Effect of Latanoprost on Regional Blood Flow and Capillary Permeability in the Monkey Eye FREE

LATANOPROST (13,14-dihydro-17-phenyl-18,19,20-trinor-prostaglandin F_2α [PGF_2α]-isopropyl ester), the active principal in latanoprost eyedrops (Xalatan, Pharmacia & Upjohn), is a selective prostaglandin F receptor agonist that is used for the reduction of the intraocular pressure in the treatment of glaucoma. Latanoprost has been shown to reduce the intraocular pressure effectively during long-term treatment in patients with open-angle glaucoma and ocular hypertension.^{1 - 4} While PGF_2α as well as PGF_2α–isopropyl ester have been shown to induce marked conjunctival hyperemia in healthy volunteers and in patients with glaucoma,^{5 - 6} (pp447-458) latanoprost induces significantly less hyperemia.^{7 - 8}

Prostaglandin F_2α is a known vasodilator but can also cause vasoconstriction (eg, in the brain).^{9 - 12} The question has, therefore, arisen whether latanoprost could induce vasoconstriction in the eye, particularly in the posterior pole, that could be detrimental and aggravate the glaucomatous disease.¹³ Conversely, it may also be of interest to know whether latanoprost induces increased blood flow in the eye, and particularly in the posterior pole, which could be advantageous in the treatment of glaucoma. Finally, since naturally occurring prostaglandins have been implicated to play a role in ocular inflammation, it is of interest to investigate whether latanoprost has any effect on the capillary permeability. The present study determined the vascular effects of latanoprost in the primate eye to address the points raised.

Animals were treated according to the tenets of the Association for Research in Vision and Ophthalmology Statement on the Use of Animals in Ophthalmic and Vision Research, and the experiments were approved by the local ethics committee for animal experimentation. Ten cynomolgus monkeys of either sex, weighing 2 to 4 kg, were used for the experiments. Anesthesia was induced with ketamine hydrochloride, 25 mg/kg body weight intramuscularly. Polyethylene tubings were inserted into both femoral arteries for registration of arterial blood pressure and collection of blood samples, and into both femoral veins for intravenous (IV) injections. A Ringer-acetate solution was infused IV at a rate of 10 mL/h. A catheter for injection of the microspheres was inserted into the left heart ventricle through the right brachial artery. Anesthesia was maintained by IV injections of a diluted pentobarbital sodium solution, 15 mg/mL. The animals underwent a tracheotomy and were connected to a ventilator for artificial ventilation. The PO₂, PCO₂, and pH of the arterial blood were regularly checked and adjusted when necessary, by changing the ventilation, by administering a sodium bicarbonate solution IV, or both. A heating pad was used to maintain normal body temperature.

Two series of experiments were performed. In the first, 5 animals were treated with PhXA34 (13,14-dihydro-15R, S-17-phenyl-18,19,20-trinor-PGF_2α-isopropyl ester), which contains about 50% latanoprost, and the blood flow was measured 1 and 2½ hours after dose administration. The capillary permeability was determined during 2 postdosing periods: 30 minutes to 2½ hours and 1½ to 2½ hours. In the second series of experiments, 5 animals were treated with latanoprost and the blood flow was determined 3, 4½, and 6 hours after dose administration. The capillary permeability was determined between 5 and 6 hours after dose administration.

Regional blood flow was measured using microspheres (mean±SEM size, 15.5±0.1 µm) labeled with cobalt 57, ruthenium 103, or niobium 95 (New England Nuclear Life Science Products, Boston, Mass) as described previously.¹⁴ A bolus of approximately 10⁶ microspheres per injection was introduced into the left ventricle of the heart. A reference blood sample was collected from a femoral artery in test tubes under free flow conditions for 1 minute from the start of the injection of the microspheres. By using microspheres with different radioactive labels, it was possible to determine the blood flow at different occasions in each experiment. At the end of the experiment, the animal was euthanized and the eyes were dissected as described.

