Author + information
- Received July 17, 2009
- Revision received September 30, 2009
- Accepted October 6, 2009
- Published online January 1, 2010.
- Mario Gössl, MD⁎,
- Daniele Versari, MD⁎,
- Heike A. Hildebrandt⁎,
- Thomas Bajanowski, MD§,
- Giuseppe Sangiorgi, MD¶,
- Raimund Erbel, MD∥,
- Erik L. Ritman, MD, PhD†,
- Lilach O. Lerman, MD, PhD‡ and
- Amir Lerman, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Amir Lerman, Division of Cardiovascular Diseases, Mayo Clinic Rochester, 200 First Street SW, Rochester, Minnesota 55905
Objectives Our aim was to investigate the role of coronary vasa vasorum (VV) neovascularization in the progression and complications of human coronary atherosclerotic plaques.
Background Accumulating evidence supports an important role of VV neovascularization in atherogenesis and lesion location determination in coronary artery disease. VV neovascularization can lead to intraplaque hemorrhage, which has been identified as a promoter of plaque progression and complications like plaque rupture. We hypothesized that distinctive patterns of VV neovascularization and associated plaque complications can be found in different stages of human coronary atherosclerosis.
Methods Hearts from 15 patients (age 52 ± 5 years, mean ± SEM) were obtained at autopsy, perfused with Microfil (Flow Tech, Inc., Carver, Massachusetts), and subsequently scanned with micro-computed tomography (CT). The 2-cm segments (n = 50) were histologically classified as either normal (n = 12), nonstenotic plaque (<50% stenosis, n = 18), calcified (n = 10) or noncalcified (n = 10) stenotic plaque. Micro-CT images were analyzed for VV density (number/mm2), VV vascular area fraction (mm2/mm2), and VV endothelial surface fraction (mm2/mm3). Histological sections were stained for Mallory's (iron), von Kossa (calcium), and glycophorin-A (erythrocyte fragments) as well as endothelial nitric oxide synthase, vascular endothelial growth factor, and tumor necrosis factor-alpha.
Results VV density was higher in segments with nonstenotic and noncalcified stenotic plaques as compared with normal segments (3.36 ± 0.45, 3.72 ± 1.03 vs. 1.16 ± 0.21, p < 0.01). In calcified stenotic plaques, VV spatial density was lowest (0.95 ± 0.21, p < 0.05 vs. nonstenotic and noncalcified stenotic plaque). The amount of iron and glycophorin A was significantly higher in nonstenotic and stenotic plaques as compared with normal segments, and correlated with VV density (Kendall-Tau correlation coefficient 0.65 and 0.58, respectively, p < 0.01). Moreover, relatively high amounts of iron and glycophorin A were found in calcified plaques. Further immunohistochemical characterization of VV revealed positive staining for endothelial nitric oxide synthase and tumor necrosis factor-alpha but not vascular endothelial growth factor.
Conclusions Our results support a possible role of VV neovascularization, VV rupture, and intraplaque hemorrhage in the progression and complications of human coronary atherosclerosis.
Atherosclerosis is a progressive, inflammatory disease of the vascular wall that begins at an early age as fatty streaks and progresses into raised lesions and mature atheromas in adulthood, often containing calcium hydroxyapatite deposits (1). A high correlation between the histological extent of the atherosclerotic plaque and its calcium deposition has been shown in the coronary vessels for all ages and both sexes making the calcification a surrogate measure of coronary atherosclerosis (2). Traditionally, the outer wall, which includes the adventitial layer and the vasa vasorum (VV), has been considered to play a passive role in atherogenesis. In contrast to this view, previous reports suggest that the adventitial layer may play a significant role in maintaining vessel integrity, and may contribute to the initiation and progression of the atherosclerotic as well as the remodeling process (3). Indeed, experimental studies demonstrated that manipulation of the adventitia, and more specifically of the VV, could lead to changes in the intimal layer, resembling the atherosclerotic process (4). Using micro-computed tomography (CT) technology in large animal models, we have demonstrated that early atherosclerosis is associated with neovascularization of the VV (5). Among other mechanisms, VV may contribute to progression of atherosclerosis by virtue of mediating, localizing, and promoting the influx of lipids and infiltrates of inflammatory cells, which are characteristic of unstable plaques (6,7). Moreover, fragility of newly formed VV and tendency to leak or break may mediate intraplaque hemorrhage, which is a major event occurring frequently during the development of the atherosclerotic lesion (8) that can progress acutely to plaque rupture, eventually resulting in a thromboembolic event or acute occlusion of the vessel (9,10).
