Flow induced inflammation mediated artery wall remodeling in the formation and progression of intracranial aneurysms

Abstract

Background:

Unruptured intracranial aneurysms (UIAs) are relatively common lesions that may cause devastating intracranial hemorrhage and thus cause considerable suffering and anxiety to those affected by the disease or increased likelihood of developing it. Advances in our knowledge of the pathobiology behind IA formation, progression, and rupture have led to preclinical testing of drug therapies that would prevent IA formation or progression and, in parallel, novel biology based diagnostic tools to estimate rupture risk are approaching clinical use. Artery wall remodeling, triggered by flow and intramural stresses and mediated by inflammation, is relevant to both.

Methods:

This review discusses the basis of flow driven vessel remodeling and translates that knowledge to the observations made on the mechanisms of IA initiation and progression on studies using animal models of induced IA formation, study of human IA tissue samples, and study of patient-derived computational fluid dynamic models.

Results:

Flow conditions leading to high wall shear stress (WSS) activate pro-inflammatory signaling in endothelial cells that especially through macrophage chemoattractant protein 1 (MCP1) recruits macrophages to the site exposed to high WSS. This macrophage infiltration leads to protease expression that disrupts the internal elastic lamina and collagen matrix, leading to focal outbulging of the wall and IA initiation. For the IA to grow, collagen remodeling and smooth muscle cell (SMC) proliferation are essential, since the fact that collagen does not distend much prevents the passive dilation of a focal weakness to sizable IA. Chronic macrophage infiltration of the IA wall promotes this SMC mediated growth and is a potential target for drug therapy. Once the IA wall grows, it is subjected to changes in wall tension and flow conditions as a result of the change in geometry and has to remodel accordingly to avoid rupture. Flow affects this remodeling process.

Conclusions:

Flow triggers an inflammatory reaction that predisposes the artery wall to IA initiation and growth and affects the associated remodeling of the UIA wall. This chronic inflammation is a putative target for drug therapy that would stabilize UIAs or prevent UIA formation. Moreover, once this coupling between IA wall remodeling and flow is understood, data from patient-specific flow models can be gathered as part of the diagnostic work-up and utilized to improve risk assessment for UIA initiation, progression, and eventual rupture.

Introduction

Unruptured intracranial aneurysms (UIA) are found increasingly often as incidental findings during intracranial MR- or CT-angiography imaging due to better availability of these studies²¹. Since incidentally found UIAs may later rupture causing devastating aneurysmal subarachnoid hemorrhage (aSAH)²¹, many patients with incidental UIAs are anxious and want their aneurysm treated. Current treatment options are all interventions with non-negligible risk of morbidity and even mortality^45,54. As a consequence, physicians treating UIAs are challenged with the assessment of whether the rupture risk of an incidental UIA justifies the risks associated with treatment²². This task is complex and demanding since multiple factors impacting risk of UIA rupture have been identified^{72, 69, 31, 22} and no absolute threshold values have been identified for any of these established risk factors to discriminate stable UIAs from those that progress towards rupture⁵⁰.

UIAs are relatively frequent lesions, with 3% or higher prevalence in the past middle age population²¹. The clearly lower prevalence of UIAs in children or young adults in population-based studies and clinical series⁶⁰ together with the fact that formation of new UIAs (so called de novo aneurysms) is observed during follow-up of patients⁵¹, demonstrates that UIAs are not innate lesions but develop during life. This implies that UIA formation is the end-result of degenerative cerebral artery wall remodeling. Understanding the biology of this remodeling process is the key to identification and rational management of persons at risk of UIA formation, as well as of those who have been diagnosed with UIAs. The fact that many, if not most UIAs remain unruptured during life-long follow-up⁴³, demonstrates that there is also adaptive remodeling that can stabilize the UIA wall and ensure sufficient strength to withstand the mechanical stress imposed on the aneurysm wall by blood pressure and flow⁵⁸. Understanding the mechanisms mediating the destructive and adaptive remodeling of the cerebral artery and aneurysm wall will open the door for the design and development of pharmaceutical or other biological therapies that would inhibit UIA formation and progression towards rupture, thus offering new hope for those at risk of UIA formation (e.g. persons with familial predisposition to aSAH²¹) and in some cases an alternative to invasive UIA treatment.

