Functional and Morphologic Imaging of Coronary Atherosclerosis in Living Mice Using High-Resolution Color Doppler Echocardiography and Ultrasound Biomicroscopy

Journal of the American College of Cardiology © 2005 by the American College of Cardiology Foundation Published by Elsevier Inc.

Vol. 46, No. 4, 2005 ISSN 0735-1097/05/$30.00 doi:10.1016/j.jacc.2005.04.053

Functional and Morphologic Imaging of Coronary Atherosclerosis in Living Mice Using High-Resolution Color Doppler Echocardiography and Ultrasound Biomicroscopy Johannes Wikström, MSC,* Julia Grönros, MSC,* Göran Bergström, MD, PHD,† Li-ming Gan, MD, PHD*† Göteborg, Sweden This study aimed to establish non-invasive methods of assessing coronary artery morphology in normal and atherosclerotic mice in vivo. BACKGROUND Coronary flow velocity reserve (CFVR) has been shown to correlate with coronary minimal lumen diameter (MLD) in patients with coronary artery stenosis. In mice, there are no existing non-invasive imaging techniques allowing quantitative measurement of the coronary artery morphology and function. METHODS Systemic hemodynamic effects of adenosine were studied in seven C57BL/6 mice. In 17 C57BL/6 mice, CFVR was measured in the mid left coronary artery (LCA) using either hypoxia- or adenosine-induced coronary hyperemia. Further, in another 10 atherosclerotic low-density lipoprotein receptor (LDLR)⫺/⫺ mice, the hypoxia-induced CFVR was performed and proximal LCA MLD was measured using ultrasound biomicroscopy (UBM). Histologic sections of the LCA were collected. RESULTS The adenosine dose of 160 ␮g/kg/min induced maximal coronary hyperemia without any systemic hemodynamic effects. Adenosine and hypoxia-induced CFVR values averaged at 2.0 ⫾ 0.1 and 1.9 ⫾ 0.3, respectively, in C57BL/6 mice (p ⫽ NS). In LDLR⫺/⫺ mice, CFVR and MLD ranged between 1.4 to 2.9 ␮m and 190 to 370 ␮m, respectively. Histology revealed proximal lumen-narrowing plaques in the LCA. Significant correlation was found between hypoxia-induced CFVR and the MLD (p ⬍ 0.005, R2 ⫽ 0.8707). CONCLUSIONS The CDE and UBM technique can be used to measure atherosclerosis-related lumen narrowing of the LCA in living mice. These non-invasive techniques may provide us with novel tools for following up disease status in mouse coronary arteries in a quantitative manner. (J Am Coll Cardiol 2005;46:720 –7) © 2005 by the American College of Cardiology Foundation OBJECTIVES

Genetically modified mice provide a powerful tool for understanding the molecular mechanisms and pathogenesis of human atherosclerosis (1). Numerous mouse strains are available today with phenotypes relevant to human cardiovascular diseases (1,2). The common use of these novel mouse strains has prompted the development of techniques for assessing the cardiovascular function and morphology in intact, living mice (3). Coronary atherosclerosis is the major cause of cardiac death in humans. Usually, coronary artery morphology is studied using invasive techniques such as angiography. Recently, several imaging techniques have been emerging as promising non-invasive alternatives, such as electron-beam computed tomography (4) and magnetic resonance imaging (5). However, in mice, which have extremely small coronary From the *Department of Physiology, Institute of Physiology and Pharmacology, and the †Department of Clinical Physiology, Cardiovascular Institute, The Sahlgrenska Academy, Göteborg University, Göteborg, Sweden. This work was supported by grants from the Swedish Medical Research Council, the Swedish Heart-Lung Foundation, the Ake Wiberg Foundation, the Memorial Foundation of Lars Hierta, the Magnus Bergvall Foundation, the Sahlgrenska University Hospital Research Foundation, the Lundberg Foundation, AstraZeneca, R&D, and performed with assistance of the SWEGENE Centre for Mouse Physiology. Manuscript received December 20, 2004; revised manuscript received April 14, 2005, accepted April 25, 2005.

