelium14,15 make this fluid a likely source for and rat. Comp Biochem Physiol. 1986;85B:585–587. the cornea. It has to be kept in mind that the DM/endothelium 12. Rodriguez MAR, Oderiz MLV, Hernandez JL, Lozano JS. Determi- in the present study represents mixed specimens in which the nation of vitamin C and organic acids in various fruits by HPLC. J Chromatogr Sci. 1992;30:433–437. 1 endothelium made up only about ⁄10 by volume (as judged 13. Riley MV, Schwartz CA, Peters MI. Interactions of ascorbate and from bovine corneal sections, not shown), and so the endothe- H2O2: implications for in vitro studies of lens and cornea. Curr Eye lial concentration could have been higher than that reflected in Res. 1986;5:207–216. the numbers. 14. Bode AM, Vanderpool SS, Carlson EC, Meyer DA, Rose RC. Ascor- The ascorbate content in bovine tears also is unknown, bic acid uptake and metabolism by corneal endothelium. Invest Ophthalmol Vis Sci. 1991;32:2266–2271. but the lacrimal gland revealed much higher values than the 15. DiMattio J. Ascorbic acid entry into cornea of rat and guinea pig. serum (0.09 mg/g and 0.0008 mg/ml, respectively). By analogy, Cornea. 1992;11:53–65. in humans, considerably higher ascorbate values have been 16. Paterson CA, O’Rourke MC. Vitamin C levels in human tears. Arch found in tears than in plasma, in basal tears even within the Ophthalmol. 1987;105:376–377.

Downloaded from iovs.arvojournals.org on 09/26/2021