The capillary permeability in the different tissues of the eye was determined by measuring the extravascular plasma-equivalent albumin space.^{15 - 16} Human serum albumin labeled with iodine 125 (¹²⁵I) or iodine 131 (¹³¹I) (Institutt for Energiteknikk, Kjeller, Norway) was used as a tracer, and free iodine was separated from labeled albumin immediately before the experiment using gel filtration (Sephadex G-25M, PD-10; Amersham Pharmacia Biotech AB, Uppsala, Sweden). The amount of free iodine in the infusion solution was less than 1%. Heparinized blood was collected from the animal and incubated for 1 hour at 37°C with a chromium 51 (⁵¹Cr)–labeled sodium cromate solution (New England Nuclear Life Science Products). The blood was centrifuged, and the supernatant discarded. The erythrocytes were washed with isotonic saline solution and centrifuged 6 times to remove unbound ⁵¹Cr. The final suspension of erythrocytes contained less than 1% free ⁵¹Cr.

In the first series of experiments, an IV bolus injection of ¹²⁵I-albumin was given 30 minutes after topical instillation of PhXA34, followed by an IV infusion to maintain a stable plasma level until the end of the experiment. Similarly, but 1½ hours after instillation of PhXA34, ¹³¹I-albumin was injected IV and then infused to maintain a stable plasma level until the end of the experiment; finally, the labeled erythrocytes were injected IV at 2 hours, ie, 30 minutes before the end of the experiment. Thus, it was possible to determine the extravascular plasma-equivalent albumin space during 2 periods, 30 minutes to 2½ hours and 1½ to 2½ hours after instillation of PhXA34. The regional blood flow to the different tissues of the eye was measured 1 and 2½ hours after instillation of PhXA34 in the same animals. In the second series of experiments, the regional blood flow was measured 3, 4½, and 6 hours after topical instillation of latanoprost, and the capillary permeability was measured between 5 and 6 hours after instillation of latanoprost using ¹³¹I-albumin and ⁵¹Cr-erythrocytes essentially as previously described.

Blood samples for the determination of PO₂, PCO₂, and pH of arterial blood were obtained 5 minutes before each microsphere injection. In addition, blood samples were obtained at about 15-minute intervals during the experiment to follow the concentration of the radioactively labeled albumin and erythrocytes in the blood and for hematocrit determination. The animals were euthanized with an overdose of pentobarbital. Aqueous humor was withdrawn, and the eyes were quickly enucleated, rinsed, and placed on ice. Tissue samples of the upper eyelid and the conjunctiva were collected. Periocular tissue adhering to the globe was removed, and the eye was then dissected into 6 parts: the retina, choroid, posterior sclera, iris, ciliary body, and anterior sclera with the cornea. Tissue and blood samples were weighed, and the radioactivity was counted in a multichannel gamma spectrometer (model 1282 Compu-Gamma; Wallac, Turku, Finland). The protein concentration in the aqueous humor was determined using a protein assay (Bio-Rad Laboratories Inc, Brussels, Belgium).¹⁷

The blood flow to the different tissues of the eye was calculated from the quotient of the radioactivity of microspheres in the tissue and the reference blood sample multiplied by the reference blood flow (volume of the reference blood sample collected during 1 minute). The blood flow of the eyelid and the conjunctiva was calculated per weight of tissue, whereas the blood flow of the other tissues is given as blood flow of the whole tissue. The total plasma-equivalent albumin space (intravascular and extravascular) was calculated by dividing the radioactivity of ¹²⁵I and ¹³¹I, respectively, in the tissue samples with the corresponding radioactivity in the reference blood sample obtained at the end of the experiment, and multiplied by the relative volume of plasma in the blood sample (1−hematocrit). The intravascular plasma volume was calculated from the ⁵¹Cr radioactivity in the tissue sample and blood sample and the hematocrit. The intravascular plasma volume is equal to the intravascular plasma-equivalent albumin space. The extravascular plasma-equivalent albumin space in the tissue samples was then calculated by subtracting the intravascular plasma volume from the total plasma-equivalent albumin space.

The test drugs, latanoprost and PhXA34, were synthetized at the Department of Medicinal Chemistry, Glaucoma Research Laboratories, Pharmacia & Upjohn, Uppsala, Sweden. PhXA34, the 15R,S epimeric mixture, contained about 50% of the R epimer (latanoprost) and was used at a higher concentration than latanoprost as the 15S epimer exhibits only about 10% of the pharmacological activity of latanoprost (J.S. and G.S.,unpublished data, 1992). The doses used in the experiment were 6 µg of latanoprost and 10 µg of PhXA34 (equivalent to about 6 µg of latanoprost), which is about 4 times higher than the clinical dose of latanoprost (1.5 µg) in Xalatan eyedrops. The prostaglandin analogues were dissolved in 0.9% sodium chloride with 0.5% polysorbate 80 as solubilizer. The contralateral control eye received the same volume of the vehicle only.