Aneurysm formation requires active collagen remodeling – not just disruption of elastic laminas

The wall of normal intracranial arteries is composed of a luminal endothelial layer (EC), underneath which is the basal lamina composed of matrix proteins, then the internal elastic lamina composed of elastic fibers, followed by the media layer composed of smooth muscle cells (SMCs), elastic laminas, and collagen fibers ^{e.g. 59,24}. The adventitia layer starts at the outer rim of the media, and is composed of collagen fibers and fibroblast cells. All the components of the artery wall have specific functions, of which resistance to mechanical stretch and maintenance of structural integrity are almost entirely due to the elastic laminas and the medial and adventitial collagen fibers ^{e.g. 59, 71}. Unlike the elastic laminas that distend when loaded to absorb and then release part of the energy of the pulse wave^{59, 71}, collagen fibers have little capacity for extension prior to failure³⁰. The collagen fibers in the medial layer of the artery wall are undulated in the unloaded state and straighten during normal physiological cyclic loading³⁶. The degree of tortuosity of the collective collagen fibers in the medial and adventitial layers determine how much the artery can distend and protects the artery from overstretching. Loss of elastic laminas is characteristic of aneurysms^{24, 25, 58} and shifts greater load bearing to the collagen fibers which diminishes the tortuosity of the collagen fibers and limits the capacity of the vessel to dilate⁵⁸.

Since collagen fibers are relatively inextensible compared with elastin^{30, 71}, growth of an aneurysm, sometimes up to a size multiple times the original diameter of the parent artery, requires remodeling of the collagen fibers in the aneurysm wall⁵⁸. Since collagen is the main load bearing structure in the vessel wall once the elastic laminas are lost, the end result of this collagen remodeling determines the strength of the aneurysm wall, and thus also the conditions that suffice for rupture⁵⁸. Not surprisingly, there is high variation in the strength of the aneurysm wall when measured under standardized laboratory conditions, with both robust and vulnerable groups within the UIAs and, on average, ruptured intracranial aneurysms (RIA) having weaker walls than UIAs⁵⁸. Of great interest is the striking observation that remodeled collagen fibers in the aneurysm wall are oriented according to flow and the wall shear stress (WSS) sensed by the EC layer¹¹. This raises the question of whether collagen remodeling in the aneurysm wall is guided by flow, a concept further supported by the observation that flow conditions in the aneurysm fundus associate with strength of the aneurysm wall¹².

What makes these observations especially intriguing is the implication that by guiding the collagen remodeling of the aneurysm wall, flow would eventually determine how prone the aneurysm is to rupture in addition to its impact on aneurysm formation and growth.

Flow as the driver of outward remodeling in vessels

The idea that flow would guide the remodeling of blood vessel structure similarly to how mechanical load guides the remodeling of bone trabeculae is not new and in fact proposed as early as the 60’s⁴⁷ if not earlier and is the underlying premise in studies of the growth and remodeling in blood vessels³⁷. Surgical arteriovenous fistulas created to enable hemodialysis, provide clear evidence that this flow induced vessel remodeling takes place in humans^{23, 64}. In these fistulas, the lack of capillary bed resistance produces an abnormally high flow which leads to outward remodeling to a size multiple times larger than original, especially in the draining vein⁶⁴ subjected also to higher than intended pressure in addition to abnormally high flow. Experimental models of surgically created AV-fistulas demonstrate that this flow induced outward remodeling occurs through a coordinated sequence of protease expression followed by synthesis of new collagen^{32, 73}. Underlying deficiency in elastic laminas accelerates this flow induced outward vessel remodeling⁷³.

