Author + information
- Received December 1, 2010
- Revision received February 2, 2011
- Accepted February 18, 2011
- Published online May 1, 2011.
- Nitesh Gadeela, MD⁎,
- Jack Rubinstein, MD⁎∥,
- Umesh Tamhane, MD⁎,
- Ruiping Huang, PhD⁎,
- Dorothy R. Pathak, PhD‡,
- Hazel-Ann Hosein, PhD§,
- Michael Rich, BSc§,
- Gaurav Dhar, MD⁎ and
- George S. Abela, MD, MSc⁎,†,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. George S. Abela, B208 Clinical Center, Michigan State University, 138 Service Road, East Lansing, Michigan 48824
Objectives The purpose of this study was to determine if cholesterol crystals can injure the endothelial surface by their jagged edges altering vasoreactivity and contributing to no-reflow after intervention.
Background After plaque rupture, cholesterol crystals are released into the circulation and flow downstream contacting the arterial wall.
Methods Both carotid arteries from 22 rabbits were placed in a dual perfusion chamber and challenged with norepinephrine followed by acetylcholine and nitroprusside. Arterial diameters were measured before and after exposure to cholesterol crystals or microspheres and compared with diameters of normal control arteries. Arteries were examined by light, confocal, atomic force and scanning electron microscopy.
Results Pre-exposure mean arterial diameter was 2.33 ± 0.27 mm. With baseline norepinephrine there was vasoconstriction of 0.82 ± 0.19 mm, 0.79 ± 0.18 mm, and 0.83 ± 0.16 mm in lumen diameter in the crystal, microsphere, and control groups, respectively. After cholesterol crystals or microspheres, vasoconstriction was significantly less for cholesterol crystals but not for microspheres (0.71 ± 0.28 mm and 0.81 ± 0.15 mm; p < 0.02 and p = 0.68). After acetylcholine in the same artery, there was significantly less dilation before versus after crystals (0.49 ± 0.24 mm vs. 0.38 ± 0.22 mm, p = 0.04) but not with microspheres or in the control group. There was no significant difference after nitroprusside in any group, suggesting endothelial injury. By scanning electron microscopy, cholesterol crystals were found embedded in the intima with endothelial cell tears whereas the microsphere treatment and control groups had minimal or no injury (93% vs. 31% vs. 14%; p < 0.01). By atomic force microscopy, surface roughness was significantly greater with cholesterol crystals compared with microspheres or in control arteries (p < 0.05).
Conclusions Cholesterol crystals damaged the endothelium and reduced vasodilator response, potentially aggravating myocardial ischemia after interventions.
Atherosclerotic plaque rupture is considered the seminal event in most acute cardiovascular syndromes (1–3). However, the subsequent release of plaque material laden with cholesterol crystals traveling downstream may produce intimal injury that impairs vascular function that worsens ischemia and contributes to no-reflow after intervention (4–13). Recently, we have demonstrated that plaque rupture is associated with cholesterol crystals extensively perforating the intima and their release into the arterial lumen (14,15). Once the plaque cap is torn, the contents, composed mostly of cholesterol crystals, are released into the arterial circulation and embolize distally, constituting a major component of the occlusive thromboatheroma. During interventions, this material can then shower and lodge in the distal vascular bed (4–9,16). The acute coronary occlusion responsible for myocardial infarction is the result of thrombosis (17) variably associated with local (18) and distal coronary vasoconstriction (11). Moreover, vasospasm has been shown to be more frequently present in patients with acute coronary syndromes, as well as those who have interventional procedures with no-reflow.
The vascular endothelium normally regulates vasodilation via nitric oxide (19,20). Although nitric oxide and acetylcholine (ACh) are both able to dilate coronary arteries, the latter depends on an intact endothelium to subsequently release nitric oxide and in cases where the endothelium is damaged will result in unopposed vasoconstriction. Thus, we hypothesize that flow of jagged crystals downstream from the site of plaque rupture especially during interventions can injure the intimal surface thus decreasing the endothelial-dependant vasoreactivity. This study was designed to investigate this process using normal rabbit carotid arteries exposed to cholesterol crystals in an ex vivo circuit to evaluate the endothelial response with pharmacological challenge.
