1 Introduction

Moyamoya disease (MMD) is an occlusive cerebrovascular disease that is characterized by idiopathic and progressive stenosis of the distal portion of major intracranial arteries with abnormal basal collaterals (“moyamoya” vessels) [1]. The internal carotid artery (ICA) and its major branches (anterior carotid artery [ACA] and middle cerebral artery [MCA]) are mainly involved, but basilar artery (BA) and posterior cerebral artery (PCA) stenosis is also detected in some patients, especially during follow-up [2]. When the vascular changes are associated with well-recognized conditions, it is called moyamoya syndrome (MMS). As the disease is defined by the radiological morphology of vessels, MMD may not be a homogeneous entity but may instead comprise a heterogeneous group of conditions.

A diagnosis is made based on symptoms and radiologic evaluation. The vessel morphology and presence of infarction are evaluated using brain magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA) [3, 4]. Transfemoral cerebral angiography is the gold standard modality to confirm the diagnosis. The regional blood flow status of the brain is usually checked with perfusion MRI or single photon emission computed tomography (SPECT).

The prevalence of MMD shows a bimodal age pattern: once in mid-childhood and again during the late 40s [3]. The presentation and treatment policy of pediatric and adult MMD are different. Hemorrhagic presentation is very rare in children, but it is not uncommon in adult patients [5].

The common sites of intracranial hemorrhage are the basal ganglia, thalamus, and near the lateral ventricle wall, so possible role of moyamoya vessels has been suspected. Some reports have shown that in cases with no moyamoya vessel dilatation, other arteries such as anterior choroidal artery or posterior communicating artery supplying the hemorrhage area showed dilatation [6]. According to morphometric evaluation using autopsy specimens, the dilated arteries in MMD seem to have fibrosis and attenuation of the media, with irregular segmentation of the elastic lamina. Under the condition of hemodynamic stress or aging, the attenuated walls in these dilated arteries may be predisposing factor for microaneurysm formation. The rupture of the microaneurysm is hypothesized to cause bleeding in MMD patients [7].

For pediatric patients, indirect revascularization is the primary treatment of choice, but in adults, it is controversial whether indirect revascularization is equivalent to direct bypass procedures [5, 8,9,10,11,12,13,14].

Much work has been done to elucidate the pathophysiology of MMD over the years. Due to the inaccessibility of the involved vessel during surgery, blood, cerebrospinal fluid (CSF), and urine have been used for molecular and genomic analyses. The attempts to establish an animal model of the disease have yet to be successful, but recent technical advances in genomic analysis have shed light on the field. This chapter covers the current knowledge of the pathophysiology of MMD in various aspects: proteins (growth factors and cytokines), cells, autoimmunity, biomechanics, and genes.

2 Pathology

Autopsy specimens of the stenotic arteries revealed that the fibrocellular thickening of the intima, an irregular undulation (“waving”) of the internal elastic lamina, and attenuation of media are the main histopathological findings of MMD. Smooth muscle cell hyperplasia is typically found in the thickened intima. These smooth muscle cells are considered to be the synthetic type and to migrate from the media [15,16,17]. In contrast to atherosclerosis, inflammatory cell infiltration and the presence of lipid-laden foamy macrophages are not found in the stenotic vessels of MMD.

The moyamoya vessels, the basal collaterals that are dilated perforating arteries, show fibrin deposits in the wall, fragmented elastic lamina, attenuated media, and the formation of microaneurysm [16, 18].

3 Cytokines

As the hallmark of the vessel pathology is the proliferation of smooth muscle cells in the intima of the arteries, the involvement of growth factors has been suspected. In addition, both the formation of basal collaterals during disease progression and the extensive neovascularization that is seen after indirect bypass surgery, which involves the mere relocation of a vascularized tissue on the surface of the brain cortex, imply the potential role of an aberrant “abundance” of angiogenic factors in MMD patients [19]. Hence, one of the earliest studies to investigate the pathophysiology of MMD examined the expression levels of various cytokines that were quantified in the CSF, serum, or tissue of MMD patients.

