The Vasodilating Effect of Acetazolamide and Dorzolamide Involves Mechanisms Other Than Carbonic Anhydrase Inhibition

Abstract

purpose. Carbonic anhydrase inhibitors reduce intraocular pressure, which may protect the optic nerve from ischemia. However, carbonic anhydrase inhibitors have also been shown to dilate the blood vessels in the retina and the optic nerve head. The purpose of the present study was to investigate whether CO₂, H⁺, or factors other than carbonic anhydrase inhibition are involved in this vasodilating effect.

methods. Porcine retinal arterioles with preserved perivascular retinal tissue were mounted in a myograph for isometric force measurements. After precontraction with the prostaglandin analogue U46619, concentration–response experiments were performed with acetazolamide and dorzolamide before and after removal of the perivascular retina. The experiments were performed at normal pH and during acidosis, during normocapnia and hypercapnia, as well as in the nominal absence of CO₂ and HCO₃ ⁻.

results. The maximum relaxation was significantly lower and the EC₅₀ significantly higher during normal pH compared with acidosis (P = 0.002 and P < 0.0001, respectively), but neither the maximum relaxation nor EC₅₀ was changed by hypercapnia (P = 0.054 and P = 0.57, respectively). The findings confirmed that carbonic anhydrase–induced vasodilation depends on the perivascular retinal tissue and that dorzolamide produces significantly more pronounced relaxation than does acetazolamide. EC₅₀ of carbonic anhydrase inhibitor–induced vasorelaxation and the maximum relaxation of dorzolamide were unchanged in the nominal absence of CO₂ and HCO₃ ⁻ (P = 0.65 and P < 0.0001, respectively).

conclusions. The vasodilating effect of carbonic anhydrase inhibitors on porcine retinal arterioles depends on the perivascular retinal tissue and acidosis, but not on hypercapnia. The effect involves mechanisms other than carbonic anhydrase inhibition.

Inhibition of carbonic anhydrase is an important principle for the treatment of glaucoma. The effect is assumed to be due to an inhibition of aqueous humor production in the ciliary body with a resulting reduction in the intraocular pressure that protects the eye from ischemic damage. ^{1 2 3} However, evidence suggests that the effect may also be due to a relaxation of the blood vessels in and around the optic nerve head, so that the perfusion of these structures is improved and tissue ischemia is prevented. ^{1 4 5} Thus, carbonic anhydrase inhibitors (CAIs) have been shown to have a vasodilating effect on the blood vessels in the retina and the optic nerve from animals both in vitro ¹ and in vivo, ⁵ and in the human retina in vivo, ² an effect that has been shown to be independent of NO. ⁶ The enzymatic effect of carbonic anhydrase is to hydrate carbon dioxide to carbonic acid, which dissolves spontaneously into protons and bicarbonate according to the equation: CO₂+H₂O↔H₂CO₃↔ H⁺+HCO₃ ⁻. However, since the concentration of metabolites on each side of the equation is unknown in vivo, it is also unknown in which direction and driven by which metabolites the reaction may be catalyzed by CAI to induce vasodilation. Previous experiments on retinal arterioles from experimental animals have shown that hypercapnic and normocapnic acidosis have different effects on the retinal metabolism. ⁷ The two acidosis forms induce the same degree of vasodilation in pial ⁸ and retinal ⁹ arterioles, which may indicate that the mechanism of action of vasodilation is through H⁺, but other in vivo experiments in pigs have shown a lack of vasodilating effect of metabolic acidosis induced by intravenous infusion of NH₄Cl. ¹⁰ However, it is likely that mechanisms other than the enzymatic effect on carbonic anhydrase may contribute to the vasodilating effect of CAI, ¹¹ and recent experiments have suggested that the perivascular retinal tissue may be involved in the effect. ¹²  

We therefore investigated the vasodilating effect of the CAIs acetazolamide and dorzolamide on porcine retinal arterioles in vitro during normocapnia and hypercapnia combined with normal pH and acidosis and in the presence of perivascular retinal tissue and after this tissue had been removed. In addition, experiments were performed in the absence of substrate to the carbonic anhydrase enzyme. 

