Introduction
CVS is thought to be a severe complication of SAH. However, the pathogenesis of CVS is not completely understood, and no definitive treatment has been established. Once aneurysm rupture occurred, blood pours into the subarachnoid space even to the brain parenchyma and ventricles. The intracranial pressure rises sharply and might increase enough to affect cerebral perfusion and cause global ischemia. Due to CVS, maximal 7–10 days after onset of SAH, the presence of blood in the subarachnoid spaces triggered and associated with DCI, persistent neurological deficits and long-term neurological disability. DCI is related to the development of CVS, as it is the most important adverse prognostic factor of outcome and a major cause of morbidity and mortality in SAH patients [
1]. The pathogenesis of DCI is hypothesized to be multifactorial, including angiographic vasospasm, ischemia, microthrombosis and microcirculation constriction [
2‐
4]. Due to DCI is among the most important adverse prognostic factors for outcome after SAH [
4], it’s of great necessity to explore new targets dealing with the progress of pathology in DCI based on previous SAH animal models [
5‐
7]. In our previous study [
8], we supported the hypothesis that gap junction blockers may relieve the CVS after SAH via cerebral angiography and morphologic study, suggested that gap junctions may play an important role in the pathogenesis of CVS.
Gap junction channels are formed by members of a family of proteins known as connexins [
9]. Among them, Cx43 is the most abundant and the major connexin in vascular smooth muscle [
10]. Liao Y et al. [
11] reported that the conditional knockout of Cx43 in vascular endothelial cells resulted in hypotension, indicating that altered Cx43 expression may be related to irregular vasomotion. In our previous study [
8,
12], we reported that Cx43 levels were higher and that vasospasm occurred in the BAs at 7 days after SAH in animals. But, the specific role of Cx43 in DCI after SAH is not well understood.
SAH-induced CVS most often occurs between day 3 and day 7 [
13‐
15], indicating that it may be associated with one or more byproducts of blood breakdown in the subarachnoid space, where the exterior surfaces of the cerebral vessels are exposed to blood and byproducts of its breakdown. The breakdown of blood results in the release of OxyHb-derived free radicals, which lead to inhibition of ATP-dependent calcium pumps [
16,
17], alterations in vasomotor tone including the release of vasoactive eicosanoids and endothelin from the vessel wall, inhibition of endothelium-dependent relaxation, scavenging of NO and possible influences on the development of CVS [
18]. OxyHb-mediated CVS is dependent on the activation of PKC enzymes [
19], which are broadly involved in vital cellular functions, including smooth muscle contraction [
20].
Several studies have revealed that in vasospastic cerebral arteries, PKC activity is significantly enhanced in the membrane fraction [
21], and the time courses of the progression of CVS and PKC activation are well correlated, as revealed by an enzyme immunoassay study [
22]. However, the specific mechanism of PKC activation leading to CVS is unclear. Joshi CN et al. [
23] demonstrated an important role of Cx43 in the proliferation of vascular SMCs: increased Cx43 expression was significantly reversed in the presence of PKA, PKG and PKC inhibitor in thoracic aorta SMCs. In addition, S Nishizawa et al. proposed that PKC isoforms potentially play a role in the initiation and maintenance of delayed CVS [
24] implicating the PKC pathway in Cx43 expression and Cx43-mediated GJIC. But, to date, there is no information available regarding the possible involvement of PKC in Cx43 alterations and gap junction remodeling on cerebrovascular SMCs under spastic conditions. The purpose of the present study was to determine the actual role of Cx43 on DCI after SAH and its involvement in the PKC pathway to lay the groundwork for developing therapeutic strategies for SAH.
In the current study, the time-course changes in Cx43 both in vitro and in vivo were explored. The exact location of Cx43 expression in BA walls was detected using a laser-scanning confocal microscope. By specifically knocking down Cx43 in BAs using siRNA interference, we correlated the important role of Cx43 with DCI in a common SAH model. In addition, by designing an experimental therapeutic study using specific PKC inhibitors, we provide evidence that Cx43-mediated GJIC enhancement and DCI may be modulated via the PKC pathway, which could be used as reference data for developing and analyzing neuroprotective strategies in further studies.