Both carotid arteries from 22 New Zealand White rabbits were used to evaluate the vasoreactive response to pharmacological challenge of the arteries before and after exposure to jagged and sharp tip cholesterol crystals (n = 24) as compared with exposure to smooth edge microspheres (n = 11) and phosphate buffered saline (PBS) in normal control arteries (n = 5). In 4 additional arteries, technical problems precluded obtaining vasoreactivity data (2 with microspheres and 2 control arteries) but those were available for morphological analysis. Rabbits were euthanized by a dose of buthanesol solution (1 ml intravenously; Virbac AH, Inc., Fort Worth, Texas) in the ear vein. Carotid arteries were then rapidly dissected by a mid-neck incision, gently removed, and placed in PBS. The adherent tissues were carefully removed under a dissecting microscope and then mounted in a dual perfusion chamber (21). The study was approved by Michigan State University committee on animal use following National Institutes of Health guidelines.
Warm PBS (37°C) was circulated in each of the divided perfusion chambers and through the 2 arterial segments (Fig. 1). After a 20-min period of stabilization, the arteries were pre-constricted with norepinephrine (NE) (1 × 10−6 mol/l) and then challenged with ACh (1 × 10−5 mol/l) followed by nitroprusside (SN) (1 × 10−5 mol/l) for 20 min each to evaluate for both endothelial and smooth muscle reactivity (see timeline in Fig. 1).
Vasodilation of the arteries was measured from the diameter of each artery at 5 different locations. These measurements were captured every 10 s during the experiment by a digital camera (Pulnix TMC-7, Pulnix America, Inc., Sunnyvale, California) and digitized images analyzed by a customized software package that was written to measure lumen dimensions using border recognition of the arterial wall. At the end of baseline pharmacological challenge, either crystals or microspheres were injected from the side port into each of the arterial segments and circulated for 30 min while only PBS was circulated in the control arteries.
Crystals were made as previously reported (22). Briefly, the crystals were grown in a substrate composed of trimethyl silicate gel made by dissolving trimethyl silicate in water at pH = 5 and specific gravity of 1.03 with 30 ml of methanol and then set for 72 h. One gram of cholesterol powder (5-Cholesten-3-beta-Ol; 3-beta-hydroxy-5-cholestene [C27H46O]: molecular weight = 386.7; 95% to 98% pure) (Sigma, St. Louis, Missouri) was added to 200 ml methanol and then added to the gel to form crystals within 48 h. These were then passed through filter paper and cholesterol crystals allowed to dry in air. The crystals ranged in length from 200 to 300 μm. After being injected into the artery and allowed to circulate for a period of 30 min, they were washed out of the system with PBS and the artery was allowed to stabilize for 20 min before the post-exposure pharmacological challenge. These synthetic crystals are almost identical in size and shape to those seen in human plaques (14,15,23).
The polystyrene microspheres (Polysciences, Inc., Warrington, Pennsylvania) used ranged from 200 to 300 μm to match the size of the crystals. Microspheres were injected, circulated, and washed out with PBS before pharmacological challenge in a manner similar to the crystals.
Microscopy to detect intimal injury
Scanning Electron Microscopy
At the conclusion of the circuit run, the arteries were perfusion-fixed with 4% glutaraldehyde for 30 min. After 24 h, 5-mm long segments of artery were chosen randomly from the artery and cut longitudinally to expose the intimal surface and vacuum dehydrated over 12 h. As previously reported, this method avoids ethanol dehydration to prevent dissolving the cholesterol crystals (14,23). Samples were dried in a vacuum chamber (Speed Vac SC110, Savant Instruments, Inc., Farmingdale, New York) evacuated by a pump (VP110, Franklin Electric, Bluffton, Indiana). Arterial segments were then mounted on stubs and gold coated in a sputter coater (EMSCOPE SC500, Quorum Technologies, Sussex, United Kingdom). The arterial lumen was then examined for intimal damage by cholesterol crystals and microspheres using a JEOL scanning electron microscope (SEM) (Model JSM-6300F, JEOL Ltd., Tokyo, Japan). Images were evaluated by a blinded examiner.
Intimal Injury Score
Injury was scored as 0 for no injury; +1 for localized endothelial cell injury without disruption; +2 for localized injury with disrupted endothelial cells; and +3 for loss of endothelial cells.
