- Email: [email protected]
Pump Function of the Human Corneal Endothelium
Pump Function of the Human Corneal Endothelium Effects of Age and Cornea Guttata DAYLE H. GEROSKI, PhD, MAMORU MATSUDA, MD, RICHARD W. YEE, MD, HENRY F. EDELHAUSER, PhD
Abstract: The specific binding of tritiated ouabain to endothelial Na/K ATPase was used to quantitate the density of pump sites in the human corneal endothelium. Donor eyes, unsuitable for use in keratoplasty, were obtained from the Wisconsin Lions Eye Bank. The endothelium of each donor eye was examined using wide-field specular microscopy, and the specular micrographs were traced and digitized for the determination of cell density. Ouabain binding was measured in matched pairs of isolated endothelial sheets. A total of 26 pairs of donor eyes, ranging in age from 11 through 91 years, were studied. Twenty pairs, determined to have normal endothelia, were found to have a constant pump site density which was independent of donor age. Six donor pairs had moderate guttata; in this group pump site density was significantly increased. These results indicate that, although pump site density is normally constant in the human corneal endothelium, conditions which increase endothelial permeability, such as guttata, can cause a compensatory increase in pump site density and presumably pump function. [Key words: cornea, cornea guttata, corneal endothelium, endothelial pump function, Fuchs' dystrophy, sodium-potassium adenosine triphosphatase.] Ophthalmology 92:759-763, 1985
A delicate balance of fluid movement is normally maintained across the posterior corneal surface. The corneal endothelium mediates this balance by acting as a barrier that limits the movement of fluid into the hydrophilic stroma, and by actively removing the fluid which manages to leak through its somewhat leaky barrier-the well-known barrier and pump functions of From the Departments of Ophthalmology and Physiology, Medical College of Wisconsin, Milwaukee, Wisconsin. Dr. Matsuda is a Visiting Scientist from the University of Osaka, Osaka, Japan. Presented at the Eighty·ninth Annual Meeting of the American Academy of Ophthalmology, Atlanta, Georgia, November 11-15,1984. Supported in part by NEI Research Grants EY·00933 and EY·05609 and Ophthalmic Research Core Grant EY-01931. Dr. Edelhauser is a 1982 Olga K. Wiess Research Scholar for Research to Prevent Blindness, Inc. Reprint requests to Dayle H. Geroski, PhD, Department of Physiology, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, Wisconsin 53226.
the endothelium. As long as the leak rate is exactly matched by the rate of the metabolic pump, stromal hydration is maintained at 78% water, and corneal transparency is maintained. Clinically, corneal edema will develop when the rate of fluid leak exceeds endothelial pump capacity. Though the barrier and pump functions of the corneal endothelium are the well established mediators of stromal hydration and corneal transparency, little is known about how these mechanisms are altered in disease. Until recently, our understanding of endothelial function as it relates to the pathogenesis of corneal edema has been largely inferential, based primarily on measurements of corneal thickness. Though pachymetry provides a valuable index of endothelial function, corneal thickness is a function of both endothelial barrier and pump functions; and increased corneal thickness can result from alterations in either or both of these functions. With the advent of fluorophotometry, it has become possible to clinically document alterations in endothelial 759
OPHTHALMOLOGY
•
JUNE 1985
•
VOLUME 92
•
NUMBER 6
Fig 1. Specular photomicrograph of normal donor cornea. Donor age 87 years, cell density 2237 cells/mm2. Each division on calibration grid represents 50 I'm.
Fig 2. Specular photomicrograph of donor cornea having moderate guttata. Donor age 66 years, cell density 1730 cells/mm2. Each division on calibration grid represents 50 I'm.
barrier function. Using this technique, alterations in endothelial permeability have been quantitated following penetrating keratoplasty,l in bullous keratopathy,2 and in Fuchs' dystrophy. 3 Though the functional status of the endothelial pump has been deduced from such measurements, alterations in endothelial pump capacity have yet to be documented. In these experiments, the specific binding of tritiated ouabain to endothelial Na/K ATPase (an essential component of the endothelial pump mechanism) was used to compare the density of Na/K pump sites in normal human donor endothelia to that in tissue having cornea guttata.
MATERIALS AND METHODS Matched pairs of intact donor eyes, unsuitable for use in keratoplasty, were obtained from the Wisconsin Lions Eye Bank. For the 26 pairs used in this study, the donor age range was 11 through 91 years, with the mean (±SE) donor age being 58.8 ± 4.7 years. The mean (±SE) time elapsed from death to the experimental use of the tissue was 21.0 ± 2.8 hours, within normal eyebank storage limits. The corneal endothelium of each donor eye was examined and photographed using a Keeler-Konan widefield specular microscope. Upon examination, 20 of the 26 donor pairs were found to have normal endothelia (Fig 1), while 6 pairs were determined to have early Fuchs' dystrophy with moderate guttata (Fig 2). For the determination of cell density, the specular photomicrographs were enlarged to a final magnification of X400.