Thus, it may be speculated that VV neovascularization has a role in the initiation and progression stages of atherosclerosis, by promoting cell proliferation, migration, and matrix production in the plaque. The current study was designed to test the hypothesis that segmental VV neovascularization and its complications of VV rupture and intraplaque hemorrhage correlate with the development and progression of human coronary atherosclerotic plaques.
The study protocol was reviewed and approved by the Mayo Foundation Institutional Review Board. Fifty coronary arterial segments (from 20 coronary arteries: 10 left anterior descending coronary arteries, 5 right coronary arteries, 5 left circumflex arteries) were obtained at autopsy from 15 patients. Patients' clinical data were obtained by reviewing the clinical files and documented by codes to avoid any identifying details. Artery segments were histologically classified as normal when no atherosclerotic plaque was present (n = 12). Segments with mild, nonlumen-compromising plaque (<50% lumen diameter stenosis) were classified as nonstenotic plaque (n = 18). If lumen-compromising atherosclerotic plaque was present on histology (>50% lumen diameter stenosis), the segments were either classified as noncalcified stenotic (n = 10, <50% plaque area calcification) or calcified stenotic plaque (n = 10, >50% lumen diameter stenosis and >50% plaque area calcification). Calcification was defined as plaque calcification.
Specimen preparation and imaging
During autopsy the hearts were harvested with specific care for the tissue surrounding the arteries in order to avoid damage to the adventitia. The coronary arteries were injected with Microfil (Flow Tech, Inc., Carver, Massachusetts) and prepared for micro-CT scanning as described in detail previously (11).
Data analysis was performed using Analyze software (Biomedical Imaging Resource, Rochester, Minnesota). In each segment, 35 to 40 cross sections at 400-μm intervals were chosen for region of interest analysis and averaged (12). In each micro-CT cross section, the vessel wall area (defined by the outer border of the adventitia) was determined as described in detail before (5,11). VV parameters were determined as described in detail before (13). A connectivity tool, included in the Analyze software, was applied on the 3-dimensional images of the arterial segments. As shown in Figure 1, this tool enables the identification of the vascular “tree” of each VV and tracking of its origin. That way, internal VV (originating directly from the main lumen) can be differentiated from external VV (originating from a major coronary branch) (11). The subsequent 2-dimensional image analysis ensures including only those VV that are actually perfused (i.e., connected to the systemic circulation) and excluding vessels that are in close proximity to the adventitia but are actually not VV.
After embedding in paraffin, cross sections of the arteries were mounted on slides and stained with hematoxylin and eosin. Parallel histology cross sections were used to confirm the accuracy of adventitial outer border tracing and VV identification in the micro-CT images as described in detail previously (11).
von Kossa Calcium Staining
Sections were deparaffinized, hydrated to distilled water, and placed in 6% sliver nitrate for 30 min in an incubator. Afterward, the sections were rinsed in distilled water and incubated in 5% sodium thiosulfate for 5 min. The sections were washed in running tap water for 5 min and counterstained in 0.1% nuclear fast red for 3 min. After rinsing in distilled water, the sections were mounted and visualized under a microscope (Fig. 2).
Glycophorin A Staining
Erythrocytes fragments within calcified and noncalcified plaques were identified using antibodies against glycophorin A (CD235a, 1:100, DakoCytomation, Carpinteria, California) on deparaffinized histology sections (Fig. 3). Normal human bone marrow sections were used for positive control studies.
Mallory's Iron Staining
After deparaffinizing and hydrating in water, the sections were stained in potassium ferrocyanide/hydrochloric acid solution (50 ml 5% potassium ferrocyanide, 50 ml 5% hydrochloric acid) for 10 min. The sections were rinsed in distilled water, stained in nuclear fast red for 5 min, and washed in running water for 2 min. Finally the sections were mounted and visualized under a microscope.
