Role of the Media in Vascular Injury: Atherosclerosis and Dissection

Abstract

Although a metabolic role for endothelium in toxic injury has been well established, a similar role has not been as thoroughly explored for the vascular media. In this study two forms of vascular medial involvement in toxic injury are examined. Early atherosclerotic plaques are studied by immunohistochemistry for an α class glutathione-S-transferase (GST) isozyme known as hGSTA4-4, which has preferential metabolic activity for α, β-unsaturated aldehydes derived from lipid peroxidation, especially 4-hydroxy-2-nonenal. Findings in human plaque indicate that hGST A4-4 is highly upregulated in vascular smooth muscle cells (VSMCs) within the plaque and in the medial VSMCs underlying plaque. Endothelial cells, while not expressing hGST A4-4 distant from plaques, were found to express the isozyme in cytoplasm overlying plaque. In a series of second experiments, we illustrate a developmental model of dissecting aortic aneurysm (DAA) obtained by administering semicarbazide, an inhibitor of the little-studied VSMC enzyme semicarbazide-sensitive amine oxidase (SSAO), to pregnant rats during the last trimester of development. Newborn rats consistently developed DAA which is characterized by splitting of media of ascending thoracic aorta and extensive blood collections surrounding the vessel. These experimental examples emphasize the potential role of the media in toxic insults to blood vessels. Also, the potential importance of toxic injury to developing blood vessels by in utero exposure to xenobiotic substances is illustrated.

Introduction

It is well known that the vascular endothelium forms a protective barrier between blood components and various tissues. It is also widely assumed that the endothelium functions in metabolism and transport of a host of substances, including exogenous toxins that find their way into the blood. The potential role of the media, however, in toxic injury and detoxification processes has been largely ignored. The purpose of the work shown here, therefore, is to give two discrete experimental examples of how vascular smooth muscle and vascular media of large elastic arteries, and potentially of other, smaller arteries as well, are involved in different forms of toxic injury.

Atherosclerosis is the major cause of morbidity and mortality in the western world. The “endothelial injury” theory of atherosclerosis holds that damage to endothelium is a prime initiating event in formation of the early atherosclerotic plaque. There is considerable evidence that oxidative stress contributes to this endothelial injury through the formation of reactive oxygen species (Shi et al., 2000) and lipid peroxidation. A number of reactive α, β-unsaturated aldehydes derived from lipid peroxidation such as 4-hydroxy-2-nonenal (4-HNE) or acrolein (Uchida, 1999) may act as “second cytotoxic messengers” in these primary oxidative reactions, or trigger the formation of reactive oxygen species themselves (Uchida, 2000).

To protect against the deleterious effects of such oxyradicals, both non-enzymatic and enzymatic antioxidant defense systems exist. Non-enzymatic defenses include low molecular weight compounds such as Vitamins A and E, ascorbate, urate, and reduced glutathione. A second line of defense includes enzymes such as catalase, superoxide dismutase, and the glutathione-dependent enzyme system, where the key protecting enzyme is glutathione peroxidase (Yang et al., 2001). In addition, a group of immunologically related α-class mammalian glutathione S-tranferases (GSTs) that utilize 4-HNE as their preferred substrate has been proposed to be a major cellular defense system against oxidative injury by products of lipid peroxidation, especially α, β-unsaturated aldehydes such as 4-HNE (Cheng et al., 1999).

In earlier studies, we showed that a rat α-class GST isozyme having high catalytic efficiency for the conjugation of 4-HNE and acrolein to GSH is localized in smooth muscle cells of aorta and is upregulated in a rat experimental model that mimics early atherosclerosis induced by the cardiovascular toxin, allylamine (3-aminopropene). In this model, the vasculotoxic effects of allylamine have been shown to be related to its metabolism to acrolein both in vivo and in vitro (Nelson and Boor, 1982; Misra et al., 1995). Alpha class GSTs have also been shown in bovine pulmonary microvessel endothelial cells (He et al., 1996), suggesting that GSTs may be involved in defense of these cells against oxidative stress. Further research is needed, however, to define the role of GSTs during vascular injury induced by oxidative stress, and in the early human atherosclerotic plaque.