arteries and high heart rates (HR), the coronary circulation constitutes a great challenge for these available imaging techniques. During the past years, our group has developed a non-invasive technique for imaging mouse coronary artery flow in vivo using color Doppler echocardiography (CDE) (6). Further, Zhou et al. (7) showed recently that proximal left coronary artery (LCA) morphology could be visualized in wild-type mice by using a novel ultrasound biomicroscopic (UBM) technique. Coronary flow velocity reserve (CFVR) can be assessed either invasively by Doppler guidewire technique or noninvasively by transthoracic CDE. The CFVR is calculated as the ratio between hyperemic and resting coronary blood flow velocities, and has been used to predict coronary stenosis and minimal lumen diameter (MLD) in atherosclerotic coronary arteries (8,9). Little is known about coronary atherosclerosis in mice, but it seems to develop first in the most proximal parts of the left and the right coronary artery (10). In the present, study we aimed to develop a CFVR protocol using either adenosine or hypoxia-induced coronary hyperemia in mice. Further, to investigate the relevance of the method, we applied the CFVR protocol in a well-known mouse model of atherosclerosis to test whether CFVR can be used to

Wikström et al. Imaging of Coronary Atherosclerosis in Mice

JACC Vol. 46, No. 4, 2005 August 16, 2005:720–7

Abbreviations and Acronyms BP ⫽ blood pressure CDE ⫽ color Doppler echocardiography CFR ⫽ coronary flow reserve CFVR ⫽ coronary flow velocity reserve HR ⫽ heart rate LCA ⫽ left coronary artery LDLR ⫽ low-density lipoprotein receptor MLD ⫽ minimal lumen diameter UBM ⫽ ultrasound biomicroscope

predict MLD in vivo. Histology was used to qualitatively evaluate the extension of mouse coronary atherosclerosis, whereas the in vivo coronary MLD was examined quantitatively using the novel UBM technique.

METHODS Experimental protocols and animals.

ESTABLISHMENT

Fifteenweek-old female C57BL/6 mice were used to: 1) determine the dose-response effect of adenosine infusion on blood pressure (BP) and HR (C57BL/6DoseResponse, n ⫽ 7), and 2) evaluate the possible difference between adenosine-induced (C57BL/6Adenosine, n ⫽ 11) and hypoxia-induced (C57BL/ 6Hypoxia, n ⫽ 6) CFVR as measured in vivo using CDE.

OF THE CFVR PROTOCOL IN WILD-TYPE MICE.

APPLICATION OF THE CFVR PROTOCOL IN ATHEROSCLE-

Thirty-eight-week-old male low-density lipoprotein receptor (LDLR)⫺/⫺ mice (n ⫽ 10) on a mixed background of 129SV and C57BL/6 mice were used to evaluate whether CFVR could be used to predict MLD. Coronary artery atherosclerosis was evaluated qualitatively using histology, and MLD was assessed in vivo using UBM. The serum lipid profile was also analyzed. The C57BL/6 and LDLR⫺/⫺ mice were purchased from Charles River Laboratories, Sulzfeld, Germany, and Taconic M&B A/S, Copenhagen, Denmark, respectively. All animals had more than 1 week of acclimatization before onset of the experiments. The mice were housed at a constant temperature (24°C) in a room with 12-h dark/light cycles and had free access to pellet diet and tap water. The C57BL/6 mice received a normal chow diet (Lactamin AB, ¨ ngelholm, Sweden). The LDLR⫺/⫺ mice were fed a A Western diet to accelerate atherosclerotic lesion progression. The diet contained 0.5% cholesterol and 20% fat (Lactamin AB). At 30 weeks, the mice were returned to standard chow for a further 8 weeks to avoid potential effects of exaggerated hypercholesterolemia (11). All experiments were performed in accordance with national guidelines and approved by the Animal Ethics Committee, Göteborg University. Anesthesia. Animals were fully anaesthetized with 0.7% to 1.5% isoflurane (Abbot Scandinavia AB, Solna, Sweden) during all experiments. A rectal thermometer connected to a thermo-regulating lamp and an electric-heated blanket ROTIC MICE.

