OUP user menu

Different pathways with distinct properties conduct dilations in the microcirculation in vivo

Cor de Wit
DOI: http://dx.doi.org/10.1093/cvr/cvp340 604-613 First published online: 10 October 2009


Aims Conduction of vasomotor signals along the vessel coordinates the behaviour of vascular cells and is attributed to the spread of hyperpolarizations through gap junctions. Intriguingly, conducted dilations encompass larger distances than can be expected by passive electrotonic spread. Because distances are quite distinct for different dilators, we hypothesized that separate pathways with distinct properties are involved.

Methods and results We characterized local and conducted responses elicited by acetylcholine (ACh) and adenosine (Ado) in the murine microcirculation in vivo. Local (and remote) ACh dilations were nearly abrogated by blockade of KCa channels (charybdotoxin), but dilations to Ado were abolished by the KATP blocker glibenclamide. Bupivacaine, a blocker of Na+ and K+ channels, and similarly the blockade of inwardly rectifying K+ channels (barium) revealed different conduction mechanisms, as the remote dilation to Ado, but not ACh, was abrogated. Surprisingly, expression of connexin37 (Cx37) was not detected in Cx40-deficient arterioles, although abundantly expressed in endothelium of wild-type arterioles. In contrast to the wild-type mice, the amplitude of conducted ACh and Ado dilations decreased similarly with distance in Cx40-deficient mice. Recordings of membrane potential in vivo showed endothelial hyperpolarization by ∼10 mV in response to ACh, whereas Ado did not alter endothelial membrane potential.

Conclusion Distinct pathways conduct responses along the vessel wall which involve dissimilar K+ channels and connexins in initiation and spreading. Most likely, the endothelium is the preferential conduction pathway activated by ACh, whereas in the case of Ado the smooth muscle serves as the signalling pathway. However, in arterioles nearly devoid of Cx40 and Cx37, ACh responses can likewise be conducted along the smooth muscle.

  • Ca2+-dependent K+ channels
  • Conducted responses
  • Inwardly rectifying K+ channels
  • Connexin40
  • Connexin37

1. Introduction

The coordination of the behaviour of vascular cells is a prerequisite for vascular function, i.e. to accomplish the task of matching blood supply to tissue needs. Diameter changes in the microcirculation that remain constrained to the vicinity of the stimulus have only limited impact on blood flow due to upstream resistance segments. However, if locally initiated vasomotor signals are communicated upstream along the length of the vessel (‘ascend’) and thereby coordinate diameter changes a wide range of flow in conjunction with a precise control of tissue perfusion can be achieved. Different mechanisms have been suggested to integrate vessel behaviour of which direct cell–cell coupling has recently attracted major attention. Vascular cells are interconnected by gap junctions consisting of connexins which permit ions to flow down an electrochemical driving force and second messengers to diffuse along a gradient.1 Specifically, ion flow and electrotonic conduction provide a fast signal transmission mechanism to change arteriolar diameter quickly along a certain length of the vessel.2,3

In fact, discrete stimulation of arterioles in the microcirculation elicited changes of membrane potential not only at the application site but also at distant sites which were associated with dilation or constriction corresponding to the diameter changes at the stimulation site.46 These so-called conducted responses are mostly studied using acetylcholine (ACh) due to its potency to elicit them and the pronounced distance that the dilation spreads along the vascular wall.4,5,7,8 Its potency is most likely related to the strong endothelial hyperpolarization elicited by ACh which is brought about through the activation of Ca2+-dependent K+ channels, specifically KCa3.1 in the murine microcirculation.9,10 The strong endothelial hyperpolarization, the length, and the orientation of individual endothelial cells as well as their abundant expression of gap junction forming connexins, suggest that it is indeed the endothelium which transmits the signal along the vessel. Of the large family of connexins, endothelial cells express mainly connexin40 (Cx40) and connexin37 (Cx37) in arterioles.11,12 Interestingly, the lack of Cx40 attenuates the conduction of ACh-induced dilations in mice7,13 and the fairly selective destruction of the endothelium by a light-dye technique likewise impaired the conduction highlighting the importance of the endothelial cell layer as a signalling pathway if responses are initiated by ACh.11