Materials and Methods

Solutions

The solutions used were as follows: (1) PSS, containing (in mM): NaCl 119, KCl 4.7, MgSO₄ 1.17, NaHCO₃ 25.0, KH₂PO₄ 1.18, CaCl₂ 1.6, EDTA 0.026, glucose 5.5, HEPES 5.0; (2) KPSS, same as PSS, with the NaCl replaced with equimolar amounts of KCl to achieve a K⁺ concentration of 124 mM; (3) PSS_NA, normocapnic acidosis. PSS with NaHCO₃ reduced to 8 mM, HEPES increased to 10 mM, and pH maintained at 7.0; (4) PSS_HC, hypercapnia: PSS with NaHCO₃ increased to 45.0 mM, NaCl reduced to 74.0 mM; (5) PSS without NaHCO₃, with NaCl increased to 144 mM and HEPES increased to 10.0 mM. Additional solutions were the same as those listed, apart from the omission of CaCl₂ (e.g., PSS^0.0: same as PSS but without CaCl₂). 

Vasoactive Compounds

Acetazolamide (Sigma-Aldrich, Vallensbaek, Denmark) was dissolved in equal amounts of 99.5% dimethyl sulfoxide (DMSO, Sigma-Aldrich) and 96% ethanol to a concentration of 1 M. The solution was titrated with NaOH from pH 5.5 to 7.2 and was stored as a stock solution at 4°C. Immediately before use, the solution was diluted in the same DMSO/ethanol solution to a concentration of 10⁻¹ M and was further diluted in demineralized water to 10⁻² to 10⁻⁶ M for the concentration–response experiments. 

Dorzolamide was generously provided by MSD Denmark (Glostrup, Denmark). The compound was dissolved in DMSO/ethanol to 1 M and further diluted in demineralized water similar to the dilution of acetazolamide, except that the primary 1 M solution was titrated from pH 3.5. 

The prostaglandin analogue 9,11-dideoxy-11α, 9α-epoxymethanoprostaglandin F 2α (U46619), purchased from Cayman Chemical Company (Ann Arbor, MI), was dissolved in demineralized water to 0.001 M and was diluted to 10⁻⁶ M in the tissue chamber. 

The vasorelaxing compound papaverine (SAD, Copenhagen, Denmark) was dissolved in demineralized water to 0.01 M, and the L-type calcium channel blocker nifedipine (Sigma-Aldrich, Broendby, Denmark) was dissolved in ethanol 99% to 10⁻² M, and subsequently diluted in demineralized water to concentrations of 10⁻³, 10⁻⁴, 10⁻⁵, 10⁻⁶, and 10⁻⁷ M. 

Procedure

Eyes of Danish pigs (bred from Danish Land Race, Danish Yorkshire, Danish Duroc, and Danish Hampshire lines) approximately 6 months of age and weighing 85 to 90 kg were collected from the local abattoir immediately after the animals had been exsanguinated. The eyes were immersed in PSS^0.0 at 4°C, and were transported to the laboratory within 60 minutes. One vessel from each pig was used. All experiments adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

The eye was 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^0.0 between the layers with a blunt syringe (Rycroft 40 × 22 mm Steriseal; Maersk Medical Ltd., Redditch, UK). The detached retina was removed from its attachment at the optic nerve head with Vanna’s scissors, and was transferred to a Sylgard-coated Petri dish to be extended and fixed by pins. The Petri dish was placed under a dissection microscope (Stemi SV11; Carl Zeiss Meditec, Oberkochen, Germany), and with a microblade (BD, angled 30°, D. J. Instrumenter A/S, Gentofte, Denmark) a 2-mm-long segment of the arteriole was dissected within 1 disc diameter of the optic disc and with a patch of 1 to 2 mm of retinal tissue preserved on each side of the arteriole. The preparation was transferred to the myograph chamber containing PSS^0.0 at 5°C using a spoon to minimize the traumatic impact on the vessel. 