Arterial segments were stained for cholesterol crystals using fluorescent dye (cholesteryl BODIPY-C12, Invitrogen, Eugene, Oregon) (14). To demonstrate the relationship of cholesterol crystals to the intimal surface, the endothelium was counterstained with eosin. Samples were then transferred to a slide incubator chamber filled with PBS and visualized. Unstained plaque samples were also examined to detect native tissue fluorescence. Fluorescence images of plaque were acquired using a Zeiss Pascal laser scanning microscope (Carl Zeiss, Inc., Jena, Germany). Fluorescence was collected sequentially using the argon 488-nm laser line for excitation of BODIPY and collecting green fluorescence with a 505 to 530 nm bandpass filter and using the HeNe 543-nm laser line for excitation of eosin and collecting red fluorescence with a 560-nm long pass filter.
Atomic Force Microscopy
The effect of perfusion on intimal surface roughness was also evaluated using atomic force microscopy. One-centimeter long segments of perfused arteries were mounted on a steel stub and images were collected on a Nanoscope IV instrument (Veeco, Santa Barbara, California) equipped with a J scanner. Images were collected in tapping mode, using commercially available silicon cantilevers. All scans were collected at a rate of 0.5 Hz using a low pass filter, and scan lines were erased. Two roughness parameters were collected, namely, Ra the arithmetic average of the deviations from the center plane, and RMS, the standard deviation of Z values within a given area. The measurements were collected in a blinded fashion to the operator.
Light microscopy was performed in the standard fashion. After fixation in 4% glutaraldehyde and serial ethanol dehydration, tissues were embedded in paraffin and cut at 5-μm thick sections while stepping into the whole block every 2 mm. Arterial sections were stained with hematoxylin and eosin or trichrome and examined for presence of intimal injury under a light microscope (Leitz Laborlux12, Oberkochen, Germany).
Statistical analysis was performed using SAS (SAS Institute, Inc., Cary, North Carolina) and InStat 3 (Graph Pad, San Diego, California). Data are reported as mean ± SD. The change in arterial diameter with NE, ACh, and SN challenge was compared in the same artery before and after cholesterol crystals, microspheres, or PBS circulation using a paired t test. The roughness measurements between untreated control and treated arteries with crystals and microspheres were compared using a 1-way analysis of variance followed by Tukey multiple comparisons test for individual comparisons between the subgroups. A Fisher exact test was used for categorical variables. A 2-sided p value <0.05 was considered statistically significant.
A total of 40 carotid arteries were studied after exposure to PBS at 37°C followed by cholesterol crystals (n = 24), microspheres (n = 11), or only PBS in controls (n = 5). The mean diameter of all the arteries at baseline was 2.33 ± 0.27 mm. After PBS, the baseline mean vasoconstrictor response to NE was 0.82 ± 0.19 mm, 0.79 ± 0.18 mm, and 0.83 ± 0.16 mm in lumen diameter for crystal-, microsphere-, and PBS-treated arteries, respectively. After crystal or microsphere exposure, the diameter reduction was 0.71 ± 0.28 mm with crystals and 0.81 ± 0.15 mm with microspheres (p = 0.02 and p = 0.68, respectively). The baseline mean vasodilatory response to ACh after NE pre-constriction was 0.49 ± 0.24 mm as compared to 0.38 ± 0.22 mm after crystals (p < 0.04), and no significant change was observed after microspheres or in the PBS control groups (Fig. 2). Also, there was no significant difference in vasodilatory response after SN exposure at baseline and after treatment with crystal, microsphere, and PBS. All arteries returned to their baseline state after vasodilation.
SEM of the arterial intima was performed in 35 arterial segments (Table 1). This demonstrated various degrees of intimal injury evidenced by tears and cholesterol crystals embedded in the endothelium (Fig. 3). Circular imprints with a “roller like” effect were rarely detected in microsphere-treated arteries, otherwise only minimal disruption of the intima was noted. With the exception of 1 small site, there was no intimal injury in PBS-perfused control arteries. There were significantly more sites of intimal injury observed with cholesterol crystals than with microspheres and in control arteries (p < 0.01), but no difference was noted between microsphere and control arteries. Intimal injury was focal and detected mainly at sites with the presence of cholesterol crystals. These data complement the physiological findings with respect to the severity of injury. Also, no microspheres were detected by SEM in any of the examined samples whereas cholesterol crystals were detected in all treated samples (Table 2). However, microspheres were detected by confocal microscopy.