However, smooth muscle cells also express connexins and form an additional cell-to-cell signalling pathway.12,14 This pathway is implicated in the transmission of vasoconstrictions locally initiated by receptor stimulation (norepinephrine) or depolarization by focal application of potassium solution.15 Although less well defined with respect to which connexins form the low-resistance conductance channels therein the localized destruction of the smooth muscle cell layer along the pathway abrogates the conduction of vasoconstrictions.15 Additionally, the smooth muscle layer provides an alternative pathway to conduct ACh dilations in the hamster microcirculation demonstrating that dilations can likewise be transmitted,5,15 although this was not observed in larger feed arteries.4

The distances which are encompassed by conducted dilations are notably large, suggesting that amplification mechanisms regenerate the conducting signal and KCa3.1 may contribute herein.10 Intriguingly, the distances as well as connexin-dependency have very recently been shown to differ for different vasodilators.13,16 Therefore, we hypothesized that different dilators initiate conducted responses by distinct mechanisms and that dilator signals are conducted along the endothelial and/or the smooth muscle cell layer which is possibly reflected by distinct properties. To test these hypotheses, we evaluated and compared the properties of conducted responses elicited by ACh and adenosine (Ado) with respect to initiation, amplification, and connexin-dependency in the murine microcirculation using wild-type (wt) and Cx40-deficient mice. Additionally, we performed membrane potential measurements to identify the vascular cell which is initially affected by these vasodilators and studied the expression of endothelial Cx37 in these mice in complementary experiments.

2. Methods

The investigation conformed with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No.85-23, revised 1996) and was approved by local authorities (Landwirtschafts- und Umweltministerium Schleswig-Holstein, V312-72241.122-2). Experiments were performed in wt and Cx40-deficient mice17 in a C57/BL6 genetic background. Intravital microscopy,18 membrane potential measurements,19 and immunohistochemistry16 were performed in the cremaster muscle as described (see Supplementary material online).

2.1 Experimental protocols

In each animal one or two second-order arterioles were studied. Conducted vasomotor responses were initiated by locally confined stimulation with ACh (1 mmol/L), Ado (10 mmol/L), cromakalim (1 mmol/L), or KCl (3 mol/L) delivered through glass micropipettes positioned in close proximity of the arterioles by a pressure pulse (140 kPa, 50–300 ms). If a response at the application site (local) was obtained, the same stimulus was reapplied and upstream, remote sites at distances up to 1100 µm were recorded. Each substance was studied in duplicate in a single vessel and their mean taken as single observation. In a number of experiments, this protocol was repeated after addition of an ion channel blocker to the superfusate: charybdotoxin (ChTx, 1 µmol/L) was applied to block Ca2+-dependent K+ channels, glibenclamide (1 µmol/L) to block ATP-dependent K+ channels, bupivacaine (100 µmol/L) for non-specific Na+- and K+-channel blockade, mepivacaine (100 µmol/L) to solely block Na+ channels, and barium (75 µmol/L) to block inwardly rectifying K+ channels (KIR). Except for ChTx (applied for 20 min before repeating the protocol) blockers were present throughout the experiment. All substances were dissolved in water except for glibenclamide (0.5% DMSO) and cromakalim (50% DMSO). Experiments were performed in the presence of a blocker of NO synthase (30 µmol/L N-nitro-l-arginine, l-NA) and cyclooxygenase (3 µmol/L indomethacin) which were added to the superfusate 30 min before the experimental protocol was started. At the end of each experiment, arteriolar maximal diameters were determined by simultaneous superfusion of sodium-nitroprusside, ACh, and Ado (each 30 µmol/L) before the animal was killed with pentobarbital.

