INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

IN RECENT YEARS it has become clear that the cerebral vasodilation to hypercapnic acidosis is partly dependent on a nitric oxide (NO)-dependent mechanism and partly due to a direct effect of the hypercapnia on the smooth muscle (4). The NO-dependent component is seen in some species in vivo where arginine analogs reduce the vasodilation induced by hypercapnic acidosis. It has been suggested that NO derived from the neuronal nitric oxide synthase (NOS) has a permissive effect for the vasodilator response to hypercapnic acidosis, whereas NO from the endothelium seems to play no role (4). It is, however, interesting that in mice where neuronal NOS is knocked out, the response to hypercapnic acidosis is as prominent as in wild-type mice and not affected by inhibition of NOS (9). This indicates that also NO-independent mechanisms may be important under in vivo conditions. Aalkjær and Peng (3) recently reported that the in vitro effect of hypercapnic acidosis on the concentration-response curve to vasopressin cannot be ascribed to NO or an endothelium-dependent mechanism, thereby substantiating previous reports based on in vitro experiments which suggested that the endothelium plays little role for the response to hypercapnia (26, 28). There is thus excellent agreement between the in vivo and in vitro findings to suggest that a direct effect of hypercapnic acidosis on the vascular smooth muscle plays a role for the relaxation. The mechanism responsible for this relaxant effect of hypercapnic acidosis on the smooth muscle cells is not known.

Hypercapnic acidosis has, however, been associated with hyperpolarization of smooth muscle cells in cerebral small arteries (7, 22), and it has therefore been suggested that hyperpolarization is responsible for this. On the other hand, patch-clamp experiments have indicated that, for a given membrane potential, reduction of pH reduces the smooth muscle calcium current through voltage-operated calcium channels (8, 13), and it has been proposed that this calcium channel-blocking effect could be responsible for the vasodilation. In either case, however, a reduction of intracellular Ca²⁺ concentration ([Ca²⁺]_i) is expected. Whether this occurs is not known, although recently it was shown (5) that in rat mesenteric arteries the acute inhibition of a potassium-induced contraction by hypercapnic acidosis could be explained by a reduction of [Ca²⁺]_i. With respect to normocapnic acidosis, no hyperpolarization may be seen (7) in cerebral arteries, and the mechanism responsible for the dilation to normocapnic acidosis is completely unknown.

The aim of this study was therefore to test the hypothesis that the direct effect of both hypercapnic and normocapnic acidosis on the smooth muscle in steady state can be explained by a reduction of [Ca²⁺]_i mediated by a hyperpolarization.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Preparation

Measurements of Force and [Ca²⁺]_i or pH_i

For measurements of [Ca²⁺]_i, the artery was excited with light from a 75-W xenon lamp, which alternatingly passed through either a 343-nm band-pass filter or a 380-nm band-pass filter (bandwidth 5 nm). The emission from the preparation was collected at 10 Hz as epifluorescence through a 500- to 530-nm band-pass filter and a <720 nm cut-off filter. The emission signals, when exciting with 343 and 380 nm light, respectively, were stored on a computer during experiments together with force measurements. After the experimental protocol was completed, the signals were calibrated as previously described (11). In preliminary experiments, we tested the effect of the experimental protocol on arteries not loaded with fura 2 and found that the background fluorescence was unaffected by the experimental protocol.

For measurement of pH_i, the artery was excited with light from a 75-W xenon lamp using 436- and 490-nm band-pass filters (bandwidth 5 nm). The emission from the preparation was collected through a 530- to 585-nm band-pass filter and a <720-nm cut-off filter every 10 s, and the ratio of emission at the two excitation wavelengths was calculated after subtraction of background fluorescence. The ratio was calibrated with nigericin as described previously (1, 24). In the BCECF experiments the background was <1% of the signal.