The second example to be addressed in this report is the dissecting aortic aneurysm (DAA). Dissection is the sudden, dramatic tearing, or splitting, of the medial layers of an artery resulting in false lumens, occlusion of distal vessels (such as coronary arteries), rupture, hemorrhage, and rapid or sudden death (Evangelista et al., 2005). It has long been recognized that certain connective tissue metabolic disorders, most notably Marfan syndrome, carry a high risk of developing dissecting aneurysms, however recognition of the occurrence of dissection as an isolated disease, and concern about this lethal vascular disease that frequently runs in families and affects younger persons, is growing (Helliker and Burton, 2003).

The largest elastic artery—the aorta—is by far the most common site for dissection, and DAA almost universally involves the thoracic aorta, localizing to the aortic arch, with frequent direct extension to arteries of the head, neck, and upper extremities. Little is known about the mechanism of formation of DAA, although, as alluded to above, it is a frequent cause of death in patients with Marfan syndrome, and surgical therapeutic strategies have been successfully devised for patients with this devastating disease. While several animal models are currently available and are used in the study of abdominal aortic aneurysm (Stehbens, 1999; Manning et al., 2002), no reproducible experimental model exists for the study of isolated DAA.

Our earlier study in weanling rats suggested that inhibitors of the little-studied enzyme semicarbazide-sensitive amine oxidase (SSAO) have a severe deleterious effect on normal vascular development, especially of aorta (Langford et al., 1999). SSAO expresses high activity in smooth muscle cells, and aorta is exhibits highest activity of any isolated tissue. A recent microarray study has revealed that the SSAO homologue, a membrane-bound copper amine oxidase known as vascular adhesion protein-1 (VAP-1) is downregulated in human dissecting ascending aortic aneurysms (Muller et al., 2002).

Stimulated by these previous data, we sought to extend our knowledge of the deleterious effects of SSAO inhibition on vascular development to potential in utero effects. In the present study, by administering the known SSAO inhibitor, semicarbazide, to timed-pregnant rats during the critical period of aortic development, we have developed a model of DAA with striking morphologic similarities to the disease in humans. This model of in utero exposure to relatively non-toxic levels of a xenobiotic could be used to explore how underlying embryologic changes in the aortic wall might lead to the disastrous vascular pathology known as dissection.

These studies serve to raise awareness of the metabolic potential of the vascular wall, and suggest the possibility that environmental factors may adversely affect blood vessels through prenatal influences.

Materials and Methods

Immunohistochemical Localization of GST A4-4 in Human Aorta and Atherosclerotic Plaque

Full-thickness segments of ascending aortic arch were procured fresh from five human donor hearts (donor ages: 16–25 years; mean age: 22; 4 male, 1 female) during routine cardiac transplant procedures under an approved human subject protocol (University of Texas Medical Branch IRB#99 = 395). Tissue was rapidly fixed in 10% neutral-buffered formalin solution and early atherosclerotic plaques (0.1–0.3 cm in greatest dimension) were identified grossly; plaque and adjacent normal aorta were routinely processed for light microscopy, sectioned at 4 μm, and stained with Movat’s pentachrome stain. Immunohistochemical demonstration of hGSTA4-4 utilized a highly specific polyclonal IgY reagent raised in chicken against full-length recombinant hGSTA4-4 (kindly supplied by Dr. P. Zimniak). Standard Streptavidin/horseradish peroxidase (Pharmingen, San Diego, CA) and diaminobenzidine immunohistochemical methods were used; negative controls consisted of omission of primary antibody and use of pre-immune serum as primary antibody.

Chemicals and Animal Treatment

Semicarbazide hydrochloride was purchased from Sigma Chemical Co (St. Louis, Mo). Timed pregnant Sprague–Dawley rat dams were treated with either physiological saline (control, n = 7) or SSAO inhibitor semicarbazide by intraperitoneal injection daily on days 14 to 20 of gestation. A dose response was established from 0.10 to 24.50 mg/kg; lesion incidence was greater than 90% at doses above 6.0 mg/kg/day. Newborn rats were sacrificed on day one of postnatal life by carbon monoxide inhalation of the mother. These experiments were approved by the Animal Care and Use Committee, the University of Texas Medical Branch at Galveston (Protocol 8812178).