Confocal microscopy was performed in 11 arterial samples (Table 2). Cholesterol crystals could be detected in 10 of 11 (90%), whereas microspheres were detected in only 7 of 11 (64%) (p = 0.12). The surface characteristics of the smooth microsphere versus sharp crystal geometry may have influenced the surface adherence (Fig. 4).
Light microscopic examination did not reveal any obvious morphological injury in examined sections (n = 38) of either crystal- or microsphere-treated or in control arteries. No crystals or microspheres were detected in any of the histological sections by light microscopy.
Fifteen perfused arterial segments for cholesterol crystals (n = 5), microspheres (n = 5), and control subjects (n = 5) were evaluated at 5 distinct areas for each set of perfused arteries, yielding a sample set of 25 roughness measurements for each intervention. Atomic force microscopy demonstrated a significant increase in surface roughness over the intimal surface in the cholesterol crystal–treated compared with the microsphere-treated and PBS-perfused arteries (p < 0.01) (Figs. 5 and 6⇓⇓). Increased surface roughness was diffuse over the entire arterial surface exposed to cholesterol crystals whereas intimal tears by SEM were visible only in focal areas.
This study demonstrated that circulating cholesterol crystals could injure the endothelial lining of the arterial wall, diminishing normal vasoreactivity. It has already been demonstrated that debris released during rupture contain great amounts of cholesterol crystals (6–8,14). These are often intermixed with other debris (e.g., platelets, fibrin) occluding the distal arterial bed causing impaired flow and reduced myocardial perfusion. Thus, released cholesterol crystals do flow downstream from the site of plaque rupture. Moreover, cholesterol crystals have been isolated from both atherectomy specimens and aspirates obstructing coronary arteries during acute coronary syndromes, confirming their presence during the acute event (8,15). Furthermore, we had previously demonstrated that cholesterol crystallizes under certain physical conditions (i.e., increase cholesterol saturation and drop in temperature) that may trigger plaque rupture (24). However, regardless of mechanism, the release of cholesterol crystals into the circulation could be clinically relevant with respect to worsening myocardial ischemia due to endothelial injury. In our in vitro model, endothelial injury with cholesterol crystals was observed using various microscopic techniques, and this was associated with physiological dysfunction of decreased vasoreactivity to ACh but not to SN, indicating endothelial injury.
This study also distinguished the effects of cholesterol crystals from the less traumatic-shaped microspheres, emphasizing the importance of the geometry and sharpness of the crystals. Although by SEM, there was evidence of some mild endothelial injury with microspheres, that was not enough to cause a significant change in the physiological behavior of the endothelium. Also, fewer microspheres were detected by microscopy compared with crystals, suggesting that the microspheres rolled away with the flow in the perfusion chamber whereas crystals tended to stick to the wall because of their geometric shape. The advantage of using the rabbit carotid artery is that it is similar in diameter to medium-size human coronary arteries (∼2 mm).
The no-reflow phenomenon has been described as the absence of distal coronary blood flow after coronary interventions (12,13). There has been much controversy regarding the exact mechanism of no-reflow in the coronary arteries after interventional procedures. However, given our observations, it would be reasonable to infer that 2 mechanisms contribute to the no-reflow, one is vasospasm from endothelial injury and the other is obstruction by the particulate materials including cholesterol crystals (15,16). Moreover, recent studies have demonstrated that cholesterol crystals also trigger an inflammatory response not dissimilar to that described with uric acid crystals (25–27). This could contribute to muscle inflammation and injury (28). Also, it has been demonstrated that endothelium-dependant coronary vasomotor function, tested through ACh-induced vasoreactivity, independently predicts long-term cardiovascular risk after acute cardiovascular events (29). Other examples of end organ damage due to cholesterol crystals include showering from carotid plaques causing amaurosis fugax (30), renal damage after invasive vascular procedures (6,31), and livedo reticularis or gangrene in peripheral arteries (32).