Increased levels of growth factors such as basic fibroblast growth factor (bFGF), hepatocyte growth factor (HGF), hypoxia-inducing factor-1α (HIF-1α), granulocyte colony-stimulating factor (GCSF), transforming growth factor-β (TGF-β), and vascular endothelial growth factor (VEGF) have been observed in the CSF and serum of MMD patients [20,21,22,23,24].

The cytokines associated with angiogenesis and vascular repair have also been evaluated. Matrix metalloproteinases-9 (MMP-9) was increased, and the tissue inhibitor of metalloproteinase (TIMP1) was decreased in the serum of MMD patients, suggesting a disruption in the balance between the two. The imbalance may cause aberrant vascular smooth muscle cell dynamics, which can lead to MMD [20, 21].

Increased levels of the cellular retinoic acid-binding protein-1 (CRABP-1) were detected in the CSF of MMD patients [23]. CRABP-1 is one of the proteins that mediates the biological activity of retinoic acid (RA) and may provide a link between RA signaling and MMD pathogenesis. As retinoids are known to negatively regulate growth factor-stimulated vascular smooth muscle cell proliferation and differentiation, it can be hypothesized that increases in CRABP-1 may inhibit the RA signaling action, resulting in an increase in vascular smooth muscle cell proliferation [23].

It should be noted that although an abundant amount of data implies an important role for the various cytokines in MMD, there is possibility that the differences in the expression levels of cytokines may actually be a compensatory response to the pathology rather than the cause.

4 Circulating Progenitor Cells

Investigating the vessel wall cells (endothelial-lineage or smooth muscle type) of MMD patients may be a direct approach to evaluate the pathogenesis of the disease. However, it is nearly impossible to obtain the cells directly from patients. Fortunately, vascular progenitor cells can be derived from the peripheral blood of patients, thus providing an experimental cell model (Fig. 6.1) [25].

Fig. 6.1
figure 1

Vascular progenitor cells. (a) Early endothelial progenitor cells (EPCs), also called colony-forming units or cell clusters (×200). This cell cluster is composed of a central core of round cells that are surrounded by spindle-shaped cells. (b) Late EPCs, also known as endothelial outgrowth cells (×40). The cells are arranged in cobblestone-like formations. (c) Smooth muscle progenitor cells (SPCs) (×40). These cells appear elongated and have a typical hill-and-valley appearance. Adapted from Journal of Korean Neurosurgical Society [25]

Full size image

Endothelial progenitor cells (EPCs) have been evaluated extensively in many vascular diseases before MMD because these cells are known to originate from the bone marrow and help maintain the vasculature and blood flow in infarcted areas [26,27,28,29,30]. EPCs are commonly characterized by the expression of the surface proteins CD34, CD133, and vascular endothelial growth factor receptor-2 (VEGFR-2). Some studies also consider CD45 and CD114 to be additional markers for defining EPC [31, 32]. Two studies have reported on the EPC count in the peripheral blood of MMD patients. One study, which included mainly adult patients, showed that MMD patients had more EPCs than the normal controls did [29], whereas another study with pediatric patients demonstrated a decreased number of EPCs in patients [33]. The contradictory results may be due to differences in the patients’ age groups or in the markers used to isolate EPCs [34]. Additionally, the number of colony-forming units (CFU) of EPCs was shown to be decreased and that of outgrowth cells was increased in MMD patients; these changes were more profound in those patients with advanced diseases. Further, the tube formation ability was significantly decreased in the late EPCs from MMD patients, suggesting the dysfunction of the EPCs of MMD patients [29, 33]. More direct evidence on the contribution of EPCs in MMD was provided when cells expressing CD34 and VEGFR2 were found in the thickened intima of distal ICA samples collected from MMD patients [35].