Determination of pH dependence of dissociation constant of fura 2. The dissociation constant (K_D) for binding Ca²⁺ to fura 2 is probably dependent on pH (6, 14). To assess this in greater detail, the K_D for the intracellularly trapped fura 2 was determined at different pH_i values. In these experiments a second- or third-order branch of the superior mesenteric artery (internal diameter ~200 µm) was used because we have previously established a protocol for determination of K_D in mesenteric small arteries (12) exposed to ionomycin and nigericin; for each vessel the ratio of fluorescence at the two excitation wavelengths was determined at zero Ca²⁺ (R_min), at 10 mM Ca²⁺ (R_max), and at an intermediate calcium concentration. This protocol was repeated at three different pH values. In addition, the K_D of fura 2 was determined in the test tube at different pH values. Figure 1 shows that the intracellular K_D determined at three pH values were close to the curve describing the pH dependence of K_D in the test tube, and on this background we used the equation that best fitted all the points in Fig. 1 to compensate for the effect of pH on K_D. In RESULTS (where indicated), the fura 2 fluorescence signals are calibrated in terms of [Ca²⁺]_i using the equation for the relationship between pH and K_D shown in the legend to Fig. 1 and using the mean pH_i values from the experiments where pH_i was measured. Another potential problem is a pH dependence of the other parameters (R_max, R_min, and ) in the calibration equation (6). Ziegelstein et al. (29) citing a personal communication (Robert J. Gillies) mentioned that R_max of fura 2 was affected by a low pH. However, in the present study R_min, R_max, and  (the ratio of fluorescence with excitation at 380 nm with zero Ca²⁺ and 10 mM Ca²⁺) were not affected by pH in the range from 6.6 to 7.3 (data not shown).

View larger version (9K): [in this window] [in a new window]   Fig. 1.   Relationship between dissociation constant (KD) of fura 2 and pH. Circles, measurements of fura 2 in smooth muscle cells; squares, measurement of KD in test tube. Vertical bars show SE where this exceeds the size of the symbols and n was 5-11 vessels. Equation of curve fitted to data points was KD = 13,343  3,462pH + 227pH2.

Measurement of Force and Membrane Potential

Protocol

4

4

View larger version (7K): [in this window] [in a new window]   View larger version (7K): [in this window] [in a new window]   Fig. 2.   Effect of changes of extracellular pH (pHe) on intracellular pH (pHi). Graphs show pHi as a function of pHe for hypercapnic (A; pH reduced by increasing CO2 from 5 to 10% and 20%, respectively) and normocapnic (B; pH reduced by decreasing from 25 to 15 mM and 8 mM, respectively) acidosis as indicated. Open symbols, in presence of 104 M NG-nitro-L-arginine (L-NNA); closed symbols, in presence of 104 M L-NNA and 0.5 U/l AVP. Bars indicate ±SE, n = 4-9 vessels.

View larger version (16K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   Fig. 3.   Traces showing measurements of force and intracellular Ca2+ concentration ([Ca2+]i) (A) and force and membrane potential (B and C). Arginine vasopressin (AVP) was present as indicated, and 104 M L-NNA was present from arrow and for rest of experiment. Control solution had a pH of 7.45, 5% CO2, and a [] of 25 mM. Where indicated pH was reduced to ~7.0 by reducing [] or by increasing CO2. A: during relaxation after first stimulation with AVP, [Ca2+]i and tension fell to a low level and then suddenly increased to a higher, more steady level. This was seen occasionally when AVP was washed out. B: L-NNA causes depolarization and force development. C: normocapnic acidosis (15 mM ) is associated with depolarization and relaxation.

Solutions and Chemicals

Other chemicals used were fura 2-AM, BCECF-AM, and pluronic F-127 from Molecular Probes. Nigericin, ionomycin, cremophor, and L-NNA were obtained from Sigma Chemical. AVP was purchased from Sandoz.