760
Individual cells in clusters of 100 to 150 were traced and subjected to computer digitization. Endothelial Na/K ATPase "pump sites" were quantitated utilizing the specific binding of tritiated ouabain to these sites in sheets of isolated endothelium. These techniques, which have been previously described,4 are summarized here. After specular microscopy, the corneas are deepithelialized and excised. Each cornea of the matched pair is pre-incubated for one hour in potassium-free bicarbonate Ringer's (3rC). This pre-incubation was found to be necessary to achieve reproducible levels of ouabain binding in human tissue. This addition to the methods represents the only distinction from the procedure originally reported for rabbit tissue. Since phosphorylation of the Na/K ATPase is a prerequisite for glycoside binding, 5 it is likely that the pre-incubation provides a temperature-reversal period for endothelial metabolism, which is effectively shut off at the cold storage temperature (4°C). It is also known that the aqueous humor potassium concentration rises in the postmortem eye. 6 Since extracellular potassium has been shown to retard the rate of glycoside binding, 7 it is probable that the pre-incubation serves to remove the excess potassium from the endothelial extracellular space. Following the one-hour pre-incubation, each cornea was incubated for three hours (37°C) in potassium-free bicarbonate Ringer's containing 2 X 10-7 M tritiated ouabain. One cornea of the matched pair additionally received 10-4 M unlabeled ouabain to determine the nonsaturable component of endothelial ouabain uptake. At the end of the incubation, the corneas were rinsed, an 8 mm central button was trephined, and the endo-
GEROSKI.
4000
N
et aI
•
ENDOTHELIAL PUMP FUNCTION
8.0
•
•
Q 6.0
3000
N
•
~
Cii 2000 u
•• •
E
~ 4.0 (/)
.2!
•
•
~
E E ......
F •
•
•
n --
~~
•
•
'ji;
0.
• •
i
• • •
•
I: (4.4:!:O.21
••
E
0: 2.0 20
40
60
80
100
DONOR AGE (yrs) Fig 3. Endothelial cell density vs. donor age for donor corneas having normal endothelia. The correlation between donor age and cell density was significant (P < 0.05), and the calculated regression line is shown. Predicted cell density (cells/mm2) = 3384 - 10.7 (donor age).
thelium (plus Descemet's) of this button was removed as an intact sheet. Each sheet of endothelium was digested, and ouabain uptake was measured using liquid scintillation. Ouabain, bound by each pair of endothelial sheets, was calculated as the difference in tracer uptake measured in the absence and presence of 10-4 M unlabeled glycoside.
20
Central endothelial cell density as a function of donor age for the 20 pairs of donor corneas having normal endothelia is shown in Figure 3. In this group of donor corneas, a significant (P < 0.01) negative correlation was found between donor age and central cell density: predicted cell density = 3384 cells/mm2 - 10.7 (donor age). In contrast to this significant decrease in cell density with increasing donor age, the amount of ouabain bound by the corneal endothelium, and hence the density of Na/K ATPase pump sites, was found to remain constant and independent of donor age (Fig 4). Over a range of donor ages which extended from 11 through 91 years, Na/K pump site density was measured as 4.4 ± 0.2 X 109 pump sites/mm2 of endothelium (mean ± SE). It should be noted that this figure most likely represents an underestimate of pump site density. Because of the limited availability of human tissue, it is not possible to perform the detailed kinetic and equilibrium analyses of glycoside binding as were done for rabbit endothelium. 4 For this reason, human binding studies were carried out at a single glycoside concentration of 2 X 10-7 M. Previous studies have demonstrated that at this concentration non-specific binding is minimal and pump sites are 90% saturated in the rabbit endothelium.4 Thus, the actual density of Na/K pump sites in the human corneal endothelium is probably somewhat higher than the values reported here. Central endothelial cell density versus donor age for
60
80
100
DONOR AGE (yrs) Fig 4. Endothelial Na/K pump site density vs. donor age for donor corneas having normal endothelia. Pump site density was found to remain constant (4.4 ± 0.2 X 109 sites/mm 2) and independent of donor age. Mean ± SE density is represented by shaded region.
4000
N
RESULTS
40
3000
E E ...... .!!! 2000 Cii u
xx X
20
40
60
DONOR AGE (yrs)
x X
80
100
Fig 5. Endothelial cell density vs. donor age for donor corneas having guttata. Solid line represents the regression line of cell density vs. donor age for normal corneas.
the 6 donor pairs having guttata is shown in Figure 5. Since the presence of central guttata in these corneas interferes with the specular reflection from these regions, cell density could only be determined in guttata-free regions somewhat peripheral to central cornea. The cell densities shown are thus an overestimate of central cell density. Nevertheless, a significantly (P < 0.05) lower cell density was found in corneas having guttata compared with age-matched normals. In Figure 6, Na/K pump site density in normal human endothelia is compared with that measured in endothelia having guttata. Pump site density was found to be significantly (P < 0.05) increased to 6.2 ± 0.72 X 109 sites/mm2 in the donor corneas having guttata. Thus, in guttata, endothelial Na/K pump site density was found to be increased when compared with normals, even though endothelial cell density in these corneas was reduced. 761
OPHTHALMOLOGY
•
JUNE 1985 •
8.0 •
·p
Fig 6. Comparison of endothelial Na/K pump site densities in normal corneas to those having guttata. Histobars represent the mean (±SE) densities measured in 20 normal endothelial pairs and in 6 endothelial pairs having guttata. The difference in pump site densities between these groups was significant (P < 0.05).
