Variable Involvement of the Perivascular Retinal Tissue in Carbonic Anhydrase Inhibitor–Induced Relaxation of Porcine Retinal Arterioles In Vitro

Abstract

purpose. Inhibition of carbonic anhydrase in the eye is an important treatment modality for reducing the intraocular pressure in glaucoma. However, evidence suggests that carbonic anhydrase inhibition also exerts a relaxing effect on the vessels in the optic nerve, and it has been suggested that this vasorelaxing effect is a result of an interplay between the perivascular tissue and constituents in the retinal vascular wall. However, the exact nature of this interplay is unknown.

methods. Isolated porcine retinal arterioles and arterioles with preserved perivascular retinal tissue were mounted in a myograph. After precontraction with the prostaglandin analogue U46619, the vasorelaxing effect of the carbonic anhydrase inhibitors methyl bromopyruvate, ethyl bromopyruvate, acetazolamide, and dorzolamide were studied.

results. All the examined carbonic anhydrase inhibitors induced a significant relaxation of retinal arterioles. There was no significant difference between the effect of the different carbonic anhydrase inhibitors in the presence of perivascular retinal tissue. However, in the isolated retinal arterioles the vasodilating effect of dorzolamide was significantly lower, and the vasodilating effect of acetazolamide almost disappeared.

conclusions. A further elucidation of the mechanisms of action of carbonic anhydrase–induced dilation of retinal arterioles may contribute to a better understanding of the regulation of retinal blood flow. The perivascular retinal tissue may play a significant role in diameter control of retinal arterioles.

Inhibition of carbonic anhydrase in the eye is an important treatment modality for glaucoma. ¹ The therapeutic effect is assumed to be due to a decrease in the formation of aqueous humor with a consequent reduction of the intraocular pressure and protection of the intraocular tissue from ischemia. ^{2 3 4} However, evidence suggests that carbonic anhydrase inhibition also increases the perfusion of the optic nerve and the retina, both in vitro and in vivo. ^{3 5 6} This effect has been shown to be independent of nitric oxide suggesting a mechanism of action that is independent of the vascular endothelium. ⁷  

Carbonic anhydrase catalyzes the association of water and carbon dioxide to carbonic acid, which in turn undergoes spontaneous dissociation into bicarbonate and hydrogen ions. However, since the concentrations of metabolites driving the reaction locally in the tissue are unknown, the normal direction of this reaction and thereby the metabolic effects of carbonic anhydrase inhibition are also unknown. Carbonic anhydrase is abundantly present in the perivascular retina, and it is conceivable that the vasorelaxing effect of carbonic anhydrase inhibitors is due to an interplay between the perivascular tissue and the vascular wall, probably involving different isoenzymes. ^{8 9 10} However, the exact nature of this interplay is unknown. 

In the present study, isolated porcine retinal arterioles and arterioles with preserved perivascular retinal tissue were mounted in a myograph, and after precontraction of the vessel with the prostaglandin analogue U46619, the vasorelaxing effect of the carbonic anhydrase inhibitors acetazolamide, dorzolamide, methyl bromopyruvate, and ethyl bromopyruvate were studied. 

Methods

Solutions

PSS^1.6 (in mM): NaCl 119, KCl 4.7, MgSO₄ 1.17, NaHCO₃ 25, KH₂PO₄ 1.18, CaCl₂ 1.6, EDTA 0.026, glucose 5.5, and HEPES 5.0. PSS^0.0: same as PSS^1.6 apart from omission of CaCl₂. KPSS: same as PSS^1.6, but with NaCl replaced with an equimolar concentration of KCl, resulting in a [K⁺] of 125 mM. 

Drugs

9,11-Dideoxy-11α, 9α-epoxymethanoprostaglandin F 2α (U46619), acetazolamide, methyl bromopyruvate, ethyl bromopyruvate, and dl-amino-5-phosphonovaleric acid (DL-APV) were all purchased from Sigma-Aldrich Denmark (Brøndby, Denmark). Dorzolamide was generously provided by MSD Denmark (Glostrup, Denmark). 