2.2 Statistics and calculations

Vascular tone is given as resting divided by maximal diameter. Internal diameters were measured and changes normalized to: % of maximal response = (DTrDCo)/(DMaxDCo) × 100, where DTr is the diameter after treatment, DCo the control diameter, and DMax the maximal possible diameter that is the maximally dilated diameter for dilations or the minimal luminal diameter (zero) for constrictions. The temporal characteristic of the responses was considered by calculation of the time from stimulus application to attainment of peak diameter (time-to-peak) and from stimulation to reattainment of baseline diameter (response duration). Comparisons within groups were performed using paired t-tests, and, for multiple comparisons, probability values were corrected according to Bonferroni. Data between groups were compared with ANOVA followed by post hoc analysis of the means. Differences were considered significant at a corrected error probability of P < 0.05. Data are presented as mean ± SEM.

3. Results

3.1 Basal data

A total of 54 arterioles with maximal diameters between 24 and 55 µm (mean: 37 ± 1) were studied in 36 wt mice. All preparations used to study conducted dilations were treated with l-NA (30 µmol/L) and indomethacin (3 µmol/L) in the superfusate to eliminate effects of NO and prostaglandins. Under these conditions, the arterioles exhibited a varying degree of tone ranging from 0.13 to 0.88 (mean: 0.40 ± 0.03). However, resting tone and maxima were not different between groups before application of ion channel blockers (Table 1). Glibenclamide, bupivacaine, mepivacaine as well as barium increased resting tone, i.e. decreased diameter, in l-NA and indomethacin treated vessels (Table 1). ChTx did not change diameter in a limited number of experiments. In one group of animals, conduction upon local K+-depolarization was studied in which bupivacaine was applied without inhibition of NO synthase and cyclooxygenase. Bupivacaine decreased diameters significantly also in these untreated preparations (Table 1).

View this table:
Table 1

Arteriolar resting and maximal diameters and effect of ion channel blockers

3.2 Conducted dilations upon stimulation with ACh and Ado

Brief local pressure ejection of ACh (130 ± 10 ms) or Ado (250 ± 20 ms) induced transient dilations with similar maximal amplitudes at the application site (Figure 1). This maximum was attained within 10 ± 1 s after ACh stimulation and lasted for 40 ± 3 s. Upon Ado stimulation, the maximum was attained more slowly (15 ± 1 s, P < 0.001 vs. ACh) and the response duration was prolonged 51 ± 4 s (P < 0.05 vs. ACh). Responses initiated by both stimuli conducted rapidly to distant, upstream sites. At 1100 µm the peak amplitude was reached without significant lag (ACh: 10 ± 1 s, Ado: 15 ± 1 s). However, the duration of the response was shorter at 1100 µm in both cases (ACh: 30 ± 3 s, Ado: 36 ± 3 s, both P < 0.01 vs. local). Although the amplitude of the response initiated by ACh did not decrease with distance, it decayed significantly after stimulation with Ado (Figure 1) suggesting that different mechanisms are involved.

Figure 1

Conducted responses upon stimulation with ACh and adenosine. Time course of dilations initiated by focal brief pressure ejection of ACh or adenosine at the local (A) and distant, upstream sites (B: 550 µm, C: 1100 µm). The maximal amplitude of the response is depicted over distance in (D). Locally induced transient dilations conducted to distant sites without measurable delay for both stimuli. In contrast to ACh, the dilation induced by adenosine decayed with distance although the local amplitude was similar. Dilation is expressed as % of maximal response, vertical dashed lines indicate agonist application, presence of l-NA and indomethacin. Thirty arterioles in 28 mice, **P < 0.01 vs. ACh (ANOVA); ##P < 0.01 and ###P < 0.001 vs. local site (t-test).