Effect of high flow in cerebral arteries – focal outward remodeling leading to aneurysm formation

Manipulation of cerebral blood flow by uni- or bilateral ligation of carotid arteries in laboratory animals predisposes to segmental or focal dilation and UIA formation at sites of the cerebral vasculature exposed by the procedure to abnormally high flow, especially if coupled with hypertension and defective collagen synthesis^{69, 42, 2} (Figure 1). These induced aneurysms develop preferably at bifurcations (reviewed in Aoki et al,²) (Figure 1), as their human counterparts. Although de novo IA formation is occasionally seen in the clinical practice after uni- or bilateral carotid occlusion due to disease or medical procedure, overall the prevalence of UIAs in patients with occlusive carotid artery diseases (3%)⁴¹ seems similar to the general population²¹. This apparent controversy between the animal models and clinical observations can be explained by slower development of the clinical disease, which allows time for collateral circulation to develop. Moreover, difference in the magnitude of flow enhancement obtained with carotid ligation in animals compared to carotid occlusion in humans is likely relevant. A more direct proof of flow induced vessel remodeling causing aneurysm initiation in humans is obtained from patients with arteriovenous malformations of the brain (bAVMs). BAVMs cause similar arteriovenous shunting as the surgically created AV-fistulas, and many of the bAVM feeding arteries that are exposed to abnormally high flow develop segmental ectasia or saccular aneurysms^{49, 67}. Of special interest is the observation that some of these bAVM associated IAs spontaneously regress once the shunting lesion is obliterated and flow normalized⁴⁹, demonstrating that these IAs are needed initiated and maintained by abnormally high flow.

Figure 1.

The model of induced formation of intracranial aneurysms (IA) established and developed by Nobuo Hashimoto, demonstrates in practice how increase in flow predisposes to IA formation. In this model unilateral carotid artery ligation (A) coupled with induced hypertension and impaired collagen remodeling leads to IA initiation (A-D) at the contralateral bifurcation (marked with a black box) of the olfactory artery (OA) and the anterior cerebral artery (ACA) that are exposed to increased flow (marked with black arrows in D) due to the demand of the artery bed supplied by the ligated CA (A). The site of IA initiation is characterized by disruption of the elastic lamina (demonstrated with Elastica van Gieson staining in C) and by thinning but not complete absence of the smooth muscle cell layer (demonstrated with alpha-smooth muscle actin staining in D). This ectatic (outward) vessel remodeling leading to IA initiation occurs mainly at apices of bifurcations (reviewed by Aoki et al. 2011 in J Biomed and Biotechnol), unless injury to the elastic laminas is exacerbated by elastase infusion to the CSF.

Inflammation as the mediator of flow induced outward remodeling and aneurysm initiation

Classical work by Langille in the 1980s demonstrated the capacity of artery wall to respond to both increases and decreases in flow through an initial acute vasomotor response, followed by wall remodeling ^{e.g 48}. This adaptation to flow requires an intact endothelium⁴⁸. It is now understood that the surface proteoglycan layer (glycocalyx) on the endothelium surface is the primary sensor of flow related forces⁶⁶. In particular, flow across the endothelial cells generates a wall shear stress (WSS), a frictional force per unit area tangential to the flow.

WSS is primarily dependent on blood viscosity and flow velocity (Figure 2). If the vessel diameter remains constant, increased volumetric flow rate increases the velocity, which in turn increases WSS (Figure 2). The largest increases in WSS is generally seen at outer walls of curved vessel segments, highly constricted regions of arteries, and the apex regions of bifurcations. At these sites the local velocity is elevated due to the coupled nature of flow and geometry (reviewed in Tanweer et al.⁶⁵). Using surgically created high flow bifurcations, Meng et al. showed that the changes predisposing to aneurysm formation, namely IEL disruption and degeneration of the SMC layer, develop at the sites exposed to high WSS and to a positive WSS gradient (WSSG)⁵². Since then, several studies using various animal models have concluded that IA initiation occurs in regions exposed to high WSS with a positive WSS gradient²⁰.