Acetazolamide was dissolved in dimethylsulfoxide (DMSO) to a 1-M stock solution. After dilution for the experiments the DMSO concentration never exceeded 1%. The remaining drugs were dissolved and diluted in distilled water to 1-M stock solutions. 

Tissue

Eyes from Danish pigs (bred from Danish Landrace, Danish Yorkshire, Danish Duroc, and Danish Hampshire) approximately 6 months of age and weighing 85 to 90 kg were collected from the local abattoir immediately after the animals had exsanguinated. The eyes were immersed in PSS^1.6 at 5°C, and were transported to the laboratory within 45 minutes. 

The eyes were bisected by a frontal section through the equator. From the posterior half of the eye, the retina was gently detached from the pigment epithelium by injection of PSS^1.6 between the layers using a blunt syringe. The detached retina was removed from its attachment at the optic nerve head with Vanna scissors and was transferred to a paraffin-coated Petri dish to be extended and fixed by pins. During all the following handling and experiments 50 μM of the glutamate NMDA receptor antagonist DL-APV was added to the dissection and chamber fluids to avoid irreversible relaxation of the vessels, probably due to glutamate released from the perivascular retinal tissue. ¹¹ Subsequently, a 2-mm-long segment of the arteriole located at the first branching level was dissected under a microscope. The vessels either had a 2-mm rim of retinal tissue preserved on each side of the vessel or were isolated from the perivascular retinal tissue. 

Myograph

A myograph (model 610; Danish Myo Technology, Aarhus, Denmark) with four chambers each containing a volume of 5 mL was used. Two steel jaws were placed centrally in each chamber. One of the jaws was attached to a micrometer screw so that the distance between the jaws could be regulated, whereas the other jaw was attached to a force transducer with a sensitivity of 0.01 mN. 

Mounting

The dissected arteriolar segment was transferred to the myograph chamber containing PSS^0.0 at 5°C to minimize the traumatic impact on the vessel. The arteriolar segment was mounted on two Wolfram wires with a diameter of 25 μm by using a stereo microscope. First, one of the wires was tied to the upper screw on the jaw attached to the micrometer. Subsequently, with forceps (Dumont no. 5), the vessel was drawn over the free end of the wire, which was subsequently tied to the lower screw on the jaw. The other wire was guided through the lumen along the first wire and was subsequently attached to the force transducer. The heating element of the myograph was turned on to increase the temperature to 37°C and bubbling of the tissue bath with a mixture of 95% atmospheric air and 5% CO₂ commenced. The force transducer of the myograph was set to register the tone of the vessel every second, and these data were continuously recorded and stored on a computer. 

Normalization

Normalization was performed to ensure that results obtained from arterioles with different diameter would be comparable. ¹² In short, the vessel diameter was increased in four steps in PSS^0.0, and the passive tensions (corresponding to transmural pressures between 0 and ∼70 mm Hg) were determined. This diameter–tension relationship was exponential, and the intercept between this curve and a straight line based on the Laplace equation (wall tension = transmural pressure × radius) with the transmural pressure set to 70 mm Hg was calculated. Using the built-in micrometer screw, the jaws of the myograph were adjusted to 93.5% of the intercept length, at which the vessels develop the maximum tension (i.e., the optimal passive stretch on the vascular smooth muscle cells). 

Experimental Protocol

After mounting and normalization, PSS^0.0 was replaced by PSS^1.6 in the chamber, and the vessels equilibrated for 30 minutes until a stable tone was achieved. Subsequently, the arteriole was precontracted with the prostaglandin analogue U46619 (10⁻⁶ M), and after 5 to 10 minutes when the tone had stabilized, either control buffer or one of the carbonic anhydrase inhibitors acetazolamide, dorzolamide, methyl bromopyruvate, or ethyl bromopyruvate, were added in increasing concentrations in five steps between 10⁻⁶ and 10⁻² M. The tone was measured after it had stabilized and at least 3 minutes after the previous concentration shift. At the end of each experiment, the viability of the arteriole was tested by adding KPSS for 3 minutes. In 39% of the experiments, the mounted arteriole did not react with an increase in the tone of at least 0.25 N/m. These vessels were regarded as nonviable at this time and were discarded. Finally, in the viable vessels the passive tone of the arterioles was measured after addition of 10⁻⁴ M papaverine. 