Vascular Endothelial Growth Factor (VEGF), Tumor Necrosis Factor (TNF)-Alpha, and Endothelial Nitric Oxide Synthase (eNOS) Staining
The slides were cut at 4 μm. Antigen retrieval was performed using a Black and Decker rice steamer (Towson, Maryland) and Dako Antigen Retrieval Solution. The slides were then blocked with 5% hydrogen peroxide for 15 min. Factor VIII antibody (1:50) was left on for 30 min. The Dako LSAB2 kit was used for the secondary and tertiary steps. Dako DAB chromogen was used. After this, the slides were placed in the appropriate antibody (eNOS pre-dilute, VEGF 1:250, TNF-alpha 1:50) for 60 min. The Dako LSAB2-AP kit was used for the second half of staining followed by the permanent red chromogen.
Glycophorin A and Iron Scoring
The percentage of vessel wall/plaque area showing glycophorin A and iron deposits was graded semiquantitatively by 2 independent observers using a scale from 0 to 4, with higher scores indicating higher percentages (1 = <25%, 2 = >25 but <50%, 3 = >50 but <75%, 4 = >75%).
Continuous data are expressed as mean ± standard error of the mean. One-way analysis of variance, followed by a Tukey-Kramer post hoc test with correction for multiple comparisons, was used to identify the statistical differences among groups. Individual group comparisons were performed by an unpaired Student t test. Statistical significance was accepted for a value of p < 0.05.
Patients and arterial segments characteristics
Clinical data are summarized in Table 1. The cause of death was trauma (n = 8), sepsis (n = 6), and suicide (n = 1).
VV spatial density was significantly higher in coronary segments with nonstenotic and noncalcified stenotic plaques as compared with normal and calcified stenotic plaque segments (Table 2). Calcified plaque segments actually had a spatial VV density similar to normal segments. These data conceivably also translated into significantly higher VV vascular area fractions and VV endothelial surface fractions in nonstenotic and noncalcified stenotic plaques compared with normal and calcified stenotic plaque segments. Adding up the endothelial surface area of all VV within the vessel wall showed that as much as 45% and 67% of the calculated main luminal endothelial surface area is available through the VV in the vessel wall of nonstenotic and noncalcified stenotic plaques (significantly more than in normal segments [17%]). Although VV density and fractions were not different in calcified stenotic plaque segments compared with normal segments, with 32% there was still significantly more endothelial exchange surface available through the VV in the calcified than in normal vessel wall (Table 2).
There was a significantly higher ratio of external/internal VV in noncalcified stenotic plaque segments compared with all other segments. The ratio already tended to be higher in nonstenotic plaques compared with normal segments (Fig. 4).
Using van Kossa (calcium deposits), Mallory's (iron, hemosiderin), and glycophorin-A staining (erythrocyte fragments), we were able to identify intraplaque hemorrhage likely through VV rupture in noncalcified and calcified plaques (Figs. 2 and 3). Interestingly, plaque calcification was associated with iron and glycophorin A staining suggesting that recurrent intraplaque hemorrhage is associated with plaque calcification (Fig. 2). VV density correlated negatively with the area of calcification (r = 0.42, p < 0.001) (Fig. 4).
Further characterization of VV revealed staining for the eNOS, the inflammatory marker TNF-alpha, but not VEGF (Fig. 5). There was no significant difference in the stains described in the preceding text between the different plaque types, except for a higher tendency for positive TNF- alpha staining of VV within diseased coronary artery segments.
Glycophorin A and iron scores
As shown in Figure 3 and Table 2, the nonstenotic and noncalcified stenotic plaque segments had significantly higher scores of glycophorin A and iron staining than normal and calcified stenotic plaque segments. The Kendall-Tau-beta rank correlation coefficient for VV density and iron score were 0.65 (p < 0.01) and 0.58 for VV density and glycophorin A score (p < 0.01), respectively.
