Evidence both L-type and non-L-type voltage-dependent calcium channels contribute to cerebral artery vasospasm following loss of NO in the rat

Evidence both L-type and non-L-type voltage-dependent calcium channels contribute to


Introduction
Cerebral arteries operate in a dynamic state of partial constriction (myogenic tone), providing the capacity to constrict or relax in response to changing levels of intraluminal pressure, shear stress and nerve activity.Myogenic tone is an intrinsic property of the smooth muscle, helping maintain constant total cerebral blood flow and adapting blood flow locally to meet metabolic demand.. Myogenic constriction is driven primarily by membrane depolarization leading to Ca 2+ influx (Davis et al., 1999;Hill et al., 2001), possibly with a contribution via stretch-activated calcium sensitization (Schubert et al., 2008).Myogenic tone is often superimposed by vasomotion in the form of synchronised oscillations in smooth muscle cell membrane potential (Em), Ca 2+ and tension.
Although the physiological function of vasomotion in general is unclear, it may help to maintain a constant blood supply in many tissues, including the brain (Haddock et al., 2005).
One key influence of basal NO release in the middle cerebral artery appears to be suppression of both myogenic tone (Golding et al., 2001;Peng et al., 1998;Zimmermann et al., 1997) and vasomotion (Dirnagl et al., 1993;Haddock et al., 2002).This influence appears largely due to activation of BK Ca (Mandala et al., 2007;Yuill et al., 2010).So block of NO generation and/or BK Ca provides a means to mimic an aspect of endothelial dysfunction that is an early feature of cardiovascular disease, including disease conditions that predispose to vasospasm (Jewell et al., 2004;Vanhoutte et al., 2009).These conditions will also mimic the loss of NO observed after subarachnoid haemorrhage, where scavenging of NO by haemoglobin (Martin et al., 1985) causes profound vasospasm (Toda et al., 1991).Significantly, enhanced vasomotion (or vasospasm) can lead to a reduction of cerebral capillary blood flow and thus compromise of neuronal function (Biswal et al., 1996;Pluta, 2005).
Several mechanisms, including the rho-kinase pathway, can contribute to the development and maintenance of constriction in smooth muscle, alongside calcium entry.However, in many vascular beds, including the cerebral vasculature changes in smooth muscle intracellular Ca 2+ ([Ca 2+ ] i ) concentration are critical for myogenic tone and vasomotion (Haddock et al., 2002;Haddock et al., 2005;Yuill et al., 2010).[Ca 2+ ] i increase involves release from intracellular stores and entry from the extracellular space via voltage-gated Ca 2+ channels and non-selective cation channels, such as transient receptor potential channels (TRPC).Ca 2+ influx through voltage-gated Ca 2+ channels (VGCC) leads to global increases in smooth muscle cell [Ca 2+ ] i and constriction, and high voltage activated (L-type) Ca 2+ channels appear central in this sequence (McCarron et al., 1997;Moosmang et al., 2003;Nelson et al., 1990).These channels are expressed widely in vascular smooth muscle and their open probability increases over a physiologically relevant range (circa -50 to -30 mV) (Lacinova, 2005;Smirnov et al., 1992).Low voltage activated or T-type Ca 2+ channels are also expressed in vascular smooth muscle of resistance arteries (Braunstein et al., 2008;Clozel et al., 1997;Kuo et al., 2008;Navarro-Gonzalez et al., 2009;Perez-Reyes, 2003).But although they are normally active in the range circa -60 to-40 mV, the characteristic rapid inactivation of these channels argues against a significant role at physiologically relevant membrane potentials in the vasculature.Despite this, they have been implicated in the maintenance of vascular tone in a variety of arteries, including rat cremaster (VanBavel et al., 2002), rat basilar (Navarro-Gonzalez et al., 2009) and middle cerebral (Lam et al., 1998) arteries, and direct measurements have shown high voltage-activated but nifedipine-insensitive Ca 2+ currents, pharmacologically indistinguishable from T-type currents, in both guinea-pig and rat terminal mesenteric arteries (Morita et al., 1999;Morita et al., 2002).
We recently reported that middle cerebral arteries develop intense and sustained constriction, associated with a very rapid form of vasomotion, when NOS and/or BK Ca channels were blocked, to mimic endothelial dysfunction.Furthermore, both constriction and vasomotion depended on calcium entry via VGCCs and the oscillations in E m were temporally linked to changes in smooth muscle [Ca 2+ ] i (Yuill et al., 2010).The temporally linked oscillations in E m , [Ca 2+ ] i and tension were similar to the widely described phenomenon of vasomotion, but displayed a much higher frequency (~1Hz as opposed to ~0.1-0.2Hz,Yuill et al., 2010).The intense "vasospastic vasomotion" was reversed by inhibition L-type Ca 2+ channels and clearly involved a complex action of NO that appeared to include stimulation of BK Ca channels and a cGMP-independent closure of VGCCs (Yuill et al., 2010).However, although a central role for VGCC, NO and BK Ca was apparent, the importance of other ionic currents that might contribute to the rapid depolarizing oscillations was unclear.
Thus, the aim of the present study was to characterize the ionic mechanisms responsible for rhythmic oscillations in E m and tension in rat isolated middle cerebral arteries following inhibition of BK Ca channels and NOS.We probed channels that may lead to both depolarization (calcium, sodium and chloride channels) and repolarization (potassium channels).Our data suggest a novel role for both smooth muscle T-type Ca 2+ channels and several potassium conductances in the both "vasospastic vasomotion" and the underlying maintenance of vasoconstriction.