Figure 2.

Wall shear stress (WSS) is a flow induced force per unit area tangential to the direction of the flow and dependent on the rate of flow as illustrated in A. Therefore, WSS at the apices of arterial bifurcations increases if flow increases, as demonstrated by the computational fluid dynamic modeling of WSS in a patient-derived geometry of internal carotid artery bifurcation with low (1.86ml/s, B) and high (2.72ml/s, C) flows. Figure 2. Wall shear stress (WSS) is a flow induced force per unit area tangential to the direction of the flow and dependent on the rate of flow as illustrated in A. Therefore, WSS at the apices of arterial bifurcations increases if flow increases, as demonstrated by the computational fluid dynamic modeling of WSS in a patient-derived geometry of internal carotid artery bifurcation with low (1.86ml/s, B) and high (2.72ml/s, C) flows. While the shear stress is a vector that has both magnitude and direction, only one component is non-zero in the case shown in A) and in B and C, the magnitude of the vector is shown in the form of contour plots. WSS gradient (WSSG) is the difference between WSS at to different site on the vessel wall (WSS1-WSS2).

Nitric oxide (NO) produced by endothelial NO synthetase (eNOS) is the primary mediator of the EC triggered vasodilation in response to high flow⁷⁵. Increase in WSS induces eNOS production in ECs, and in addition to relaxing the medial smooth muscle cells (SMCs)⁵⁶. This increase in NO downregulates the expression of pro-inflammatory adhesion molecules macrophage chemotactic protein 1 (MCP-1) and vascular cell adhesion molecule 1 (VCAM1) that are otherwise concomitantly upregulated in ECs in response to WSS⁷⁵. Experiments with NOS knockout mice demonstrate how blocking NOS signaling predisposes to IA formation through increase in macrophage infiltration, presumably through increased MCP1 expression³ (Figure 3). Recent results suggest that mechanical stretch can induce the MCP1 expression also in adventitial fibroblasts, which promotes the transmural migration of macrophage⁴⁴. Colocalization of high WSS with wall stretch would thus lead to an amplified MCP1 expression and macrophage infiltration with increased likelihood of eventual IA initiation⁴⁴.

Figure 3.

High flow leads to increase in WSS and the associated increase in vessel caliber will increase the wall tension. High WSS is sensed by the endothelium through the glycocalyx (light blue) and leads to the induction of macrophage chemoattractant protein 1 (MCP-1) in the endothelial cells. The mechanical stretch of the artery wall is in turn sensed by intramural cells and can induce MCP-1 expression in fibroblasts. MCP-1 plays a crucial role in the recruitment of circulating monocytes to the site of vessel remodeling, which appear to occur through the luminal endothelium and the artery wall in mice but may occur mainly through adventitial capillaries in larger animals. These infiltrating macrophages that migrate through the artery wall following the chemotactic MCP-1 signaling from the adventitia or lumen, secrete proteases that disrupt the internal elastic lamina (IEL) and the collagen matrix. Through this mechanism high flow can induce IEL disruption, a key step in aneurysm initiation

The crucial role of MCP1 in aneurysm initiation is demonstrate by experiments with MCP1 knockout mice, in which macrophage infiltration was nearly absent in cerebral arteries exposed to high flow and IA initiation and formation were reduced by more than half⁴. The key role of macrophage infiltration in IA initiation has been confirmed by several other studies using variations of the classical induced IA formation model and clodronate to deplete macrophages^{40, 68, 44} or manipulation of macrophage activation through PPARgamma⁶².

In the rodent models of induced IA formation, macrophages infiltrate the artery wall through the luminal endothelium during IA initiation⁴⁴. Following this, they migrate through the wall eventually reaching the adventitia following a chemotactic gradient generated by MCP1 expression in adventitial fibroblasts⁴⁴. Since transendothelial migration to the vessel wall has to occur through the internal elastic lamina (IEL) (Figure 3) and once in the wall the macrophages produce proteases causing collagen degradation⁵, it is logical that the macrophage infiltration eventually leads to IEL disruption as well as to disruption of the collagen matrix that is the load bearing structure of the wall once elastic laminas are lost (please see above).

