Influence of Hypercholesterolemia and Adventitial Inflammation on the Development of Aortic Aneurysm in Rabbits
Abdominal aortic aneurysms are characterized by intimal atherosclerosis, disruption and attenuation of the elastic media, and a variable adventitial inflammatory infiltrate. We have developed an animal model of this disorder to evaluate the contribution of hypercholesterolemia, medial injury, and adventitial inflammation to aneurysmal dilatation. To accomplish this, we used periaortic application of calcium chloride, which induced both medial injury with calcification and endothelial injury. Ultrasonography was used to demonstrate the dilatation and thickening of the aortic wall. Over the first 3 weeks after periaortic application of 0.25 mol/L CaCl2, the external aortic diameter increased from 3.5±0.5 to 4.2±0.8 mm, but the ID remained unchanged. This apparent wall thickening was accompanied by vascular remodeling, and biochemical changes included ≈50% reduction in tissue hydroxyproline concentration and increased activity of gelatinases (matrix metalloproteinase [MMP]-2 and MMP-9). Independently, cholesterol feeding to induce hypercholesterolemia or the concomitant periaortic application of thioglycollate had little effect on the histological, biochemical, or diameter changes. Together, hypercholesterolemia and thioglycollate were associated with rapid aortic dilatation in CaCl2-treated animals but not controls: after 3 weeks, the ID and OD had doubled, the OD increasing from 3.5±0.4 to 7.1±0.4 mm, P=.005. The remarkable feature that accompanied this dilatation was the infiltration of cells, mostly foamy macrophages, into the adventitia, with a further reduction in hydroxyproline concentration. Adventitial inflammation may provide the critical stimulus to dilatation of an aorta with preexisting intimal and medial injury.
- Received December 15, 1995.
- Revision received April 25, 1996.
In 1988 Gertz et al1 described a method of inducing an aneurysm in the rabbit common carotid artery by the periarterial application of calcium chloride. In most species, including humans, the common carotid artery, like the aorta, is an elastic artery. However, whereas aneurysms of the abdominal aorta are a common clinical problem, aneurysms of the carotid artery are relatively uncommon. Since there are few good experimental models of aortic aneurysm,2 we hoped to induce experimental aortic aneurysms with calcium chloride.
In humans, abdominal aortic aneurysms are usually a degenerative disorder of later life, being associated with smoking, hypertension, and atherosclerosis in addition to having a familial tendency. Histology of the aneurysmal wall shows atherosclerosis, medial disruption, and atrophy with fibrous replacement and a variable spectrum of inflammation in the media and adventitia.3 Fibrotic changes in the adventitia may cause thickening of the aortic wall, but intimal hyperplasia and dense calcification were not observed.3 This description differs from that of the carotid artery aneurysm induced by calcium chloride.1 In this model, the increasing OD of the common carotid artery was attributable principally to a threefold increase in wall thickness with extensive intimal hyperplasia superimposed on a modest dilatation. Dense calcification of the elastic tissue of the media was the focus of an inflammatory reaction similar to that observed in giant cell arteritis. These changes in the rabbit carotid artery occurred in response to the periarterial application of 0.5 mol/L CaCl2; the same concentrations of MgCl2 or NaCl did not evoke a similar response.1
In smaller animals (rats), fusiform aortic aneurysms of more characteristic pathological appearance have been induced by the intraluminal administration of elastase.4 This experimental model provides support for the hypothesis that destruction of medial elastic tissue is critical to aneurysmal dilatation. An inflammatory response follows elastase infusion, the extent of which correlates with aneurysmal dilatation.5 This inflammatory response may be augmented by the intraluminal administration of plasmin and thioglycollate; again, the extent of the inflammatory response was associated with the magnitude of aortic dilatation.6 In humans, there is a similar association between the aneurysm diameter and the extent of adventitial inflammation.3 Many consider atherosclerosis to be the cause of abdominal aortic aneurysm.7 Aortic aneurysms have not been reported in hypercholesterolemic rabbits except when multiple fusiform aneurysms were observed as an uncommon consequence of aortic balloon injury.8
The purpose of this study was to investigate whether adventitial insult, in the form of calcium chloride application, could cause abdominal aortic aneurysms in rabbits and how the course of any aneurysmal dilatation was influenced by hypercholesterolemia or thioglycollate.