Control Experiments

Additional control experiments were performed for both isolated arterioles and arterioles with preserved perivascular retinal tissue. The following conditions were tested. (1) Osmolarity: concentration–response experiments were performed with the carbonic anhydrase inhibitor replaced by sucrose 10⁻⁶to 10⁻² M (n = 13); (2) solute: the carbonic anhydrase inhibitor replaced by 50 μL dimethylsulfoxide (DMSO), corresponding to a concentration of 1% in the myograph chamber (n = 10); and (3) DL-APV: concentration–response experiments with acetazolamide in which DL-APV was omitted (n = 11) or replaced by 10⁻⁴ M ibuprofen (n = 12). DMSO induced a slight (0.18 ± 0.07 N/m) but significant (P = 0.04) vasorelaxation in experiments with preserved perivascular retinal tissue, but otherwise the control experiments did not identify any biasing factors. 

Data Analysis

The active tone was defined as the precontraction tone subtracted by the tone after addition of papaverine. For further analyses, the tone measured at each concentration of carbonic anhydrase inhibitor was normalized to the precontraction tone (100%) and the tone after addition of papaverine (0%) for the given experiment. 

The tone response was assessed as the maximum relaxation—that is, the normalized tone produced at the highest concentration of carbonic anhydrase inhibitor (10⁻² M)—and as the median effective concentration (EC₅₀). To calculate the EC₅₀, the data points were fitted to the Michaëlis-Menten formula (Prism, ver. 4.02, GraphPad, San Diego, CA). 

The Kruskal-Wallis test was used to test for differences in EC₅₀ and the maximum relaxation among the different carbonic anhydrase inhibitors, both on isolated arterioles and on arterioles with preserved perivascular retinal tissue. 

The Mann-Whitney test was used to test for differences in EC₅₀ and the normalized tone produced at each concentration of carbonic anhydrase inhibitor among isolated vessels and vessels with preserved perivascular retinal tissue. 

Results

The isolated arterioles had a significantly higher active tone (0.78 ± 0.04 N/m) than the arterioles with preserved perivascular retinal tissue (0.57 ± 0.06 N/m, n = 24; P < 0.01, n = 24). 

Figure 1shows an example of the effect of acetazolamide on an isolated retinal arteriole (Fig. 1A)and an arteriole with preserved perivascular retinal tissue (Fig. 1B) . It appears that the vasorelaxing effect of high concentrations of acetazolamide depended on the presence of the perivascular tissue. 

Figure 2shows the normalized concentration response curves for the four studied carbonic anhydrase inhibitors with and without perivascular retinal tissue, and Table 1shows the maximum relaxation and the EC₅₀ values obtained from these curves. 

In retinal vessels with preserved perivascular retinal tissue there was no significant difference between the EC₅₀ and maximum relaxation for the different carbonic anhydrase inhibitors (P = 0.17 and P = 0.99, respectively). However, in isolated arterioles there was a significant difference between both the EC₅₀ and the maximum relaxation (P < 0.001 and P < 0.001, respectively) for the different carbonic anhydrase inhibitors. The difference in EC₅₀ was due to a significantly reduced sensitivity to dorzolamide and acetazolamide (P = 0.01 and P = 0.004, respectively), and a trend to increased sensitivity to methyl- and ethyl bromopyruvate (P = 0.078 and P = 0.055, respectively) in the isolated arterioles. The difference in the maximum relaxation was due to a significantly reduced response to acetazolamide (P = 0.004) in the isolated arterioles. However, at the next highest concentration (10⁻³ M) the vasorelaxing effect of both dorzolamide and acetazolamide were significantly reduced in the isolated arterioles (P = 0.007 and P = 0.015, respectively) compared with the arterioles with perivascular tissue. 

