Molecular, Cellular, and Genetic Determinants of Sporadic Brain Arteriovenous Malformations

Brain arteriovenous malformations (AVMs) are a significant cause of hemorrhagic stroke in children and adults.1 They can occur as part of hereditary syndromes such as capillary malformation–AVM syndrome and hereditary hemorrhagic telangiectasia, in which they result from mutations in genes with known or plausible roles in angiogenesis and vascular remodeling such as RASA1 and ACVRL1.2-6 Less is known about the cause of sporadically occurring AVMs, which account for the vast majority of lesions in the general population. The formation of these sporadic lesions is theorized to result from several possible diverse elements that could contribute to a common AVM phenotype, including gain- or loss-of-function genetic mutations,7,8 inflammation,9-11 impaired blood-brain barrier integrity,12,13 or a response to injury phenomenon.14 We are currently investigating sporadic AVM pathogenesis on these related fronts. GENETICS The process by which AVMs develop, progress, and eventually hemorrhage is poorly understood. However, aberrations in blood vessel formation and segregation during embryonic development or adult life are thought to be primordial factors.15 In vascular development, the establishment of a vascular identity (be it arterial or venous) initiates from molecular signals that result in functional, and subsequently structural, changes.16,17 Several hierarchical signaling pathways promote or inhibit divergent endothelial cell fates, including hedgehog,18 vascular endothelial growth factor (VEGF),19,20 notch,21 transforming growth factor-β,22,23 and the ephrin ligand-receptor pathway,24 among others.25 These pathways are crucial regulators of vascular assembly, differentiation, and boundary formation and have complex interactions during both development and adult life. Syndromes with mendelian inheritance patterns such as hereditary hemorrhagic telangiectasia are associated with known mutations in the transforming growth factor-β superfamily signaling pathways that result in AVMs. This provides the first clue that a causative genetic basis for the sporadic lesions is possible. However, patients with sporadic AVMs have failed to demonstrate germline mutations or structural variation, and causative somatic mutations in candidate genes (ACVRL1, ENG, SMAD4) have not been identified.26 Several studies have characterized various single-nucleotide polymorphisms in these genes, although their role in affecting gene function and gene expression or conferring disease susceptibility is not well elucidated.27-30 Further study with available tools, namely next-generation sequencing technology, holds the potential to aid in the discovery of genetic alterations not previously identified and to overcome many of the limitations of conventional techniques such as hypothesis-driven genotyping and DNA microarray.31,32 In other diseases of the central nervous system in which both sporadic and hereditary forms occur, advanced genomic techniques have been used to identify mutations in genes acting in similar pathways (Table 1). Analysis of next-generation sequencing data has been able to identify focal genetic alterations in sporadic hemangioblastoma,33 port wine stain,34 and meningioma,35,36 which are shared with known germline mutations in their syndromic forms. This confirms the long-standing hypothesis that these sporadic phenotypes are genetically related to their hereditary counterparts.TABLE 1: Relationship in Gene Alterations Between Sporadic Lesions and Their Associated SyndromesCould sporadic AVMs be linked to the mutations that are known to cause hereditary hemorrhagic telangiectasia or other vascular malformation syndromes? Two main advantages of next-generation sequencing are being used to answer this question. First, a comprehensive and unbiased analysis of the genome is able to detect alterations in gene pathways outside of those typically implicated in angiogenesis and vasculogenesis yet result in a similar AVM phenocopy. Second, next-generation sequencing has the sensitivity to detect rare mutations in heterogeneous AVM samples. Surgical AVM samples are a diverse mix of vascular cells (endothelial cells, pericytes, and smooth muscle cells), circulating cells (lymphocytes), glial cells, and fibroblasts, with the exact pathological cell population being unknown (Table 2).39-42 Therefore, if the potential causative mutation occurs in only a single cell type (ie, endothelium), the purity of the sample analyzed is low, and the chance of finding that mutation in any single reading of the DNA sequence is proportionally low. Next-generation sequencing overcomes this limitation by reading DNA sequences many times over, with “deep sequencing” of several hundred–fold coverage being used to effectively identify low-frequency events in heterogeneous samples.33,45 With the spectacular 