(Circulation. 1997;96:1624-1630.)
© 1997 American Heart Association, Inc.
Articles |
From the University of Alabama at Birmingham, Departments of Medicine, Vascular Biology, and Hypertension Program (C.R.W., J.S., S.-J.C., L.A., P.W.S., Y.-F.C., S.O.), Pathology (V.D.-U.), Physiology (S.O.), and Nephrology (C.N., P.W.S.); the Department of Veterans Affairs Medical Center (P.W.S.); and the Center for Free Radical Biology (C.R.W., V.D.-U.), Birmingham, Ala.
Correspondence to C. Roger White, PhD, Departments of Medicine, Vascular Biology, and Hypertension Program, 1046 Zeigler Research Bldg, 703 S 19th St, Birmingham, AL 35294-0007. E-mail card029{at}uabdpo.dpo.uab.edu
| Abstract |
|---|
|
|
|---|
Methods and Results Ten-week-old male and female Sprague-Dawley rats with intact gonads underwent balloon injury to the right common carotid artery. Female rats were randomized to receive either daily subcutaneous injections of 17ß-estradiol (17ß-E2; 20 µg · kg-1 · d-1) or vehicle over the course of the study. Vessel morphology was assessed 2 weeks after injury. Significant neointima formation was observed in vehicle-treated males. This response was blunted in vehicle-treated and 17ß-E2supplemented females. Intima-to-media ratios were 1.28±0.23 (males), 0.72±0.07 (vehicle-treated females), and 0.49±0.07 (17ß-E2supplemented females). To test whether reductions in neointimal lesion formation were related to the functional reendothelialization of the damaged vessel, endothelium-dependent relaxation was tested in isolated ring segments from the three experimental groups. Vessels were precontracted with phenylephrine followed by cumulative administration of acetylcholine, an endothelium-dependent vasodilator. Maximum relaxation to acetylcholine was 8.13±1.70% (males), 22.06±4.36% (vehicle-treated females), and 46.47±3.48% (17ß-E2supplemented females). The enhanced endothelium-dependent relaxation of rings from 17ß-E2supplemented females correlated with reduced neointimal proliferation in this group. The concentration of nitric oxide metabolites in plasma correlated positively with plasma 17ß-E2 concentration.
Conclusions These results suggest that estrogen protects against neointimal injury in the balloon-injured rat, at least in part, by facilitating the reendothelialization of the damaged vessel.
Key Words: hormones balloon vasculature
| Introduction |
|---|
|
|
|---|
Transluminal balloon injury is a commonly used paradigm for the study of mechanisms of response to vascular damage and atherosclerosis.6 7 8 Balloon injury models are characterized by the formation of a concentric fibromuscular lesion that encroaches on the arterial lumen.6 This injury response is characterized by the proliferation of smooth muscle cells in the intima. These cells also adopt a secretory phenotype, resulting in excessive production of extracellular matrix. In addition, platelets and macrophages adhere to the vessel wall and release cytokines and growth factors. These stimulate further smooth muscle cell proliferation and act as chemoattractants for other cell types, which in turn become incorporated into the neointimal lesion.6 9
Recent reports indicate that estrogen protects against neointimal hyperplasia resulting from balloon injury in the rat carotid artery10 and rabbit iliac artery.11 In the balloon-injured, nongonadectomized male rabbit, estrogen treatment resulted in reduced neointimal thickening, [3H]thymidine incorporation, and DNA content of injured vessels compared with rabbits not receiving estrogen. These results suggest that estrogen inhibited cell proliferation in this model of vascular injury. The protective effects of estrogen were similar to those exerted by angiopeptin, a somatostatin analogue, which has previously been shown to inhibit neointimal hyperplasia.12 Estrogen may also protect against the development of atherosclerosis. In animal models of hypercholesterolemia, estrogen replacement therapy prevents the development of atheromatous lesions and defects in endothelium-dependent relaxation via mechanisms unrelated to changes in plasma lipoprotein concentration.13 14
Pharmacological interventions can prevent or inhibit the neointimal response after vascular injury. Angiopeptin, a somatostatin analogue, administered by a Wolinsky porous balloon inhibits myointimal proliferation in the balloon-injured rabbit aorta.12 Interestingly, hyperplasia was not inhibited at the site of local delivery but rather at regions of the aorta downstream from the site of application. The authors suggested that the protective downstream effect of angiopeptin may be due to healing of the damaged endothelium.12 Endothelial cell seeding techniques have also been used to test whether the reendothelialization of the damaged vessel limits the neointimal response. In a model of bilateral iliac artery balloon injury, vessels seeded with venous endothelial cells demonstrated a higher degree of reendothelialization than nonseeded control vessels.15 However, both treated and untreated vessels demonstrated similar extents of neointimal hyperplasia.15
The goal of the present study was to determine whether estrogen protects against the vascular injury response by stimulating the regrowth and functional recovery of the endothelium. Our results indicate that estrogen inhibits neointima formation and partially restores the endothelium-dependent vasodilator response of the balloon-injured rat carotid artery. It is hypothesized that these protective effects of estrogen are mediated by reendothelialization of the damaged region of the artery and that enhanced release of NO plays an important role in this process.
