Introduction

Aneurysmal subarachnoid hemorrhage (SAH) impacts approximately 10 to 21 in 100,000 individuals annually [1–3]. An accurate rupture rate for un-ruptured cerebral aneurysms has also been reported [4]. In a society that is continually aging, the rupture rate for aged persons is relatively high and can be expected to increase [2, 5, 6]. Following aneurysmal hemorrhage, 28 to 40 % of patients die within a few days of SAH, due to the initial damage or to the re-rupture of an aneurysm [5, 7]. Overall mortality and morbidity are greater than 50 % [1, 5, 7]. A major influence on the outcome of post-operated SAH patients is cerebral vasospasm (CVS) [6, 8, 9]. Even when aggressive treatment of CVS is initiated, the reported incidence of symptomatic CVS can still range from 10 to 50 % [6, 10–13]. Vascular smooth muscle contraction is the main cause of CVS, and the use of vasodilator medicines is therefore a simple, direct way to reduce or treat it [11, 13–18]. But these medicines do not control all types of CVS and other factors may be involved. Four percent of non-vasospasm patients suffer delayed spasm, with their areas of infarction more likely to be watershed areas. The mechanisms of this type of vasospasm-unrelated infarction are considered to be microcirculatory dysfunction, a cerebral thromboembolic event, or a cortical spreading depression [19].

One possible mechanism of untreatable CVS is phenotypic transformation of vascular smooth muscle cells (SMC) [20, 21]. A healthy arterial middle layer consists mainly of SMC. So-called differentiated SMC are well known to control vasoconstriction and vasodilation. But differentiated SMC has varied potential for proliferating, secreting cytokines, and migrating into the subendotherial layer, depending on their environment (Fig. 1). This transformation of differentiated SMC into dedifferentiated SMC is called the “phenotypic transformation of SMC.”

Fig. 1
figure 1

Temporal change in the arterial wall in subarachnoid hemorrhage. a Initial damage due to subarachnoid hemorrhage. b Switching of smooth muscle cells for phenotypic transformation. c Vasospasm and arterial wall thickening. d Release of vasospasm

Full size image

Phenotypic transformation of SMC has been studied in atherosclerotic disease and ischemic cerebrovascular disease, in the coronary artery [22, 23], the aorta and iliac artery [24, 25], kidney [26, 27], the cervical artery [28–30], and in cultured SMC [22, 31]. Cerebrovascular dysfunction in these diseases could result from phenotypic transformation of cerebrovascular smooth muscle cells and impaired blood flow to metabolically active brain regions [21].

In this review, we describe the phenotypic transformation of SMC after SAH from macroscopic reactions in humans and extending down to the level of the gene.

Insult due to Aneurysm Rupture

At the onset of SAH, nearly all patients suffer severe headaches that are induced by elevated intracranial pressure (ICP) and by direct hematoma stimulation of pain receptors located near the vessels and the dura matter (Fig. 1a). Concurrently, acute, severe hypertension is produced in order to maintain cerebral blood flow that opposes the elevated ICP and as a reaction to pain. Acute, severe hypertension can induce edema in the vessel wall, with resulting fragmentation of the elastic tissue and myonecrosis [32]. Several authors have reported that a subarachnoid clot and a sudden rise in the ICP are important factors causing breakdown of the blood-arterial wall barrier, while the effect of the clot itself is of greatest significance [33–35]. They also reveal that opening of the interendothelial junctions is one important mechanism of CVS [35]. Severe hypertension is caused by an elevation in adrenergic factors; norepinephrine induces the proliferation and phenotypic transformation of SMC via α1 and β1 receptors [31].