3.3 Initiation of the conducted responses

We have previously demonstrated that local and conducted dilations upon ACh and Ado are mostly independent of NO synthase and cyclooxygenase.16,18 Therefore, the contribution of NO and prostaglandins was not evaluated and all preparations were treated with l-NA and indomethacin. Moreover, ACh-induced dilations have been shown to depend strongly on the activation of Ca2+-dependent K+ channels (KCa) in the murine microcirculation.10,20,21 ChTx (1 µmol/L), a blocker of KCa3.1 (IK1) and KCa1.1 (BK), was used to determine their role in conducted dilations. As expected, ChTx strongly impaired ACh dilations at local and distant sites (Figure 2A, C, E). In contrast, local and conducted dilations initiated by Ado were not attenuated and conducted dilations tended to be enhanced, although the effect was not significant. Interestingly, local dilations in response to Ado were nearly abolished by blockade of ATP-dependent K+ channels (KATP; glibenclamide, 1 µmol/L). Consequently, conducted responses were not observed anymore (Figure 2B, D, F), suggesting that the activation of KATP channels is able to initiate a conducted dilation. This was further verified by local application of an activator of KATP channels (cromakalim). Local pressure ejection of cromakalim elicited a dilation which also conducted rapidly with a decaying amplitude along the vessel (Figure 2G). These responses were completely blocked by glibenclamide. Microapplication of the solvent of cromakalim (50% DMSO) was without effect on diameter (local: 0.3 ± 3%). In contrast, glibenclamide neither attenuated local nor conducted dilations induced by ACh (Figure 2).

Figure 2

Different K+ channels initiate responses upon stimulation using ACh and adenosine. Time course of dilations in response to ACh and adenosine (Ado) at the local site (A and B) and 1100 µm upstream (C and D) at resting conditions and after inhibition of KCa channels by charybdotoxin (ChTx, 1 µmol/L) or KATP channels by glibenclamide (Glib, 1 µmol/L). Peak amplitude of the dilation is depicted over distance for ACh (E), adenosine (F), and cromakalim (G). ChTx strongly impaired local and conducted dilations to ACh, but not to adenosine. In contrast, adenosine dilations were abrogated by glibenclamide. Focal application of cromakalim elicited conducted dilations which were blocked by glibenclamide. In cases without an identifiable maximum, diameter was averaged for 15 s after stimulation and taken instead. Dilation in % of maximal response, agonist application: vertical dashed lines, presence of l-NA and indomethacin. ChTx n = 3, glibenclamide n = 5, and cromakalim n = 4 arterioles and mice, **P < 0.01 and ***P < 0.001 vs. control (ANOVA), #P < 0.05 vs. local site (t-test).

3.4 Mechanisms which contribute to the conduction process

To further distinguish between different mechanisms involved in the conduction process, we applied different ion channel blockers including two which block voltage-dependent Na+ channels (NaV). A relatively specific blocker of NaV (mepivacaine, 200 µmol/L) attenuated the local dilation in response to ACh by ∼25% without altering the temporal characteristics (Figure 3A). Despite this small local attenuation, the response conducted to distant sites without impairment (Figure 3C, F). Ado-induced dilations at local and conducted sites remained fully intact in the presence of mepivacaine (Figure 3B, D, G). Bupivacaine does not only block NaV but also several K+ channels, including KCa1.1, voltage-dependent K+ channels and possibly KIR.22 Similar to mepivacaine, bupivacaine reduced the local dilation to ACh-stimulation by ∼20% (Figure 3A). Nevertheless, the amplitude of the local dilation did not further diminish during conduction to remote sites (Figure 3C, F), suggesting that the attenuating effect was purely local and without impact on the conduction process itself. In marked contrast, the local dilation upon Ado was fully intact but remote dilations were completely abrogated in the presence of bupivacaine (Figure 3B, D, G).