Myograph

Two myographs (model 610; Danish Myo Technology, Aarhus, Denmark) each with four chambers containing a volume of 5 mL were 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, and the other jaw was attached to a force transducer with a sensitivity of 0.01 mN. 

Mounting

The vessel was mounted on two Tungsten wires with a diameter of 25 μm, guided by a stereo microscope (Stemi 100; Carl Zeiss Meditec). First, one of the wires was tied to the upper screw on the jaw attached to the micrometer. Subsequently, with a 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. Finally, the heating element of the myograph was turned on to increase the temperature to 37°C. The force transducer of the myograph was set to register the tone of the vessel every second and store the data on a computer. 

Normalization

Normalization was performed to ensure that results obtained from arterioles with different diameter were comparable. ¹³ 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 approximately 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, we adjusted the jaws of the myograph to 93.5% of the intercept length, at which the vessels show the maximum tension (i.e., the optimal passive stretch on the vascular smooth muscle cells). ¹³ At this length, the diameter of the vessels ranged between 130 and 170 μm. After the arteriole had stabilized, the starting tone to be indexed as 0% tone for that vessel was recorded. Finally, the fluid was changed to PSS for the concentration response experiments. 

Experimental Conditions

The vasodilating effect of acetazolamide and dorzolamide were studied under five different conditions:

•  
  Normal pH aimed at 7.4 and normocapnia produced by PSS bubbled with a mixture of 95% atmospheric air and 5% CO₂;
•  
  Acidosis aimed at pH 7.0 and normocapnia produced by PSS_NA bubbled with 95% atmospheric air and 5% CO₂;
•  
  Acidosis at pH 7.0 of hypercapnia produced by PSS bubbled with 80% atmospheric air and 20% CO₂;
•  
  Normal pH 7.5 and hypercapnia produced by PSS_HC bubbled with 80% atmospheric air and 20% CO₂;
•  
  Normal pH 7.5 in PSS without NaHCO₃ ⁻ and bubbled with 79% N₂, 21% O₂, thus omitting substrates (CO₂ and HCO₃ ⁻) to the carbonic anhydrase reaction.

Preparation

With a pH electrode (PHM210, Radiometer Analytical, Lyon, France), the pH was measured alternately in the different chambers at the beginning and the end of an experiment and continuously in one of the chambers throughout the experiment. Levels were recorded immediately before changes in the composition of the chamber fluids. 

The chambers were bubbled continuously from a gas cylinder with the outlet set to a pressure of approximately 1 bar. In each myograph chamber, the air stream was adjusted to achieve the intended pH while avoiding large bubbles that would cause artifactual movements of the vessel. 

Continuous measurements showed that pH varied between 7.15 ± 0.05 in the acidosis experiments and between 7.6 ± 0.1 in the normal pH experiments. 

Experimental Procedure

Each experiment consisted of a concentration–response experiment with preserved perivascular retinal tissue, removal of the perivascular retina, and finally a repetition of the concentration–response experiment without perivascular retinal tissue. 

Concentration-Response Experiment.

The arteriole was precontracted with 10⁻⁶ M U46619, and if the vascular tone increased more than 0.2 N/m compared with the condition in PSS^0.0, the vessel was considered to be viable and the experiment was continued. After a stable tone was achieved, acetazolamide or dorzolamide were added in increasing concentrations (i.e., 10⁻⁶, 10⁻⁵, 10⁻⁴, 10⁻³, and 10⁻² M). Each new concentration was added when the tone had stabilized for at least 3 minutes after addition of the previous concentration. 

Removal of the Perivascular Retinal Tissue.

After the first concentration–response experiment, the fluid in the chamber was changed twice, and the perivascular retina was removed from the vessel with a forceps (Dumont no. 5). Finally, the chamber fluid was changed twice. 

Repetition of the Concentration-Response Experiment.