The role of EPCs in MMD was further analyzed in a recent study that compared the gene expression profiles of EPCs from MMD patients and normal controls. A downregulation of retinaldehyde dehydrogenase 2 (Raldh2), which is one of the enzymes in the physiological process to convert RA from retinol, was noted. Based on the previous link between RA signaling and MMD, the effect of Raldh2 on the function of EPCs was analyzed. It was shown that decreased levels of Raldh2 in normal EPCs resulted in poor tube formation, similar to EPCs from MMD. Furthermore, the disrupted tube formation capacity was restored by supplementation with exogenous RA in both MMD EPCs and the normal EPCs with downregulated Raldh2 [36].

The possible role of smooth muscle progenitor cells (SPCs) in MMD has recently been evaluated to determine the origin of smooth muscle cell hyperplasia in the thickened intima. SPCs differ from EPCs in the appearance of the outgrowth cells, namely, “hill-and-valley” for the former and “cobblestone” for the latter (Fig. 6.1) [37]. Additionally, SPCs stain positive for smooth muscle-specific markers, such as smooth muscle actin-α, smooth muscle myosin heavy chain (MHC), and calponin [38]. The SPCs from MMD expressed lower levels of platelet-derived growth factor receptor (PDGF)-α, MHC, and calponin than did normal controls. Furthermore, in the tube formation assay, MMD SPCs made more irregular and thick tubes, reminiscent of the pathologic arteries in MMD patients (Fig. 6.2) [39]. These findings suggest that SPCs of MMD patients may have a defect in the cell maturation process that results in dysfunctional vasculogenesis.

Fig. 6.2
figure 2

The networks formed by smooth muscle progenitor cells (SPCs). Photomicrographs showing immunofluorescence staining and tubule formation of SPCs on Matrigel. The cells are labeled with red fluorescent dye (PKH26). Compared with the SPCs obtained from a healthy volunteer (a), the SPCs obtained from a moyamoya disease patient (b) have rather irregularly arranged tubules of varying sizes. In some areas, thickened tubules are noted. Adapted from Journal of Korean Neurosurgical Society [25]

Full size image

A recent study investigated the interaction between the precursors of the two major cellular components’ vessel walls. The results showed that EPCs are mostly responsible for the dysfunction of both EPCs and SPCs in MMD and that the enhanced migration of the SPCs to the EPCs is mediated by the release of chemokine (C-C motif) ligand 5 (CCL5) by the EPCs [40].

5 Immunologic Factors

The role of immune function in the pathogenesis in MMD has been hypothesized based on its association with immune diseases [41]. Many studies have reported on cases of MMS associated with Graves’ disease, an autoimmune disease that causes hyperthyroidism [42]. A recent meta-analysis on the literature has clearly demonstrated a close relationship between thyroid function and MMS [43]. Most of the cases were females, and most of the patients were suffering from active hyperthyroidism when they were diagnosed with MMS. Moreover, recurrence or aggravation of the MMS symptoms was more common in patients with poorly controlled thyroid function [43]. The fact that abnormal cerebrovascular autoregulation associated with sympathetic nerve excitation has been known to occur in patients with hyperthyroidism suggests a potential mechanism underlying the role of autoimmunity in MMD pathogenesis [44]. However, the possibility that hemodynamic stress caused by the thyrotoxicosis instead of autoimmune dysfunction still stands as the cause of vascular changes in these patients.

Antiphospholipid syndrome (APS) consists of various conditions that are caused by the presence of antiphospholipid (APL) autoantibodies, and systemic lupus erythematosus (SLE) is a representative disease in APS. MMS patients with SLE or increased APL autoantibodies have been reported [45]. Also some SLE patients with “lupus headaches” have been investigated for the possibility of underlying cerebral ischemia and MMS-like changes [46].

Studies to search for more direct evidence of the involvement of immunity in MMD have been performed over the last several years. An increase in the number of thyroid-related autoantibodies (but with normal thyroid function) in MMD patients has been demonstrated in both Asians and Caucasians. The recent reports are different from previous ones, in which the patients showed both hyperthyroidism and increased autoantibodies, thus providing supportive evidence that autoimmunity, instead of thyroid dysfunction, may be the main contributor in MMS [42].