Flow triggered inflammation mediated remodeling in aneurysm growth

Once the IEL is lost due to the infiltration of inflammatory cells, and the collagen matrix damaged to the point of allowing the outward bulging of the initial aneurysm, the aneurysm may start to enlarge. This however requires that the aneurysm wall actively grows and synthesizes new matrix (discussed above). Mere prolonged proteolytic injury by infiltrating macrophages without concomitant remodeling of the collagen matrix would not lead to a substantial growth, but rather a rupture from a small, blister-like bleb sometimes encountered in the clinical practice.

Smooth muscle cells (SMCs) of the IA wall synthesize the new collagen^25,26 required for this active IA wall growth. These mural cells express receptors for several growth factors secreted by macrophages, e.g. transforming growth factor beta (TGFb) and platelet derived growth factor B (PDGF-B) that stimulate SMC matrix synthesis and proliferation²⁷. That the macrophage activation constantly stimulates the SMCs, in addition to protease secretion, can explain why the aneurysm wall grows. Once initiated, the macrophage activation in the aneurysm wall amplifies itself through an autocrine feedback loop in which prostaglandin E2 (PGE₂) produced by cyclo-oxygenase 2(COX2) activates the transcription factor nuclear factor kappa b (NFkB) in other macrophages, leading to increased expression of macrophage chemotactic protein 1 (MCP1) as well as COX2 by them^6,7 (Figure 4). This stimulates the recruitment of more macrophages to the aneurysm wall, as well as NFkB activation in them. This amplification loop may explain the observation that wall remodeling initiated by high flow conditions can continue even if the pathological flow is normalized⁵³.

Figure 4.

Once initiated, the aneurysm will not grow unless there is active growth of the aneurysm wall and synthesis of new collagen. While loss of the elastic laminas (in purple) enables the artery wall to distend until the normally unloaded, curly collagen fibers (in green) are distended, further growth of the aneurysm after initiation requires remodeling of the collagen since collagen does not distend. During this collagen remodeling, not only the orientation of the fibers but also the type of collagen in them changes (reviewed by Etminan et al. 2016 in Nat Rev Neurol). In parallel to the collagen remodeling, change in the phenotype of smooth muscle cell (SMC, in red) occurs in the IA wall, in part related pro-inflammatory signaling by infiltrating inflammatory cells (blue) that activates in SMCs NFkB transcription factor essential for artery wall remodeling and aneurysm formation.

Clinically one of the most pertinent questions regarding the biology of aneurysm formation, is how to stabilize small inceptions of aneurysms or small aneurysms and prevent them from growing to a size where rupture becomes increasingly likely^{21, 72, 31}. The presence of MCP1 expression and macrophages^{34, 8, 25, 55} in human IA walls implicates that the same molecular mechanisms as in IA initiation might be at play in IA progression, at least in the formation of daughter aneurysms or so-called secondary pouches (Figure 7) that tend to form in the IA sacs at regions with high WSS¹³. In addition to high WSS areas, MCP1 can be also induced in IA regions with low WSS⁸. This is in line with the observation that inflammation of the IA wall showed an associative trend with both high WSS and low WSS 14.

Figure 7.