Animals and tissue isolation
Male Wistar rats (200-250 g) were killed by cervical dislocation followed by decapitation, following institutional guidelines for animal welfare and schedule 1 of the Animals (scientific procedures) Act 1986.The brain was removed and immediately placed in ice-cold Krebs solution.
Segments of the middle cerebral artery (~2 mm long) were dissected and stored in ice-cold Krebs for use within 30 min, with similar size vessels used in all experimental groups

Experimental protocols
Segments of middle cerebral artery (internal diameter ~150 m) were mounted in a Mulvany-Halpern myograph (model 400A, Danish Myotechnology) in Krebs solution containing (mM): NaCl, 118.0, NaCO 3 , 24; KCl, 3.6; MgSO 4 7H 2 O, 1.2; glucose, 11.0; CaCl 2 , 2.5; gassed with 20% O 2 , 5% CO 2 and balance N 2 and maintained at 37C.After equilibration for 20min, vessels were tensioned to 1-1.5mN (approximates wall tension at 60mmHg).Smooth muscle tension was recorded with an isometric force transducer and Powerlab software (ADI, Australia).Vessel viability was assessed by addition of exogenous K + (15-55 mM, total K + concentration); only vessels developing tension of 3 mN were used, following this Endothelial cell viability was assessed by the ability of the protease activated receptor 2 activating peptide; SLIGRL (20 M) (Alexander et al., 2008) to relax U46619 induced tone (100 nM) by ≥ 75 %, vessels with less relaxation were discarded.In some experiments, endothelial cells were removed by gently rubbing the luminal surface with a human hair; subsequent relaxation of <15 % to SLIGRL (20 µM) was considered as successful removal and further abrasion often lead to damage of smooth muscle cells.L-NAME (100 M), indomethacin (10 M) and iberiotoxin (100 nM) were added throughout the experiment (to block NO synthase (NOS), cyclooxygenase and BK Ca channels, respectively), unless otherwise stated.In combination, these drugs gave a robust and sustained constriction (similar to vasospasm), and an associated rapid vasomotion.Similar responses were recorded in each case after inhibition of NOS alone, but vasomotion was more variable between preparations.Indomethacin had no effect on oscillations, but was included in the experimental cocktail to minimize any potential for confounding thromboxane signaling after NOS inhibition (Benyo et al., 1998;McNeish et al., 2007).Recordings were assessed in the presence of: the T-type (mibefradil 100 nM and NNC 55-0396 300 nM) and L-type (nifedipine 1 M) Ca 2+ channel blockers, the K Ca channel blockers, apamin (K Ca 2.3 (SK Ca ), 50 nM), TRAM-34 (K Ca 3.1 (IK Ca ), 1 M), iberiotoxin (BK Ca , K Ca 1.1, 100nM) and charybdotoxin (K Ca 3.1, BK Ca , 100 nM), the K IR channel inhibitors BaCl 2 (30 M) and CsCl 2 (10 mM), the Na + /K + -ATPase inhibitor, ouabain (1 M) and the voltage-gated K + channel inhibitor, 4-aminopyridine (4-AP, 3 mM).Papaverine (150 M) was added at the end of each experiment to assess overall tone.All blocking drugs were incubated for at least 20 min before data was recorded to ensure maximal effect.In most experiments smooth muscle membrane potential (E m ) and tension were measured simultaneously as previously described, using glass microelectrodes (filled with 2 M KCl; tip resistance, 80-120 M) to measure E m (Garland et al., 1992).