Calculations and Statistics

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Effect of Hypercapnic and Normocapnic Acidosis on [Ca²⁺]_i and Force

4

4

4

View larger version (35K): [in this window] [in a new window]   Fig. 4.   Effect of hypercapnic (HA1 and HA2) and normocapnic acidosis (NA1 and NA2) on [Ca2+]i and tension. Effect of hypercapnic and normocapnic acidosis on tension (top) and [Ca2+]i (bottom). In all cases, effect of acidosis was significantly different from zero for both tension and [Ca2+]i. In control solution, tension was ~1.5 and 2.8 N/m for L-NNA-treated and L-NNA + AVP-treated arteries, respectively, [Ca2+]i was ~325 and 430 nM for L-NNA-treated and L-NNA + AVP-treated arteries, respectively. Open bars, moderate acidosis (7.2), hatched bars, severe acidosis (7.0). Error bars indicate ±SE; n = 6 vessels for moderate acidosis and n = 8 vessels for severe acidosis.

The reduction of tension during acidosis was associated with a reduced [Ca²⁺]_i both with normocapnic acidosis and with hypercapnic acidosis (Fig. 3A and 4) and both with moderate (solutions HA₁ and NA₁) and severe (solutions HA₂ and NA₂) acidosis, irrespective of whether the vessel was activated with AVP. However, unexpectedly the further reduction of tension seen with severe acidosis relative to moderate acidosis was not associated with a further reduction of [Ca²⁺]_i. On the other hand, if the values for [Ca²⁺]_i were computed without correction of K_D for the effect of pH_i, the uncorrected decrease in [Ca²⁺] with severe acidification was significantly greater than the decrease in [Ca²⁺]_i with moderate acidosis (unpaired t-test).

To evaluate whether the responsiveness to [Ca²⁺]_i was indeed affected by acidosis, calcium concentration-response curves were made in AVP-activated arteries. Figure 5 shows the results of these experiments. It can be seen that normocapnic acidosis reduced the responsiveness to [Ca²⁺]_i. The tension corresponding to 300 nM [Ca²⁺]_i was 1.67 ± 0.25 and 1.20 ± 0.26 N/m for control solution and solution NA₁, respectively (P < 0.05).

View larger version (10K): [in this window] [in a new window]   Fig. 5.   Effect of normocapnic acidosis on relationship between tension and [Ca2+]i. After extracellular and intracellular calcium stores were depleted, arteries were exposed to 2 U/l AVP in cumulatively increasing concentrations of extracellular calcium (0.05, 0.1, 0.2, 0.4, 0.8, 1.6, 2.5 mM). Each symbol represents mean of 5 experiments and bars indicate ±SE. Open symbols, control solution; filled symbols, moderate normocapnic acidosis.

Effect of Hypercapnic and Normocapnic Acidosis on Membrane Potential and Force

4

4

Table 1 shows that moderate hypercapnic acidosis either had no significant effect on membrane potential or was associated with a small hyperpolarization, whereas strong hypercapnic acidosis caused hyperpolarization irrespective of whether L-NNA was present or the vessel was activated with AVP. Normocapnic acidosis on the other hand depolarized resting arteries and arteries exposed to L-NNA despite reduced force development, whereas in the presence of AVP, the relaxation was not associated with any significant change in membrane potential.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The main conclusions from this study are that the relaxation of cerebral small arteries to hypercapnic acidosis and normocapnic acidosis is explained in part by a reduction of [Ca²⁺ ]_i. However, whereas hyperpolarization may contribute to this reduction during hypercapnic acidosis, a membrane potential-independent mechanism seems to play a major role in the reduction of [Ca²⁺]_i. In addition, it seems likely that acidosis reduces the responsiveness of the contractile machinery to [Ca²⁺]_i. Before discussing this in detail, we will discuss a few other points that have become apparent from this study.