The study used New Zealand White rabbits, and all procedures were performed according to a protocol approved and licensed by the Home Office. Rabbits (2.5 to 3.5 kg) were anesthetized with fluanisone 4 mg·kg−1 and fentanyl 0.08 mg·kg−1 IM (Hypnorm, Roche) and midazolam 0.2 mg·kg−1 IV (Hypnovel, Janssen). Two centimeters of the proximal infrarenal abdominal aorta were exposed by careful dissection via a midline incision and “painted” for 15 minutes with sterile solution containing either sodium chloride or calcium chloride, which was applied gently with a cotton wool tip during this period. Calcium chloride (0.01, 0.1, 0.25, or 0.5 mol/L) and sodium chloride (0.5 mol/L) were prepared in deionized, distilled water, the final pH being adjusted to 7.0 with sodium hydroxide, except for the lowest concentration of calcium chloride, which was prepared in 0.14 mol/L NaCl. For some experiments, sodium thioglycollate (0.05 mol/L) was included in the solution. After painting of the aorta, in most cases, abdominal contents were replaced, the incision was closed, and the animal was allowed to recover. The experiments were performed in two phases: phase 1, variation of the concentration of CaCl2 applied, and phase 2, concomitant application of thioglycollate with or without cholesterol feeding. The schedule of experiments is given in Table 1⇓.
Acute, Nonrecovery Studies
Some animals were killed by an overdose of pentobarbitone 60 mg·kg−1 (RMB Animal Health) 30 minutes after the periaortic application of salt solutions; after the painted region was carefully marked, the infrarenal aorta was excised and dissected free from adherent tissue, and the vessel was transported to the laboratory in oxygenated Krebs solution at 4°C for studies of endothelium-dependent relaxation. Other animals were given 10 mL IV of a solution containing 200 mg Evans blue dye (Gurr) and 1 g BSA (Cohn fraction V, Sigma). After 10 minutes to allow the dye to circulate, the animals were killed with intravenous pentobarbitone. They were immediately perfusion-fixed by flushing of the circulation via cannulas in the distal abdominal aorta with isotonic saline at a pressure of 100 cm H2O, with drainage through a jugular vein catheter. After 1.5 minutes, the initial perfusion solution was replaced with 4% glutaraldehyde, and the vessels were perfusion-fixed in situ for 20 minutes. After the treated zone of the aorta was marked, the vessel was excised, postfixed in 1% glutaraldehyde, and transported to the laboratory for assessment of distribution of Evans blue and sectioning.9
Organ Bath Studies of Endothelium-Dependent Relaxation
Five-millimeter rings of treated (proximal infrarenal) aorta and untreated control (distal infrarenal) aorta were suspended between metal stirrups, the upper one being connected to a pressure transducer, in 10 mL oxygenated Krebs solution at 37°C in a tissue organ bath, as previously described for saphenous vein.10 The aortic rings were allowed to equilibrate for 20 minutes and then were stretched to 2 g, after which a further equilibration period of 20 minutes was allowed. The aortic rings were then exposed to 30 μmol/L phenylephrine (Sigma), and the resultant contraction was noted. After the system was flushed and the aortic ring allowed to relax for 10 minutes, the process was repeated with 0.5-g increments of stretch until the optimum tension-response was obtained, usually between 2 and 3 g. The aortic rings were precontracted with phenylephrine to ≈70% of maximum contraction and then exposed to increasing doses of acetylcholine (10−9 to 10−5 mol/L) to evoke endothelium-dependent relaxation. At the end of the experiment, the viability of the medial smooth muscle was confirmed by observation of a relaxation to sodium nitroprusside 10 μmol/L (Sigma).
Endothelial Permeability Studies
Abdominal aorta excised from Evans blue–treated animals was divided into treated and untreated regions. Frozen sections (10 μmol/L) were cut transversely through the aorta and were photographed by fluorescence microscopy with a filter set with excitation illumination of 450 to 490 nm and emission at wavelengths >520 nm. Photomicrographs were obtained by an observer without knowledge of whether the periarterial application was sodium- or calcium-based.
Sections of painted and unpainted aorta immediately adjacent to those used for organ chamber studies were fixed for routine histological examination with hematoxylin and eosin, elastic van Gieson, and Mallory's trichrome staining.