| Methods |
|---|
|
|
|---|
Balloon Injury Procedure
Male and female rats underwent balloon injury of the right
common carotid artery as described previously.10 Briefly,
rats were anesthetized with sodium pentobarbital (50
mg/kg), and the right carotid artery was isolated by a middle
cervical incision. The distal right common carotid artery and the
region of the bifurcation were exposed. A Fogarty 2F balloon catheter
(Baxter V. Mueller) was inserted into the external carotid artery and
advanced into the thoracic aorta. The balloon was inflated with saline
to distend the common carotid artery and was pulled back to the
external carotid artery. The catheter was advanced and then withdrawn
for five additional passes to denude the vessel wall of
endothelium. The catheter was then removed, and the
external carotid artery was tied off before closure of the wound. The
left carotid artery was not damaged and served as a control.
Estrogen Supplementation
Female rats with intact gonads were randomly assigned to one of
two treatment groups. The first group (n=6) received daily injections
of 17ß-E2 (20 µg/kg in 100 µL cottonseed oil
SC daily), and the second group (n=9) received vehicle (100 µL
cottonseed oil SC daily). 17ß-E2 was obtained from Sigma
Chemical Co. Male rats (n=7) also received daily vehicle treatment.
Vessel Reactivity Studies
Two weeks after balloon injury of the right carotid artery,
heparinized rats were euthanized with an overdose of sodium
pentobarbital (75 mg/kg IP).
Endothelium-dependent relaxation of isolated carotid
arteries was assessed 2 weeks after injury. Right and left carotid
arteries were excised and placed in prewarmed and
oxygenated Krebs-Henseleit solution and cleansed of fat and
adhering tissue. The undamaged left carotid artery served as a control
in functional studies of the balloon-injured vessel. Individual ring
segments (2 to 3 mm in width) were cut from injured and control
vessels. Each vascular ring was mounted on two stainless steel hooks
and suspended from a force-displacement transducer in a water-jacketed
tissue bath. The ring was anchored to a hook at the base of the
chamber, and a passive load of 1.5 g was applied to the vessel.
Ring segments were bathed in a Krebs-Henseleit solution of the
following composition (mmol/L): NaCl 118, KCl 4.6,
NaHCO3 27.2, KH2PO4 1.2,
MgSO4 1.2, CaCl2 1.75, Na2EDTA
0.03, and glucose 11.1 (pH 7.4). The solution was continuously aerated
with a 95%O2/5% CO2 gas mixture and
maintained at 37°C. Changes in isometric tension were measured with
capacitive force transducers (Radnoti Glass Technology, Inc). Real-time
data were acquired and digitized with an IBM-compatible computer and
stored for later analysis with commercially available software
(Experimenter's Workbench). Dose-response profiles for different
experimental conditions were analyzed and tested to determine
differences in contraction and relaxation responses.