Data analysis and statistical procedures
Results are expressed as the means.e.mean of n animals.Tension values are given in mN (always per 2 mm segment) and E m as mV.During the vasospastic vasomotion E m is expressed as the mean E m over a random 10s period of the rapid vasomotion where possible we have also reported the size of the depolarizing oscillations in mV.Vasodilatation is expressed as percentage reduction of the total vascular tone (myogenic tone plus vasoconstrictor response induced by either U46619 or the combination of L-NAME and iberiotoxin, as appropriate), quantified by relaxation with papaverine (150 M).Graphs were drawn and comparisons made using either Student's t-test, or one-way ANOVA with Tukey's post-hoc test using Prism software (Graphpad, USA).P≤0.05 was considered significant.
In the presence of L-NAME and indomethacin, the BK Ca channel inhibitor, iberiotoxin (100 nM), evoked further depolarization (to E m -35.7±1.1 mV, n=40; P<0.05) and constriction (to 5.7±0.2mN, n=41; P<0.05), associated with a marked increase in the amplitude of oscillations in E m temporally linked to tension (Figure 1B).The oscillations in E m and tension had a frequency of 0.84±0.02Hz and 0.80±0.05Hz and amplitude of 22.6±1.3mV and 0.19±0.02mN, respectively (n=41).All subsequent experiments were performed in the presence of L-NAME, indomethacin and iberiotoxin unless stated.Removal of the endothelium abolished SLIGRL-mediated relaxation (20 µM), but failed to affect oscillations in E m (frequency of 0.76±0.10Hz; amplitude of 19.4±4.3 mV) and tension (frequency of 0.66±0.10Hz; amplitude of 0.16±0.05mN, n=3).
The oscillations in E m and tension were not modified by the K Ca 2.3 blocker apamin (50 nM), either alone (Figure 3C) or in the additional presence of TRAM-34 (Figure 3D).Likewise, charybdotoxin (100 nM) alone (Figure 4B) or with apamin (Figure 4C) did not affect either oscillations in E m and tension (Figure 4D and E), or mean tension and E m .