Figure 3

Bupivacaine discriminates conducted dilations between ACh and adenosine. Time course of dilations in response to ACh and adenosine (Ado) at the local site (A and B) and 1100 µm upstream (C and D) at resting conditions and after bupivacaine (Bupi, 100 µmol/L) or mepivacaine (Mepi, 200 µmol/L). (F and G) Peak amplitude of the dilation over distance. Only bupivacaine abolished conducted dilations upon adenosine, but not ACh. Local application of K+ solution (3 mol/L) induced a biphasic response which also conducted to distant sites (E). The amplitude of the constriction remained unaffected by bupivacaine (H), whereas the transient dilation was abrogated (E). Values in % of maximal response, agonist application: vertical dashed lines. In cases without an identifiable maximum, diameter was averaged for 15 s after stimulation and taken instead. Experiments in the presence of l-NA and indomethacin except for the investigation of K+ responses (E and H). Mepivacaine n = 6, bupivacaine n = 5, K+ responses n = 6 arterioles and mice, *P < 0.05, **P < 0.01, ***P < 0.001 vs. control (ANOVA), #P < 0.05 vs. local site (t-test), §P < 0.05 vs. bupivacaine at time point 10 s (t-test).

We further studied the effect of bupivacaine on the conduction of vasoconstrictions which were initiated by local K+ depolarization achieved by short (60 ± 6 ms) pressure ejection of K+ solution (3 mol/L) in the absence of indomethacin and l-NA. K+ induced a rapid, short constriction at the local site followed by a transient dilation. This biphasic response conducted to remote sites without delay and decayed with distance (Figure 3E, H). In the presence of bupivacaine, the constriction and its conduction were unaffected (Figure 3H), but the transient dilation was abolished at all sites (Figure 3E). The abrogation of the dilation suggests that KIR channels are involved and blocked by bupivacaine. This was further evaluated by the application of a more specific blocker of KIR. Barium (75 µmol/L) did not alter significantly local or conducted responses to ACh (Figure 4A). In contrast, responses elicited by Ado were not attenuated at the local site, but strongly reduced at distant sites (Figure 4B). Thus, barium differentially affected conducted dilations for ACh and Ado and mimicked the effect of bupivacaine which suggests that discriminable pathways are involved.

Figure 4

Barium attenuates only conducted dilations in response to adenosine and lack of Cx40 reduced conducted ACh dilations. Maximal amplitude of dilations upon ACh and adenosine are depicted over distance. In the presence of barium (75 µmol/L) local and conducted ACh responses remained unaffected. In contrast, barium reduced conducted dilations initiated by adenosine without significant alteration of the local response (A and B, six arterioles in six mice). In Cx40-deficient mice ACh dilations decayed with distance (C). Thus, the decay of the amplitude of ACh- and adenosine dilation (D) was similar in Cx40−/− (six arterioles in three animals). Data for wt animals are replotted from Figure 1 for comparison. Presence of l-NA and indomethacin, **P < 0.01 and ***P < 0.001 vs. control (t-test), #P < 0.05, ##P < 0.01, and ###P < 0.001 vs. local site (t-test).

3.5 Expression and role of connexins

Conducted responses were also studied in three Cx40-deficient mice in six arterioles with a maximal diameter of 41 ± 2 µm. All preparations were treated with l-NA and indomethacin and arteriolar tone amounted to 0.36 ± 0.05. Local application of ACh elicited a local dilation which rapidly conducted to distant sites. However, as demonstrated previously,7 the amplitude of the remote dilation was significantly attenuated in Cx40-deficient mice in marked contrast to wt (Figure 4C). Ado induced local dilations and conducted responses with decaying amplitude which resembled responses in wt. In contrast to wt (Figure 1), dilations upon ACh and Ado were thus similar at all sites in Cx40-deficient mice (Figure 4D).

The expression of Cx37 was studied using immunohistochemistry. Cx37 was densely expressed in endothelial cells and located at cell borders in the aorta (Figure 5A) and in arterioles of different sizes in the cremaster muscle (Figure 5C, E) in wt animals. Surprisingly, Cx37 was not detected in endothelial cells in arterioles of Cx40-deficient mice (Figure 5D, F). However, Cx37 could still be identified at endothelial cell borders in the aorta although staining appeared to be less bright (Figure 5B).