(jI
Q 6.0 )(
N
E
~CII 40 • .2!
'iii
a. E
(f. 2.0
NORMAL GUTTATA (20)
(6)
DISCUSSION The population of endothelial cells in the human cornea declines constantly throughout life. 8- 10 Despite this constant loss of cells, normal thickness and transparency are maintained. It is only when cell density has declined to several hundred cells/mm2 that corneal decompensation occurs. II These observations indicate that the corneal endothelium possesses a very large reserve capacity in terms of the density of cells required for the maintenance of normal transparency. Results of the present study confirm and further define this functional reserve capacity. Na/K ATPase, an essential component of the endothelial fluid pump,12-15 has been localized to the lateral endothelial cell membrane. 16 Therefore, as endothelial cells are lost in the normal aging process, the pump sites associated with these cells are similarly lost. The data presented in this study demonstrate that the remaining endothelial cells adapt to this loss by generating new pump sites. The human corneal endothelium can, in this manner, maintain normal pump site density and, presumably, pump capacity. Fluorescein permeability studies l7 .18 have demonstrated that the permeability of the endothelium to fluorescein also remains constant with age, indicating that endothelial barrier function is normally preserved. Thus, in the normal aging human corneal endothelium, normal pump and barrier functions are maintained despite a declining population of cells. In comparison to the normal loss of cells with aging, the presence of cornea guttata places additional demands on the functional reserve of the corneal endothelium. In guttata, not only is cell density reduced to a greater extent than normally occurs in aging, but also the endothelial barrier is compromised. As the endothelial cells are stretched over the excrescences in Descemet's membrane, the junctional complexes between cells loosen and may eventually disappear. 19 The consequences of these ultrastructural changes are manifest as an increased endothelial leak rate. Measurements of endothelial waterO and fluorescein permeabilities3• 17 have documented an increased leak in cornea guttata patients. Results of
762
VOLUME 92
•
NUMBER 6
the present study describe the increased endothelial pump capacity which enables normal corneal thickness to be maintained in early Fuchs' dystrophy. The genesis of additional endothelial pump sites and thus increased Na/K pump site density in cornea guttata would appear to be an efficient and effective mechanism that enables endothelial pump capacity to adapt to long-term alterations in endothelial barrier function. Similar long-term adaptations in Na/K pump site densities have been described in other tissues. Chronic potassium loading, for example, has been shown to result in an increased pump site density in the renal cortex. 21 Similarly, salt-stressed marine birds are able to use sea water for the maintenance of positive water balance because of the adaptations in pump site density which occur in the avian salt gland. 7 Such studies indicate that long-term adaptation of transport processes is most efficiently effected by the addition (or removal) of pump sites to the cell membrane. Short-term regulation of pump function, on the other hand, is largely mediated by alterations in the rate at which existing pumps are operating. 22 Since pump capacity is a function of both the number of pumps and their activity (turnover), alterations in pump capacity can, and probably do, involve each of these factors. Previous studies have demonstrated that, in early Fuchs', endothelial permeabilities to wate~ and fluorescein l7 are increased to twice normal. Na/K pump site density, by comparison, increases by 50%. Since normal thickness is maintained in these patients, pump and leak must be balanced. This would suggest that, in addition to an increase in pump site density, an increased pump site turnover number is also associated with cornea guttata. In summary, the results of this study indicate that the human corneal endothelium possesses reserve, adaptive pump capacity which strives to preserve the balance between endothelial leak and pump rates. This reserve capacity preserves normal pump function despite a loss of endothelial cells, as occurs in the aging human cornea. Further, beyond the preservation of normal pump function, reserve pump capacity can increase the endothelial pump function to well above normal as exemplified by early Fuchs' dystrophy. The mechanism for these adaptations involves the generation of new pump sites and, most likely, an increase in turnover rate as well. It is likely that, when this reserve capacity is exceeded, as occurs in late Fuchs', corneal decompensation ensues.
REFERENCES 1. Sawa M, Araie M, Tanishima T. Permeability of the corneal endothelium to fluorescein-a follow-up of keratoplasty cases. Jpn J Ophthalrnol 1982; 26:326-37. 2. Ota Y. Endothelial perrneability to fluorescein in corneal grafts and bullous keratopathy. Jpn J Ophthalmol 1975; 19:286-95. 3. Burns RR, Bourne WM, Brubaker RF. Endothelial function in patients with cornea guttata. Invest Ophthalmol Vis Sci 1981; 20:77-85. 4. Geroski DH, Edelhauser HF. Quantitation of Na/K ATPase pump
GEROSKI, et al
5.
6. 7.
8.
9. 10. 11. 12. 13.