Discussion
This study provides the first demonstration, that rhythmic oscillations in membrane potential and tension as well as the associated spasm in rat isolated middle cerebral arteries following inhibition of BK Ca channels and/or NOS reflect Ca 2+ influx via T-type Ca 2+ channels, in addition to L-type Ca 2+ channels.These data extend our previous observation that NOS inhibition leads to L-type Ca 2+ channel opening and arterial spasm, characterized by sustained constriction and superimposed by rapid vasomotion (Yuill et al., 2010).Vasomotion was of much higher frequency than previously recorded in other vascular beds, and as such we refer to it as 'vasospastic' vasomotion.We also provide evidence for modulation of both the Ca 2+ dependent vasomotion and constriction through Na/K ATPase and a K + conductance.
Consistent with previous work, our data suggest that calcium influx underlies rhythmic oscillations in the constricted rat middle cerebral artery, and that oscillations in membrane potential and tension are linked to oscillations in intracellular [Ca 2+ ] i as well as spasm in both middle cerebral and basilar arteries (Haddock et al., 2002;Navarro-Gonzalez et al., 2009;Yuill et al., 2010).Reducing extracellular calcium diminished the amplitude of oscillations, led to relaxation and paradoxically depolarized the membrane.The relaxation presumably reflected the reduction in peak E m associated with the reduced amplitude of oscillations in E m .Some small oscillations in tension did persist and may reflect vasomotion-independent intracellular calcium release, as reported in some arterial beds: for review see Haddock et al., 2005(Haddock et al., 2005).As oscillations were insensitive to the calcium-dependent Cl -channel inhibitor, DIDS and the voltage-dependent Na + channel blocking agent, tetrodotoxin they certainly appeared to be mediated exclusively by calcium conductance.These data contrast with the basilar artery, where inhibition of Cl -channels in this larger artery abolished calcium-dependent oscillations, leading to hyperpolarization and relaxation (Haddock et al., 2002), and parenchymal arterioles, where calcium-dependent oscillations were blocked with tetrodotoxin (Filosa et al., 2004).However, our data are consistent with human pial arteries, where tetrodotoxin was also without effect against oscillations in diameter (Gokina et al., 1996).In the present study, oscillations in muscle membrane potential were also resistant to direct damage of the endothelium, suggesting this monolayer may not influence rapid vasomotion associated with arterial spasm.
In the rat middle cerebral artery, opening L-type Ca 2+ channels is essential for vasoconstriction and vasomotion to develop, because inhibition of these channels abolishes vasomotion and fully reverses tone (Yuill et al., 2010).Surprisingly, in the present study under similar vasospastic conditions, oscillations in E m , and the associated oscillations in tension were abolished and followed by relaxation after calcium influx through T-type Ca 2+ channels was blocked with mibefradil (at a concentration selective for block of T-type Ca 2+ channels; 100 nM) or with a T-type Ca 2+ channel selective, non-hydrolysable analogue of mibefradil, NNC 55-0396, (Huang et al., 2004).Block of oscillations with the putative T-type Ca 2+ channel blockers was not associated with a net hyperpolarization and complete relaxation, contrasting with the L-Type Ca 2+ channel blocker nifedipine.So the effect of mibefradil or NNC 55-0396 is unlikely to reflect a non-specific effect against L-type channels.Furthermore, although mibefradil apparently reduced blood pressure and myogenic tone by an action on L-type calcium channels (Moosmang et al., 2006), the lower concentration of mibefradil used in the present study is relatively specific against T-type Ca 2+ channels.In fact, mibefradil seems only to inhibit the L-type Ca 2+ channels after tissue metabolism (Wu et al., 2000).The non-hydrolysable analogue of mibefradil NNC 55-0396 is selective for Ttype Ca 2+ channels, with no reported block of L-type Ca 2+ channels even in concentrations as high as 100 µM (Huang et al., 2004).Mibefradil has also been reported to block both Cl - (Bernd et al., 1997) and Na + (Eller et al., 2000;Guatimosim et al., 2001) channels, but again with much higher concentrations (µM) than employed in the current study.Furthermore, the fact that blockers such as DIDs and TTX had no effect against vasospastic vasomotion makes an action of mibefradil against these channels extremely unlikely.Therefore, the differential effect of mibefradil and NNC-0396 compared to the selective blocker L-type Ca 2+ channel blocker nifedipine indicate a critical role for T-type Ca 2+ channels in vasospastic vasomotion.Further, these channels contribute significantly to overall constriction in the middle cerebral artery.
Vasospastic vasomotion only developed once the smooth muscle cells depolarized to circa -40 mV, so it may be that T-type channels involved in vasomotion have gating properties similar to high voltage activated Ca 2+ channels.This is surprising, as by definition T-type Ca 2+ channels activate at low potentials and then quickly inactivate (Lacinova, 2005).However, our data are consistent with studies reporting T-type Ca 2+ channels that influence vascular tone and have properties similar to high-voltage activated Ca 2+ channels, (Navarro-Gonzalez et al., 2009).Both T-and L-type Ca 2+ channels are expressed in rat basilar (Navarro-Gonzalez et al., 2009) and middle cerebral arteries (Kuo et al., 2008), and in each artery the CaV 3.2 (or T-type) is the most abundant VGCC alpha subunit expressed.Human recombinant T-type Ca 2+ channels (Kaku et al., 2003) and T-type Ca 2+ channels co-expressed with auxiliary subunits (Wyatt et al., 1998) do have gating properties similar to high voltage activated channels, so channels in the middle cerebral artery may be similar.