Other mechanisms that influence the onset of CVS should also be considered. Blast-induced traumatic brain injury can induce cerebral vasospasm without SAH. In an experiment involving arterial lamellae, this phenomenon occurred within an hour after the stimulus. Injured tissues displayed altered intracellular calcium dynamics, leading to hypersensitivity to contractile stimulus with endothelin (ET)-1 [36]. Blast force-dependent, prolonged hypercontraction and vascular smooth muscle phenotype switching occurred the next day. At the time of aneurysm rupture, the parent artery suffers stretching due to antagonism by the burst of ruptured blood. Interestingly, during clipping surgery, suction of the arterial wall or manipulation of the artery can produce the same phenomenon as is seen in blast injury in an artery.

Insult due to aneurysmal rupture, severe hypertension, and manipulation due to surgery can all produce phenotypic transformation of SMC and lead to CVS.

Inflammation

Subarachnoid hematoma induces an inflammatory reaction in vessels and in brain parenchyma. Human SAH cases in patients who developed angiographically demonstrated vasospasm showed significantly elevated C-reactive protein (CRP) levels in serum and cerebrospinal fluid (CSF), and increased CRP measurements were closely associated with poor clinical outcome [37]. Temporal changes in CSF interleukin(IL) -1β, IL-6, and tumor necrosis factor-α (TNF-α) were evaluated in cases of human SAH [38]. Inflammatory cytokines are closely associated with the development of increased blood flow velocities in the basal cerebral vessels, as recorded by transcranial Doppler sonography. Moreover, intrathecal secretion of inflammatory cytokines was significantly increased in patients with poor clinical outcomes [39]. In particular, elevation of IL-6 significantly correlated with poor outcome. Also, mast cells migrate into the muscular layer of cerebral arteries after human SAH [40]. Mast cells contain heparin, histamine, hydrolytic enzymes, and serotonin in metachromatic cytoplasmic granules that induce inflammation and catabolism.

Expressions of IL-1β and TNF-α were measured in lipopolysaccharide-stimulated monocytes derived from human SAH cases [41]. Activation of IL-1β was significantly higher in patients with symptomatic vasospasm than in patients without vasospasm, while the IL-1β activation index correlated with the degree of postoperative angiographic vasospasm. Specific analyses of inflammation were undertaken with animal models and cultured cells. Neutrophil recruitment in the brain is IL-1 dependent, and transendothelial migration across IL-1–stimulated brain endothelium triggers neutrophils to acquire a neurotoxic phenotype that causes the rapid death of cultured neurons [42]. These migrated neutrophils also released decondensed deoxyribonucleic acid (DNA) associated with proteases, which comprises neutrophil extracellular traps. Transcription of inflammatory genes (IL-1β, IL-6, TNF-α, chemokine C-X-C motif ligand 1 and 2, and chemokine C-C motif ligand 20) increased at around 1 to 3 h after the SAH [43]. Transcription of extracellular matrix-related genes (matrix metalloproteinase (MMP) 8, MMP 9, MMP 13, and inducible nitric oxide synthase (NOS)) also increased at around 6 h and expression declined slowly [43]. Leukocyte–endothelial cell interaction during the development of vasospasm and leukocyte migration into the subarachnoid space release proinflammatory cytokines, oxygen radicals, and ETs [44]. These effects increase leukocyte migration, deplete nitric oxide (NO)-related vasodilation, and compound inflammation-induced arterial contraction and chronic vasospasm.

In 1999, a full-length complementary DNA (cDNA) for the previously unknown clone DVS 27, whose expression was highly up-regulated, was isolated from the cerebral artery cDNA library by hybridization in a canine SAH model [45]. This gene is now named IL-33 and is in focus as a therapeutic target in cardiovascular diseases [46]. The IL-33 is a member of the IL-1 family of cytokines that promote T helper cell type 2 immune responses by signaling through the ST2L (a trans-membrane receptor) and IL-1 receptor accessory protein dimeric receptor complex. IL-33 also affects endothelial cells, inducing IL-6, IL-8, and NO production, promoting angiogenesis and increasing vascular permeability. IL-1β has been shown to cause progressive proliferation of human infragenicular SMC over a 4-day period and an increase in fibronectin production [47].