The values for resting [Ca²⁺]_i were higher than what is often reported for resting [Ca²⁺]_i in other vascular smooth muscle. However, we have previously compared the [Ca²⁺]_i values in mesenteric and cerebral small arteries (using the same K_D in the two vascular beds) and found that [Ca²⁺]_i in unstimulated cerebral arteries is unstable and higher than that in the mesenteric vessels. The membrane is also more depolarized in cerebral small arteries (in the present study the resting membrane potential was about 50 mV) compared with mesenteric small arteries [membrane potential about 60 mV under similar conditions (10, 16)], and the cerebral small arteries have tone under nonstimulated conditions in contrast to the mesenteric arteries. Although we cannot entirely exclude that the K_D we use is inappropriate for the cerebral arteries, we would expect the [Ca²⁺]_i to be higher in the cerebral small arteries than, for example, in mesenteric small arteries. Another point pertaining to the calibration of the fura 2 signal is the effect of pH_i on the K_D. It seemed from our measurements that the relationship between pH_i and K_D was similar in the extracellular and intracellular environment, and we used the curve that best fitted all points to correct K_D. It is clear, however, that this correction is associated with substantial statistical error. First, a certain error is associated with fitting the curve relating pH and K_D. Second, there is an error associated with the determination of the pH_i value used for the correction, which includes an error in the calibration of the BCECF signal. In addition it should be pointed out that determination of the pH_i dependence of K_D was made in mesenteric small arteries, but the relationship was used in cerebral small arteries.

Another observation that was not central to the main topic of the study was the marked depolarization, increase of [Ca²⁺]_i, and force development when L-NNA was applied to the unstimulated vessels. An NO-mediated hyperpolarization has been suggested by Tare et al. (23). A hyperpolarizing effect that can be ascribed to NO has not been a consistent finding, and its importance is still debated (19). However, in the guinea pig coronary artery (18) and in the rabbit mesenteric artery (17), L-NNA did cause depolarization consistent with inhibition of a basal release of hyperpolarizing NO. In the cerebral circulation the evidence is less conclusive, since in the basilar artery of the rabbit, NO causes no hyperpolarization (19), although arginine analogs do reduce the hyperpolarizing response to acetylcholine (20). The 10- to 20-mV depolarization seen in the present study when L-NNA was applied under resting conditions and the associated ~40 nM increase of [Ca²⁺]_i and of tension strongly suggest that in rat middle cerebral arteries, a basal release of NO reduces tension through hyperpolarization and consequent reduction of [Ca²⁺]_i, although this does not exclude a calcium-independent effect of NO on the tone in these vessels.

The main purpose of this study was to assess the role of the membrane potential and [Ca²⁺]_i for the vasodilation to acidosis. Given the substantial influence of NO on membrane potential and [Ca²⁺]_i in this preparation as discussed above, the first consideration is whether NO or perhaps another endothelium-dependent substance is important for the response. This issue is, however, settled as outlined in the introduction, and there is agreement in the literature that under the conditions of these experiments, there is no contribution from NO or the endothelium to the relaxation to acidosis. Because we were interested in the effect of acidosis on basal myogenic tension, most of the experiments were made in the presence of L-NNA, where the myogenic tension is greater and it is easier to demonstrate effects of acidosis.

The relaxant effect of both hypercapnic and normocapnic acidosis on the smooth muscle was shown to be associated with a reduction of [Ca²⁺]_i, with respect to both the myogenic tension and the AVP-induced tension and irrespective of whether the [Ca²⁺]_i values were computed taking into account the effect of pH on K_D. From Fig. 5 it can be seen that a reduction of 30-50 nM [Ca²⁺]_i from about 300-400 nM is likely to cause a reduction of tension. Recently (5), the acute effect of hypercapnic acidosis in potassium-activated rat mesenteric small arteries was shown to be associated with a reduction of [Ca²⁺]_i. There is thus good evidence that a reduction of [Ca²⁺]_i explains in part the reduction in tension to acidosis.

However, we also provided evidence that the responsiveness to [Ca²⁺]_i plays a role for the effect of acidosis. First, the additional reduction in tension with severe acidosis appeared not to be associated with a further reduction of [Ca²⁺]_i (Fig. 4). As discussed above, the calibration of the fura 2 signal is difficult and critically dependent on corrections of K_D for pH as pointed out in RESULTS. However, the changes in [Ca²⁺]_i with severe acidosis are consistent with the corresponding changes in membrane potential. Furthermore, direct evidence that a reduction of the responsiveness of the contractile apparatus to [Ca²⁺]_i plays a role comes from the calcium concentration-response experiments. In these experiments we used only moderate acidosis where the correction of K_D for pH is minimal and not important for the conclusion. These experiments show that a reduced responsiveness to [Ca²⁺]_i can play a role for the response to acidosis and are consistent with the conclusions from Fig. 4.

