Vasa vasorum (von Willebrand factor positive) channel density in the media and adventitia of the infrarenal aortic wall. Data presented as median and interquartile range. Aneurysmal aortae had significantly increased medial and adventitial vasa vasorum densities compared with controls. Unlike wall cellularity measures (Fig. 11), there was no difference in vessel density between different sized aneurysms. (Data courtesy of Dr Jun Li).
Atherosclerosis is a chronic condition characterised by the formation of lipid-rich plaques within the walls of medium and large arteries (Ross, 1993; Lusis, 2000) and underlies many forms of vascular disease, including abdominal aortic aneurysm (AAA). The development of all vascular disease phenotypes is dependent on multiple genetic and environmental determinants (Lusis, 2000) though the relative contribution of each of these risk factors may vary with different vascular disease phenotypes. This is certainly true for AAA, which appears to have particularly strong associations with male gender, tobacco smoking (Singh et al., 2001) and a family history of AAA (Verloes et al., 1995; Ogata et al., 2005). Interestingly, diabetes, a well-established occlusive atherosclerotic disease risk factor is not associated, and may even be protective for AAA (Shantikumar et al., 2010).
Abdominal aortic aneurysm is a common condition being responsible for 1.3% of all deaths in 65-85 year old white males (Sakalihasan et al., 2005). The reported prevalence rates vary depending on factors such as the threshold aortic diameter used to define an aneurysm and age strata (relevant when reviewing rates from screening studies). If the most widely accepted value of 30mm is applied, 1% of women and 4.2% of men between the ages of 50 to 79 years are affected in a predominantly white (north American) population (Lederle et al., 2000). In British males, over the age of 65 years, this rate has been reported to be a high as 7.6% (Scott et al., 1995), although there are indications that incidence may be falling internationally (Sandiford et al., 2011). An abdominal aortic aneurysm is typically defined as being localised in the infrarenal abdominal aorta and may either extend up to involve the renal ostia, or down to involve the aortic bifurcation and into common iliac arteries (Sakalihasan et al., 2005).
These epidemiologic and anatomical features are important when considering the pathobiology of AAA. While the initial pathobiology is atherosclerotic, at some stage it diverts to a distinctive dilating, rather than aortic occlusive, phenotype. A schema outlining the key pathogenic components was proposed by Thompson and colleagues in 1995 (Holmes et al., 1995), a modified interpretation of which is shown in Figure 1. This chapter will briefly describe each of the pathohistological features.
2. Histology of the abdominal aorta
2.1. Histological features in the aging arteries
The abdominal aorta is a transitional elastic artery. In the prenatal abdominal aorta the intima consists of an endothelial monolayer and a scant pad of sub-endothelial fibrous connective tissue upon an intact internal elastic lamina (IEL). The only openings between the intima and underlying media are fenestrations in the IEL, which are spanned by elastic tissue trabeculae. The media is composed of approximately 28-30 concentric lamellar subunits (Wolinsky & Glagov, 1969), one lamellar unit consisting of an elastic lamina and the smooth muscle and extracellular matrix contents of the adjacent interlamellar zone (Fig. 2). The medial elastic laminae are approximately two thirds the thickness of the internal elastic lamina and are interconnected by a network of finer elastic fibres. Collagen and elastin fibres are uniformly dispersed around the circumferentially orientated smooth muscle cells (Keech, 1960; Karrer, 1961; Wolinsky & Glagov, 1964).
Within the aortic adventitia a network of vasa vasorum is present that originates from adjacent intercostal, lumbar and mesenteric arteries (Heistad & Marcus, 1979). The vasa vasorum has been shown to only penetrate the media of vessels with greater than 29-30 lamellae (Wolinsky & Glagov, 1969) and consequently the media of the human infrarenal aorta is vastly avascular, despite the presence of the adventitial vasa vasorum (Fig. 3). However, if the luminal diffusion is insufficient to supply nutrients to the intima, vasa vasorum becomes an important alternative nutrient supply (Heistad & Marcus, 1979). Consequently, as the intima thickens, the vasa vasorum extends channels into the outer medial lamellae (Fig. 3B). Significantly however, unlike vasa vasorum in larger vessels such as the thoracic aorta, which are well integrated into the lamellar structure of the wall during its vasculogenesis, these new vessels appear to disrupt the medial structure during their invasion of the wall, as described below in section 2.4.