Discussion
Our present study showed that Cx43 protein was increased in the in vitro CVS model with OxyHb incubation. Moreover, our dye transfer assay also revealed enhanced GJIC in BASMCs in culture incubated with OxyHb. Our previous study demonstrated that Cx43 expression may up-regulate contractions in rabbit BAs via a GJIC-dependent mechanism evaluated by a scrape-loading method [
49], we further confirmed here. The enhancement in GJIC may lead to exaggerated contractile responses among more coupled SMCs. Thus, spasmogenic signals such as red blood cell breakdown and the metabolic products spreading into the subarachnoid space could propagate more rapidly and forcefully from the ruptured vessel wall to a more extensive area, causing a prolonged contractile response. In addition, in our previous study, we showed that heptanol, a gap junction inhibitor, significantly inhibited the sustained contraction of rabbit BA rings induced by OxyHb in a dose-dependent manner [
12]. The implication of gap junctions, especially Cx43, in OxyHb-induced CVS was further demonstrated here.
Both specific inhibitors of PKC we selected (CHE/GF) have been used to study the involvement of PKC in signal transduction pathways [
27,
28]. By treatment with the two PKC inhibitors, the increased Cx43 protein and the enhanced GJIC were both reversed, implying that OxyHb-induced Cx43 alterations and enhanced GJIC may be modulated via the PKC pathway. However, previous studies have reported that PKC is composed of a family of serine–threonine kinases consisting of three major classes with at least 11 different isoforms [
50]. During the pathological process of canine basilar artery cells in previous CVS research [
24], PKCδ translocates from the cytosol to the membrane in early stage, and PKCα is subsequently translocated, suggesting that PKCδ plays a role during initiation and PKCα in the maintenance of CVS. However, as the two PKC inhibitors used here are both non-selective, the specific isoforms of PKC and their functional roles leading to Cx43 alterations are presently unknown and their identity remains to be elucidated. Furthermore, Richards et al. [
51] proposed that PKC spatially and temporally controls GJIC via phosphorylation of Cx43 at serine 368, meanwhile as PKC activity was significantly enhanced in the membrane fraction of vasospastic cerebral arteries, the enhanced GJIC revealed here may be associated with the phosphorylation level of Cx43 at serine 368. Interestingly, given that the total expression of Cx43 protein also increased significantly in our observations, it seems that the modulation of GJIC may result from the ratio of pS368-Cx43 to total Cx43, which remains to be explored.
In our previous study [
8], we found that vasospasm was most intense and coincidentally that the expression of Cx43 peaked on day 7. Together with our observation of the time phase change of Cx43 on SMCs, our findings further determined the involvement of increased Cx43 in SAH-induced CVS both in vivo and in vitro. On the other hand, DCI is thought to be caused by the combined effects of cortical spreading ischemia, arteriolar constriction and thrombosis. Here, DCI was investigated by performing intravital fluorescence microscopy and MRI measurements by PWI in rats. We directly visualized the post-SAH microcirculation in vivo by intravital fluorescence microscopy. This technique is advantageous over ex vivo studies for investigating arterioles in the living brain, allowing us to observe vasospasms and dynamic interactions in the entire vascular tree of MCA. Our main findings of the DCI investigation were that CBF significantly decreased and over 60% of arterioles derived from the MCA showed pearl string-like constrictions on day 7 after SAH. As reducing vessel diameter by 30% could reduce flow by ~ 80%, leading to DCI, our results may explain why DCI is observed after SAH in both experimental animals and patients [
51‐
53]. It suggesting that changes in Cx43 might be implicated in the pathophysiological events of DCI since both DCI and increment of Cx43 coincidentally occurred on day 7.
After knockdown of Cx43 via RNA interference, MCA deficiency and reduced CBF were both eased while control siRNA did not significantly relieve the deficits. In a previous study [
54], corrosion casts of arterioles showed that they exhibited tapered narrowing with folding after SAH. Additionally, the width of the arterioles was significantly lower at 3 and 7 days after SAH. Morphometric examination by light microscopy showed that the internal diameter of the arterioles was significantly smaller and that this change was associated with a significant increase in wall thickness at any depth from the brain at 7 days after SAH. This condition had improved by 14 days after SAH, suggesting that arteriole constriction occurs after SAH and may contribute to DCI. A previous study [
55] used orthogonal polarization spectral (OPS) imaging and found that in SAH patients, capillary density was significantly lower and the small arteries and arterioles at the cortical surface exhibited vasospasm that was not detected by angiography or transcranial Doppler sonography. These changes indicated that CVS may be associated with a significant decrease in capillary perfusion. Although Cx43 expression was restored by pretreatment with PKC inhibitors or Cx43 siRNA, resulting in the significant alleviation of the deleterious effects of SAH, the microvascular spasms were not completely reversed (Fig.