•
ENDOTHELIAL PUMP FUNCTION
sites in the rabbit corneal endotheliurn. Invest Ophthalrnol Vis Sci 1984; 25:1056-60. Tobin T, Baskin SI, Akera T, Brody TM. Nucleotide specificity of the Na+-stirnulated phosphorylation and eH) ouabain-binding reactions of (Na+ + K+)-dependent adenosine triphosphatase. Mol Pharrnacol 1972; 8:256-63. Bito LZ, Salvador EV. Intraocular fluid dynarnics. II. Postrnortern changes in solute concentrations. Exp Eye Res 1970; 10:273-87. Ernst SA, Mills JW. Basolateral plasrna rnernbrane localization of ouabain-sensitive sodiurn transport sites in the secretory epitheliurn of the avian salt gland. J Cell Bioi 1977; 75:74-94. Laing RA, Sandstrorn MM, Berrospi AR, Leibowitz HM. Changes in the corneal endotheliurn as a function of age. Exp Eye Res 1976; 22:587-94. Bourne WM, Kaufrnan HE. Specular rnicroscopy of hurnan corneal endothelium in vivo. Am J Ophthalmol 1976; 81:319-23. Blatt HL, Rao GN, Aquavella JV. Endothelial cell density in relation to morphology. Invest Ophthalmol Vis Sci 1979; 18:856-9. Mishima S. Clinical investigations on the comeal endothelium. Am J Ophthalmol 1982; 93:1-29. Brown SI, Hedbys BO. The effect of ouabain in the hydration of the cornea. Invest Ophthalmol 1965; 4:216-21. Langham M, Kostelnik M. The effect of ouabain on the hydration and the adenosine triphosphatase activity of the cornea. J Pharmacol Exp Ther 1965; 150:398-405.
14. Trenberth SM, Mishima S. The effect of ouabain on the rabbit comeal endothelium. Invest Ophthalmol 1968; 7:44-52. 15. Geroski DH, Kies JC, Edelhauser HF. The effects of ouabain on endothelial function in human and rabbit corneas. Curr Eye Res 1984; 3:331-8. 16. Kaye GI, Tice LW. Studies on the cornea. V. Electron microscopic localization of adenosine triphosphatase activity in the rabbit cornea in relation to transport. Invest Ophthalmol 1966; 5:22-32. 17. Waltman SR, Kaufman HE. In VfVO studies of human comeal endothelial permeability. Am J Ophthalmol 1970; 70:45-7. 18. Bourne WM, Nagataki S, Brubaker RF. The permeability of the corneal endothelium to fluorescein in the normal human eye. Curr Eye Res 1984; 3:509-13. 19. Iwamoto T, DeVoe AG. Electron microscopic studies on Fuchs' combined dystrophy. I. Posterior portion of the cornea. Invest Ophthalmol1971; 10:9-28. 20. Stanley JA. Water permeability of the human cornea. Arch Ophthalmol 1972; 87:568-73. 21. Rodriguez HJ, Hogan WC, Hellman RN, Klahr S. Mechanism of activation of renal Na+-K+:ATPase in the rat: effects of potassium loading. Am J Physiol 1980; 238:F315-23. 22. Trachtenberg MC, Packey OJ, Sweeney T. In vivo functioning of the Na+,K+-activated ATPase. Curr Topics Cell Regul 1981; 19: 159-217.
Discussion by
Joel Sugar, MD Over the years various means of assessing the cornea have developed. The specular microscope, I the pachymeter, 2 and the fluorophotometer3,4 have all added to our ability to evaluate corneal functional capabilities. Using these techniques, Bourne and co-workers5,6 have been able to show both increases and decreases in endothelial barrier function with decreasing cell density in some disorders. Nonetheless, clinical ophthalmologists, especially corneal surgeons, have continued to be mystified by the patient whose cornea is thin and clear despite exceedingly low endothelial cell density. In the present study the authors demonstrate a laboratory technique for estimating pump function through measurement of Na/K ATPase pump sites. They then go on to show that with advancing age, despite decreasing endothelial cell density, pump site density remains constant. Even more fascinating is their finding that in cornea guttata, in the face of even lower cell densities, pump site densities are higher than normal. This demonstrates a possible adaptive mechanism for the maintenance of corneal deturgescence. How this can occur remains uncertain. As cell density decreases and pump site density remains constant or increases, the number of pump sites per cell obviously must increase. Perhaps increased cell surface area or cell-to-cell contact area with surrounding cells induces generation of pump sites. This, however, would not explain the disproportionate increase in
From the University of Illinois Eye and Ear Infirmary, Chicago.
pump site density in cornea guttata, which remains unexplained.
It is interesting also to speculate whether the corneas with
guttata were potential future Fuchs' dystrophy corneas or not, and whether pump site density analysis can distinguish cornea guttata from asymptomatic Fuchs' dystrophy or show that these are just different points along the spectrum of the same disease. These intriguing studies raise many questions. What would happen to pump site density with trauma, especially in an animal like man with limited or absent endothelial regenerative capacity? What happens is other disease states? Can pump site density increases be induced pharmacologically? References 1. Maurice OM. Cellular membrane activity in the comeal endothelium of the intact eye. Experientia 1968; 24:1094-5. 2. Mishima S, Hedbys BO. Measurement of corneal thickness with the Haag-Streit pachometer. Arch Ophthalmol 1968; 80:710-3. 3. Jones RF, Maurice OM. New methods of measuring the rate of aqueous flow in man with fluorescein. Exp Eye Res 1966; 5:20820. 4. Waltman SR, Kaufman HE. In vivo studies of human corneal endothelial permeability. Am J Ophthalmol 1970; 70:45-7. 5. Bourne WM, Brubaker RF. Decreased endothelial permeability in the iridocorneal endothelial syndrome. Ophthalmology 1982; 89: 591-5. 6. Burns RR, Bourne WM, Brubaker RF. Endothelial function in patients with cornea guttata. Invest Ophthalmol Vis Sci 1981; 20:77-85.