When aneurysms grow, wall tension (WT) increases. This can be seen from a simple force balance in an idealized spherical shell using the Law of Laplace A). In a sphere, the WT acts around the circumference with similar loads acting on the components of the wall such as collagen fibers and intramural cells. Although true intracranial aneurysms are rarely close to spherical (B, a sphere superimposed on the 3D geometry of an actual aneurysm in a patient), similar relationships exist between the local wall tension and geometry. This in turn means that in a growing aneurysm, as the diameter increases the wall has to remodel to withstand higher WT. Moreover, since WT can also be defined through wall thickness and intramural stress (force pulling the wall components a part) as shown, it can be concluded that simple increase in wall thickness will reduce intramural stress. In a wall with intact functional smooth muscle cells (SMC)(C), this can occur for example through collagen remodeling and proliferation of SMCs that may lead to significant focal increase in wall thickness (t1 vs t2 in C, HE stained aneurysm wall). However, a wall with few healthy SMCs is less able to adapt (HE stained example of a decellularized wall in D). This may explain at least in part the high variation observed in the strength of aneurysm wall tissue samples when stretched to failure. The stress-strain curve adapted from the measurements performed by Robertson et al. 2011 and applicable for most aneurysm walls is shown in D (intramural stress determined as a 1st order Piola-Kirkhoff tensor and the blue area corresponding to the range of curves measured from individual samples with the exception of few outliers in the original data). Of note is the dependence of intramural stress from wall thickness which means that thinner regions (t1) of the same aneurysm wall can be stretched to the maximum and close to rupture while thicker regions (t2) in the same aneurysm wall are not yet maximally loaded.

In animals, activation of NFkB at the site of vessel remodeling is necessary for the IA to form⁹. This NFkB activation occurs through the COX2- PGE₂-EP2 -NFkB pathway as discussed above^6,7. Genome wide gene expression analysis of ruptured and unruptured human IA walls has shown upregulation of multiple NFkB regulated genes in ruptured IA walls⁴⁶, strongly implying that macrophage induced NFkB activation is relevant in the wall remodeling of established human IAs similarly to the experimental models of induced IA initiation and formation^9,7. Presence of COX2 expression in the human IA wall^34,9, as well as of PGE₂ receptor subtype EP2, further implies that the COX2- PGE₂ -EP2-NFkB-COX2 signaling pathway is involved in the growth and wall remodeling of human IAs⁷ similar to the experimental models of induced UIA formation. The seminal observations that drugs inhibiting COX2 activity, such as aspirin³⁵ or non-steroidal anti-inflammatory drugs seem to reduce the growth of UIAs in patients²⁸ further supports this concept.

Hyperplastic remodeling as a response to increase in mechanical stretch – role of inflammation

Once the initiated aneurysm starts to grow, its wall will be subjected to progressively higher wall tension (Figure 5). For the IA to remain unruptured, the wall has to adapt to the increased mechanical load. In general, the artery wall adapts to chronically increased mechanical load through proliferation of the medial SMCs and collagen deposition, as seen in wall changes associated with hypertension³⁸. In cases where this remodeling process is insufficient, the intramural cells may be exposed to supra physiological stretch. The dynamics of this repair process following overstretch of this kind has been well studied in arterial balloon dilation injury models^16,63 and shown to depend on macrophages¹⁷ and NFkB activation in SMCs⁷⁴.

Figure 5.

Smooth muscle cells (SMCs) of the IA wall that synthesize the new collagen express receptors for several growth factors secreted by macrophages, e.g. transforming growth factor beta (TGFb) and platelet derived growth factor B (PDGF-B) that stimulate SMC matrix synthesis and growth. That the macrophage activation constantly stimulates the SMCs, in addition to protease secretion, can explain why the aneurysm wall grows. Once initiated, the macrophage activation in the aneurysm wall amplifies itself through an autocrine feedback loop in which prostaglandin E2 (PGE₂) produced by cyclo-oxygenase 2(COX2) activates NFkB in other monocytes/macrophages leading to upregulation of macrophage chemotactic protein 1 (MCP1) as well as COX2. This stimulates the recruitment of more macrophages to the aneurysm wall, as well as to NFkB activation in them, and maintains the chronic inflammation driving aneurysm growth.