For these studies, the abdominal aorta was imaged by ultrasonography with an Aloka SSD500 scanner with a 7.5-MHz linear transducer before laparotomy. With the rabbit lying in the right lateral position, a longitudinal image of the abdominal aorta was obtained. The aortic ID and OD were measured at a point 3 cm proximal to the iliac bifurcation. Diameters were measured to the nearest millimeter; readings from repeated images (three or four), which did not vary by >1 mm, were used to calculate the mean diameter. The aorta was imaged before laparotomy and at weekly intervals thereafter. After recovery, the animals were allowed to feed and drink ad libitum; some animals were fed a chow enriched in cholesterol (0.5%). At weekly intervals, the animals were weighed, their aortas imaged by ultrasonography, and aortic ID and OD recorded. From cholesterol-fed rabbits, blood was obtained from the ear vein for measurement of total cholesterol.
Histological and Functional Studies
Animals were killed at 3 to 12 weeks. The aortas were dissected carefully to remove adherent tissue and processed for use in the organ chamber, permeability studies, and histological studies as described for the acute nonrecovery studies. In addition, unfixed samples of aorta were frozen in freezing isopentane or liquid nitrogen for future immunohistochemical and biochemical analysis. The histology studies were more extensive than for the acute nonrecovery aortas. Sections were stained with hematoxylin and eosin, Masson's trichrome, elastic van Gieson, and oil red O in addition to immunochemical staining of macrophages with a monoclonal antibody, RAM11 (Dako), and smooth muscle cells with a monoclonal antibody specific for smooth muscle actin, IA4 (Dako). Some sections were stained with a sheep polyclonal antibody against human gelatinase A, which recognizes rabbit gelatinase A; this antiserum was a kind gift from Dr G. Murphy, Strangeways Laboratory, Cambridge.
The sections were assessed for endothelial coverage, intimal thickening, presence of internal elastic lamina, medial calcification, characteristics of elastic lamellae, smooth muscle cells, adventitial thickness and cellularity, and the presence of inflammatory cells (lymphocytes/macrophages). The intensity of inflammation was assessed by grading of the number of inflammatory cells from 0 (no cells) to 4+ (>100 cells per high-power [×40] field). Semiquantitative assessment of intimal proliferation, intimal-to-medial thickness ratio, and wall thickness was obtained as described by Weidinger et al,11 which permits physiological, biochemical, and histomorphometric measurements to be made in the same tissue and therefore minimizes the number of animals used. Briefly, excised aortic rings were fixed and embedded, and three cross sections stained with hematoxylin and eosin were prepared from each aorta. Intimal and medial thicknesses were measured at 24 polar coordinates with a digitizing tablet, the ratio of intimal to medial thickness was calculated at each coordinate, and the mean value for each ring was determined.
Metalloproteinase Activity and Collagen Content
Full-thickness (intima, media, and adventitia) frozen aorta (100 to 300 mg net weight) was diced finely and homogenized with a Polytron in 6 vol ice-cold 1 mol/L NaCl containing 2 mol/L urea, 0.1% EDTA, 0.1% Brij 35, 0.2 mmol/L phenylmethylsulfonyl fluoride, and 50 mmol/L Tris/Cl−, final pH 7.6. The homogenate was centrifuged at 100 000g for 1 hour at 4°C. The supernatant was dialyzed against 50 mol/L Tris/Cl−, pH 7.6, containing 0.15 mol/L NaCl, 10 mol/L CaCl2, 0.1% Brij 35, and 0.2 mmol/L phenylmethylsulfonyl fluoride before being used for gelatin zymography. The precipitate was washed well with water and then dried to constant weight before preparation for amino acid analysis by hydrolysis in 6N HCl at 110°C for 48 hours. Amino acid analysis was performed on an LKB4151 Alpha Plus amino acid analyzer, with norleucine as an internal standard. The protein content of the supernatant was estimated by a dye binding assay (Coomassie protein reagent, Pierce Chemical Co), and 10 μg protein was used for gelatin zymography.12 Samples were activated with 1 mmol/L p-aminophenylmercuric acetate at 37°C for 1 hour before electrophoresis. Semiquantitative assessment of gelatinase activity was made by laser densitometry of zymograms. Monoclonal antibodies to matrix metalloproteinase [MMP]-1 (collagenase), MMP-2 (gelatinase A), and MMP-9 (gelatinase B) (from Oncogene Science SA) were used for the identification of metalloproteinases by Western blotting.