Morphologic Study of 1-Day-Old Rats

Intraperitoneal injection of 1, 2, and 3 ml 10% neutral-buffered formalin was performed immediately, and thoracic and newborn’s abdominal organs were removed en bloc and immersion fixed in 10% neutral-buffered formalin. Following routine dehydration and routine processing, three levels were stepped sectioned at (1) the aortic root at level of aortic valve; (2) the mid- aortic arch; (3) the thoracic carotid artery. This near-serial cross sectioning allowed for examination of the entire aortic arch and branches. Histological sections were stained with routine hematoxylin and eosin (H + E), as well as Movat (for collagen and elastin).

Results

GST A4-4 in Human Vessel and Human Atherosclerosis

Aortic samples obtained freshly from young organ donors were found to exhibit predominantly normal histology, with focal, small, early atherosclerotic plaques 1–3 mm in diameter. These plaques were morphologically characterized as small, stage 3 (1–2 mm in greatest dimension, foam cell, proliferating vascular smooth muscle cells,) or stage 4 (rare cholesterol clefts) fibrous plaques by standard histopathologic criteria (Stehblens and Lie, 1995). Stage 2 plaques (lipid streaks) were not examined by immunohistochemistry.

Immunohistochemical demonstration of hGSTA4-4 of normal aortic wall adjacent to plaques showed scattered staining in normal vascular smooth muscle cells of media and no staining evident in normal endothelium. Within plaques (Figure 1), an intense increase in staining was seen in the cytoplasm in vascular smooth muscle cells, within foam cells, and in the cytoplasm of endothelial cells overlying plaque, indicating induction of the isozyme. In addition, the vascular smooth muscle directly beneath plaques demonstrated an evident increase in staining for hGSTA4-4.

General Observations in DAA Experiments

No differences in maternal weight, stillbirths, litter size, or mean pup weight was found among groups born to dams given semicarbazide. Remarkably, no mortality, or overt morbidity, was observed in pups at any time and growth curves (1–28 days; data not shown) did not differ between groups. Lesions in aorta appeared to resolve by fibrosis of medial dissection at 28 days without pathologic consequences or clinically observable effects (data not shown).

Morphologic Observations: Dissecting Aortic Aneurysms

In newborn, 1-day-old pups, striking dissection of blood in the thoracic aorta occurred in 100% of pups born to dams treated with the highest doses; at a dose level of 6.125 mg/kg the incidence of dissection was greater than 90% (Table 1). At the 2 lowest doses (0.10 and 1.153 mg/kg) only one aortic dissection was found. Dissections were massive, greatly increased the diameter of outer adventitia of the aorta, and frequently extended into the vessels of the aortic arch and superior to the carotid arteries (Figure 2). The dissection of blood involved the outer third of aortic media, and was consistently located in the interstitial space between media and adventitia. The blood was limited by a markedly displaced (dilated) adventitia or surrounding structures, thus forming a “false lumen.” Complete tearing of the media was seen focally in a minority of cases (17–30%). No evidence of blood was seen in the pericardium (hemopericardium) or thoracic cavities (hemothorax).

Movat stain revealed that irregular areas of reduced, frayed elastic lamellae were found focally in vascular media of all vessels examined. In some areas adjacent to the false lumen we observed complete tearing of the media. In contrast, in other regions of the dissecting aneurysm within the same aorta, we observed considerable external dilation with an intact medial layer and narrowing of the original aortic lumen. No intimal lesions, stenosis, or inflammatory cells were found. Focal, small areas of mural necrosis characterized by homogenous pink staining of vascular smooth muscle were-observed occasionally in H+E stain.