6 Biomechanical Factors

The regional predilection of MMD to only involve the distal portion of major intracranial arteries (ICA and BA) is a distinguishing characteristic of the disease [44]. However, most studies on the pathophysiology of MMD have only been able to investigate systemic factors, which does not address the question of why certain arteries are involved. Biomechanical factors, including hemodynamic changes and the properties of the intracranial vasculature, have been hypothesized as one of the etiologies [4]. Many techniques, such as perfusion MRI, SPECT, and transcranial Doppler, have been used to predict the regional blood perfusion and hemodynamic changes in MMD, but the data were only useful for detecting the “resulting” changes in regional cerebral perfusion [47,48,49].

Computerized mathematical models have been used to analyze the hemodynamics of cerebral arteries [50]. Only one study is available in the literature, and it utilized computational modeling to elucidate the underlying mechanism of the regional specificity of MMD [51]. Based on data obtained for cardiovascular diseases, the authors hypothesized that a chronic, constant state of relatively low shear stress causing turbulent flow to the vessel wall may result in the stimulation and proliferation of smooth muscle cells of the intima of the vessel. After establishing models of ICA and BA with the respective distal branches of ACA/MCA and PCA, the shear stress at various points, including the predisposing areas of stenosis at the bifurcation of distal branches, was determined. The stimulation results revealed that the predisposed areas showed lower values of shear stress than do regions of ICA or BA that are more distant from the bifurcation. Although the computational models have a critical limitation in recapitulating the real in vivo phenomenon, this study is noteworthy because it shows the potential biomechanics to explain the regional predilection of MMD. A retrospective clinical investigation on the delayed involvement of PCA suggested that a smaller angle between PCA and BA is significantly associated with the progressive stenosis of PCA [52]. This is in line with the computational modeling, as a smaller angle between the vessels may result in less shear stress.

7 Genetics

A strong predisposition for ethnicity and the existence of familial cases (10–15%) suggest a genetic susceptibility of the disease [34, 44]. Different modes of inheritance have been proposed, including an autosomal dominant pattern with incomplete penetrance in several parent-child cases and an autosomal recessive pattern in sibling pedigrees [53]. Various technical methods have been applied and have provided clues (candidate loci or gene) for the genetic etiology since the late 1990s. The interpretation of the genetic findings was difficult because the results were infrequently replicated in following studies. Additionally, genetic analysis on Caucasians has yet to be successful in identifying associated genes. Recent high-throughput genomic analysis has yielded a gene (Ring finger 213 [RNF213]) with a strong robust association, but the mechanism of the mutation in the pathogenesis of MMD remains unknown.

Linkage studies using genome-wide markers or previously reported loci have revealed several candidate loci: 3p24-26, 6q25, 8q21-22, 12p12-13, and 17q25 [54,55,56,57]. Only the 17q25 locus was replicated, and further genome-wide linkage analysis narrowed the candidate locus to 17q25.3. Many case-controlled association studies were performed and screened genes based on various hypotheses about the pathogenesis of the disease. As the possible contribution of autoimmune dysfunction in MMD has been suggested, the association of human leukocyte antigen (HLA)-related genes was investigated first. A significant association with HLA B51 and HLA B51-DR4 was shown in Japanese patients and HLA B35 in Korean patients [58,59,60]. In European patients, the HLA DRB1*03, DRB1*13, DRA*02, DRB*08, and DQB1*03 antigens were more frequent in patients than they were in controls [61].

The next set of genes of interest was based on the increase in the amount of growth factors and cytokines in the blood, CSF, urine, or tissues of patients. Some studies found associations with single-nucleotide polymorphisms (SNPs) in PDGFR-β promoter or TGF-β [62]. Additionally, associations with SNPs in genes related to MMPs and their tissue inhibitors have been demonstrated (TIMP2, MMP2, MMP3 genes or their promoters) [22, 63].