Results are expressed as mean±SD except for endothelium-dependent relaxations, for which results are reported as mean±SEM and differences were assessed by ANOVA. Other differences between groups were assessed by Student's t test for paired data and the Mann-Whitney U test for unpaired data. Statistical significance was taken as a probability value of <.05. Measurements of aortic diameters in vivo and in vitro were compared by the method of Bland and Altman.13
Measurement of Aortic Diameter
In 10 animals, the ultrasonographic diameter measured in vivo was compared with the diameter obtained from histomorphometric analysis of aortic sections after pressure perfusion fixation of the aorta (Fig 1⇓). All measurements were from 3-week experiments and included treatment of the aorta with 0.5 mol/L NaCl, 0.25 mol/L CaCl2, and 0.5 mol/L CaCl2. Ultrasonography overestimated the OD compared with morphometry (mean difference, 0.4 mm), but the limits of agreement were narrow (±0.2 mm) and did not increase with increasing diameter. The two measures of ID were similar (mean difference, 0.1 mm), but the limits of agreement (±0.3 mm) were wider than for OD. Therefore, ultrasonography provides a reliable method for monitoring aortic diameter in vivo, although there is a constant overestimation of aortic OD by 0.4 mm compared with morphometric analysis.
Attempts to Create Experimental Aneurysms With Periaortic Application of 0.5 mol/L CaCl2
The conditions described by Gertz et al1 to create carotid artery aneurysms in rabbits were used in an attempt to fashion abdominal aortic aneurysms, with aortic ID and OD being monitored by ultrasonography. For the adult New Zealand White rabbit, the ID of the abdominal aorta was 2.6±0.4 mm and the OD was 3.5±0.5 mm (n=16). No change in either aortic ID or OD was observed for up to 12 weeks after periaortic application of 0.5 mol/L NaCl (the ODs are given in Table 2⇓). After periaortic application of 0.5 mol/L CaCl2 (n=8), there was a gradual increase over 3 weeks in aortic OD to 5.0±0.6 mm (P=.018 compared with original diameter), after which time there was no further change in aortic OD up to 12 weeks (Table 2⇓). There was no significant change in aortic ID (3.0±0.5 mm at 3 weeks, 2.9±0.4 mm at 12 weeks) during the same time period. Periaortic application of 0.5 mol/L CaCl2 appeared to cause thickening of the aortic wall but not aneurysmal dilatation.
This apparent thickening of the aortic wall was confirmed by histology and measurement of intimal-to-medial thickness ratio. The full thickness of the aortic wall after 3 weeks was 120±28 μm and 286±45 μm in animals treated with 0.5 mol/L CaCl2 and 0.5 mol/L NaCl, respectively. After 3 weeks, the intimal-to-medial thickness ratio was 0.88±0.13 and <0.05 in aortas treated with 0.5 mol/L CaCl2 and 0.5 mol/L NaCl, respectively. Three weeks after periaortic application of 0.5 mol/L CaCl2, gross histological changes were observed, with dense calcification of the media and circumferential intimal hyperplasia (Fig 2d⇓). In the retroperitoneal fat beyond the aortic wall, granulomatous foci were present. The cells in the thickened intima stained positively with antibodies to smooth muscle cell actin (not shown).
Titration of the Concentration of Calcium Chloride
We were concerned that after application of 0.5 mol/L CaCl2, the extensive calcification of the aortic media and the new connective tissue in the intima might provide a rigid framework that would prevent aortic dilatation. We commenced “titrating” the concentration of periaortic calcium chloride, initially using each treatment in two rabbits. The aim of this titration was to establish a condition that provided medial injury without the extensive medial calcification and intimal hyperplasia observed in response to 0.5 mol/L CaCl2. Transverse sections of the aorta 3 weeks after treatment with 0.01, 0.1, 0.25, and 0.5 mol/L CaCl2 are shown stained with hematoxylin and eosin in Fig 2⇑. Untreated aortas and aortas treated with 0.5 mol/L NaCl did not appear much different from the aorta treated with 0.01 mol/L CaCl2 (Fig 2a⇑). There was scant evidence of either calcification or intimal hyperplasia in the aortas treated with 0.1 mol/L CaCl2, although the inner elastic lamellae appeared to be extended (Fig 2b⇑). The aortas treated with 0.25 mol/L CaCl2 showed evidence of medial injury and patchy intimal hyperplasia. The overlying endothelium appeared to be intact. In the media, there was partial calcification of the elastic lamellae, which were straight and extended, and the smooth muscle cells were attenuated. The external elastic lamina also was straight and extended. Inflammatory cells were uncommon and confined to occasional foci of dense calcification at the medial/adventitial junction.