The results are expressed as the arithmetical mean value±SEM, and have been statistically analyzed with the matched pair t test. P≤.05 was considered statistically significant. Statistical power calculations demonstrated that a 50% change in blood flow of the intraocular tissues could be detected at the P=.05 level with 80% power of the test for n=5, while the corresponding figure for the extraocular tissues was 75% to 100%. The corresponding figures for the changes in extravascular plasma-equivalent albumin space were 20% to 50% for the intraocular tissues and 40% to 100% for the extraocular tissues.

The regional blood flow in the eye 1 to 6 hours after topical application of PhXA34 or latanoprost is provided in Table 1. In the eyelid, conjunctiva, and anterior sclera, there was a tendency toward lower blood flow in the experimental eye compared with the control eye at 1 hr after dose administration, although the difference was not statistically significant (P=.11, P=.40, P=.18, respectively). With time, the blood flow of these tissues tended to become higher in the experimental eyes than in the control eyes, and the difference reached statistical significance in the anterior sclera (Table 1). In the posterior sclera, the blood flow was low in the experimental and control eyes throughout the experiment, with a tendency toward higher blood flow in the experimental eye throughout the postdosing period. The blood flow of the intraocular tissues, the iris, the ciliary body, the choroid, and the retina did not change 1 to 6 hours after instillation of the test drugs, with the exception of the ciliary body, in which a small but statistically significant increase was found at 2½ hours after application of PhXA34 (Table 1). The mean arterial blood pressure at the time of the blood flow determinations was 94.3±5.7 mm Hg at 1 hour, 91.2±3.2 mm Hg at 2½ hours, 75.8±3.4 mm Hg at 3 hours, 75.4±2.7 mm Hg at 4½, hours and 77.8±4.2 mm Hg at 6 hours after dose administration. The 1 and 2½-hour values are based on the first series of experiments; the 3-, 4½-, and 6-hour values are based on the second series of experiments.

The values of the extravascular plasma-equivalent albumin space measured 30 minutes to 2½ hours, 1½ to 2½ hours, and 5 to 6 hours after topical application of PhXA34 or latanoprost are provided in Table 2. There was no difference between the experimental and control eyes in the extravascular plasma-equivalent albumin space of any of the tissues studied. Thus, the capillary permeability to albumin was not increased by the administration of latanoprost or PhXA34 to the eye. The largest extravascular plasma-equivalent albumin space in the control and experimental eyes generally was found at the end of the 30 minutes to 2½-hour period after dosing, which is logical since the labeled albumin had time to circulate 1 hour longer than during the other periods and reflects the physiological albumin turnover in the tissues. The protein concentration of the aqueous humor 2½ hours after topical application of PhXA34 was 0.09±0.01 and 0.11±0.03 mg/mL in the experimental and control eyes, respectively. The corresponding figures 6 hours after topical application of latanoprost were 0.19±0.08 and 0.36±0.28 mg/mL in the experimental and contralateral control eyes, respectively.

Latanoprost, a selective prostaglandin F receptor agonist, has been shown to induce significantly less surface hyperemia of the eye than PGF_2α–isopropyl ester in animals and humans.^{8 ,18 - 21} In vitro concentrations of latanoprost acid exceeding 10⁻⁶ mol/L have been shown to cause mild to moderate constriction of arteries and veins, whereas lower concentrations tend to cause vasodilation.^{21 - 22} Since PGF_2α is a known constrictor of some blood vessels (eg, in the brain),^{9 - 12} and since PGF_2α can cause leakage of capillaries and postcapillary venules,²³ it was considered important to investigate the vascular effects of latanoprost in the eye, although it has previously been shown that PGF_2α–isopropyl ester causes increased blood flow in the anterior uvea of rabbits and monkeys.⁶ (pp155-170)

When evaluating the results of the present study, it should be remembered that the study consists of 2 separate series of experiments. In the first series, 5 animals were followed up for 2½ hours after administration of PhXA34; in the second series, another 5 animals were followed up for 6 hours after the administration of latanoprost. Thus, it is not possible to make comparisons of the blood flow sequentially from 1 to 6 hours after dose administration. It should also be noted that the mean arterial blood pressures were lower in the second series of experiments compared with the first series. Since no baseline value was obtained before the administration of latanoprost or PhXA34, the experimental eye should always be compared with the contralateral control eye. A prerequisite for such a comparison is that there is no effect of the drug in the contralateral control eye. We have seen no indications of such an effect in the contralateral eye during treatment with latanoprost in animals (J.S. and G.S., unpublished data, 1991), and effects in the contralateral eye have not been seen with PGF_2α after single-dose administration.²⁴ It is thus adequate to use the contralateral vehicle-treated eye as a control.