The current study demonstrates a significant difference in the pattern of VV distribution in human coronary arteries in different stages of atherosclerosis. VV spatial density was higher in artery segments with nonstenotic plaques and further increased in segments with noncalcified stenotic plaque. Overall, VV neovascularization in nonstenotic and stenotic plaques was strongly correlated with the histological markers of intraplaque hemorrhage. Nevertheless, once the plaques showed significant calcification, VV spatial density significantly decreased to levels not different from normal coronary segments. Indeed, there was a negative correlation between VV density and the area of calcification within calcified coronary artery sections in micro-CT. However, compared with normal segments, calcified plaque segments had higher glycophorin A and iron scores indicating past intraplaque hemorrhages. Hence, the current study further supports the association between adventitial VV neovascularization and the progression and complication phases of the atherosclerosis process.
The current study is in accord with previous autopsy studies and supports the notion of a role of intraplaque hemorrhage (iron and glycophorin A deposits), by virtue of VV rupture or an injury to the endothelial lining of the main coronary lumen, in the mechanism of plaque calcification. The mechanism of this process may be speculated upon: the immature VV vasculature tends to leak and together with rupture of fragile VV neovasculature may lead to intraplaque hemorrhage. This may support the development of calcified plaque segments and may be a natural attempt to stabilize the plaque after subclinical intraplaque hemorrhage that did not lead to a plaque rupture with major coronary occlusion and sudden cardiac death.
The atherosclerotic process has a heterogeneous distribution on the surface of the vascular tree, as observed in autopsies as well as in coronary angiography and intravascular ultrasound studies (14). The mechanism of this nonuniform expression can be partially explained by different shear stress forces in regions like bifurcations and curves (15). However, hemodynamic mechanisms may not completely explain the heterogeneity of plaque distribution within arterial segments and even at the same cross section. Hence, it has become important to identify cofactors for plaque initiation and progression as well as complications in order to assess the plaque burden accurately and modify the natural history of the disease. Utilizing micro-CT, we have previously demonstrated, in the human and porcine circulation, the heterogeneity of adventitial VV density among different vascular beds in relation to their known susceptibility to develop atherosclerosis (16,17). In addition, we found that VV spatial density is increased in hypercholesterolemic porcine coronary arteries compared with normal pigs, suggesting a significant role for the adventitial VV in the early atherosclerotic remodeling process before vasofunctional alterations and plaque formation (18).
Vessels with mild atherosclerosis exhibit proliferation of VV in the adventitia that results in increased VV count and density. Based on our 3-dimensional analysis, VV neovascularization is pronounced at the level of external VV; the ratio of external VV to internal VV never becomes <1 during the atherosclerotic process although it has been demonstrated that growth of internal VV is stimulated by neointima formation, likely secondary to higher oxygen demand of the tissue (19). This remodeling of the adventitia is further increased in segments with evident plaque. Although VV neovascularization may serve as a compensatory mechanism to counteract vessel wall ischemia, the extensive VV network may function as a conduit for further entry of macrophages and inflammatory factors that may potentially promote the progression of inflammation and plaque formation. Indeed, inhibition of angiogenesis has been shown to reduce macrophages in the plaque and around the VV (20). Further characterization of VV in our studies using immunohistochemistry showed expression of eNOS in all plaque types and expression of TNF-alpha especially in diseased segments. We did not observe expression of VEGF, which may have been affected by the age and processing of the specimen.
The neovascularization process, which accompanies the evolution of the atherosclerotic process, leads to the formation of many new vessels and therefore to an increased probability of intraplaque hemorrhage. This process is also enhanced by the tortuosity and frailty of the newly-formed microvessels. Intraplaque hemorrhage is associated with increase in the size of the necrotic core and lesion instability in coronary plaques (6). However, before the final plaque rupture, the deposition of blood products in the plaque interstitium (21) may be responsible for further stimulation of the inflammatory process and may be followed by organization and fibrosclerosis, leading eventually to calcium deposition.
Coronary artery calcification is pathognomonic of atherosclerosis and its extent is related to the plaque burden and not to the degree of obstruction (22). Identification and quantification of arterial calcification is a reliable method of predicting the risk of cardiovascular events, especially when the calcium score is adjusted to age and sex (23). Arterial calcification is known to start in the early stages of atherosclerosis and to represent an active process of mineralization and hydroxyapatite crystals deposition (24). Moreover, a recent study demonstrated that microcalcification is associated with plaque vulnerability and positive remodeling (25). However, the mechanism and the source of the calcification process, whether it is part of the damage (e.g., inflammation) or part of the repair process involving osteogenic endothelial progenitor cells (26), is still under investigation.