Discussion

Atherosclerosis and the Human Plaque

Oxidative stress, with resultant lipid peroxidation, has been implicated in the pathogenesis of numerous diseases, including atherosclerosis, diabetes, cancer, and rheumatoid arthritis, as well as in drug-associated toxicity, and aging. There is increasing evidence that aldehydes produced during lipid peroxidation reactions are causally involved in many of the pathophysiologic effects associated with oxidative stress in cells and tissues (Uchida, 1999).

A subgroup of α-class mammalian GSTs, which utilize 4-HNE as their preferred substrate, has been proposed to be a major cellular defense system against oxidative injury (Cheng et al., 1999; Yang et al., 2001). Our previous studies indicated that GSTs play an important protective role in vascular smooth muscle cells exposed to allylamine, a primary unsaturated amine that is thought to exert its toxic effect both in vivo and in vitro through conversion to acrolein by the action of semicarbazide-sensitive amine oxidase, or SSAO (He et al., 1998). In these early studies, transfection with rGST8-8, a 4-HNE metabolizing isozyme, afforded protection from allylamine or acrolein exposure, and inhibition of GSTs resulted in enhanced toxicity. We also know that GSTs catalyze the first step of detoxification of acrolein to mercapturic acid by conjugation with GSH (Sanduja et al., 1989). Hence, the clear protective effect of mGSTA4 transfection on allylamine toxicity is likely due to augmentation of this detoxification pathway.

4-Hydroxynonenal is a relatively stable end product of lipid peroxidation that is generated by most common oxidative toxicants. GSTA4-4 has high activity toward α, β-unsaturated aldehydes, so in this study, it is likely that GSTA4-4 is acting in a protective role against aldehydes generated during lipid peroxidation, such as 4-HNE or acrolein. 4-HNE and H₂O₂ have been shown to cause apoptosis in a variety of cell lines by activating signal transduction pathways with some common features, such as activation of JNK, and the caspase cascade; GSTs play a role in protection of these cells from apoptosis (Camandola et al., 2000; Yang et al., 2001, 2002). In the vascular wall, oxidative stress-induced endothelial cell apoptosis may be an initiating event in atherogenesis, but the biochemical mechanisms regulating endothelial cell apoptosis in response to modified lipids or oxidative stress are incompletely understood (Choy et al., 2001).

Our human studies of hGSTA4-4 in atherosclerosis have confirmed for the first time the expression of GST enzymic protein in normal vascular wall. It is of interest, however, that our immunocytochemical localization studies in normal aorta demonstrated hGSTA4-4 reactivity in the media only; normal aortic endothelium showed no evidence of cytoplasmic, or other staining. Microvascular endothelium from pulmonary vessels has previously been found to express GSTs including an isozyme with characteristics of mGSTA4-4 (He et al., 1996; Hagan et al., 2000; Januzzi et al., 2004), suggesting that differing endothelia may vary in expression. Further studies are necessary to clarify the localization and function of GST isozymes, including GSTA4-4, in various vascular beds.

The aortic plaques which we studied immunohistochemically represent the earliest stages of human atherosclerosis (1–3 mm), before excessive lipid is stored in the plaque or the plaque is complicated by calcification, plaque discontinuity, or gross, secondary thrombosis. We interpret the occurrence of intense staining for hGSTA4-4 in the proliferative vascular smooth muscle cells and foam cells as a reactive, protective phenomenon in the face of oxidative stress and lipid peroxidation, and the consequent production of highly reactive unsaturated alkenals such as 4-HNE. In addition, the appearance of intense cytoplasmic staining in endothelial cells overlying plague argues that endothelial induction of hGSTA4-4 occurs early as a protective mechanism adopted by endothelial cells exposed to oxidative stress during atherogenesis.

Overall, these studies suggest an important role for GST defense mechanisms in protecting the vascular wall against oxidative stress, and oxidative atherogenic injury. By manipulating such defense mechanism, it may be possible in the future to prevent or delay the progression of the commonest human vascular disease, atherosclerosis.