Figure 5

Endothelial expression of Cx37 is strongly reduced in Cx40 deficient mice. Staining for Cx37 in the aorta (viewed en face: A and B) and in the cremaster muscle (CF). In wt, endothelial cell borders in the thoracic aorta (A) and in cremaster arterioles (C and E) are brightly stained demonstrating abundant expression of Cx37. In aortas of Cx40-deficient mice (B), Cx37 is also detectable; however, staining is less bright. In contrast to wt, Cx37 is not detectable in cremaster arterioles in Cx40-deficient mice (D, F). Images are representative for three experiments in each genotype, scale bars are 50 (C and D), 20 (A, B, F), and 10 µm (E).

3.6 Membrane potential changes in endothelial cells

Finally, we performed membrane potential measurements in vivo as described previously10,19 to determine changes of endothelial membrane potential upon ACh or Ado. Hitherto, 33 endothelial cells were studied in second-order arterioles in 21 animals. Microelectrodes contained carboxyfluorescein and cells were stained during impalement which allowed the identification of endothelial cells by their longitudinal orientation with respect to the vessel axis. Non-stained cells were excluded from data analysis. Resting membrane potential in endothelial cells amounted to −44 ± 2 mV. Focal, brief ejection of ACh through a micropipette induced a rapid hyperpolarization at all stimulation durations (0.1 s: from −42 ± 2 to −52 ± 2 mV, n = 37; 1 s: from −46 ± 3 to −57 ± 2 mV, n = 32; 3 s: from −45 ± 3 to −59 ± 4 mV, n = 9). However, the maximal amplitude was not significantly different for these stimulation durations, but the hyperpolarization was prolonged for longer stimulation durations (0.1 s: 4.9 ± 0.4 s; 1 s: 7.0 ± 1.0 s; 3 s: 10.8 ± 2.3 s, P < 0.01 vs. 0.1 s). In marked contrast, endothelial membrane potential was virtually not altered by Ado applied for similar durations (Figure 6). In fact, application of Ado was interspersed between ACh stimuli demonstrating the non-responsiveness to Ado despite pronounced and robust responses to ACh (representative trace in Figure 6A).

Figure 6

Only ACh hyperpolarizes endothelial cells. Endothelial membrane potential was measured in vivo using sharp electrodes. Cells were stained during measurement by carboxyfluorescein which allowed cell identification. Arterioles were stimulated by brief pressure ejection of ACh or adenosine using micropipettes. (A) Single trace depicting membrane potential over time. Pressure application of ACh (dashed lines with black circles) at varying durations (numbers above circles: stimulation time in seconds) induced transient endothelial hyperpolarizations. In contrast, adenosine application (grey circles) did not change membrane potential. (B) Changes in membrane potential are depicted after pressure application (dashed line) of ACh for 0.1, 1, and 3 s (n = 37, n = 32, and n = 9, respectively) or adenosine (0.1, 1, and 3 s: n = 21, n = 29, and n = 39, respectively). ACh significantly hyperpolarized endothelial cells, however, membrane potential was not altered by adenosine regardless of the application duration. Measurements were obtained in 33 cells in 21 animals, n gives number of stimulations, **P < 0.01, ***P < 0.001 vs. resting potential 1 and 2 s after stimulation (t-test).

4. Discussion

The present data demonstrate that both, ACh and Ado, initiate responses that conduct along the vessel wall, however, they exhibit unique, distinct features. These include the (i) K+ channels that are activated by the agonists to initiate the response, (ii) distance that is covered by the spreading response, (iii) mechanisms that augment the dilation while conducting, (iv) dependency on specific connexins, and (v) ability to elicit endothelial cell hyperpolarization. Together these unique properties suggest that distinct pathways conduct the responses. Because only ACh is able to hyperpolarize endothelial cells, we suggest that the primary pathway in case of ACh is the endothelium which is supported by the dependency on Cx40. Most interestingly, deficiency of Cx40 is associated with a strong attenuation of the expression of the second important endothelial connexin (Cx37) in cremaster arterioles. Under these conditions, namely lack and/or strong decrease of Cx40 and Cx37 expression, the conducted response initiated by ACh resembles the Ado response which is most likely due to the incapability of the endothelial cell layer to support signalling along the vessel. We therefore suggest that the smooth muscle cell layer acts as the signalling pathway for conducted dilations initiated by Ado and for ACh in Cx40-deficient animals.