One paper evaluating the effectiveness of erythropoietin in SAH showed that erythropoietin improved cerebral blood flow and microcirculatory flow. The mechanism of this effect was the suppression of caspase-3 expression, an increase in the hematocrit, and the expression of endothelial nitric oxide synthase [48]. Additionally, bilirubin oxidation products play a role in the vascular remodeling that occurs following aneurysmal SAH [49].

Inflammatory responses play a major role in the development of several types of vasospasm. These insults throw the switch for phenotypic transformation of SMC (Fig. 1b).

Thrombin

Thrombin and oxyhemoglobin are present at high concentrations in the CSF after SAH. Both increase ET1, which is a strong inducer of CVS and is secreted from endothelial cells and SMC [50]. Bilirubin oxidation products also altered cultured SMC morphology and metabolic activity, both of which are related to vascular remodeling [49].

In the rat SAH model, expression of thromboxane A2 (TX A2) receptors in the smooth muscle layer of basilar arteries increased significantly 48 h after SAH, and TXA2 expression of intraparenchymal microvessels increased as compared to the sham group [51]. TXA2 triggers vascular smooth muscle cell mitogenic activity, connective tissue synthetic activity, and contractile activity [52]. Also, TXA2 produces platelet aggregation. TXA2 mediates both arterial vessel wall proliferation and vasospasm. Inhibition of thrombin activity leads to the amelioration of cerebral vasospasm and suppression of mitogen-activated protein kinase (MAPK) diphosphorylation in a rabbit SAH model [53]. This suggests that thrombin and its related signal transduction, including the MAPK cascade, appear to play an important role in the pathogenesis of cerebral vasospasm after SAH.

For the prevention of vasospasm, heparin inhibited the sustained activity of mitogen-activated protein kinase-1 and mitogen-activated protein kinase and prevented DNA synthesis induced by thrombin, angiotensin II, endothelin-1, and lysophosphatidic acid [54]. Heparin suppressed phenotypic changes in SMC associated with proliferation in cell culture [55]. Unfractionated heparin complexes with oxyhemoglobin that block the activity of free radicals, including reactive oxygen species, antagonize endothelin-mediated vasoconstriction, smooth muscle depolarization, and inflammatory, growth, and fibrogenic responses [56].

SMC Proliferation and Growth Factors

Subendothelial granulation after SAH was reported by Conway and McDonald in 1972 [57]. Intimal changes were observed on day 4 of SAH, and subendothelial proliferation survived at 4 months after angiographically severe vasospasm. Arteriolar morphology showed a large decrease in SMC length and increases in width [32]. The increased number of cells in the media and adventitia of arteries subject to vasospasm was evaluated in dog basilar arteries in a double-hemorrhage SAH model [58].

Ohkuma et al. investigated vascular remodeling that occurred during cerebral vasospasm after SAH in a canine model [59]. They showed that increased beta-actin messenger ribonucleic acid (mRNA) expression, and its structural changes in the 3′ untranslated region in the vasospastic basilar artery were prominently seen 7 and 14 days after SAH, accompanied by an increased area of tunica media in the basilar artery. The results suggest that vascular remodeling occurs and takes part in luminal narrowing during cerebral vasospasm. On the other hand, MacDonald and Weir dispute this phenomenon; they found no significant increase in the area of the tunica media in the cross-sectional areas of dog basilar arteries seven days after SAH [32]. The increase in wall mass is not severe enough to be the primary cause of narrowing of the lumen, but it may reflect other remodeling changes that do somehow contribute to vasospasm [58].

SMC and fibroblast proliferation of arteries surrounded by blood clot was observed after SAH in the mouse model, and this cellular replication was observed in conjunction with platelet-derived vascular growth factor (PDGF) protein at the sites of thrombus (Fig. 1c) [60]. PDGF-AB concentrations in CSF from human SAH cases were significantly higher than those from non-SAH patients and normal controls. In human pial arteries, a localized thrombus stimulated vessel wall proliferation. The proliferation was blocked by neutralizing antibodies directed against PDGFs. Vascular endothelial growth factor (VEGF) in CSF also increased markedly in SAH patients. Hypoxia and VEGF increase both nonmuscle myosin heavy chain (MHC) and smooth muscle MHC in adult ewes. These factors go on to induce arterial remodeling [61].