In 2011, two groups reported the RNF213 gene located in chromosome 17q25.3 as the strongest susceptibility gene for MMD in East Asian (Japanese and Korean) patients using two different approaches (genome-wide association study [GWAS] or whole-exome sequencing [WES]) [64, 65]. A GWAS detected a variant of the RNF213 (p.R4810K) in 95% of familial MMD patients, with 80% in sporadic cases, and demonstrated that the variant strongly increased the risk to develop MMD, with an odds ratio (OR) greater than 190 [65]. Another study utilizing genome-wide linkage analysis and WES in eight multigenerational families revealed that all of the probands had this variant. The variant was also detected in 1.4–2.4% of normal Asian controls, thus providing a possible explanation for the extreme ethnic gradient in disease prevalence [64]. The potentially strong contribution of the variant was further supported by the fact that patients with homozygous mutations of the p.R4810K variant have an earlier age of onset and more severe phenotypes than do those with heterozygous mutations [66]. Recent studies have also revealed novel variants in RNF213 in a small number of Caucasian and Chinese cases [67].

RNF213 encodes for a cytosolic protein with a really interesting new gene (RING) finger domain and a pair of two ATPase associated with a variety of cellular activities (AAA) positive ATPase modules and is speculated to have ubiquitin ligase activity [68]. The functional consequence of the mutation in RNF213 is under intensive evaluation. In an in vitro functional study, the mutation had no effect on the transcription level or ubiquitin ligase activity of the protein [64]. However, in a more recent study, the vascular endothelial cells derived from induced pluripotent stem cells were compared between normal controls and both MMD patients and RNF213 mutant carriers [69]. The latter group showed decreased angiogenic activity compared with the former. Additional experiments have shown the association of RNF213 with securin, the interferon (IFN)-beta signaling pathway, and PI3 kinase-AKT pathway in endothelial cells [69]. A series of in vivo experiments using genetically engineered animals have been performed. Knockdown of RNF213 in zebrafish resulted in abnormal vascular development, showing irregular wall formation and abnormal sprouting vessels [64]. However, RNF213-deficient mice did not show any anatomical or histopathological evidence of MMD in the brain vasculature under physiological conditions [70]. Another experiment with mice harboring the p.R4810K mutation was not successful in developing MMD under normal conditions [71]. Recently, the environmental setting was added as a crucial component in the experiment by exposing the RNF213-deficient mice to a chronic, ischemic insult. Post-ischemic angiogenesis was found to be enhanced in RNF213-deficient mice in response to chronic hind limb ischemia [72]. Nevertheless, because the role of RNF213 as the causative mutation has not been demonstrated, despite its robustness, the possibility that another causative gene is still to be found should be considered by evaluating the entire genome using whole-genome sequencing.

Investigation of the genes known to be related to other vascular diseases (e.g., coronary artery disease, stroke) or associated diseases in MMS (achalasia) has shown other potential genes. Smooth muscle alpha-actin (ACTA2) is a major component of the contractile apparatus of smooth muscle cells, and mutations of the ACTA2 gene have been known to cause familial thoracic aortic aneurysm and dissections. It was also found to be associated with pseudo-moyamoya angiopathy [73]. Mutations of the guanylate cyclase soluble subunit alpha-3 (GUCY1A3) gene have been found in MMD patients with achalasia. Loss of function mutations of this gene lead to alterations of the nitric oxide pathway in smooth muscle cells, thus supporting the possible pathogenetic mechanism of MMD [74].

Conclusions

MMD is probably caused by a combination of complex etiologies, genetics, and environmental factors. Genetic susceptibility, including RNF213 and many other factors such as the RA pathway, autoimmune process, environmental factors, and hemodynamic stress, may be related to increased levels of growth factors and cytokines. Through mechanisms that are still to be elucidated, the systemic factors and conditions may lead to the proliferation and migration of SMC into the vessel wall, thereby causing thickening of the intima. The low shear stress in the bifurcation regions may lead to the involvement of only the distal ICA and BA in MMD. Establishing an animal model will shed light on integration of the various, independent factors.