Isometric tension was measured in isolated ring segments of control and balloon-injured carotid arteries. After an equilibration period of 45 minutes, ring segments were depolarized with KCl (70 mmol/L) to determine the contractile capacity of the vessel. Rings were then thoroughly washed and allowed to equilibrate for an additional 45 minutes. All subsequent experiments were performed in the presence of indomethacin (5 µmol/L) to eliminate the effects of cyclooxygenase-derived vasoactive molecules. In initial studies, contractile responses of balloon-injured vessels were tested by cumulative administration of PE (10-9 to 3x10-6 mol/L) or 5-HT (10-8 to 10-4 mol/L). NO-dependent relaxation was tested in PE-contracted ring segments by exposure to ACh (10-9 to 3x10-6 mol/L), an endothelium-dependent vasodilator. In other studies, PE-contracted ring segments were exposed to SNP (10-9 to 3x10-5 mol/L) to elicit endothelium-independent relaxation.
Vessel Morphometry
After completion of the in vitro blood vessel bioassay, each
ring was placed in 10% buffered formalin solution. Vessels were
embedded in paraffin and serially sectioned (5 µm) for
morphometric analysis. Tissue was stained with
hematoxylin-eosin. Neointimal proliferation was assessed in
thin sections of balloon-injured and control carotid arteries 2 weeks
after injury. The undamaged left carotid artery served as a control in
these studies. Morphometric analysis of each
arterial ring segment was performed with a computer-based
Bioquant II Morphometric system. At least three sections of each vessel
were examined, and measurements were averaged for statistical
analysis. The cross-sectional areas of the media and
neointima were measured. The degree of
neointima formation of the injured carotid artery was
expressed as the absolute area of neointima and the ratio
of the neointimal area to the medial area. All morphometric
measurements were performed by the same individual, who was blinded to
the treatment groups.
Plasma Estradiol Assay
Blood samples were collected when the animals were killed to
monitor the effect of 17ß-E2 treatment on circulating
levels of the hormone. Plasma 17ß-E2 levels were
determined by radioimmunoassay with a commercially available kit
(Diagnostic Products Corp). Assay sensitivity was 8.0
pg/mL, and intra-assay and interassay coefficients of variation
for estradiol were 5.3% and 6.4%, respectively.
Plasma NO3- and
NO2- Measurements
To determine whether the protective effects of estrogen were
related to a stimulatory effect of the hormone on NO formation, the
plasma concentrations of NO3- and
NO2-, the primary metabolites of NO, were
determined as described previously.16 NO has a relatively
short half-life in plasma, but NO2- and
NO3- are stable products. The Griess
reagent (1% sulfanilamide/0.1%
N-[1-naphthyl]ethylenediamine dihydrochloride) was
used in the measurement of NO3- and
NO2-. Because this reagent reacts only with
nitrite, plasma NO3- was first reduced to
NO2- by incubation of plasma samples with
Escherichia coli rich in nitrate reductase. After a 2-hour
incubation period with nitrate reductase, the Griess reagent was added
to plasma samples and incubated at room temperature for 10 minutes.
After this incubation period, the absorbance of samples was read at 540
nm. Plasma NO2- concentrations were then
estimated by comparing absorbance values with those obtained from a
standard curve for NaNO2 (0 to 200
µmol/L).
Statistical Analysis
All results are expressed as mean±SEM. Two to three
observations for a given treatment were obtained from each animal.
These values were averaged so that a single mean value is reported for
each animal. Dose-response profiles for the different experimental
treatments were analyzed and tested to determine differences in
contraction and relaxation responses by use of the StatView statistical
analysis program. Unpaired observations were assessed by
one-way ANOVA and multiple-range tests. Pearson's correlation
analysis was used to compare treatment responses in the
experimental groups.
| Results |
|---|
|
|
|---|
|
In vitro studies were designed to determine whether estrogen treatment
influenced vascular reactivity in balloon-injured and control carotid
artery ring segments of males, vehicle-treated females, and
17ß-E2supplemented females. Results of cumulative PE
dose-response experiments showed that the force-generating capacity of
balloon-injured ring segments was similar in all groups despite
differences in vessel morphology (Fig 1
).
Furthermore, the EC50 for maximum contraction of
balloon-injured ring segments was similar in all treatment groups (Fig 2
). Similar responses were observed when
ring segments were exposed to cumulative concentrations of the
vasoconstrictor 5-HT (Fig 3
). Although
tension development of balloon-injured vessels was less than that of
undamaged carotid ring segments for each treatment group (data not
shown), control and injured vessels displayed similar sensitivities to
PE (Fig 2
). These results indicate that vasoconstrictor responses of
balloon-injured carotid ring segments are similar in the three
treatment groups.
|
|
|
Additional experiments were designed to determine whether
estrogen treatment resulted in the functional
reendothelialization of the damaged vessel.