Discussion

In the present study, it was shown that the vasorelaxing effect of the carbonic anhydrase inhibitors acetazolamide, dorzolamide, methyl bromopyruvate, and ethyl bromopyruvate on porcine retinal arterioles in vitro was similar in the presence of perivascular retinal tissue. This is in accordance with clinical studies showing no difference between the vasorelaxing effect of different carbonic anhydrase inhibitors administered intravenously ⁴ where the effect is exerted on vessels embedded in a normal retinal environment. In addition, the similar effect of different carbonic anhydrase inhibitors on retinal arterioles with preserved perivascular tissue makes it unlikely that the different effects observed on isolated retinal arterioles had any relation to the preconstriction of the arterioles with U46619, or the slight vasorelaxing effect of the dimethylsulfoxide used to dissolve acetazolamide for the experiments. 

The vasorelaxing effect of carbonic anhydrase inhibitors was observed in a concentration range at which acetazolamide and dorzolamide in clinical studies have been shown to induce vasorelaxation ^{2 13 14} and at a concentration comparable to that obtained after infusion of 500 mg acetazolamide—for example, in therapy for acute glaucoma. ¹⁵ Similar findings have been found in animal experiments, both in vitro ^{3 4} and in vivo. ^{5 6} However, the findings are highly suggestive that the effective concentration of the carbonic anhydrase inhibitor is much higher than that found in experiments of binding kinetics in vitro. ¹⁶ It might therefore be conjectured that the observed effect involves mechanisms other than inhibition of carbonic anhydrase or perhaps is related to a reduced sensitivity of the compounds due to a masking of binding sites. 

The maximum effect and the potency of the tested carbonic anhydrase inhibitors appears to be similar in both in vivo and in vitro experiments. ^{3 5 6} Therefore, it is unlikely that blocking of carbonic anhydrase in erythrocytes that are absent in the nonperfused vessels studied in in vitro experiments is involved in the effect. The vasorelaxation observed in vitro or after systemic administration of carbonic anhydrase inhibitors can therefore be assumed to be due to an effect on the vascular smooth muscle cells alone or on these cells in an interplay with other cellular elements in the surrounding tissue. Previous studies suggest that this effect is not via NO from the endothelium, ⁷ and the results of the present study support the view that the perivascular retinal tissue is involved. First, the resting tone of the isolated vessels was higher than that of vessels with perivascular tissue suggesting that a vasorelaxing factor released from the perivascular tissue had been eliminated. ¹¹ Second, removal of the perivascular tissue had significant effects on the vasorelaxing effect of the studied carbonic anhydrase inhibitors. However, this response depended on the type of the carbonic anhydrase inhibitor. Thus, removal of the perivascular retinal tissue resulted in an insignificant increase in the tone response and the sensitivity of methyl- and ethyl bromopyruvate. Since the effect of these compounds is mainly on the cytosolic isoenzyme I located in the vascular smooth muscle cells ¹⁷ or the vascular endothelial cells, ⁸ it is likely that the removal of the perivascular retina eliminates a factor interfering with these cells. Conversely, the vasorelaxing effect of dorzolamide and especially acetazolamide was significantly reduced in isolated arterioles, as evidenced by both a reduction in the maximum relaxation and an increase in EC₅₀. It is therefore highly likely that the vasorelaxing effect of these drugs depends on carbonic anhydrases in the perivascular retina. This effect may have involved different isoenzymes such as cytosolic isoenzyme II in pericytes and retinal Müller cells and membrane-bound isoenzyme XIV located in retinal astrocytes and Müller cells. ^{8 9 10}  

In several studies extracellular acidosis has been shown to have a tone-relaxing effect on retinal arterioles. ^{4 12 18} However, other findings indicate that the tone-relaxing effect of carbonic anhydrase inhibitors is independent of the extracellular acidosis. ⁶ This notion suggests either that the vasodilating effect of extracellular acidosis is related to a simultaneous presence of intracellular alkalosis ⁴ or that the vasodilation is due to a direct effect of carbonic anhydrase inhibitors on the vascular smooth muscle cells. 