763
Abstract: The specific binding of tritiated ouabain to endothelial Na/K ATPase was used to quantitate the density of pump sites in the human corneal endothelium. Donor eyes, unsuitable for use in keratoplasty, were obtained from the Wisconsin Lions Eye Bank. The endothelium of each donor eye was examined using wide-field specular microscopy, and the specular micrographs were traced and digitized for the determination of cell density. Ouabain binding was measured in matched pairs of isolated endothelial sheets. A total of 26 pairs of donor eyes, ranging in age from 11 through 91 years, were studied. Twenty pairs, determined to have normal endothelia, were found to have a constant pump site density which was independent of donor age. Six donor pairs had moderate guttata; in this group pump site density was significantly increased. These results indicate that, although pump site density is normally constant in the human corneal endothelium, conditions which increase endothelial permeability, such as guttata, can cause a compensatory increase in pump site density and presumably pump function. [Key words: cornea, cornea guttata, corneal endothelium, endothelial pump function, Fuchs' dystrophy, sodium-potassium adenosine triphosphatase.] Ophthalmology 92:759-763, 1985
A delicate balance of fluid movement is normally maintained across the posterior corneal surface. The corneal endothelium mediates this balance by acting as a barrier that limits the movement of fluid into the hydrophilic stroma, and by actively removing the fluid which manages to leak through its somewhat leaky barrier-the well-known barrier and pump functions of From the Departments of Ophthalmology and Physiology, Medical College of Wisconsin, Milwaukee, Wisconsin. Dr. Matsuda is a Visiting Scientist from the University of Osaka, Osaka, Japan. Presented at the Eighty·ninth Annual Meeting of the American Academy of Ophthalmology, Atlanta, Georgia, November 11-15,1984. Supported in part by NEI Research Grants EY·00933 and EY·05609 and Ophthalmic Research Core Grant EY-01931. Dr. Edelhauser is a 1982 Olga K. Wiess Research Scholar for Research to Prevent Blindness, Inc. Reprint requests to Dayle H. Geroski, PhD, Department of Physiology, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, Wisconsin 53226.
the endothelium. As long as the leak rate is exactly matched by the rate of the metabolic pump, stromal hydration is maintained at 78% water, and corneal transparency is maintained. Clinically, corneal edema will develop when the rate of fluid leak exceeds endothelial pump capacity. Though the barrier and pump functions of the corneal endothelium are the well established mediators of stromal hydration and corneal transparency, little is known about how these mechanisms are altered in disease. Until recently, our understanding of endothelial function as it relates to the pathogenesis of corneal edema has been largely inferential, based primarily on measurements of corneal thickness. Though pachymetry provides a valuable index of endothelial function, corneal thickness is a function of both endothelial barrier and pump functions; and increased corneal thickness can result from alterations in either or both of these functions. With the advent of fluorophotometry, it has become possible to clinically document alterations in endothelial 759
OPHTHALMOLOGY
•
JUNE 1985
•
VOLUME 92
•
NUMBER 6
Fig 1. Specular photomicrograph of normal donor cornea. Donor age 87 years, cell density 2237 cells/mm2. Each division on calibration grid represents 50 I'm.
Fig 2. Specular photomicrograph of donor cornea having moderate guttata. Donor age 66 years, cell density 1730 cells/mm2. Each division on calibration grid represents 50 I'm.
barrier function. Using this technique, alterations in endothelial permeability have been quantitated following penetrating keratoplasty,l in bullous keratopathy,2 and in Fuchs' dystrophy. 3 Though the functional status of the endothelial pump has been deduced from such measurements, alterations in endothelial pump capacity have yet to be documented. In these experiments, the specific binding of tritiated ouabain to endothelial Na/K ATPase (an essential component of the endothelial pump mechanism) was used to compare the density of Na/K pump sites in normal human donor endothelia to that in tissue having cornea guttata.
MATERIALS AND METHODS Matched pairs of intact donor eyes, unsuitable for use in keratoplasty, were obtained from the Wisconsin Lions Eye Bank. For the 26 pairs used in this study, the donor age range was 11 through 91 years, with the mean (±SE) donor age being 58.8 ± 4.7 years. The mean (±SE) time elapsed from death to the experimental use of the tissue was 21.0 ± 2.8 hours, within normal eyebank storage limits. The corneal endothelium of each donor eye was examined and photographed using a Keeler-Konan widefield specular microscope. Upon examination, 20 of the 26 donor pairs were found to have normal endothelia (Fig 1), while 6 pairs were determined to have early Fuchs' dystrophy with moderate guttata (Fig 2). For the determination of cell density, the specular photomicrographs were enlarged to a final magnification of X400.