Compared to the thoracic segment and other mammalian aortas of the same thickness, human abdominal aortic media contains less lamellar units, and therefore medial elastic laminae. Consequently, the vessel appears to be stiffer, due to a higher collagen-to-elastin ratio. The flow on effect of this is that the mean tension per lamellar unit is higher in the human abdominal aorta (Wolinsky & Glagov, 1969; Wolinsky,1970). As will be described below, loss of load bearing elastin, along with the avascular nature of the abdominal aortic media, are key pathologic features of AAA. Therefore the vessels initial lower elastin content and lack of medial vasa vasorum may make it structurally vulnerable to degenerative alteration and the subsequent development of aneurysmal disease later in life (Dobrin, 1989; He & Roach, 1994).
2.2. Early life elastic lamina degeneration, the nidus of atherosclerosis?
After birth one of the first changes in aortic structure consists of disruption of the IEL (Fig. 4). These spontaneous, transversely orientated, elastic lamina defects have been reported within the arteries of newborn and adolescent humans, particularly those vessels known to be prone to the subsequent development of atherosclerosis, such as the aorta (Meyer & Lind, 1972), coronary (Levene, 1956; Moon, 1957; Sims, 1985; Ikari et al., 1999) and carotid (Meyer et al., 1980)] arteries. Critically, Meyer and colleagues demonstrated that elastic lamina defects occur in very young humans, at sites prone to the eventual development of atherosclerosis (Meyer & Lind, 1972; Meyer et al., 1980).
The intima also undergoes early cellular alteration, with the formation of a smooth muscle cell layer between the endothelium and IEL. The origins of these smooth muscle cells has been considered controversial, with some authors suggesting that they are an expansion of a resident intimal cell population (Stary, 2000), while others have argued that they originate from the media and migrate into the intima (Schwartz, 1997; Willis et al., 2004). The most significant contributor to intimal smooth muscle thickening is likely to be a medial origin, as evidenced by the fact that the intima is thicker above regions of IEL disruption in children (Fig. 4). This is further supported by the ApoE deficient mouse model of atherosclerosis, which clearly shows that early intimal smooth muscle cell accumulation is due to medial smooth muscle cell migration through defects in the IEL (Jones et al., 2005). The thickening intima typically forms two distinct layers, a deep musculoelastic layer and a less compact superficial subendothelial layer (Figs 5 & 6C).
The literature also reports the internal elastic lamina beneath the thickened intima and subsequent atherosclerotic plaques as becoming ‘frayed’, ‘reduplicated’ or undergoing laminated elastosis. All of these terms refer to intimal elastic tissue deposition and as such should be considered as a distinct, though often co-localised, processes to the disruption of the original IEL. The key cells in producing intimal elastic tissue are smooth muscle cells (Kojimahara, 1988) with a possible contribution by the endothelial monolayer (Jones & Stehbens, 1995) and macrophages (Krettek et al., 2003).
Intimal elastic tissues appear to have an increased affinity for lipoproteins compared with the ‘original’ elastic laminae (Adams & Tuqan, 1961; Urry, 1975; Meyer et al., 1980). The lipophillic features of intimal elastic tissue (Fig. 5) appears to be a key early contributor to intimal lipid retention (Williams & Tabas, 1995).
Calcification has been shown to be localised to the elastin adjacent to the break edges of spontaneous IEL defects in the arteries of children (Meyer & Lind, 1972; Meyer et al., 1980). While calcium containing matrix vesicles have been shown to preferentially accumulate in microfractures within elastic laminae adjacent to abrupt, haemodynamically induced, elastic tissue defects in experimental animals (Jones & Stehbens, 1995). As such failure of the elastic laminae may facilitate accumulation of granulovesicular debris, including calcospherites, and thereby contribute to vascular connective tissue calcification.