This autopsy study represents a selected patient cohort; an inherent selection bias cannot be excluded. Due to the sample preparation process, no molecular mechanistic studies can be performed on the same arterial segment to explore the molecular mechanisms of VV neovascularization and plaque formation. Furthermore, as a correlative study, the results cannot determine causation; in particular, even if it is likely that calcium deposition followed repeated plaque VV rupture/hemorrhage, it is possible that a primitive fibrosclerotic process, induced by the inflammatory process, may induce the deposition of calcium and make the plaque less metabolically active and less in need of nourishment, causing eventually a decrease in VV density. However, it is unlikely that the increased amount of vessels found in the mature, noncalcified plaque can spontaneously regress. Alternatively, they could be compressed by surrounding fibrosclerotic tissue, thus becoming invisible for micro-CT, which detects only patent microvessels connected to the circulation. Also, the observed evidence for VV rupture in the form of iron and glycophorin A deposits may have occurred after the process of plaque calcification; the current study design did not allow looking at different time points of plaque progression. In the current study we did not determine iron deposition using micro-CT technology as shown previously in rodent studies (27).
The current study supports a role for coronary VV not only in the initiation of atherosclerosis but throughout the atherosclerotic process, including the progression phase (i.e., plaque growth and remodeling) and complication phase (i.e., intraplaque hemorrhage and calcification). The strong association of iron and glycophorin A deposits in calcified plaques may indicate that recurrent VV rupture and intraplaque hemorrhage promotes plaque calcification. The segmental heterogeneity of VV and early evidence of intraplaque hemorrhage further support the role of the VV in the focal complication of coronary atherosclerosis. In vivo imaging techniques assessing VV neovascularization in atherosclerosis may help to further define plaque vulnerability and guide medical therapy in the future (28).
This work was supported by the National Institutes of Health (R01 HL63911, K-24 HL69840-02, RO1 HL65432, RO1 EB000305, DK73608, HL085307, HL77131, and DK77013) and the Mayo Clinic College of Medicine. Dr. Amir Lerman is an Established Investigator of the American Heart Association.
- Abbreviations and Acronyms
- computed tomography
- endothelial nitric oxide synthase
- tumor necrosis factor
- vascular endothelial growth factor
- vasa vasorum
- Received July 17, 2009.
- Revision received September 30, 2009.
- Accepted October 6, 2009.
- American College of Cardiology Foundation
- Agatston A.S.,
- Janowitz W.R.,
- Hildner F.J.,
- Zusmer N.R.,
- Viamonte M. Jr..,
- Detrano R.
- Herrmann J.,
- Lerman L.O.,
- Rodriguez-Porcel M.,
- et al.
- Virmani R.,
- Kolodgie F.D.,
- Burke A.P.,
- et al.
- Virmani R.,
- Burke A.P.,
- Farb A.,
- Kolodgie F.D.
- Moreno P.R.,
- Purushothaman K.R.,
- Fuster V.,
- et al.
- Stone P.H.,
- Coskun A.U.,
- Kinlay S.,
- et al.
- Herrmann J.,
- Best P.J.,
- Ritman E.L.,
- Holmes D.R.,
- Lerman L.O.,
- Lerman A.
- Moulton K.S.,
- Vakili K.,
- Zurakowski D.,
- et al.
- Virmani R.,
- Kolodgie F.D.,
- Burke A.P.,
- Farb A.,
- Schwartz S.M.
- Eggen D.A.,
- Strong J.P.,
- McGill H.C. Jr.
- Doherty T.M.,
- Asotra K.,
- Fitzpatrick L.A.,
- et al.
- Shimada Y.,
- Yoshiyama M.,
- Kobayashi Y.,
- et al.
- Gossl M.,
- Modder U.I.,
- Atkinson E.J.,
- Lerman A.,
- Khosla S.
- Coli S.,
- Magnoni M.,
- Sangiorgi G.,
- et al.