Both acetazolamide and dorzolamide are known to inhibit carbonic anhydrase isoenzyme II, and especially dorzolamide has a strong selectivity for this isoenzyme. ^{1 17 19} Dorzolamide has been synthesized to possess physicochemical properties that allow its uptake into the eye by diffusion after instillation in the conjunctiva. Systemic absorption of dorzolamide after this route of administration is low, ²⁰ and it has been shown that short-term treatment with this drug does not affect the blood flow in the retina and the optic nerve. ^{21 22 23} However, it has also been shown that after a few days of local treatment with dorzolamide, the blood flow of the optic nerve is increased as a result of absorption to the posterior part of the eye. ^{24 25 26} The findings of the present study suggest that this effect is mediated by the perivascular retinal tissue. However, the experiments also indicate that the effect of acetazolamide is different from that of dorzolamide which may reflect that the mechanisms of action of these two compounds differ. Acetazolamide has recently been shown to exert a direct effect on vascular smooth muscle cells through an opening of calcium-activated potassium channels, an effect that is conceivably due to intracellular alkalosis. ¹⁶ Opening of these channels leads to inactivation of voltage-gated calcium channels and a consequent relaxation of the vascular smooth muscle cells leading to vasorelaxation. However, this is not the case for acetazolamide in porcine retinal arterioles as these show almost no relaxation in response to acetazolamide in the absence of perivascular tissue. Thus, a more detailed knowledge of the molecular mechanism of action of different carbonic anhydrase inhibitors is needed to understand completely the effect of these compounds on retinal vascular tone. This knowledge might be obtained by including studies of intracellular pH in individual cells of the walls of retinal arterioles and the perivascular retinal tissue. It is conceivable that a full understanding of the vasodilating effect of carbonic anhydrase inhibitors on retinal arterioles may contribute further to the understanding of how retinal perfusion is regulated and point to possible modes of intervention to cure disturbances in the retinal blood flow. The present findings show that these efforts should include considerations of the role of the perivascular retinal tissue in the regulation of retinal blood flow. 

 Supported by MSD and the VELUX Foundation.

 Submitted for publication January 16, 2007; revised February 25, April 18, and May 16, 2007; accepted August 20, 2007.

 Disclosure: A.K. Kehler, MSD (F); K. Holmgaard, MSD (F); A. Hessellund, MSD (F); C. Aalkjaer, MSD (F); T. Bek, MSD (F)

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Representative traces showing the tone response of acetazolamide on precontracted isolated retinal arterioles (A) and on precontracted retinal arterioles with preserved perivascular tissue (B).

Figure 2.

 

Normalized concentration–response curves obtained with the carbonic anhydrase inhibitors acetazolamide, dorzolamide, and methyl- and ethyl bromopyruvate on isolated retinal arterioles and arterioles with preserved perivascular retinal tissue. *Significantly different response between isolated arterioles and arterioles with preserved perivascular retinal tissue.

Table 1.

 

Maximum Relaxation and EC₅₀

Carbonic Anhydrase Inhibitor	Isolated Arterioles	Arterioles with Perivascular Retinal Tissue	P
Maximum relaxation (%)
Acetazolamide	6.8 (4.3–9.3)	102 (100–105)	0.004*
Dorzolamide	62 (50–73)	95 (92–98)	0.010*
Methyl bromopyruvate	97 (97–98)	97 (94–100)	0.34
Ethyl bromopyruvate	100 (98–102)	95 (92–98)	0.15
	P < 0.001, †	P = 0.17
EC50 (mM)
Acetazolamide	Infinite	0.62 (0.21–1.86)	0.004*
Dorzolamide	5.50 (4.50–6.71)	1.22 (0.75–1.97)	0.010*
Methyl bromopyruvate	0.12 (0.066–0.22)	1.08 (0.69–1.69)	0.078
Ethyl bromopyruvate	0.06 (0.038–0.093)	0.67 (0.30–1.48)	0.054
	P < 0.001, †	P = 0.99

 Data are the mean ± SEM (range); n = 6.

*  Mann-Whitney test.

†  Kruskal-Wallis test.

The authors thank technician Poul Rostgaard for skillful assistance.