The enlarging IA wall can adapt to this increase in wall tension through SMC proliferation and collagen remodeling similar to the way an artery wall responds to overstretching or chronic high pressure, provided that it has healthy SMCs²⁵. Many IA walls, however, have at least focal regions that show loss of SMCs^24–26. Moreover, in many IA walls the remaining SMCs turn into foam cells due to lipid ingestion^{29, 55} which impairs their normal function⁵⁵. This can lead to a potentially dangerous scenario in which the aneurysm continues growing but has wall regions that are not able to adapt to the increase in WT caused by growth. The mechanisms of lipid accumulation in the IA wall are discussed in more detail in Ollikainen et al.⁵⁵ and Frösen et al.²⁹. Since endothelial dysfunction seems to be a key factor in promoting the accumulation of lipids in the IA wall^{29, 55, 24}, flow that regulates endothelial function is likely to affect the process of lipid-induced IA wall SMC dysfunction as well.

Clinical applications – flow modulation, flow modeling, and multimodality diagnostics

Flow conditions in the human IA sac associate with focal changes in wall structure^{39, 15}, as well as with histological changes of the IA wall, including inflammation¹⁴. Moreover, additional data from somewhat small number of cases suggests that flow conditions associate with collagen remodeling and strength of the IA wall^11,12.

Measurement of flow conditions seems to hold great potential for the identification of i) persons at risk of IA formation, ii) IAs that are likely to grow and should be followed attentively, and iii) unstable IAs that need an intervention. Flow conditions can be estimated with reasonable accuracy from 3D high resolution angiographies using computational fluid dynamics (CFD), provided that the boundary conditions needed for the flow simulations are determined reasonably (separate reviews on the topic in this issue). Although several studies have demonstrated how CFD can be used to differentiate ruptured and unruptured IAs with reasonable accuracy¹⁸ (and separate publications on the topic in this issue), from the biology point of view it is worth noting that many human IA walls have regions that have completely lost the endothelium²⁵. This will significantly impair the normal mechanobiological coupling between flow and wall remodeling^{48, 10}, and questions whether such IA wall is sensitive to flow induced remodeling at all. In order to reliably determine what kind of flow conditions will lead to an unstable UIA, follow-up studies of UIAs with CFD performed at baseline are needed. For the time being this kind of clinical series are few.

Although IAs in humans form preferentially at sites of high WSS¹, the fact that they do not form in all bifurcations exposed to high WSS and to all persons does seem to imply that a second factor is needed for IA initiation. In experimental models, concomitant elastase injury significantly promotes aneurysm formation at sites of pro-aneurysmatic flow conditions⁷⁰. Since inflammatory cells, especially neutrophils are an important source of elastase²⁴, it seems plausible that such a “second hit” in humans would take effect through activation of inflammatory cells and subsequent increase in elastase activity. Recently involvement of gut microbiome in a model of induced IA formation has been reported⁶¹. Even more recently, we have shown that severe periodontitis and gingival inflammation predisposes to IA formation and eventual aSAH in humans³³. Prior to this, presence of oral bacteria derived DNA has been shown in the IA wall⁵⁷. The biological mechanism by which gut microbiome or periodontitis predisposes to IA formation remains to be shown, but clearly other factors besides flow are significant in the initiation and progression of UIAs.

In addition to acquired factors such as periodontitis, in some cases the “second hit” can be genetic, for example in polycystic kidney disease (PCKD) in which mutation in the polycystin gene affects the mechanobiological coupling of flow-endothelial cell interaction¹⁹. Of note, flow driven vessel remodeling in AV-fistulas in mice heterozygous for defective elastin gene was significantly increased⁷³. Identification of similar inherited “secondary triggers” will facilitate the identification of persons at risk of developing IAs, especially when coupled with flow modeling (CFD). Knowledge of the mechanobiological coupling between flow and wall remodeling, and identification of these environmental or genetic “secondary triggers” may also enable more efficient primary prevention of the disease by e.g. treatment of periodontitis, or by treatment of the flow induced IA initiation with preventive drug therapy. Results from experimental models suggest that inhibitors of COX2- PGE₂-EP2 signaling^5,6 or PPARgamma inhibitors⁶², for example, could be used for this purpose. Another potential target for drug therapy could be inhibition of the proteases mediating the flow induced inflammation driven remodeling.