760
Individual cells in clusters of 100 to 150 were traced and subjected to computer digitization. Endothelial Na/K ATPase "pump sites" were quantitated utilizing the specific binding of tritiated ouabain to these sites in sheets of isolated endothelium. These techniques, which have been previously described,4 are summarized here. After specular microscopy, the corneas are deepithelialized and excised. Each cornea of the matched pair is pre-incubated for one hour in potassium-free bicarbonate Ringer's (3rC). This pre-incubation was found to be necessary to achieve reproducible levels of ouabain binding in human tissue. This addition to the methods represents the only distinction from the procedure originally reported for rabbit tissue. Since phosphorylation of the Na/K ATPase is a prerequisite for glycoside binding, 5 it is likely that the pre-incubation provides a temperature-reversal period for endothelial metabolism, which is effectively shut off at the cold storage temperature (4°C). It is also known that the aqueous humor potassium concentration rises in the postmortem eye. 6 Since extracellular potassium has been shown to retard the rate of glycoside binding, 7 it is probable that the pre-incubation serves to remove the excess potassium from the endothelial extracellular space. Following the one-hour pre-incubation, each cornea was incubated for three hours (37°C) in potassium-free bicarbonate Ringer's containing 2 X 10-7 M tritiated ouabain. One cornea of the matched pair additionally received 10-4 M unlabeled ouabain to determine the nonsaturable component of endothelial ouabain uptake. At the end of the incubation, the corneas were rinsed, an 8 mm central button was trephined, and the endo-
GEROSKI.
4000
N
et aI
•
ENDOTHELIAL PUMP FUNCTION
8.0
•
•
Q 6.0
3000
N
•
~
Cii 2000 u
•• •
E
~ 4.0 (/)
.2!
•
•
~
E E ......
F •
•
•
n --
~~
•
•
'ji;
0.
• •
i
• • •
•
I: (4.4:!:O.21
••
E
0: 2.0 20
40
60
80
100
DONOR AGE (yrs) Fig 3. Endothelial cell density vs. donor age for donor corneas having normal endothelia. The correlation between donor age and cell density was significant (P < 0.05), and the calculated regression line is shown. Predicted cell density (cells/mm2) = 3384 - 10.7 (donor age).
thelium (plus Descemet's) of this button was removed as an intact sheet. Each sheet of endothelium was digested, and ouabain uptake was measured using liquid scintillation. Ouabain, bound by each pair of endothelial sheets, was calculated as the difference in tracer uptake measured in the absence and presence of 10-4 M unlabeled glycoside.
20
Central endothelial cell density as a function of donor age for the 20 pairs of donor corneas having normal endothelia is shown in Figure 3. In this group of donor corneas, a significant (P < 0.01) negative correlation was found between donor age and central cell density: predicted cell density = 3384 cells/mm2 - 10.7 (donor age). In contrast to this significant decrease in cell density with increasing donor age, the amount of ouabain bound by the corneal endothelium, and hence the density of Na/K ATPase pump sites, was found to remain constant and independent of donor age (Fig 4). Over a range of donor ages which extended from 11 through 91 years, Na/K pump site density was measured as 4.4 ± 0.2 X 109 pump sites/mm2 of endothelium (mean ± SE). It should be noted that this figure most likely represents an underestimate of pump site density. Because of the limited availability of human tissue, it is not possible to perform the detailed kinetic and equilibrium analyses of glycoside binding as were done for rabbit endothelium. 4 For this reason, human binding studies were carried out at a single glycoside concentration of 2 X 10-7 M. Previous studies have demonstrated that at this concentration non-specific binding is minimal and pump sites are 90% saturated in the rabbit endothelium.4 Thus, the actual density of Na/K pump sites in the human corneal endothelium is probably somewhat higher than the values reported here. Central endothelial cell density versus donor age for
60
80
100
DONOR AGE (yrs) Fig 4. Endothelial Na/K pump site density vs. donor age for donor corneas having normal endothelia. Pump site density was found to remain constant (4.4 ± 0.2 X 109 sites/mm 2) and independent of donor age. Mean ± SE density is represented by shaded region.
4000
N
RESULTS
40
3000
E E ...... .!!! 2000 Cii u
xx X
20
40
60
DONOR AGE (yrs)
x X
80
100
Fig 5. Endothelial cell density vs. donor age for donor corneas having guttata. Solid line represents the regression line of cell density vs. donor age for normal corneas.
the 6 donor pairs having guttata is shown in Figure 5. Since the presence of central guttata in these corneas interferes with the specular reflection from these regions, cell density could only be determined in guttata-free regions somewhat peripheral to central cornea. The cell densities shown are thus an overestimate of central cell density. Nevertheless, a significantly (P < 0.05) lower cell density was found in corneas having guttata compared with age-matched normals. In Figure 6, Na/K pump site density in normal human endothelia is compared with that measured in endothelia having guttata. Pump site density was found to be significantly (P < 0.05) increased to 6.2 ± 0.72 X 109 sites/mm2 in the donor corneas having guttata. Thus, in guttata, endothelial Na/K pump site density was found to be increased when compared with normals, even though endothelial cell density in these corneas was reduced. 761
OPHTHALMOLOGY
•
JUNE 1985 •
8.0 •
·p
Fig 6. Comparison of endothelial Na/K pump site densities in normal corneas to those having guttata. Histobars represent the mean (±SE) densities measured in 20 normal endothelial pairs and in 6 endothelial pairs having guttata. The difference in pump site densities between these groups was significant (P < 0.05).
(jI
Q 6.0 )(
N
E
~CII 40 • .2!