Membrane potential changes induced by agonists at a local application site initiate vasomotor responses that rapidly spread along the vessel wall. Such a conducted response is well established for ACh4,5,7,8 which induces an endothelial cell hyperpolarization by the activation of KCa channels. Herein, KCa3.1 (IK1) is of major importance.10 Consequently, ChTx a non-specific blocker of IK1 strongly attenuated the response. In marked contrast, local and conducted dilations upon Ado application remained unaffected by ChTx, but were nearly abrogated by a blocker of KATP channels (glibenclamide). In line with the ability of KATP channels eliciting conducted responses, an opener of KATP (cromakalim) dilated arterioles locally and at remote sites resembling the Ado response. In both cases, the amplitude of the dilation decayed with distance and was sensitive to glibenclamide, in contrast to ACh responses.

It has been argued that Ado-elicited arteriolar dilations by direct stimulation of smooth muscle cells or through endothelial stimulation and NO release depending on the vascular bed.23 Moreover, the endothelium and an endothelial calcium increase was required to elicit a local and conducted dilation in response to Ado in the hamster microcirculation.24 However, in the present study, NO synthase and cyclooxygenase were blocked and dilations are therefore independent of NO in line with the previous observations.16 The Ado receptor subtype which elicits the activation of KATP channels and subsequent dilation was not evaluated in the present study, but in coronary arterioles and other vessels of the rat the A2A subtype of the Ado receptor family (also known as P1 receptors) is invoked.25,26 Regardless of the receptor involved, Ado-induced conducted dilations that are mediated by the activation of KATP channels.

Interestingly, the distances that dilations were conducted along the vascular wall exceeded distances that would be expected by pure electrotonic conduction.7,13 This was especially true for ACh, but likewise for Ado. In contrast, vasoconstrictions induced by K+-depolarization had nearly vanished at a distance of 1100 µm, whereas the Ado dilation amounted to about two-thirds of the local amplitude and the ACh response was not attenuated at this location. This implies that dilations are amplified specifically most likely by the activation of K+ channels. Indeed, the non-specific blocker of K+ channels (and Na+V channels) bupivacaine22 strongly attenuated remote Ado dilations without affecting the efficacy of ACh conduction. Likewise, the conduction of vasoconstrictions remained unaffected leaving it unlikely that bupivacaine blocked gap junctional coupling along the pathway. Moreover, mepivacaine which also blocks Na+V channels but not K+ channels22 was without effect on conducted dilations excluding that Na+V channels are contributing as an amplifying mechanism as recently suggested.27 After application of high K+, a delayed secondary dilation was observed most likely due to the dilution of K+ and this secondary dilation was also conducted along the vessel. Such dilations in response to intermediate K+ concentrations have been demonstrated to be due to an activation of KIR and/or the sodium pump.28,29 Interestingly, bupivacaine abrogated these responses suggesting an effect through blockade of K+ channels herein as well. The contribution of KIR acting as an amplifier specifically in Ado dilations was further verified by the strong attenuation of conducted dilations in the presence of barium which blocks KIR at the concentration used (75 µmol/L). Although these observations using merely specific blockers do not unambiguously verify a contribution of KIR to Ado responses, they do provide evidence for dissimilar and distinct mechanisms which augment the conducting dilation in case of Ado (bupivacaine- and barium-sensitive) and in case of ACh (insensitive to bupivacaine and barium). Likewise, modulation of conducted responses by endogenous NO was distinct for different conduction pathways.30