After the arteriolar tone had stabilized, U46619 was added again, and the concentration–response experiment was repeated. Finally, the chamber fluid was exchanged three times with PSS, once to KPSS for 3 minutes, to ensure that the vessel could still contract to produce a tone higher than 0.2 N/m, and further shifted with PSS three times before 10⁻⁴ M papaverine was added, and the tone was recorded to represent maximum relaxation which was indexed to tone = 0%. 

Control Experiments

The following control experiments were conducted:

•  
  Evaluation of whether acidosis itself affects vasorelaxation independent of carbonic anhydrase inhibition. Nifedipine was used as a vasorelaxing compound in the concentrations: 10⁻¹⁰, 3 × 10⁻¹⁰, 10⁻⁹, 3 × 10⁻⁹, 10⁻⁸, 3 × 10⁻⁸, 10⁻⁷, 3 × 10⁻⁷, and 10⁻⁶ M, and was tested during normocapnia with normal pH and acidosis, respectively.
•  
  Control experiment to determine the effect of the solvent. The experiments were performed during normocapnia with normal pH, without addition of CAIs, but with the ethanol and DMSO that were present after dilution of acetazolamide and dorzolamide from the stock solutions.
•  
  Time control to observe whether one concentration–response experiment with acetazolamide during normocapnia and normal pH affects a subsequent experiment. The perivascular retina was removed from the vessel before normalization so that two identical experiments on isolated retinal vessels were performed sequentially.

Data Analysis

Altogether 148 arterioles were studied, of which 78 were discarded because of a weak (<0.2 N/m) response after addition of U46619, and 11 arterioles were discarded because of sudden unexplained relaxation unrelated to the addition of vasodilators. Observations were therefore conducted in 59 arterioles. 

The tone measurements obtained every second were recorded by commercial software program (Myodaq 2.01 Multo+ and were viewed in the program Myodata 2.02; Danish Myo Technology, Aarhus, Denmark). The data was stored in spreadsheet format for further analysis (Excel, Microsoft, Redmond, WA). The vascular tone obtained after addition of 10⁻⁶ M U46619 was set to 100% and the tone obtained in PSS^0.0 (arterioles with preserved perivascular retina) and the tone after addition of papaverine (isolated arterioles) were set to 0%. In each normalized concentration response curve EC₅₀ was calculated on the basis of the Michaëlis-Menten equation using the program (GraphPad Prism 5.00; San Diego, CA). EC₅₀ could not be calculated in cases in which relaxation was only observed at the highest concentration of CAI and was therefore set to the highest concentration used (i.e., 10⁻² M). The maximum relaxation was calculated as 100% minus the normalized tone produced by the highest CAI concentration. 

Statistical Analyses

Both EC₅₀ and the maximum relaxation were compared by four-way ANOVA with the following factors: type of CAI (acetazolamide or dorzolamide), presence or absence of perivascular retinal tissue, pH (acidosis or normal pH), and CO₂ concentration (hypercapnia or normocapnia). 

In the experiment to test whether the vasodilating effect of CAI depends on the presence of substrate to the carbonic anhydrase enzyme, EC₅₀ and the maximum relaxation were compared by three-way ANOVA with the following factors: type of CAI (acetazolamide or dorzolamide), presence or absence of perivascular retinal tissue, and presence or absence of CO₂ and HCO₃ ⁻ in the chamber fluid. The maximum relaxation for dorzolamide and acetazolamide was further compared by two-way ANOVA with the following factors: presence or absence of perivascular retinal tissue and presence or absence of CO₂ and HCO₃ ⁻ in the chamber fluid. 

In the experiment to control for the effect of acidosis on nifedipine-induced vasorelaxation, EC₅₀ and the maximum relaxation were compared by two-way ANOVA for nifedipine with the following factors: normal pH or acidosis and presence or absence of perivascular retinal tissue. 

Repeated-measures ANOVA was used to test the influence of solvent and time. 