'iii
a. E
(f. 2.0
NORMAL GUTTATA (20)
(6)
DISCUSSION The population of endothelial cells in the human cornea declines constantly throughout life. 8- 10 Despite this constant loss of cells, normal thickness and transparency are maintained. It is only when cell density has declined to several hundred cells/mm2 that corneal decompensation occurs. II These observations indicate that the corneal endothelium possesses a very large reserve capacity in terms of the density of cells required for the maintenance of normal transparency. Results of the present study confirm and further define this functional reserve capacity. Na/K ATPase, an essential component of the endothelial fluid pump,12-15 has been localized to the lateral endothelial cell membrane. 16 Therefore, as endothelial cells are lost in the normal aging process, the pump sites associated with these cells are similarly lost. The data presented in this study demonstrate that the remaining endothelial cells adapt to this loss by generating new pump sites. The human corneal endothelium can, in this manner, maintain normal pump site density and, presumably, pump capacity. Fluorescein permeability studies l7 .18 have demonstrated that the permeability of the endothelium to fluorescein also remains constant with age, indicating that endothelial barrier function is normally preserved. Thus, in the normal aging human corneal endothelium, normal pump and barrier functions are maintained despite a declining population of cells. In comparison to the normal loss of cells with aging, the presence of cornea guttata places additional demands on the functional reserve of the corneal endothelium. In guttata, not only is cell density reduced to a greater extent than normally occurs in aging, but also the endothelial barrier is compromised. As the endothelial cells are stretched over the excrescences in Descemet's membrane, the junctional complexes between cells loosen and may eventually disappear. 19 The consequences of these ultrastructural changes are manifest as an increased endothelial leak rate. Measurements of endothelial waterO and fluorescein permeabilities3• 17 have documented an increased leak in cornea guttata patients. Results of
762
VOLUME 92
•
NUMBER 6
the present study describe the increased endothelial pump capacity which enables normal corneal thickness to be maintained in early Fuchs' dystrophy. The genesis of additional endothelial pump sites and thus increased Na/K pump site density in cornea guttata would appear to be an efficient and effective mechanism that enables endothelial pump capacity to adapt to long-term alterations in endothelial barrier function. Similar long-term adaptations in Na/K pump site densities have been described in other tissues. Chronic potassium loading, for example, has been shown to result in an increased pump site density in the renal cortex. 21 Similarly, salt-stressed marine birds are able to use sea water for the maintenance of positive water balance because of the adaptations in pump site density which occur in the avian salt gland. 7 Such studies indicate that long-term adaptation of transport processes is most efficiently effected by the addition (or removal) of pump sites to the cell membrane. Short-term regulation of pump function, on the other hand, is largely mediated by alterations in the rate at which existing pumps are operating. 22 Since pump capacity is a function of both the number of pumps and their activity (turnover), alterations in pump capacity can, and probably do, involve each of these factors. Previous studies have demonstrated that, in early Fuchs', endothelial permeabilities to wate~ and fluorescein l7 are increased to twice normal. Na/K pump site density, by comparison, increases by 50%. Since normal thickness is maintained in these patients, pump and leak must be balanced. This would suggest that, in addition to an increase in pump site density, an increased pump site turnover number is also associated with cornea guttata. In summary, the results of this study indicate that the human corneal endothelium possesses reserve, adaptive pump capacity which strives to preserve the balance between endothelial leak and pump rates. This reserve capacity preserves normal pump function despite a loss of endothelial cells, as occurs in the aging human cornea. Further, beyond the preservation of normal pump function, reserve pump capacity can increase the endothelial pump function to well above normal as exemplified by early Fuchs' dystrophy. The mechanism for these adaptations involves the generation of new pump sites and, most likely, an increase in turnover rate as well. It is likely that, when this reserve capacity is exceeded, as occurs in late Fuchs', corneal decompensation ensues.
REFERENCES 1. Sawa M, Araie M, Tanishima T. Permeability of the corneal endothelium to fluorescein-a follow-up of keratoplasty cases. Jpn J Ophthalrnol 1982; 26:326-37. 2. Ota Y. Endothelial perrneability to fluorescein in corneal grafts and bullous keratopathy. Jpn J Ophthalmol 1975; 19:286-95. 3. Burns RR, Bourne WM, Brubaker RF. Endothelial function in patients with cornea guttata. Invest Ophthalmol Vis Sci 1981; 20:77-85. 4. Geroski DH, Edelhauser HF. Quantitation of Na/K ATPase pump
GEROSKI, et al
5.
6. 7.
8.
9. 10. 11. 12. 13.