Cx40 is expressed abundantly in the endothelium and is crucial for intact conduction of ACh-induced dilations.7,13,16 In contrast, conduction of Ado responses remained unaltered in Cx40-deficient mice as also shown previously.13,16 Most interestingly, distances covered by the conducting dilations upon ACh and Ado were indistinguishable in Cx40-deficient mice, suggesting that they were conducted along the same pathway. Surprisingly, arterioles in the cremaster muscle were not only devoid of Cx40 but also of Cx37 as judged by immunohistochemistry. It has been previously reported that in aortic endothelium Cx37 and Cx40 mutually require each other to achieve full and effective integration of the proteins into the plasma membrane.31 In the present study, we identified Cx37 in aortic endothelium of Cx40-deficient mice, however also at a lower expression level.31 However, if Cx37 is not integrated into the membrane of endothelial cells in arterioles in Cx40-deficient mice, it is tempting to speculate that the endothelial pathway is non-functional and thus both, ACh and Ado, are conducted along the same pathway, namely the smooth muscle cell layer, in these animals. The hypothesis that ACh responses, in contrast to Ado dilations, are conducted in wt mice along the endothelium is further corroborated by our finding that only ACh, but not Ado, was able to initiate a strong endothelial hyperpolarization. This was observed in the same arterioles using two micropipettes and alternate application of these substances excluding a general non-responsiveness of the endothelium. Due to the lack or the strong impairment of functional gap junctions in the endothelium of Cx40-deficient arterioles, ACh dilations are conducted along the smooth muscle cell layer in these mice a phenomenon also observed after local disruption of the endothelial cell layer along the signalling pathway.5,15 However, this is not the case in wt mice because augmentation mechanisms are very dissimilar in wt mice as outlined above. It is worth mentioning that Ado dilations depended on KATP channel activation, however, an endothelial hyperpolarization was not observed. Therefore, we suggest that the smooth muscle is unable to transfer charge to the endothelium in sufficient amounts to hyperpolarize endothelial cells arguing against a tight myoendothelial coupling as observed before in these cremaster vessels in vivo using different experimental approaches.19

The presence of distinct pathways exhibiting dissimilar features has several implications, the most obvious being distinct regulation of conducted dilations in the microcirculation. Our finding may help to understand the previously demonstrated divergent sensitivities of conducted responses to gap junction blockers.32 In addition, Ado as well as KATP channels have been implicated in remote dilations in response to contraction of the skeletal muscle.33 These observations together with the present experiments would implicate the smooth muscle layer contributing as a signalling pathway in exercise hyperaemia in the microcirculation. However, the endothelial pathway has been experimentally proven to be involved in feed arteries residing outside the tissue, but it remains unclear which mediator invokes an endothelial hyperpolarization.34 Solving the mysteries of the suggested distinct pathways await further experimental investigation.

In summary, the present data demonstrate that ACh and Ado initiate conducted dilations which, however, can be clearly distinguished by the initiating K+ channel, pharmacological blockade of amplifying ion channels, distance encompassed, and Cx40-dependency. The fact that only ACh initiated an endothelial hyperpolarization suggests that ACh-induced hyperpolarization is conducted along the endothelial cell layer, whereas Ado responses use the smooth muscle cell layer. The understanding and identification of distinct properties of the endothelial vs. the smooth muscle pathway may help to clarify the role of conduction in functional hyperaemia which is despite its crucial importance in physiology still incompletely understood.35 Most interestingly, single Cx-deficiency alters the expression of other connexins which may result in the complete failure of an important signalling pathway in arterioles.


This work was supported by a grant of the Deutsche Forschungsgemeinschaft (WI 2071/2-1).


The author thanks Dr Stephanie E. Wölfle for acquirement and analysis of the data and Rita Meuer for technical assistance.

Conflict of interest: none declared.


View Abstract