Dissecting Aortic Aneurysm

Dissecting aortic aneurysm (DAA) is the sudden tearing of aortic media, usually originating in the thoracic aorta, and frequently resulting in sudden death (Hagan et al., 2000; Nienaber and Eagle, 2003; Januzzi et al., 2004). DAA occurs in younger adults, including teenagers (Helliker and Burton, 2003) making it distinct from aneurysms of the abdominal aorta, which are characterized by male predominance and occurrence in older individuals (Sakalihasan et al., 2005). Furthermore, DAA and abdominal aortic aneurysm differ in typical pathologic characterisistics, since DAA is associated with splits or tears of vascular media without dilatation, whereas abdominal aortic aneurysms are characterized by marked dilatation and wall thinning. Indeed, the term “aneurysm” means “widening,” a characteristic of abdominal aneurysms, but not of DAA.

DAA is relatively frequent in patients’ with Marfan syndrome, where genetic defects in fibrillin-1 are thought to lead to abnormal elastin formation as the cause of dissection (Hayward and Brock, 1997; Pereira et al., 1997). The clinical condition of Ehlers-Danlos syndrome Type IV, the vascular type, is also associated with dissection and rupture of aorta and other vessels (Pepin et al., 2000). In the majority of DAA, however, no etiology is evident.

To date, no useful reproducible DAA models have been available (Ikonomidis et al., 2003). This is in contrast to abdominal aortic aneurysm, which has been extensively modeled, predominantly through regional elastase infusion of the aorta, with consequent collagen and elastin degradation and accompanying increases in matrix metalloproteinase production (Boyle et al., 1998).

In the present study, we have illustrated a remarkable developmental form of dissecting aortic aneurysm. By treating dams with semicarbazide in the late gestational period (14–21 days in rat) a DAA is nearly universally induced at birth in rat pups. Hence, this developmental model is likely to be the result of disruption of late gestational or embryologic processes involved in vasculogenesis of the thoracic aorta. Surprisingly, this lesion is silent in the newborn, since it appears to be accompanied by no mortality or acute morbidity. We have performed preliminary experiments following the progression of the aortic dissections out to 28 days after birth (data not shown in the present study). Remarkably, no apparent chronic morbidity has been noted and, hence, these dissections appear to be clinically silent in the short term following birth. The long term consequences of these lesions are presently under study in our laboratory.

Mounting evidence suggests that degenerative changes in the structural components of the vascular wall are the precursor to dissection (Brooke et al., 2003). Degenerative changes in the media are thought to involve the complex elastin lamellae. These intricately interconnected layers are a complex mix of the large protein, elastin, laid down on a substrate of the microfibrillar protein, fibrillin (Brooke et al., 2003). The occurrence of DAA in elderly individuals with hypertension (Angouras et al., 2000) suggests that age-related loss and degeneration of elastin, with consequent weakening in normal elastic properties, may predispose to dissection. Other possibilities include alterations in any of the many complex interstitial proteins, including the collagens.

The morphologic similarities to the DAA shown in this study to that seen in Marfan syndrome argue strongly that some defect in the structure of elastin is likely to underlie the striking shearing of vascular media. Specifically, the DAA in our model appears to begin in the arch of the aorta, likely in the ascending portion based on the prevalence of lesions in that location and in the finding of through-and-through tears of media there. Similar to DAA in humans, the dissection apparently continues inferiorly down the abdominal aorta and superiorly into the carotids.

In summary, these experiments with the cellular oxidative defense enzyme GST suggest a major metabolic role for media during oxidative stress and toxic insults. Also, we have described a reproducible model of DAA in newborn rats from dams treated with the SSAO inhibitor, semicarbazide on days 14–21 of in utero development. The model has no discernible mortality or morbidity (at least in the short term), and also affords a model to examine the potential long term effects of vascular remodeling of DAA. While our initial data does not uncover a clear mechanism of DAA with regard to collagen, elastin, or other matrix defects, our working hypothesis is that DAA is caused by aberrations of elastogenesis or matrix formation during the critical embryologic period of vasculo-genesis. Through future studies, it is hoped that therapeutic strategies can be developed for the lethal vascular disease of dissection aortic aneurysm.

These experimental examples emphasize the role of the media in toxic insults to blood vessels. Also, the potential importance of toxic injury to developing blood vessels by in utero exposure to xenobiotic substances is illustrated.