Results

The results are shown in Figure 1and in Tables 1 and 2 . There was a significant increase in maximum relaxation (MR) and a significant decrease in EC₅₀ (P = 0.002 and P < 0.0001, respectively) during acidosis compared with normal pH (broken versus solid lines), whereas there was no significant difference between EC₅₀ (P = 0.57) and the MR (P = 0.054) obtained during normocapnia versus hypercapnia (Figs. 1A 1Bversus Figs. 1C 1D ). The results confirmed previous results of a significant increase in EC₅₀ (P < 0.0001) and decrease in the MR (P < 0.0001) after removal of the perivascular retina (blue versus red curves), as well as a significant decrease in the MR (P < 0.0001) but not EC₅₀ (P = 0.2215) after addition of acetazolamide compared with dorzolamide (Figs. 1A 1Cversus Figs. 1B 1D ). All four- and three-factor interactions were found to be statistically insignificant, but there was significant interaction (P = 0.003) between all combinations of two of the three factors found to significantly influence tone individually. 

The experiments to test the effect of the presence of the substrate to the carbonic anhydrase enzyme are shown in Figure 2and in Table 3 . Eliminating substrate to the carbonic anhydrase enzyme resulted in no significant change in EC₅₀ (P = 0.65), whereas the MR was significantly reduced (P = 0.0031). Subsequent two-way ANOVA showed that this difference was due to a significant reduction in the MR for dorzolamide (P < 0.0001) but not for acetazolamide (P = 0.0542). 

The effect of acidosis (MR: P = 0.008; EC₅₀: P = 0.027) and perivascular tissue (MR: P = 0.0003; EC₅₀: P < 0.0001) on vasodilation induced by nifedipine was found to be similar to the effect on vasodilation induced by the CAIs (two-way ANOVA for all comparisons). 

There was no significant difference between the repeated vasodilation responses obtained during the time control experiments (P = 0.49). The DMSO/ethanol solvent induced a significant (P < 0.0001) contraction of the vessels of 4.7% (4.3% to 5.1%; mean, range) over time in the isolated retinal arterioles, whereas this effect was not observed in the presence of perivascular retinal tissue (P = 0.24). 

Discussion

The regulation of retinal blood flow secondary to changes in retinal metabolism can be studied by measuring diameter changes of retinal arterioles on the basis of images recorded through the optics of the eye in vivo. ^{14 15} However, the observed response of the arterioles to specific interventions on retinal metabolism may be hampered by counterregulatory mechanisms that ensure the metabolic homeostasis under normal conditions, ⁴ and in in vitro experiments, even minimal amounts of perivascular retinal tissue may influence the tone produced in the arteriolar walls. ^{16 17} Therefore, the study of metabolic autoregulation in the retina can benefit from reduced experimental models where the influence of surrounding counterregulatory mechanisms can be controlled. Previous studies have shown a vasodilating effect of dorzolamide on isolated bovine retinal vessels, ¹ but this response may be specific for the bovine species. Thus, in other studies of porcine vessels the vasodilating effect of CAIs depended on the presence of the perivascular retinal tissue, and this effect varied for different CAIs. ¹² However, no study in any species has determined whether the vasodilating effect is due to inhibition of the carbonic anhydrase enzyme or to another effect of CAIs on retinal vascular tone. 

The results of the present study confirmed those of previous studies that dorzolamide has a lower EC₅₀ and produces an MR at a lower concentration than acetazolamide. ¹² This difference was not found to be significant for EC₅₀, probably because these values were truncated at the highest concentration of acetazolamide used when the response at this concentration was too small for an EC₅₀ to be calculated. The difference was observed at all the experimental conditions, supporting the notion that the two CAIs share the same mechanism of action, the difference being due to differences in the physical–chemical properties or the binding to the target site. ^{18 19} The influence of the perivascular retinal tissue on CAI-induced vasodilation has not been elucidated in detail, but the significance of specific metabolic pathways, such as adenosine-induced vasodilation stimulated by the release of glutamate in the perivascular tissue, has been confirmed in recent studies. ^{16 17}  