Endothelium-dependent relaxation was assessed in
isolated ring segments from the three experimental groups. Vessels were
first contracted with PE followed by administration of ACh, an
endothelium-dependent vasodilator. Relaxation responses
were calculated as percentage change in tension from the predose
response level. Dose-response profiles to ACh were similar in undamaged
left carotid arteries in all treatment groups (Fig 4A
). The ACh-induced relaxation of
balloon-injured carotid arteries was diminished in all treatment
groups, with ring segments of male rats displaying the most severe
impairment (Fig 4B
). Maximum relaxation in response to ACh was
8.13±1.70% (males), 22.06±4.36% (vehicle-treated females), and
46.47±3.48% (E2-supplemented females). Although
Emax was significantly increased in both groups of female
rats compared with males, the recovery of ACh-induced relaxation was
greatest in E2-supplemented females (Fig 4B
). The
Emax value correlated positively with plasma
17ß-E2 (r=.50, P<.05) (Fig 5
). The enhanced
endothelium-dependent relaxation of rings from
17ß-E2supplemented females also correlated with reduced
neointima formation in this group (r=-.59,
P<.05). In related experiments, PE-contracted ring segments
taken from balloon-injured arteries were exposed to the
endothelium-independent vasodilator SNP. Dose-response
profiles to SNP were similar in all groups (Fig 6
), suggesting that the differential
relaxation response of vessels from the three treatment groups was not
due to an altered responsiveness of the VSMC. Rather, these data
suggest that the enhanced relaxation of carotid artery rings from
17ß-E2supplemented rats is due to increased
production of endothelium-derived NO in these
vessels relative to those isolated from vehicle-treated male and female
rats.
|
|
|
Because reports suggest that increased production of NO
protects against vascular injury17 18 and that estrogen
stimulates constitutive NOS activity,19 20 21 studies were
designed to test whether NO production was increased in
17ß-E2supplemented rats. We used the Griess assay to
measure plasma concentrations of NO2- and
NO3-, the primary metabolites of NO. Increases
in plasma NO2- and
NO3- were observed in vehicle-treated and
17ß-E2supplemented rats compared with males (Fig 7
). The enhanced
NO2- and NO3- levels
correlated positively with increased plasma 17ß-E2 levels
(r=.46, P<.05) and maximum vessel relaxation
responses (r=.84, P<.01) in these animals and
inversely with neointimal area (r=-.49,
P<.05) (Fig 8
) and
intima-to-media ratios (r=-.47, P<.05).
|
|
| Discussion |
|---|
|
|
|---|
Recently, treatment of balloon-injured rats with 17ß-E2 was shown to inhibit neointimal lesion formation in gonadectomized rats of both sexes.10 Similar results have been described in a rabbit model of balloon injury in which estrogen supplementation reduced [3H]thymidine incorporation in injured vessels and inhibited neointimal cell proliferation.11 Estrogen treatment inhibits neointima formation in injured carotid arteries of gonadectomized rats.10 23 The inhibition of cell proliferation correlated with a decrease in the expression of the proto-oncogene c-myc, whose induction has been linked to the hyperplastic response of blood vessels.10
Increased production of NO attenuates vascular damage in balloon injury models.17 18 In the rat, intravenous infusion of the NO donor CAS-1609 inhibits neointimal hyperplasia and restores the vasodilator response to ACh. CAS-1609 has a stimulatory effect on endothelial cell proliferation in culture while inhibiting platelet-derived growth factorstimulated VSMC growth. These results suggest that NO minimizes vessel damage and promotes the functional recovery of endothelial cells in this model of carotid vessel injury.17
Estrogen stimulates the synthesis of NO in numerous tissues, including the uterine artery, heart, uterus, and skeletal muscle.19 Both pregnancy and estrogen supplementation enhance NOS-I and NOS-III expression, whereas NOS-II is unaffected.20 Furthermore, the 5'-flanking region of the gene for NOS-III contains an estrogen response element.24 Estrogen-mediated stimulation of NOS activity occurs by a receptor-dependent mechanism, because NOS-III expression is inhibited by the estrogen receptor antagonists tamoxifen and ICI182780.25 In the present study, plasma concentrations of NO metabolites correlated positively with plasma 17ß-E2 levels and negatively with the extent of neointima formation. Thus, increases in NO production may be a fundamental mechanism by which estrogen blunts vascular injury.