Conclusions

High flow is an initiator of UIA formation and flow conditions also drive the wall remodeling that determines whether the aneurysm will remain stable or progress and eventually rupture. Inflammatory cells mediate this flow induced remodeling. Flow is not, however, the only factor involved in UIA initiation, and neither the only factor determining UIA wall remodeling. Understanding the mechanobiological coupling of flow and aneurysm wall remodeling is the key to predict the clinical course of an aneurysm. This will become increasingly more important since drug therapies that modulate flow induced inflammatory remodeling are being developed.

Figure 6.

When the aneurysm grows, flow conditions in the aneurysm lumen change because the luminal geometry changes. This is likely to change the flow induced WSS, especially which areas of the wall are exposed to high WSS (marked with a blue line in A-C). Regions exposed to high flow are likely to undergo similar flow induced inflammation mediated remodeling as in aneurysm initiation, which in turn can explain focal growth of secondary pouches in the aneurysm wall at sites of high WSS. While in very small aneurysms a larger extent of the aneurysm wall is exposed to high WSS, in general the more the aneurysm grows less of its surface area is exposed to high WSS. In addition to this, increase in the aneurysm diameter is prone to cause regions with slower flow, which can predispose to luminal thrombosis (in pink in B-C). Luminal thrombosis stimulates aneurysm wall remodeling through multiple mechanisms reviewed in detail by Frösen et al. 2011 in Acta Neuropath. In brief, red blood cells (RBC, in red), platelets (ellipsoids in red) and inflammatory cells including neutrophils, are trapped in the thrombus and release cytokines and growth factors that stimulate pro-inflammatory phenotype and growth of smooth muscle cells (SMCs) in the aneurysm wall (D). This promotes wall remodeling that encompasses secretion of proteases that degrade the collagen matrix. In addition to SMCs, also the inflammatory cells in the luminal thrombus are a source of collagen degrading proteases. This thrombus derived chronic proteolytic injury is especially relevant in aneurysms that have a decellularized wall devoid of mural cells capable of synthesizing new collagen. Flow related factors predisposing to loss of mural cells include cytotoxic iron released from the degrading RBCs of the luminal thrombus, as well as lipids that accumulate in the aneurysm wall to mural cells causing foam cell formation and eventually cell death. This lipid accumulation seems to be related to dysfunction of the luminal endothelium, which in turn can be explained by the non-physiological flow conditions in the aneurysm lumen (E). Overall, through the effects of flow on endothelial function and thrombus formation, flow in the aneurysm lumen affects the wall remodeling by also other mechanisms than the ones relevant in aneurysm formation. Consecutive exposure of different aneurysm wall regions to changing flow conditions resulting from aneurysm growth, likely explain the high degree of heterogeneity on aneurysm wall structure observed during surgery (F,G). That flow conditions associate with wall structure observed at surgery supports this concept.

Figure 8.

IAs develop in humans mostly at arterial bifurcations in areas subjected to focal high WSS and high spatial gradients in wall shear stress (example of WSS distribution in a computational fluid dynamics model of right ICA in upper left corner). The ectatic remodeling (mechanisms presented in Figures 3–7) induced by high WSS predisposes to IA initiation and growth, which eventually will change the geometry of the aneurysm leading also to a change in flow conditions. Depending on what kind of flow conditions result from this aneurysm growth, the wall will remodel differently and focal growth (secondary pouch formation) or global growth may ensue. It is important to note that aneurysm growth will change the wall tension through the increase in size (discussed in Figure 7) in addition to inducing wall remodeling. Wall remodeling in turn determines wall structure and strength, with great variation in the remodeling process, even within the unruptured population. These three factors: wall strength, wall structure and wall load (tension) are critical determinants of the structural integrity of the wall. Since the flow in the aneurysm sac can impact all of these, it is a key determinant of what the risk of rupture and natural history of an unruptured IA will be.