Although periaortic application of 0.25 mol/L CaCl2 caused medial injury without the gross intimal hyperplasia and calcification observed in response to 0.5 mol/L CaCl2, aneurysmal dilatation of the abdominal aorta was not observed. By ultrasonography, after 3 weeks, the maximum ID of aortas treated with 0.25 mol/L CaCl2 (n=5) was 2.8±0.5 mm, and the maximum OD was 4.2±0.8 mm (Table 2⇑). No further changes in aortic diameter occurred up to 12 weeks (Table 2⇑). Moreover, attempts to stimulate aortic dilatation by hypercholesterolemia (induced by diet) did not effect significant changes in aortic diameter or thickness up to 12 weeks or histology at 3 weeks after periaortic application of 0.25 mol/L CaCl2, even though the serum cholesterol concentration had increased from 1.8±0.2 to 28.9±5.6 mmol/L (n=4).
Endothelial Injury After Periaortic Application of CaCl2
Before making further changes to the procedure in an attempt to effect aneurysmal dilatation, we considered it important to investigate whether periaortic application of 0.25 mol/L CaCl2 caused endothelial injury in addition to medial injury.
Half an hour after periaortic application of 0.25 mol/L CaCl2, endothelium-dependent relaxation in response to acetylcholine was abolished, although endothelium-independent relaxation in response to sodium nitroprusside (10 μmol/L) was preserved (n=4). Three weeks after periaortic application of 0.25 mol/L CaCl2, there was still no endothelium-dependent relaxation in response to acetylcholine, and evidence of mild impairment of the endothelium-independent response of maximum relaxation in response to 10 μmol/L sodium nitroprusside had decreased from 90±6% (t=0) to 78±7% (t=3 weeks). In contrast, although periaortic application of 0.5 mol/L NaCl caused acute abolition of endothelium-dependent relaxation in response to acetylcholine, 3 weeks later this response had recovered, and maximum relaxation (57±4%) and endothelium-independent relaxation (to sodium nitroprusside) remained normal throughout (Table 3⇓).
Similarly, 30 minutes after periaortic application of 0.25 mol/L CaCl2, there was increased permeability of the endothelium to albumin–Evans blue complexes (this is visualized as an intense fluorescence; dark gray in Fig 3b⇓). In contrast, periaortic application of 0.5 mol/L NaCl did not cause the endothelium to become permeable to albumin–Evans blue complexes (Fig 3a⇓).
Aneurysmal Dilatation Elicited by Periaortic Application of 0.25 mol/L CaCl2 and Thioglycollate
Thioglycollate often is used as a chemical activator of macrophages. Periaortic application of 0.05 mol/L thioglycollate together with either 0.5 mol/L NaCl or 0.25 mol/L CaCl2 did not effect an increase in aortic diameter or histological change (hematoxylin and eosin) after 3 weeks compared with aortas treated with 0.5 mol/L NaCl or 0.25 mol/L CaCl2 alone. When rabbits were fed a cholesterol-enriched diet, there was still no change in aortic diameter or thickness 3 weeks after periaortic application of 0.05 mol/L thioglycollate and 0.5 mol/L NaCl, although small foci of intimal hyperplasia were observed. However, in rabbits fed a cholesterol-enriched diet whose aortas had been treated with 0.25 mol/L CaCl2 and thioglycollate, ultrasonography demonstrated progressive localized dilatation over 3 weeks. Sonograms of the aorta of cholesterol-fed rabbits before and after treatment with 0.25 mol/L CaCl2 and thioglycollate are shown in Fig 4⇓. The dilatation was maximal proximally, and often the rib cage prevented imaging to the neck of the aneurysm. The maximum aortic ID had increased from 2.7±0.4 to 5.7±0.4 mm (P=.017) and the aortic OD from 3.5±0.4 to 7.1±0.4 mm (n=5) (P=.005) after 3 weeks (Table 2⇑). The diameter of the distal aorta remained unchanged. Little additional increase in the maximum aortic diameter was observed up to 12 weeks (n=2), although one more rabbit died of aortic rupture.