The topical administration of latanoprost or PhXA34 at a dose of approximately 4 times the clinical dose had minimal effects on the intraocular and extraocular blood flow. In the surface structures of the eye, the eyelids, conjunctiva, and anterior sclera, there was a tendency toward reduced blood flow at 1 hour after dose administration. This could reflect a moderate vasoconstriction and would be logical with respect to the rather high concentration of latanoprost in these structures. It can be estimated that the concentration of latanoprost in the surface structures early may have been around 10⁻⁶ to 10⁻⁵ mol/L, and that would be in the concentration range at which latanoprost can be anticipated to cause mild to moderate vasoconstriction.^{21 - 22} Subsequently, vasodilation or a tendency toward vasodilation was seen in the surface structures, which agrees well with the hyperemia seen in some patients treated with Xalatan eyedrops. This is what can be anticipated with a decreasing concentration of latanoprost in the tissues, ie, a concentration range is reached at which latanoprost may cause vasodilatation.²²

In the posterior sclera, there was no tendency toward vasoconstriction at any time, if anything there was a tendency toward vasodilation, although the low blood flow in this tissue makes the blood flow determination with the microsphere method somewhat unreliable.

We found no or only minimal effects of latanoprost or PhXA34 on the blood flow in the anterior uvea, choroid, or retina. In particular, there was not even a tendency toward vasoconstriction in the posterior segment. In the anterior uvea, there was slight vasodilation at 2½ hours after dose administration, which coincides with the submaximum concentration of latanoprost acid in the aqueous humor (approximately 10⁻⁷mol/L,²⁵ ie, the concentration range at which the drug causes moderate vasodilation). Thus, the effect of topically applied latanoprost on the intraocular blood flow was negligible or at the most minimal in the monkey eye, which is in sharp contrast to the effect of PGF_2α–isopropyl ester, which in a previous study⁶ (pp155-170) was shown to increase the blood flow in the ciliary body by about 400% and in the anterior sclera by about 500% in the monkey.

Latanoprost or PhXA34 had no effect on the plasma-equivalent extravascular albumin space in any of the tissues studied. We measured the albumin leakage at 3 different periods after administration of the drug since there could be an early, in-between, or late effect on capillary permeability, and it is not known how quickly a leakage of albumin (eg, in the iris) is washed away. In contrast, we found that topically applied PGF_2α–isopropyl ester in the cynomolgus monkey causes a slight increase of capillary permeability to albumin in the anterior uvea (J.S. and G.S., unpublished data, 1992). The difference between the present results and the results previously obtained with PGF_2α–isopropyl ester⁸ may be due to the different receptor profile of the 2 compounds, latanoprost being a much more selective prostaglandin F receptor agonist. These data are important because they clearly indicate that latanoprost does not induce capillary leakage and are thus in good agreement with the results of clinical trials with latanoprost in which various techniques have been used to measure barrier permeability.^{26 - 28}

In conclusion, latanoprost used at 4 times the clinical dose had no or a negligible effect on the intraocular blood flow in phakic monkey eyes and only moderate effects on the blood flow in the extraocular structures. As the present data were obtained after a single dose, it cannot with certainty be stated that the response would be identical during long-term treatment with latanoprost. However, latanoprost does not seem to cumulate in the eye during repeated once-daily administration,²⁵ and the most marked effects with many drugs are frequently obtained after the first dose. The effect of latanoprost on the microcirculation was studied in anesthetized animals, and the anesthesia may have affected the cardiovascular system. However, since the prostaglandin effects at the microcirculatory level most likely are based on a direct action on the arterioles, capillaries, and venules, it is unlikely that the general anesthesia would have altered the response compared with that in conscious animals, although this possibility cannot be excluded.

Accepted for publication May 13, 1999.

We thank Iréne Aspman for help with editing the manuscript.

Reprints: Johan Stjernschantz, MD, PhD, Department of Neuroscience, Unit of Pharmacology, Uppsala University, Biomedical Center, Box 593, S-751 24 Uppsala, Sweden (e-mail: Johan.Stjernschantz@neuro.uu.se).