Because voltage-sensitive calcium channels play an important role for control of [Ca²⁺]_i in cerebral arteries (18), these channels are likely to be involved in the response to acidosis. There is evidence based on patch-clamp experiments that the opening probability at a constant membrane potential of the L-type calcium channel in vascular smooth muscle is reduced at an acid pH (13, 27). Furthermore, hypercapnic acidosis causes hyperpolarization in resting cerebral arteries (7, 21), whereas normocapnic acidosis had no effect in one study (7). This means that both a membrane potential-dependent and a membrane potential-independent mechanism could be important for the relaxation to acidosis, and we have addressed this question.

To compare our data with the data in the literature (7, 21), we first measured membrane potential in arteries not exposed to L-NNA or AVP. Severe hypercapnic acidosis was associated with hyperpolarization, which is consistent with previous reports (7, 21), although with moderate hypercapnic acidosis, we found no significant hyperpolarization. Furthermore, we found a small depolarization to severe normocapnic acidosis, which is not inconsistent with a previous study (7), where a small depolarization was seen, which, however, did not obtain statistical significance. In that study (7), it was suggested that the different effect on membrane potential of hypercapnic and normocapnic acidosis might relate to a selective intracellular acidification with the hypercapnic acidosis. In the present study, however, we showed that both normocapnic and hypercapnic acidosis caused intracellular acidosis, and a different explanation for this must therefore be sought. Because both normocapnic and hypercapnic acidosis reduced pH_i, it would seem most likely that the different CO₂ and/or concentrations in the two situations may in some way be responsible for the different effects on membrane potential.

The mechanism responsible for the relaxation to acidosis may depend on the mode of preactivation (e.g., in potassium-activated vessels vasodilation mediated through membrane hyperpolarization may be difficult to evaluate). To assess whether the proposed mechanism may be relevant under different conditions, we repeated the experiments in both arteries treated with L-NNA and arteries exposed to AVP. Under these conditions the pattern was the same: hyperpolarization with hypercapnic acidosis and depolarization with normocapnic acidosis, although the changes in membrane potential did not obtain statistical significance in all situations. For example, with moderate acidosis (solution HA₁) no significant hyperpolarization was seen in L-NNA-treated arteries despite reduction of [Ca²⁺]_i and tension. It is therefore likely that mechanisms in addition to the hyperpolarizing effect may be important for the reduction of [Ca²⁺]_i. The depolarization seen with normocapnic acidosis was most pronounced with the severe acidosis and was not apparent when the vessels were activated with AVP. This suggests that normocapnic acidosis may reduce [Ca²⁺]_i and hence tension through a mechanism that is partly different from that utilized by hypercapnic acidosis. As discussed above, acidosis reduces the calcium current through L-type calcium channels at constant membrane potential, so a probable mechanism for the reduction of [Ca²⁺]_i and tension seen with the normocapnic acidosis is a calcium channel-blocking effect, which also very likely contributes to the relaxation during hypercapnic acidosis.

In conclusion, we have demonstrated that the mechanism(s) responsible for the relaxant effect of both normocapnic and hypercapnic acidosis on smooth muscle tone involves a reduction of [Ca²⁺]_i and possibly also a reduction of the responsiveness to [Ca²⁺]_i. Furthermore, during hypercapnic acidosis the reduction of [Ca²⁺]_i can in part be explained by membrane hyperpolarization. In contrast, during normocapnic acidosis [Ca²⁺]_i is reduced despite a significant depolarization, suggesting that normocapnic acidosis reduces [Ca²⁺]_i and hence force through a voltage-independent mechanism.