Indirect evidence suggests that the basal release of NO is elevated in female rabbits compared with males.26 27 Endothelium-dependent relaxation of aortic rings was impaired in ovariectomized rabbits compared with intact females, and dose-response profiles were similar in magnitude to the responses seen in male rabbits. Furthermore, the impaired response of ring segments from ovariectomized females correlated with the reduction in plasma E2 concentration.26 In another study, progesterone was shown to antagonize the facilitative effects of estrogen on endothelium-dependent relaxation of coronary arteries of ovariectomized dogs.28 It is clear, therefore, that steroid hormones differentially regulate vascular function.
Results of animal and human studies suggest that acute exposure to 17ß-E2 induces rapid changes in vessel tone and that treatment with NOS inhibitors abolishes estrogen-mediated vascular relaxation.29 30 Plasma concentrations of NO2- and NO3-, metabolites of NO, are elevated in postmenopausal women receiving hormone replacement therapy.31 In postmenopausal women, short-term estrogen replacement therapy results in increased coronary blood flow30 32 and peripheral vasodilator responses.33 34 These data suggest that protective effects of estrogen on the vasculature may be related to the enhanced production of NO. Results of the present study show that plasma 17ß-E2 levels correlate positively with plasma concentrations of NO metabolites. The ability of estrogen to attenuate neointima formation may, therefore, be related to NO-mediated inhibition of VSMC proliferation and platelet adhesion.
Several reports suggest that the regrowth of the endothelium may attenuate vascular damage in balloon-injury models.17 18 Endothelial cell growth factors may play an important role in this protective response. Intravenous treatment of balloon-injured rabbits with bFGF, an endothelial cell mitogen, resulted in a significant reendothelialization of damaged iliac arteries compared with controls not receiving bFGF. The extent of neointimal thickening was not different, however, between the two groups. Functional responses of bFGF-exposed and control vessels were tested by in vitro bioassay of endothelium-dependent relaxation. Iliac arteries of rabbits receiving bFGF treatment demonstrated enhanced ACh-mediated relaxation compared with controls. Those authors suggested that angiogenic growth factors such as bFGF may facilitate the recovery of endothelial cell function in injured vessels.35
Other mitogens, including VEGF, promote endothelial cell growth and proliferation36 37 and angiogenesis,38 39 increase endothelial barrier permeability,40 41 and modulate blood vessel tone.42 Local delivery of VEGF to balloon-injured rat carotid arteries results in diminished neointimal hyperplasia and enhanced reendothelialization of damaged vessels at periods up to 4 weeks after injury.22 VEGF treatment was associated with reductions in proliferating cell nuclear antigen immunostaining and neointimal thickening. The authors suggest that the protective effect of VEGF in balloon-injured vessels was mediated by the reendothelialization of the vessel wall.22 VEGF is secreted by a variety of cell types in the vessel wall, including VSMCs and macrophages,43 and 17ß-E2 has been shown to regulate VEGF mRNA expression in uterus and endometrial carcinoma cells.44 45 The extent of mitogen-induced endothelial cell growth and angiogenesis in vivo can be regulated by other factors. For instance, progesterone stimulates the production of thrombospondin-1, an extracellular matrix glycoprotein, which suppresses angiogenesis in vivo in human endometrium.46 Thus, the stimulatory effects of estrogen on endothelial cell growth may be regulated by other cellular mediators.