Acidosis resulted in a reduction in EC₅₀ and an increase in MR in all experimental conditions. This effect was independent of the vasorelaxing effect of the tested CAIs, since acidosis induced a similar increase in the sensitivity to dilation induced by nifedipine, which has a different mechanism of action. ²⁰ It has been shown that acidosis induces hyperpolarization of the smooth muscle cell membrane through an opening of potassium channels, but the protons may also close voltage-dependent calcium channels by a direct effect. Both of these effects lead to a reduction of [Ca²⁺]_i and to vasorelaxation. ²¹ However, other studies suggest that endothelially derived or extravascular NO may also be involved in the mechanism. ^{9 22} The contribution of these different factors to acidosis-induced vasorelaxation in retinal arterioles remains to be studied. 

Finally, the experiments showed that the influence of acidosis on the vasodilating effect of CAIs can be reproduced with another vasodilating compound, and therefore the effect is probably not linked to the effect of the carbonic anhydrase enzyme. This interpretation is confirmed by the observation that hypercapnia did not affect CAI-induced vasorelaxation, implying that the conversion of CO₂ to H⁺ and HCO₃ ⁻ had no influence on vascular tone and that elimination of the substrates CO₂ and HCO₃ ⁻ in the fluid did not affect the vasodilating effect of dorzolamide and only affected the MR of acetazolamide. The findings therefore strongly suggest that mechanisms other than carbonic anhydrase inhibition are involved in CAI-induced vasodilation, although a contributing action of the enzyme cannot be excluded. A possible explanation of the finding is that CAIs may have a dual action with inhibition of the carbonic anhydrase at low concentrations ^{11 23} and a vasodilating effect at high concentrations as previously found in clinical and experimental studies in vitro ^{1 3 12} and in vivo. ^{2 4 5 24} The findings are confirmed by recent studies showing that acetazolamide-induced vasodilation of pulmonary vessels during hypoxia is independent of carbonic anhydrase inhibition. ^{23 25} Several alternative mechanisms of action of CAIs in high concentrations have been suggested, including an effect through voltage-gated K_Ca channels, ²⁴ and an effect mediated by NO. ^{6 26} However, the existence of similar effects in the retinal circulation remains speculative, and further studies are needed to elucidate the influence of these and other mechanisms on CAI-induced vasodilation. 

In conclusion, the experiments have shown that mechanisms other than inhibition of the carbonic anhydrase enzyme are involved in the vasodilating effect of CAIs. The vasodilation induced by these compounds is increased by the perivascular retinal tissue and by acidosis through effects unrelated to the carbonic anhydrase enzyme. CAIs may have several mechanisms of action of significance for the beneficial actions of these drugs in ocular disease. 

 Supported by The Danish Medical Research Council and the Velux Foundation.

 Submitted for publication June 14, 2008; revised July 25, 2008; accepted October 27, 2008.

 Disclosure: M.S. Torring, MSD Denmark (F); K. Holmgaard, None; A. Hessellund, None; C. Aalkjaer, None; T. Be k, None

 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.

 

The effect on tone at different concentrations of CAIs. Blue curves: arterioles with perivascular retinal tissue; red curves: isolated arterioles; solid lines: normal pH; dotted lines: acidosis. (A, C) Acetazolamide; (B, D) dorzolamide; (A, B) normocapnia; (C, D) hypercapnia.

Figure 3.

 

EC₅₀ for the Curves Shown in Figure 1

Figure 4.

 

Maximum Relaxation for the Curves Shown in Figure 1

Figure 2.

 

The effect on tone at different concentrations of CAIs in the absence and presence of CO₂ and HCO₃ ⁻. Blue curves: arterioles with perivascular retinal tissue; red curves: isolated arterioles; solid lines: normal buffer; dotted lines: nominal absence of CO₂ and HCO₃ ⁻.

Figure 5.

 

EC₅₀ for the Curves Shown in Figure 2

The authors thank technician Poul Rostgaard and associate professor Mogens Erlandsen (Department of Biostatistics, University of Aarhus) for assistance in the study.