6-log drop in the cost of sequencing over the last 12 years, next-generation sequencing platforms are a powerful and cost-effective means for discovering the seemingly cryptic genetic basis of sporadic AVMs.TABLE 2: Cell Populations Implicated in Arteriovenous Malformation PathogenesisaINFLAMMATION AVMs are dynamic lesions, and even though the majority of them appear to remain stable, they are well known to disappear,46-49 grow,50-52 and even recur after removal.53-55 If genetic alterations are the predisposing events in sporadic AVM pathogenesis, inflammation and its consequences may be one propagating force that advances growth, maintains their architecture, or even results in AVM hemorrhage. Genetics and inflammation are closely tied to each other; multiple single-nucleotide polymorphisms have been associated with increased levels of proinflammatory cytokines such as tumor necrosis factor-α, interleukin-6, interleukin-1 α, and interleukin-1β.9 How this hyperzealous, local, proinflammatory milieu, presumably incited through some form of injury such as stroke or trauma, results in AVM development or propagation is yet to be determined. It is possible that these local cytokines result in the chemotaxis of neutrophils and macrophages, which are frequently identified in the vascular wall of resected AVM tissue.43 Clinically, advanced magnetic resonance imaging (ferumoxytol-enhanced magnetic resonance imaging) has been able to noninvasively identify macrophage localization to AVM vessel walls.10,56 These recruited inflammatory cells then secrete a number of signaling or enzymatically active molecules, leading to further inflammation or destabilization of the vascular wall. Local secretion of interleukin-6 and interleukin-8 increases local matrix metalloproteinase expression and increases proliferation, migration, and survival of endothelial and smooth muscle cells.57-59 Secretion of leukocyte-derived myeloperoxidase and matrix metalloproteinase-9 leads to degradation of a number of vascular basement membrane components, including collagen, fibronectin, laminin, and other adhesion molecules.60 Inflammatory cytokines are also known to increase the expression of VEGF61 and angiotensin (Ang)-2,62 which are key downstream ligands responsible for promoting angiogenic responses in neoplasia and other conditions.37,63 These downstream pathways are also possible targets for therapeutic intervention because inhibition with agents in clinical use and preclinical development has the potential to disrupt the key steps in AVM propagation by normalizing the structure and function of abnormal blood vessels.64,65 A human clinical trial of bevacizumab for brain AVM is currently underway and recruiting patients (ClinicalTrials.gov identifier, NCT02314377). Alternatively, a secondary local inflammatory state (reactionary phenomenon) as a result of pathological hemodynamic forces may be responsible for the observed local changes in cytokines and leukocytes. Altered hemodynamic shear stress has been shown to be a direct determinant of endothelial function and phenotype in a range of vascular pathophysiologies,66,67 and the turbulent high-flow nature of AVMs is unlikely to be an exception. This makes the cause-and-effect nature of inflammation and AVM pathology difficult to discern, and the pathological hemodynamics within AVM tissue likely alter the gene expression profile sufficiently to complicate the interpretation of observational studies (Figure 1).FIGURE 1: Dense cellular extravascular inflammatory infiltrate in brain arteriovenous malformations. Representative confocal microscopy images of (A) control temporal lobe and (B) resected temporal lobe arteriovenous malformation at low magnification (×10) showing endothelium (CD31 platelet endothelial adhesion molecule, green), vascular smooth muscle cells (α-smooth muscle actin [SMA], red), and Hoechst-positive nuclei (blue).An emerging area of interest in AVM-related inflammation is the contribution of the adaptive, cell-mediated immune response. The abnormal presence of T cells has been demonstrated in AVM vessels, but their nature or purpose remains an enigma.11,43 Using immunohistochemistry and flow cytometry, researchers are differentiating subtypes of T cells (eg, Th1, Th2, Th17, regulatory T cells) in previously untreated brain AVMs. Further examination of local effector secreted cytokines, intracellular cytokine staining, and transcription factors in the AVM microenvironment will be targeted on the basis of cell populations discovered. Because inflammation, hemodynamic stress, genetic variation, and angiogenesis are closely tied with one another, the exact contribution of inflammation to AVM pathogenesis remains an area of intense investigation.44 BLOOD-BRAIN BARRIER Normal cerebrovascular structure and function are dependent on coordinated signaling of multiple cell types, including endothelial cells, mural cells (vascular smooth muscle cells and pericytes), immune cells, glia, and neurons.68-70 Pericytes are the principal mural cell population at the capillary level, covering roughly 80% to 90% of the capillary wall. Pericytes fulfill a modulatory role in a number of integral cerebrovascular functionalities that are disrupted in AVMs, including regulation of brain angiogenesis, blood vessel diameter, vascular wall stability, and integrity of the blood-brain barrier.68,69,71 Not surprisingly, pericytes are reduced in murine AVM models.12,72 Taken a step further, when pericyte-deficient rodents are examined, vessels lacking pericytes display focal or diffuse dilatations with increased tortuosity and multiple findings consistent with heightened vascular fragility or permeability, including microaneurysms, overt hemorrhage, chronic leakage of circulating plasma-derived proteins, and heightened extravasation of blood-borne immune cells,68,73-79 findings that are consistently observed in human AVM tissue specimens.9,80-82 In 1 qualitative study to date, pericytes were shown to be reduced in human perinidal tissue specimens.12,72 However, pericytes are normally confined to smaller vessels, and the relative contributions of vascular smooth muscle cells and pericytes to AVM development and propagation are unclear. To date, no studies have quantified mural cells (pericytes and vascular smooth muscle cells) in AVM specimens, and the relationship between a hypothesized loss of mural cells and increased AVM permeability (ie, microhemorrhage) is unknown. Endothelial–mural cell signaling in AVM biology is not a standalone entity, and there is significant cross-talk between both cell types, with many angiogenic factors central to inflammatory pathways. For example, the effects of Ang2 on the vasculature can be either proangiogenic or antiangiogenic, depending on the presence of VEGF.83,84 In the presence of VEGF, Ang2 induces pericytes to dissociate from existing vessels, making them leaky and allowing the extravasation of other proangiogenic factors. These events lead to sprouting and the formation of new blood vessels. In the absence of VEGF, Ang2 induces loss of pericyte coverage and stimulates vessel regression, culminating in an antiangiogenic effect. Thus, because of this duality, the balance between Ang1, Ang2, and VEGF is extremely important in mediating angiogenesis and may play a central role in AVM development and progression. In addition to angiopoietin and VEGF cell signaling pathways, the recruitment of mural cells to nascent vascular tubes during angiogenesis is dependent on platelet-derived growth factor B and platelet-derived growth factor receptor β signaling,73-76,78,85-87 further highlighting the complexity of investigations needed to dissect these pathways. A reduction in AVM mural cell population not only may represent a gateway to pathological angiogenesis but also could be associated with lesion destabilization and hemorrhage (Figure 2).FIGURE 2: Cellular constituents of the vascular wall in the normal cerebral vasculature and arteriovenous malformations (AVMs). A, cerebral arteries are made up of a continuous endothelial cell lining (green) connected via tight and adherens junctions and then further surrounded by a thick vascular basement membrane (yellow) and concentric rings of vascular smooth muscle cells (blue). B, in brain AVMs, there is hyperproliferation of endothelial cells (green). Focal areas of vascular smooth muscle cell (blue) proliferation and degeneration are also noted. Thinning of the basement membrane (pale yellow) and vascular smooth muscle cell degeneration contribute to the destabilization of the vascular wall, giving rise to microscopic and macroscopic hemorrhage.CONCLUSION This multifaceted approach to investigating sporadic brain AVM pathophysiology considers the interconnected, contributory roles of diverse biological processes based on research from the past 2 decades. Ongoing investigation using human surgical specimens and refinement of animal models will continue to advance our understanding of how AVMs develop and eventually rupture.38,65,88,89 A disease model that incorporates how genetics, inflammation, and impaired blood-brain barrier integrity conspire to generate AVMs will be essential to the development of screening protocols, the ability to differentiate high- from low-hemorrhage-risk lesions, and the identification of novel pharmacological agents. Disclosures Financial support for the research in this report was provided by the Congress of Neurological Surgeons (Christopher Getch, MD, Research Fellowship) to Dr Walcott. Dr Rouleau holds a Canada Research Chair in Genetics of the Nervous System and the Wilder Penfield Chair in Neurosciences. Dr Lawton is the principal investigator of multiple projects funded by National Institutes of Health grant U54 NS065705-07. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.