The most remarkable histological change in these aneurysmal aortas at 3 weeks was the cellular infiltrate in the adventitia. There was still evidence of patchy intimal hyperplasia and medial injury, as described previously (Fig 2c⇑). The adventitia was thickened and hypercellular, with evidence of an inflammatory infiltrate, including lymphocytes (Fig 5⇓). The presence of numerous foamy macrophages in the adventitia was demonstrated by immunostaining with the monoclonal antibody RAM-11 (Fig 5⇓). Macrophages were not observed in the intima and were scant in the media. Any oil red O staining of the aortic wall was confined to the adventitial macrophages. The density of adventitial macrophages was graded as 4+, compared with 1+ for aortas in normal-chow rabbits, in which the aorta developed lesions with 0.25 mol/L CaCl2 plus thioglycollate, or 0 in cholesterol-fed rabbits, in which the aorta developed lesions with 0.5 mol/L NaCl plus thioglycollate. Other histological changes included the disruption and fragmentation of medial elastic fibers.
Changes in Collagen Content and Metalloproteinase Activity in Aneurysmal Aortas
The amount of hydroxyproline per microgram dry weight was used to monitor the insoluble collagen in the different experimental aortas. The hydroxyproline concentration of normal abdominal aorta was 565±55 nmol/μg. The hydroxyproline concentrations were 489±34 and 261±43 nmol/μg 3 weeks after periarterial application of 0.5 mol/L NaCl and 0.25 mol/L CaCl2, respectively. Therefore, there appeared to be a reduction of ≈50% in the concentration of insoluble collagen in the aortas that developed lesions with CaCl2, P=.005. Independently, cholesterol feeding or periaortic application of thioglycollate had little effect on the hydroxyproline content, but the hydroxyproline content was reduced further to 207±24 nmol/μg in the aneurysmal aortas 3 weeks after periarterial application of 0.25 mol/L CaCl2 and thioglycollate in cholesterol-fed animals (P=.042 compared with 0.25 mol/L CaCl2 only). The amount of desmosine/isodesmosine was too low to permit quantification of the elastin cross-links.
Assessment of collagenase (MMP-1) was made by Western blotting. Homogenates from CaCl2-treated aortas showed very faint bands at 55 kD, which were not visible in homogenates from NaCl-treated aorta. Gelatin zymography of aortic homogenates showed the presence of four principal lytic zones at molecular weights of 92, 85, 72, and 65 kD. On Western blotting, the 92- and 85-kD zones were recognized by antibodies to MMP-9 (gelatinase B) and the 72- and 65-kD lytic zones by antibodies to MMP-2 (gelatinase A). In normal aorta, all these bands were faintly visible (Fig 6⇓, lane 7). All treatments of the aorta were associated with increased gelatinase bands, even with periaortic application of 0.5 mol/L NaCl. Densitometry indicated that the intensity of the 72-kD zone had doubled (Fig 6⇓, lane 2). Periaortic application of 0.25 mol/L CaCl2 was associated with further increase (more than fivefold on densitometry) of the 72- and 65-kD lytic zones attributable to gelatinase A, MMP-2 (Fig 6⇓, lanes 1, 3, 5, and 6). The highest-molecular-weight lytic zones (attributable to gelatinase B, MMP-9) showed the greatest increase in intensity (threefold to fourfold on densitometry) where thioglycollate had been applied (Fig 6⇓, lanes 4 through 6). However, the profile of gelatinase activity in aneurysmal aortas (lane 6) was not very different from normal-diameter arteries injured with CaCl2. The heavy macrophage infiltrate of the aneurysmal aorta did not appear to be associated with a further increase in gelatinase B (MMP-9) activity in homogenates compared with homogenates from other aortas treated with thioglycollate. To summarize these observations, after all aortic manipulations, there was increased evidence of gelatinase A (MMP-2) activity; gelatinase B (MMP-9) activity showed the greatest increase after periaortic application of thioglycollate. The results shown in Fig 6⇓ were qualitatively similar to those in two additional experiments using homogenates from different aortas, although the intensity of bands in lane 1 (0.25 mol/L CaCl2 only) was sometimes less.