Peripheral arteries also contain VGCCs with similar biophysical properties to high voltage activated channels (-50 to -20 mV), and are pharmacologically indistinguishable from T-type Ca 2+ channels in both the guinea-pig and rat (Morita et al., 1999;Morita et al., 2002).In these small mesenteric arteries, T-type Ca 2+ channels are the predominant voltage-gated subtype (Gustafsson et al., 2001;Jensen et al., 2004), and show increased window current due to non-inactivation at physiological E m (Jensen et al., 2009).
As nifedipine abolished vasomotion, we propose that L-type channels are key for the initial depolarization and constriction (Yuill et al., 2010); whereas T-type are activated subsequently and as such are critical for the vasospastic vasomotion.Interestingly, input from T-type Ca 2+ channels seem to be important for the initial constriction in the basilar artery, whereas L-type channels are critical for vasomotion (Navarro-Gonzalez et al., 2009).But taken together, these results all suggest a functional coupling between L-and T-type Ca 2+ channels, as previously suggested in renal (Hansen et al., 2001) and mesenteric arterioles (Braunstein et al., 2008).As such, this might explain why neuroprotection in ischemic stroke is more effective in patients given blockers for more than just L-type VGCCs (Kobayashi et al., 1998).
As VGCC and hence vasospastic vasomotion are inhibited by a complex interaction between NO and smooth muscle cell BK Ca channels in middle cerebral arteries, we attempted to characterise further the vasospastic vasomotion.Endothelial cell damage did not affect the vasomotion, so we inhibited a variety of K + currents.Both inwardly rectifying and voltage-gated K + channels participate in maintenance of resting membrane potential and vascular tone (Ko et al., 2008;Nelson et al., 1995;Sobey, 2001).However, inhibition of K IR channels with CsCl or barium did not affect vascular tone, although CsCl did slightly increase E m oscillation amplitude.As Ba 2+ was without effect, this small change most likely reflected a non-selective action of CsCl.Inhibition of K v channels with 4-AP also had little effect, causing only a small increase in the amplitude of oscillations in E m and tension.This is consistent with the reported role of these channels in rat mesenteric artery where inhibition of K v increased rhythmic contractions (Gustafsson et al., 1994).
So voltage-gated K + channels did not appear to play any major role in vasospastic vasomotion.
Our data do suggest that Na + /K + -ATPase might contribute to vasomotion, as ouabain caused relaxation and reduced the amplitude and frequency of oscillations in E m , although surprisingly without affecting oscillations in tension.Ouabain can attenuate intercellular communication in smooth muscle (Harris et al., 2000;Martin et al., 2003;Matchkov et al., 2007) and the synchronized changes in vascular [Ca 2+ ] i (Koenigsberger et al., 2004) that lead to vasomotion (Chaytor et al., 1997;Matchkov et al., 2004;Peng et al., 2001).So in part, ouabain may alter membrane potential oscillations by modifying cell-cell communication.Interestingly, ouabain effects were reversed by 4-AP, again indicating that K V might contribute under some conditions to influence vasomotion.
The ability of the K Ca 3.1 channel inhibitor TRAM-34 to reduce rather than enhance the amplitude of oscillations in E m and tension was also unexpected.This effect was on the smooth muscle, as it was not altered by removal of the endothelium, and in contrast to TRAM-34, charybdotoxin, a mixed BK Ca and K Ca 3.1 inhibitor failed to modify the oscillations.One explanation, is that TRAM-34 inhibits non-selective cation channels in the smooth muscle, similar to its action in isolated immune cells (Schilling et al., 2007).Non-selective cation channels are present in rat middle cerebral artery smooth muscle cells and appear to contribute to the calcium entry and vascular tone (Marrelli et al., 2007;Welsh et al., 2002).So data with TRAM-34 suggest that non-selective cation channels may play an important role in the calcium entry events underpinning depolarization and vasomotion after NOS inhibition in the middle cerebral artery.
In summary, inhibition of either BK Ca channels and/or NOS evokes vasospasm and fast, rhythmic oscillations in E m and tension that are mediated by Ca 2+ influx via both T-type and L-type Ca 2+ channels.Our data suggest the T-type channels are active at physiologically relevant membrane potentials and can therefore make an important contribution to the control of cerebrovascular blood flow during vasospasm associated with disease states where NO synthesis or action is impaired, such as cerebral ischemia or subarachnoid haemorrhage.

Figure 1 .
Figure 1.Original traces showing simultaneous recordings of membrane potential (upper panels)

Figure 3 .
Figure 3. Original traces showing control conditions (A), the effect of the K Ca 3.1 channel blocker,

Figure 4 .
Figure 4. Original traces showing control conditions (A) and the effect of either the BK Ca and