•
ENDOTHELIAL PUMP FUNCTION
sites in the rabbit corneal endotheliurn. Invest Ophthalrnol Vis Sci 1984; 25:1056-60. Tobin T, Baskin SI, Akera T, Brody TM. Nucleotide specificity of the Na+-stirnulated phosphorylation and eH) ouabain-binding reactions of (Na+ + K+)-dependent adenosine triphosphatase. Mol Pharrnacol 1972; 8:256-63. Bito LZ, Salvador EV. Intraocular fluid dynarnics. II. Postrnortern changes in solute concentrations. Exp Eye Res 1970; 10:273-87. Ernst SA, Mills JW. Basolateral plasrna rnernbrane localization of ouabain-sensitive sodiurn transport sites in the secretory epitheliurn of the avian salt gland. J Cell Bioi 1977; 75:74-94. Laing RA, Sandstrorn MM, Berrospi AR, Leibowitz HM. Changes in the corneal endotheliurn as a function of age. Exp Eye Res 1976; 22:587-94. Bourne WM, Kaufrnan HE. Specular rnicroscopy of hurnan corneal endothelium in vivo. Am J Ophthalmol 1976; 81:319-23. Blatt HL, Rao GN, Aquavella JV. Endothelial cell density in relation to morphology. Invest Ophthalmol Vis Sci 1979; 18:856-9. Mishima S. Clinical investigations on the comeal endothelium. Am J Ophthalmol 1982; 93:1-29. Brown SI, Hedbys BO. The effect of ouabain in the hydration of the cornea. Invest Ophthalmol 1965; 4:216-21. Langham M, Kostelnik M. The effect of ouabain on the hydration and the adenosine triphosphatase activity of the cornea. J Pharmacol Exp Ther 1965; 150:398-405.
14. Trenberth SM, Mishima S. The effect of ouabain on the rabbit comeal endothelium. Invest Ophthalmol 1968; 7:44-52. 15. Geroski DH, Kies JC, Edelhauser HF. The effects of ouabain on endothelial function in human and rabbit corneas. Curr Eye Res 1984; 3:331-8. 16. Kaye GI, Tice LW. Studies on the cornea. V. Electron microscopic localization of adenosine triphosphatase activity in the rabbit cornea in relation to transport. Invest Ophthalmol 1966; 5:22-32. 17. Waltman SR, Kaufman HE. In VfVO studies of human comeal endothelial permeability. Am J Ophthalmol 1970; 70:45-7. 18. Bourne WM, Nagataki S, Brubaker RF. The permeability of the corneal endothelium to fluorescein in the normal human eye. Curr Eye Res 1984; 3:509-13. 19. Iwamoto T, DeVoe AG. Electron microscopic studies on Fuchs' combined dystrophy. I. Posterior portion of the cornea. Invest Ophthalmol1971; 10:9-28. 20. Stanley JA. Water permeability of the human cornea. Arch Ophthalmol 1972; 87:568-73. 21. Rodriguez HJ, Hogan WC, Hellman RN, Klahr S. Mechanism of activation of renal Na+-K+:ATPase in the rat: effects of potassium loading. Am J Physiol 1980; 238:F315-23. 22. Trachtenberg MC, Packey OJ, Sweeney T. In vivo functioning of the Na+,K+-activated ATPase. Curr Topics Cell Regul 1981; 19: 159-217.
Discussion by
Joel Sugar, MD Over the years various means of assessing the cornea have developed. The specular microscope, I the pachymeter, 2 and the fluorophotometer3,4 have all added to our ability to evaluate corneal functional capabilities. Using these techniques, Bourne and co-workers5,6 have been able to show both increases and decreases in endothelial barrier function with decreasing cell density in some disorders. Nonetheless, clinical ophthalmologists, especially corneal surgeons, have continued to be mystified by the patient whose cornea is thin and clear despite exceedingly low endothelial cell density. In the present study the authors demonstrate a laboratory technique for estimating pump function through measurement of Na/K ATPase pump sites. They then go on to show that with advancing age, despite decreasing endothelial cell density, pump site density remains constant. Even more fascinating is their finding that in cornea guttata, in the face of even lower cell densities, pump site densities are higher than normal. This demonstrates a possible adaptive mechanism for the maintenance of corneal deturgescence. How this can occur remains uncertain. As cell density decreases and pump site density remains constant or increases, the number of pump sites per cell obviously must increase. Perhaps increased cell surface area or cell-to-cell contact area with surrounding cells induces generation of pump sites. This, however, would not explain the disproportionate increase in
From the University of Illinois Eye and Ear Infirmary, Chicago.
pump site density in cornea guttata, which remains unexplained.
It is interesting also to speculate whether the corneas with
guttata were potential future Fuchs' dystrophy corneas or not, and whether pump site density analysis can distinguish cornea guttata from asymptomatic Fuchs' dystrophy or show that these are just different points along the spectrum of the same disease. These intriguing studies raise many questions. What would happen to pump site density with trauma, especially in an animal like man with limited or absent endothelial regenerative capacity? What happens is other disease states? Can pump site density increases be induced pharmacologically? References 1. Maurice OM. Cellular membrane activity in the comeal endothelium of the intact eye. Experientia 1968; 24:1094-5. 2. Mishima S, Hedbys BO. Measurement of corneal thickness with the Haag-Streit pachometer. Arch Ophthalmol 1968; 80:710-3. 3. Jones RF, Maurice OM. New methods of measuring the rate of aqueous flow in man with fluorescein. Exp Eye Res 1966; 5:20820. 4. Waltman SR, Kaufman HE. In vivo studies of human corneal endothelial permeability. Am J Ophthalmol 1970; 70:45-7. 5. Bourne WM, Brubaker RF. Decreased endothelial permeability in the iridocorneal endothelial syndrome. Ophthalmology 1982; 89: 591-5. 6. Burns RR, Bourne WM, Brubaker RF. Endothelial function in patients with cornea guttata. Invest Ophthalmol Vis Sci 1981; 20:77-85.
763