The detailed mechanisms of the development of arterial aneurysms in humans remain obscure.14 15 We have developed an experimental model of abdominal aortic aneurysm that provides new insight into potential pathogenetic mechanisms. Reduction in the concentration of insoluble collagen and the influx of lipid-laden macrophages into the adventitia were identified as two important factors associated with rapid onset of aneurysmal dilatation. In addition, we have validated the use of ultrasonography for the serial, noninvasive monitoring of aortic dilatation.
Medial injury, with damage to elastin and consequent alterations in the elastic properties of the aorta in response to the pulse pressure, is considered critical to aneurysmal dilatation.4 15 Whether the medial injury results from atherosclerosis, hypertension, smoking, genetic predisposition, or other causes remains a matter of debate.7 14 15 16 The ability of intraluminal elastase to induce experimental aneurysms in rats emphasizes the hypothesis that medial injury with elastolysis results in aortic dilatation.4 This hypothesis is further supported by the much reduced concentrations of elastin found in aneurysmal aorta.17 In the experimental model we describe, periaortic application of 0.25 mol/L CaCl2 causes the medial injury on which aneurysmal dilatation depends.
The association of aortic aneurysm with both Marfan's and Ehlers-Danlos type IV syndromes suggests that alteration in composition or concentration of other matrix proteins, fibrillin and collagen, also might contribute to the development of abdominal aortic aneurysm. The turnover of these matrix proteins is regulated principally by the MMP family of enzymes.18 In particular, the gelatinases A (MMP-2) and B (MMP-9) have the potential to degrade elastin, type IV collagen, and many other matrix components but not fibrillar collagen.19 20 Both of these enzymes have been identified in the aortic wall, MMP-9 being immunolocalized to macrophages in the media and adventitia, and uncontrolled activity of these gelatinases could contribute to continuing aortic dilatation.3 21 22 The expression of interstitial collagenase (MMP-1), together with degradation of collagen, fibrillin, or other matrix proteins of the aortic wall, has received less attention.
Differences in the structure and function of the carotid artery and the aorta might underlie the fact that the conditions used to create carotid artery aneurysms in rabbits, periaortic application of 0.5 mol/L CaCl2,1 did not cause aneurysmal dilatation of the aorta. Although periaortic application of 0.5 mol/L CaCl2 caused histological changes similar to those described in the carotid artery, there was no increase in aortic ID. The periaortic application of CaCl2 caused both endothelial and medial injury, the endothelial injury manifested by abolition of endothelium-dependent relaxation and increased permeability to plasma proteins. In response to this injury, there was an ≈50% reduction in the concentration of insoluble collagen (as assessed by hydroxyproline) and evidence for increased amounts and/or activity of several matrix metalloproteinases, particularly gelatinase A (MMP-2). This reduction in the concentration of collagen, with concomitant increases in gelatinase activity, was not sufficient to cause rapid aneurysmal dilatation, even when rabbits were rendered hypercholesterolemic. Similarly, aneurysms were only an occasional and late sequela (>12 months) of aortic balloon angioplasty in hypercholesterolemic rabbits.8 Rapid aneurysmal dilatation occurred in response to periaortic application of CaCl2 and thioglycollate in cholesterol-fed animals. This dilatation was associated with a >60% reduction in collagen concentration. Aneurysmal dilatation may not occur until the concentration of collagen and other matrix proteins has been reduced by a critical amount.
Hypercholesterolemia appears to have an important influence on monocyte chemotaxis and activation. The previous work of Weidinger et al23 demonstrated that 4 weeks after balloon angioplasty of the iliac artery in normocholesterolemic rabbits, the presence of macrophages in the media or adventitia was a rare finding, whereas in hypercholesterolemic rabbits there was a diffuse monocyte/macrophage infiltration of the entire vessel wall. Similarly, in our experiments using CaCl2 to effect endothelial and medial injury, periaortic application of thioglycollate did not cause macrophage infiltration of the aortic wall, except in hypercholesterolemic rabbits. In such rabbits, the infiltrate of numerous foamy macrophages into the adventitia was associated with aneurysmal dilatation. These foamy macrophages have the potential to secrete both cytokines and enzymes other than gelatinase B that could stimulate the degradation of the aortic connective tissue and hence facilitate aneurysmal dilatation.