The localisation of atherogenic lipids and granulovesicular debris within elastic laminae break edges and the intimal elastic tissue may act as chemotaxic targets for migrating smooth muscle cells and macrophages effectively ‘lighting up’ these intimal / medial gateways.
When assessing the histopathology of early atherosclerotic lesions, care should therefore be taken to differentiate between intimal elastic tissue and the original IEL. Such a distinction is clearly significant given the possible differences in cell permeability and affinity for lipophilic debris of these two forms of elastic tissue. Certain stains, such as phosphotungstic acid haematoxylin (PTAH), differentially stain these two forms of elastic connective tissue (Gillman et al., 1955), though common elastic tissue stains, such Verhoeff’s elastic stain, do not.
Of relevance to the study of abdominal aortic aneurysm histopathology the distal abdominal aorta and common iliac arteries appear to be particularly vulnerable to both IEL degeneration and intimal elastosis (Fig. 4-6).
Even at its earliest stages of development, atherosclerosis displays preferential localisation about the segmental branch ostia of the posterior abdominal aortic wall (Miller et al., 1993; Stehbens, 1995). While systemic factors, such as serum LDL levels, smoking, hypertension and diabetes (Grundy, 1995; Schaefer et al., 1995; Strong & Group, 1995; Jousilahti et al., 1999), have been implicated in the progression of atherosclerosis they fail to fully account for the
localised nature of the disease. Haemodynamic stresses, including pulse reflection and summation, may predispose various anatomical sites to the development of vascular disease (Stehbens, 1995) as may local intrinsic wall abnormalities. Early atherosclerotic alterations appear to involve the development of intimal thickenings, which are partly derived from medial smooth muscle cells (Fig. 6). Normally the intact internal elastic lamina (Meyer et al., 1980; Ikari et al., 1999) restricts intimal medial cell migration. In latter stages of atherosclerosis the structural and molecular modification of elastic tissues is believed to reduce arterial compliance leading to increased pulse pressure and pressure wave velocities in aging individuals (Urry, 1975; Robert et al., 1995).
This raises the possibility that early-life structural and functional modification of the aortic wall may directly contribute to the progression of atherosclerosis (Stehbens, 1997) rather than simply being normal physiological remodeling. As shown in figure 6, it is somewhat remarkably that the elastic lamina defect edges remain so abrupt, even in elderly aneurysmal aortae, given the considerable elastolytic activity reported in atherosclerotic tissues, and in particularly AAA (Thompson & Parks, 1996). This is undoubtedly a testament to the stability and low turnover rate of elastin, which is estimated at approximately 70 years in humans (Shapiro et al., 1991).
Nevertheless, given the proteolytic activity that occurs within developing atherosclerotic lesions (Knox et al., 1997) and aneurysms (Vine & Powell, 1991; Gargiulo et al., 1993; McMillan et al., 1995; Thompson et al., 1995), the lack of change to the elastic lamina defect edges, apparently over many decades, is noteworthy in that it may provide a stable cumulative index of connective tissue degeneration, without the confounding effect of connective tissue remodelling (as is the case with the fibrillar collagens).
Though elastolytic activity undoubtedly occurs, the histo-morphological features indicate that the elastin within condensed laminae is remarkably resistant to enzymatic degradation. This is not to say that proteolysis is not a significant pathophysiological process in the loss of elastin, for example proteolytic activity may play an indirect role by degrading structural fibrillar collagens. The effect of this would be to shift mechanical stress from these high tensile strength connective tissue components to the relatively weaker elastic tissues, resulting in their accelerated structural fatigue.
2.3. Early atherosclerotic lesions
The earliest lesion generally considered to be atherosclerotic consists focal accumulations of intracellular lipoproteins in the intima and formation of fatty streaks (Fig. 5A). Fatty streaks may be present in the aorta from early childhood and may even start to develop during the foetal life, especially in foetuses of mothers with hypercholesterolaemia (Napoli et al., 1997). The major cellular constituents of fatty streaks are monocyte-derived macrophages recruited into wall via endothelial cell mediated diapedesis. Within the intima, these macrophages may engulf the blood-derived low density lipoprotein (LDL) and become lipid-filled foam cells. By the age of puberty (12-15 years), almost all children will develop lesions containing macrophages and foam cells in the intima of the aorta. This initial accumulation of lipoprotein and formation of fatty streaks does not protrude into the lumen and are thus asymptomatic (Stary, 2000).