Results of the present studies clearly show that estrogen supplementation inhibits the vascular injury response and promotes reendothelialization, as demonstrated by enhanced endothelium-dependent relaxation in carotid arteries of normal female rats subjected to balloon injury. The subsequent localized increase in NO production in damaged carotid arteries of estrogen-supplemented rats may play an important role in limiting neointima formation, because NO is known to inhibit smooth muscle cell proliferation and vascular adhesion processes.47 Further studies are required to define the cellular and molecular mechanisms underlying the estrogen-mediated reendothelialization of the balloon-injured vessel.
| Selected Abbreviations and Acronyms |
|---|
|
| Acknowledgments |
|---|
Received January 22, 1997; revision received March 5, 1997; accepted March 7, 1997.
| References |
|---|
|
|
|---|
-dihydroequilin sulfate, a conjugated equine
estrogen, and ethynylestradiol on atherosclerosis in
cholesterol-fed rabbits. Arterioscler Thromb
Vasc Biol. 1995;15:837-846.This article has been cited by other articles:
![]() |
V. M. Miller and S. P. Duckles Vascular Actions of Estrogens: Functional Implications Pharmacol. Rev., June 1, 2008; 60(2): 210 - 241. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. M. Cho, N. Peng, J. T. Clark, L. Novak, S. Roysommuti, J. Prasain, and J. M. Wyss Genistein Attenuates the Hypertensive Effects of Dietary NaCl in Hypertensive Male Rats Endocrinology, November 1, 2007; 148(11): 5396 - 5402. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Feletou and P. M. Vanhoutte Endothelial dysfunction: a multifaceted disorder (The Wiggers Award Lecture) Am J Physiol Heart Circ Physiol, September 1, 2006; 291(3): H985 - H1002. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. G. Mishra, F. Z. Stanczyk, K. A. Burry, S. Oparil, B. S. Katzenellenbogen, M. L. Nealen, J. A. Katzenellenbogen, and R. K. Hermsmeyer Metabolite ligands of estrogen receptor-{beta} reduce primate coronary hyperreactivity Am J Physiol Heart Circ Physiol, January 1, 2006; 290(1): H295 - H303. [Abstract] [Full Text] [PDF] |
||||
![]() |
V. L. Ballard and J. M. Edelberg Harnessing Hormonal Signaling for Cardioprotection Sci. Aging Knowl. Environ., December 21, 2005; 2005(51): re6 - re6. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. R. I. Williams, T. Dawood, S. Ling, A. Dai, R. Lew, K. Myles, J. W. Funder, K. Sudhir, and P. A. Komesaroff Dehydroepiandrosterone Increases Endothelial Cell Proliferation in Vitro and Improves Endothelial Function in Vivo by Mechanisms Independent of Androgen and Estrogen Receptors J. Clin. Endocrinol. Metab., September 1, 2004; 89(9): 4708 - 4715. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. Kipshidze, G. Dangas, M. Tsapenko, J. Moses, M. B. Leon, M. Kutryk, and P. Serruys Role of the endothelium in modulating neointimal formation: Vasculoprotective approaches to attenuate restenosis after percutaneous coronary interventions J. Am. Coll. Cardiol., August 18, 2004; 44(4): 733 - 739. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. A. Nussmeier, M. R. Marino, and W. K. Vaughn Hormone replacement therapy is associated with improved survival in women undergoing coronary artery bypass grafting J. Thorac. Cardiovasc. Surg., December 1, 2002; 124(6): 1225 - 1229. [Abstract] [Full Text] |
||||
![]() |
K. J. Ho and J. K. Liao Nonnuclear Actions of Estrogen Arterioscler. Thromb. Vasc. Biol., December 1, 2002; 22(12): 1952 - 1961. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. Polvani, M. R. Marino, M. Roberto, L. Dainese, A. Parolari, G. Pompilio, S. D. Matteo, A. Fumero, A. Cannata, F. Barili, et al. Acute effects of 17{beta}-estradiol on left internal mammary graft after coronary artery bypass grafting Ann. Thorac. Surg., September 1, 2002; 74(3): 695 - 699. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. J. Ho and J. K. Liao Non-nuclear Actions of Estrogen: New Targets for Prevention and Treatment of Cardiovascular Disease Mol. Interv., July 1, 2002; 2(4): 219 - 228. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. G. Espinosa-Heidmann, I. Suner, E. P. Hernandez, W. D. Frazier, K. G. Csaky, and S. W. Cousins Age as an Independent Risk Factor for Severity of Experimental Choroidal Neovascularization Invest. Ophthalmol. Vis. Sci., May 1, 2002; 43(5): 1567 - 1573. [Abstract] |