The presence of inflammatory cells, including macrophages, has been associated with aneurysm enlargement in other experimental models and in humans.3 5 6 In biopsies from abdominal aortic aneurysms in humans, gelatinase B (MMP-9) was concentrated in macrophages in the media and at the media-adventitial border.22 Therefore, we had anticipated that zymography would demonstrate a sharp increase in gelatinase B activity in homogenates prepared from the aneurysmal rabbit aorta. However, any manipulation of the aortic adventitia is injurious: treatment with 0.5 mol/L NaCl resulted in the prompt abolition of endothelium-dependent relaxation, which had not fully recovered after 3 weeks, and 0.25 mol/L CaCl2 caused overt injury to both endothelium and media. The increased activity of gelatinases in response to these injuries and subsequent vascular remodeling provides a background against which further increases in concentration/activity of specific enzymes may be difficult to detect by zymography. The question of which macrophage functions or products might have critical effects on the process of aneurysmal dilatation cannot be answered by immunohistochemistry but could be addressed, in the future, with the use of specific inhibitors or antibodies.
We propose that after medial injury resulting from the periaortic application of CaCl2, the migration of activated macrophages into the adventitia is necessary for rapid aneurysmal dilatation. This hypothesis is concordant with other studies in experimental animals and humans in which adventitial inflammation has been associated with aneurysm growth.3 5 6 Therefore, in humans, smoking and hypercholesterolemia may stimulate aneurysm growth from both intimal and adventitial processes by both exacerbating atherosclerosis to cause medial injury and stimulating the recruitment of activated macrophages into the adventitia.
We thank the British Heart Foundation and AFG Research Trust for support. We also thank Sam Andrews, Robert Hicks, Dinah Parums, Lesley Walton, and Roger Greenhalgh for their help, advice, and encouragement.
Gertz DS, Kurgan A, Eisenberg D. Aneurysm of the rabbit common carotid artery induced by periarterial application of. calcium chloride in vivo. J Clin Invest. 1988;81:649-656.
Freestone T, Turner RJ, Coady A, Higman DJ, Greenhalgh RM, Powell JT. Inflammation and matrix metalloproteinases in the enlarging abdominal aortic aneurysm. Arterioscler Thromb Vasc Biol. 1995;15:1145-1151.
Anidjar S, Salzmann JI, Gentric P, Lagneau P, Camilleri P, Michel JB. Elastase-induced experimental aneurysms in rats. Circulation. 1990;82:973-981.
Reed D, Reed C, Stemmermann G, Hayashi T. Are aortic aneurysms caused by atherosclerosis? Circulation. 1992;85:205-211.
Weidinger FF, McLenachan JM, Cybulsky MI, Gordon JB, Rennke HG, Hollenberg NK, Fallan JT, Ganz P, Cooke JP. Persistent dysfunction of regenerated endothelium after balloon angioplasty of rabbit iliac artery. Circulation. 1990;81:1667-1679.
Kuivaniemi H, Tromp G, Prockop DJ. Genetic causes of aortic aneurysms. J Clin Invest. 1991;88:1441-1444.
Dollery CM, McEwan JR, Henney AM. Matrix metalloproteinases and cardiovascular disease. Circ Res. 1995;77:863-868.
Murphy G, Cockett MI, Ward RV, Doherty AJP. Matrix-metalloproteinase degradation of elastin, type IV collagen and proteoglycan. Biochem J. 1991;277:277-279.
Thompson RW, Holmes DR, Mertens RA, Liao S, Botney MD, Mecham RP, Weigus HG, Parks WC. Production and localization of 92-kilodalton gelatinase in abdominal aortic aneurysms. J Clin Invest. 1995;96:318-326.
Weidinger FF, McLenachan JM, Cybulsky MI, Fallon JT, Hollenberg NK, Cooke JP, Ganz P. Hypercholesterolemia enhances macrophage recruitment and dysfunction of regenerated endothelium after balloon injury of the rabbit iliac artery. Circulation. 1991;84:755-767.