The initial accumulation of lipids within the intima is then followed by progressive accumulation of both intra- and extra-cellular lipids. The origin of these lipoproteins can be classified as either blood-derived or that released from the death of resident macrophage/foam cells (Guyton, 2001). Intimal smooth muscle cells also contain lipid droplets (Stary, 2000), but, not having the oxidised LDL scavenger receptor, never develop
the foam cell appearance observed in tissue macrophages. The process of intimal lipid sequestration is pernicious and eventually results in the formation of a so-called lipid-core within the deep intima. As the size of this region increases it becomes filled with cholesterol ester crystals (Tangirala et al., 1994) and calcium deposits (Fig. 7). Beginning as regions of vesicular debris, termed matrix vesicles (Tanimura et al., 1983; Bobryshev et al., 2008), these small (approximately 20nm in diameter) calcium phosphate rich membrane bound vesicles appear to act as a nidus for subsequent vascular calcification (Fig. 7C). The lipid core is covered by a ‘fibrous cap’ composed of proteoglycan-rich extracellular matrix, smooth muscle and inflammatory cells (Stary, 2000).
The increased intimal thickness results in decreased luminal diffusion to the deep intima and inner media, producing a hypoxic zone (Pugh & Ratcliffe, 2003). This is generally considered to occur when the intima reaches 500µm in thickness (Moreno et al., 2006).
2.4. Medial atrophy
There are two histopathological features associated with the formation of a hypoxic intimal-medial zone. Firstly the lipid-laden foam cell rich lipid core appears to erode the underlying media (Fig. 7). While, in children and young adults, the initial failure of the internal and medial elastic laminae appears to be driven by biomechanical fatigue, there is little doubt that a strong proteolytic component drives the medial erosion observed in more advanced plaques. The matrix metalloproteinases (MMP), in particular MMP-9 (Fig. 8), have been strongly implicated in the pathogenesis of both atherosclerosis and AAA progression (Thompson & Parks, 1996). It is this outward remodeling of the media, and eventually the adventitia, that preserves luminal patency during the early-mid phases of atherogenesis (Ge et al., 1993; Hermiller et al., 1993).
Secondly, as a result of the deep intima hypoxia, the adventitial vasa vasorum network begins to penetrate the media (Moreno et al., 2006). In addition, the process of elastic tissue degradation produces small elastin-derived peptides, which have also been shown to stimulate angiogenesis (Nackman et al., 1997; Robinet et al., 2005). The neovascularisation process disrupts the lamellar structure of the media, with heavily vascularised regions of the media having significantly reduced elastic tissue and smooth muscle cell content (Fig. 9). Paralleling the process of medial neovasculisation, the adventitia typically undergoes progressive thickening due to the expansion and proliferation of the vasa vasorum network (Table 1).
|Vessel density (channels/0.1mm2)||Control n=11||AAA n=28||p-value|
|Media||0.1 (0-1.5)||1.6 (0.4-2.9)||<0.002|
|Adventitia||3.2 (1.3-5.2)||6.1 (2.8-9.3||<0.003|
The combined effects of intimal atherosclerosis and medial neovascularistion result in progressive medial atrophy (Crawford & Levene, 1953; Isner et al., 1986). This is associated with a loss of medial smooth muscle cells due to migration into the intima (Fig. 10) and increased cell death due to apoptosis (Lopez-Candales et al., 1997; Henderson et al., 1999; Boyle et al., 2001). This process can be demonstrated by quantifying the cellularity of the media within infrarenal aortae of varying diameters (Fig. 11). Large AAAs have approximately half the number of (smooth muscle) cells compared with normal sized aortae.