No-reflow has been associated with a worse clinical prognosis (33). Although treatment with intracoronary adenosine and calcium-channel blockers helps to re-establish arterial flow, they do not seem to have much effect on improving the size of infarction (34). Our study suggests that using other agents that dissolve cholesterol crystals may have an important impact on no-reflow. It has already been demonstrated that alcohol, statins, aspirin, and high-density lipoproteins have an effect on dissolving cholesterol crystals (14,35–37). Future studies may need to consider this novel aspect in the treatment of no-reflow.
In addition to help explaining the no-reflow phenomenon, our study may partially explain anginal and ischemic episodes that have been shown to occur in the absence of obstructive coronary artery disease (10), as well as recurrent events after an acute coronary syndrome (11). Even during an acute cardiovascular event that does not completely occlude the circulation, cholesterol crystals may still cause endothelial damage that can induce ischemia. In these cases, the coronary vasculature may seem to be relatively intact when subsequently evaluated with angiography or intravascular ultrasound.
A limitation of this study was that it was performed in vitro where the crystals were recirculated, thus amplifying intimal injury as compared to what would be expected in vivo where 1 pass would occur during an acute event or after an interventional procedure. However, the process of crystal release in vivo may also be recurring for a certain period from a ruptured plaque after the initial event. Additional studies will need to evaluate these aspects.
In summary, we have demonstrated that cholesterol crystals in the arterial circulation as released during plaque rupture can cause endothelial cell injury resulting in decreased vasoreactivity. The preserved vasodilation in response to SN suggests that this effect was due to endothelial injury. In addition to vascular dysfunction after acute coronary syndrome and no-reflow after interventions, these findings may help elucidate the mechanisms underlying previously unclear clinical syndromes including vasospastic angina.
The authors thank Ewa Danielewicz, MSc, Abigail Vanderberg, MSc, Carol Flegler, BA, and Melinda Frame, PhD at the Center for Advanced Microscopy, Michigan State University.
This work was supported by funds from Edward Sparrow Hospital, Lansing, Michigan, and Michigan State University, East Lansing, Michigan. Dr. Gadeela has reported that he has no relationships to disclose. Dr. Rubinstein has received a grant from Merck & Co., Inc. Drs. Tamhane, Huang, Pathak, Hosein, Rich, and Dhar have reported that they have no relationships to disclose. Dr. Abela has received grants from Merck/Schering-Plough & Co., Inc., and Novartis Pharmaceutical Corporation, and has received speaker fees from Merck, GlaxoSmithKline, Takeda, and CardioNet.
- Abbreviations and Acronyms
- phosphate buffered saline
- scanning electron microscopy
- Received December 1, 2010.
- Revision received February 2, 2011.
- Accepted February 18, 2011.
- American College of Cardiology Foundation
- Davies M.J.,
- Thomas A.C.
- Schaar J.A.,
- Muller J.E.,
- Falk E.,
- et al.
- Falk E.,
- Thuesen L.
- Reffelmann T.,
- Kloner R.A.
- Kotani J.,
- Nanto S.,
- Mintz G.S.,
- et al.
- Breuckmann F.,
- Nassenstein K.,
- Bucher C.,
- et al.
- Rezkalla S.H.,
- Kloner R.A.
- Abela G.S.,
- Shamoun F.,
- Vedre A.,
- et al.
- Schwartz R.S.,
- Burke A.,
- Farb A.,
- et al.
- Prieto A.R.,
- Ma H.,
- Huang R.,
- et al.
- Pervaiz M.H.,
- Huang R.,
- Narisetty K.,
- Vedre A.,
- Berger K.,
- Abela G.S.
- Halcox J.P.,
- Schenke W.H.,
- Zalos G.,
- et al.
- Fukumoto Y.,
- Tsutsui H.,
- Tsuchihashi M.,
- et al.
- Mehta R.H.,
- Harjai K.J.,
- Cox D.,
- et al.
- Ross A.M.,
- Gibbons R.J.,
- Stone G.W.,
- et al.
- Vedre A.,
- Aziz K.,
- Huang R.,
- Abela G.S.
- Abela GS, Vedre A, Janoudi A, Huang R, Durga S, Tamhane U. Effect of statins on cholesterol crystallization and atherosclerotic plaque stabilization. Am J Cardiol 2011 Apr 18 [E-pub ahead of print]; doi: 10.1016/j.amjcard.2011.02.336.