Ohkuma et al. were the first to reveal that phenotypic modulation of SMC and vascular remodeling of intraparenchymal small cerebral arteries occurred after SAH [62]. Seven to 14 days after canine experimental SAH, the amount of beta-actin mRNA, evaluated by northern blot analysis, increased in intraparenchymal perforating arteries, the structural change in the 3′ untranslated region of beta-actin mRNA detected by polymerase chain reaction analysis exhibited enhanced expression, and immunohistochemistry showed marked induction of the embryonal isoform of myosin heavy chain accompanied by decreased expression of smooth muscle myosin heavy chain. Histological morphometric analysis showed an increase in the area of the arterial wall without changes in the number of SMC nuclei (Fig. 1c). These changes may affect cerebral blood flow after SAH by inducing increased cerebrovascular resistance.

Yamaguchi-Okada et al. performed an experiment with a two-hemorrhage canine SAH model [63]. They reported that maximal contraction capacity decreased until day 21 and arterial stiffness progressed until day 28. Medial thickening and an increase in connective tissue continued until day 21 and returned to control findings by day 28 (Fig. 1d). Enhanced expression of embryonal nonmuscle MHC continued from days 7 to 21 and contractile types of smooth muscle MHC (SM1 and SM2) disappeared on days 14 and 21. They concluded that biomechanical and phenotypic changes may play a pivotal role in sustaining cerebral vasospasm.

Several growth factors cause SMC to proliferate and migrate. These changes occur not only in the subarachnoid space but also in the intraparenchymal area. Phenotypic changes remain prolonged for 14 days after the SAH.

Calcium and Potassium

In 2006, Macdonald et al. reported that calcium sensitivity decreased during vasospasm after SAH in a dog model, and other mechanisms were involved in maintaining the contraction of the basilar artery [64]. They concluded that reduced compliance seems to be due to an alteration in the arterial wall extracellular matrix rather than in the smooth muscle cells, because reduced compliance could not be alleviated by depolymerization of smooth muscle actin.

The basilar artery of a dog SAH model showed significant down regulation of the voltage gated K + channel Kv 2.2 and of the β1 subunit of the large-conductance, Ca2+-activated K + (BK) channel [65]. Changes in mRNA levels of Kv 2.2 and in the BK-β1 subunit correlated with the degree of vasospasm. K + channel dysfunction may contribute to the pathogenesis of cerebral vasospasm. Basilar artery SMC from of dog double-hemorrhage SAH increased the expression of KIR 2.1 mRNA and protein during vasospasm after experimental SAH, suggesting that this increase is a functionally significant adaptive response that occurs to reduce vasospasm [66].

Calcium and potassium influence SMC contraction, but these cations do not directly affect the phenotypic transformation of SMC.