2.5. Medial and peri-aortic inflammation
As described above, the formation of an atherosclerotic intima stimulates a neovascular response from the adventitia towards the intima. The presence of these new channels allows for direct recruitment of inflammatory cells, such as macrophages, plasma cells and T-lymphocytes, into the arterial wall. These cells secrete cytotoxic mediators, such as perforin, which induce smooth muscle apoptosis (Henderson et al., 1999; Lindeman et al., 2008) and further weakens the aortic wall. Moreover, these cells also release chemoattractants thereby stimulating further inflammatory recruitment and sustaining the inflammatory tissue reaction. In this way, medial and adventitial neovascularization establishes a chronic inflammatory state within the outer aortic wall (Figs 12 &13), resulting in continued matrix degradation and remodelling, propagating outward expansion of the aorta and consequent aneurysm formation.
In a small but significant subset (3-10%) of AAA patients this process can become so extensive that the adventitia undergoes significant macroscopic thickening, developing peri-aortic (Fig 14) and retroperitoneal fibrosis and even adhesions of adjacent abdominal organs. These are typically referred to as inflammatory aneurysms (Tang et al., 2005), but it is generally accepted that, rather than representing a separate pathobiological entity, the inflammatory AAA is an extreme manifestation of the same medial and adventitia inflammation which characterizes all abdominal aneurysms (Koch et al., 1990). While there is a spectrum of inflammatory cell infiltration, ranging from minimal (Fig. 9), moderate (Fig. 12) to extensive (Figs. 13 &14), the adventitia cellularity tends to be increased in all aneurysms compared with non-aneurysm vessels (Fig 11). This is due a combination of both cells forming neovascular channels and variable degrees of associated inflammatory infiltrate.
2.6. Intraluminal thrombus
Blood flow studies suggest that the shape of an abdominal aortic aneurysm results in disordered flow, an increased wall tension and thrombus formation within the sac (Moore et al., 1992). Chronic intraluminal thrombus (ILT) forms as a mutli-layered structure, consisting of a luminal layer of fresh red clot, a middle laminated thrombus and an actively fibrinolysed ablumenal layer (Fig 15), with variable incorporation into the aortic intima, including regions of intramural haemorrhage. The presence of ILT contributes to aneurysm progression in multiple ways (Michel et al., 2011). Firstly, when present, ILT significantly thickens the aortic wall (Fig. 16) and acts as a barrier for oxygen transport, resulting in further hypoxia-associated degeneration (Vorp et al., 1998), as described above. Secondly, ILT is a rich source of proteases and their activators, including MMPs and urokinase-type plasminogen activator (Houard et al., 2007).
The luminal zone is enriched with neutrophils, due to these cells strong affinity for the fibrin-fibronectin network. Neutrophil proteases are centrifugally propelled towards the abluminal zone and underlying aortic wall through a network of canaliculi (Adolph et al., 1997). The plasmin within the fibrinolytic abluminal zone results in MMP activation and facilates the release of matrix sequestered growth factors such as transforming growth factor beta. An association between the presence of ILT and risk of AAA rupture (Fig 16B) has been suggest by many authors but the relationship remains controversial. Recent evidence suggests that a key factor may be the degree of ILT fissuring, leading to localised regions of increased wall stress (Polzer et al., 2011).
The pathohistology of AAA involves the initial formation of intimal atherosclerosis followed by medial atrophy, neovascularisation, inflammatory cell infiltration and intraluminal thrombus formation. Elastic tissue fragility, both in terms of the initial intimal connective tissue degeneration and subsequent medial atrophy, along with the vascular inflammatory response to atherosclerosis appear to be key pathogenic features of AAA.
A special thanks to Professor Andre van Rij, Dr Ian Thomson, Professor Han-Seung Yoon and Dr Noelyn Hung for their assistance in obtaining specimens. The use of material from Dr Jun Li’s BMedSci dissertation is acknowledged and much appreciated. A special thanks to Mrs Vicky Phillips and Ms Liz Girvan for their invaluable technical assistance. On going funding support provided by the Health Research Council of New Zealand is acknowledged.