5). In our Cx43 siRNA interference experiment (Fig.
4a), the efficacy reached an ideal level prior to days 3–5. Because the development of constriction in microarterioles is a gradual process and because a previous study in rats [
56] showed that subarachnoid arterioles constrict by ~ 40% in the early phase, i.e., 5 min to 2 h after SAH in vivo, we suspect that the efficacy of Cx43 siRNA interference and the occurrence of micro-constriction are not synchronously correlated. Thus, in accordance with our proposed hypothesis, we showed that the severity of CVS and, subsequently, DCI induced by SAH were strongly related to vascular Cx43 alterations. Moreover, in addition to SMC, astrocytes are also important regulators of vasomotor reaction. The high abundance of gap junctions in astrocytes allows for the direct intercellular diffusion of ions, nutrients, and signaling molecules between these cells. Cx43 is the most abundant connexin expressed in astrocytes and thus constitutes the major connexin contributing to gap junction communication in astrocytes. Hence, SAH-induced alterations in the expression and functions of Cx43 should be systematically explored in other types of cells in future studies.
Endothelium-mediated vasodilatation is evoked by increased calcium concentrations in the endothelium, which triggers hyperpolarization by activating endothelial calcium-dependent potassium channels [
57], leading to the recognition of endothelium-dependent hyperpolarizing factor (known as EDHF). Thus, the conduction of EDHF via MEGJs to the underlying smooth muscle cell layer results in a decrease in intracellular calcium and subsequent vasodilatation. Because vasodilation signals upon focal stimulation of arterioles have been shown to be conducted along the vascular wall [
58], Hoepfl B et al. [
59] suggested that EDHF, but not NO or prostaglandins, serves as a critical mediator to initiate conducted vasodilation (CVD) upon acetylcholine administration in hamster arterioles. In the present study, Cx43 increased significantly after SAH, and it is believed that progressive up-regulation of Cx43 in SMCs may result in the development of intimal hyperplasia [
60], and increasing intimal hyperplasia mechanically dissociates endothelial–medial interactions at the MEGJ interface. Decreasing myoendothelial interactions progressively attenuates the hyperpolarization that normally passes from the ECs to medial SMCs, eventually leading to irregular vasomotion. Moreover, studies using connexin-mimetic peptides to selectively inhibit GJIC in rabbit iliac arteries suggested that Cx43 is required for propagation of EDHF within the smooth muscle layer [
61], implying the involvement of Cx43 in disruption of CVD. Although our data demonstrated that Cx43 is closely related to the CVS in pathological conditions of SAH, these findings do not explain the specific mechanism of Cx43 underlying the impaired vasodilation. Although SMCs have been reported to primarily express Cx43 [
41] and ECs to primarily express Cx40 [
62,
63], the connexin content of MEGJs remains unknown and has not been adequately explored. In contrast with Cx43 knockout mice, constitutive deletion of Cx40 results in hypertension in both anesthetized and awake mice [
64,
65]. Cx40 is also critical for transferring vasodilatory signals from ECs to the vascular media [
66]. Moreover, it has been proposed that nearly a twofold difference in the Cx43:Cx40 expression ratio gives rise to more than a 25-fold difference in dye coupling via the GJIC in vitro [
67]. Together with our observation that Cx43/Cx40 proteins form heteromeric gap junctions that are increased in SAH, we hypothesize that the Cx43:Cx40 ratio maintains a certain pattern for the maintenance of various physiological functions under normal conditions. Once SAH occurred, the connexin patterns varied, and both increased Cx43 and decreased Cx40 within the cerebral vessel wall may be responsible for the development of CVS. The present findings suggest a novel emphasis on this research direction for future studies.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.