Transient Receptor Potential Channel

Smooth muscle cells possess at least three different types of Ca2+-permeable channels, i.e., voltage-dependent Ca2+ channels (VDCC), stretch-activated Ca2+ channels, and receptor-activated Ca2+ channels, which include receptor-operated and store-operated Ca2+ channels (SOCs). Transient receptor potential channel (TRPC) 1 acted as a key signal for store-operated Ca2+ entry (SOCE), and TRPC1 up-regulation produced angiotensin II and subsequent nuclear factor-kappaB activation, which induced human coronary artery smooth muscle cell hypertrophy [67]. The function of the stromal interaction molecule 1 (STIM1)–TRPC1 complex was an essential component of SOCE and it is also involved in vascular SMC proliferation [68]. Rat hyperplastic smooth muscle cells exhibited up-regulated TRPC1 associated with enhanced calcium entry and cell cycle activity 5 days after vascular injury [69]. A blockade of TRPC1 reduced neointimal growth in human vein, as well as calcium entry and proliferation of smooth muscle cells in culture. Another human vascular smooth muscle cell study revealed a complex situation in which there is not only an interaction of plasma membrane-spanning STIM1, which is important in cell migration and TRCP 1-independent store-operated cationic current, but also a TRPC1–STIM1 interaction, a TRPC1-dependent component of store-operated current, and STIM1-independent TRPC1 linked to cell proliferation [70]. In a rat two-hemorrhage SAH model, elevated mRNA and protein expressions of TRPC1 and STIM1 were detected after SAH and peaked on days 5 and 7, following a time course parallel to the development of cerebral vasospasm [71]. Enhanced expression of proliferating cell nuclear antigen and the disappearance of SMα-actin from days 1 to 7 after SAH were revealed by immunohistochemical study. TRPC1 and STIM1 mediated the Ca2+ influx and phenotypic switching in SMC. Xie et al. examined the relationship between endotherin-1 and the TRPC family (TRPC1α, TRPC1β, TRPC2 ~7) in a dog SAH model. They found ET-1 significantly increases Ca2+ influx mediated by TRPC1 and TRPC4 or their heteromers in smooth muscle cells, which promotes the development of vasospasm after SAH [72].

The TRPC family relates positively to CVS and phenotypic transformation of SMC. This field of study will become central in the search for new mechanisms of vasospasm.

Clinical Trial for the Prevention of Phenotypic Transformation of SMC in CVS

The use of statins for CVS is controversial. In an experiment on mice, simvastatin treatment before and after SAH attenuated cerebral vasospasm and neurological deficits due to endothelial NOS up-regulation [73]. But a randomized, double-blind, placebo-controlled pilot study of simvastatin and a systematic review revealed no statistically significant support to the finding of a beneficial effect of statins in patients with aneurysmal subarachnoid hemorrhage [74, 75]. On the other hand, the Rho-kinase inhibitor, fasdil chloride, reduced the incidence of CVS in humans [76]. Long-term inhibition of Rho-kinase also suppresses intimal thickening of vein grafts in rabbits [77]. A combination of statin and fasudil prevents CVS and phenotypic transformation of SMC [78].

An antagonist of the ET receptor is effective in preventing CVS. Clazosentan at 5 mg/h had no significant effect on mortality and vasospasm-related morbidity or functional outcome in a CONSCIOUS-2 trial [79]. But in a CONCIOUS-3 trial clazosentan prevented CVS and improved outcome in SAH dose-dependently [18]. Of course, detecting the prevention of phenotypic transformation of SMC is impossible in such phase 3 trials, but clazosentan can, in theory, reduce phenotypic transformation of SMC.

Cilostazol selectively inhibits phosphodiesterase 3 present in SMC and platelets, resulting in an increase in intracellular cyclic adenosine monophosphate. Kawanabe et al. examined mouse SMC and arteries to demonstrate the dual-antagonizing effects of cilostazol on vasoconstriction and cell proliferation induced by endothelin [80]. Cilostazol blocked endothelin-induced extracellular calcium influx and produces both effects. Although cilostazol does not inhibit endothelin-induced intraorganellar calcium release, blockade of extracellular calcium influx is sufficient to blunt endothelin-induced vasoconstriction. Two prospective, randomized trials from Japan report on the effectiveness of cilostazol in cerebral vasospasm after aneurysmal subarachnoid hemorrhage [11, 12]. Both trials led to the conclusion that cilostazol significantly reduced symptomatic CVS, asymptomatic CVS, and infarction. The clinical outcome of medicated cases improved. Cilostazol has the potential to prevent phenotypic transformation of SMC and reduced CVS.

Conclusion

In this paper, we have reviewed the phenotypic transformation of SMC in vasospasm. While considerable effort was expended to evaluate the overall mechanism of CVS, we did not reach this goal. We will endeavor to evaluate the unknown mechanism of CVS and bring clarity to this complicated sequela.