721

was used to maintain normothermia in the experimental animals. Establishment of the CFVR protocol in wild-type mice. ADENOSINE INFUSION AND INVASIVE BP MEASUREMENTS. The C57BL/6DoseResponse mice were used to study BP and HR in response to adenosine infusion. A heatstretched catheter (PE50, Transonic Systems Inc, Ithaca, New York) was inserted into the right carotid artery for direct measurement of arterial BP. A catheter (0.4 (27GA) ⫻ 9.5 ⫻ 300, Sa-TE-Z Set, Temse, Belgium) connected to a syringe containing 0.5 mg/ml adenosine (ITEM Development AB, Stocksund, Sweden) was inserted into the right tail vein. The intravenous administration was performed by means of an infusion pump (syringe pump 22, Harvard Apparatus Inc., Holliston, Massachusetts) at a rate of 8 ␮l/min. After a short stabilization period, baseline BP was measured continuously for 2 min and then for 4 min per dose of adenosine. The doses infused were 80, 160, 320, and 640 ␮g/kg/min. The total volume of infusion was 192 ␮l. The mice were killed after the measurements. In the following CFVR studies in C57BL/6Adenosine mice, only one dose of adenosine was used, with a total infusion volume typically below 40 ␮l, which accounts for approximately one-fiftieth of the mouse blood volume. BASIC ECHOCARDIOGRAPHY MEASUREMENTS AND CALCU-

Before all ultrasound scanning, the hair of the mouse chest wall was carefully removed and warm ultrasound transmission gel was liberally applied to ensure optimal image quality. Echocardiography was performed using a high-frequency 15-MHz linear transducer (Entos CL15-7, Philips Medical Systems, Bothell, Washington) connected to an ultrasound system (ATL-HDI5000, Philips Medical Systems). Standard basic echocardiography measurements and calculations were performed as described previously (12).

LATIONS.

Doppler flow measurements were made in the mid-segment of the LCA. The mid LCA was visualized according to a protocol previously described with slight modifications (6). Briefly, the heart was imaged using a parasternal long-axis view with the probe lateralized and the ultrasound beam anteriorly tilted. In this image window, the entire LCA, from the aortic sinus to the distal branch site, could be visualized using CDE (Figs. 1A and 1B). The course of the LCA was typically parallel to the Doppler beam, which also facilitates Doppler measurements without any angle correction. A 6-MHz pulsed Doppler with a gate size of 0.5 to 1 mm was used. The flow velocity measurements were made at the same vessel site at baseline and during hypoxia or adenosine-induced hyperemia (Fig. 2). Mean diastolic flow velocity was obtained by outlining the diastolic phase of coronary flow signals in an off-line software program (MedArchive viewer 2.1, 1996, Secure Archive, Indianapolis, Indiana). Measurements were averaged from three consecutive cardiac cycles. The CFVR was

ASSESSMENT OF CFVR.

722

Wikström et al. Imaging of Coronary Atherosclerosis in Mice

JACC Vol. 46, No. 4, 2005 August 16, 2005:720–7

Figure 1. (A) En face inspection of left coronary artery (LCA) and the specific vascular site of flow measurement as indicated by the arrowhead. (B) A color Doppler echocardiography image corresponding to image (A), showing LCA and the flow velocity measurement site, as indicated by an arrowhead. Typical ultrasound biomicroscope images of the proximal LCA in (C) relatively non-diseased and (D) severely atherosclerotic low-density lipoprotein receptor (LDLR)⫺/⫺ mice. Small arrowheads show the measurement site of lumen diameter. Scale bar ⫽ 2 mm in panel A and 1 mm in panels C and D. AO ⫽ aorta; LA ⫽ left atrium; LV ⫽ left ventricle; PA ⫽ pulmonary artery.

thereafter calculated as the ratio between the mean diastolic flow velocity during hyperemia and baseline. One investigator (L.G.) performed all ultrasound scans of the LCA in all experimental animals to maintain good reproducibility (coefficient of variation intraobserver variability ⱕ5%). Baseline coronary flow velocity was measured in another 10 LDLR⫺/⫺ mice age 40 weeks by two different operators at two occasions to evaluate interobserver variability (coefficient of variation interobserver variability ⱕ10%). ADENOSINE AND CFVR. Doppler measurements were made at baseline and during adenosine-induced hyperemia. Infusion of 160 ␮g/kg/min of adenosine was facilitated via the

tail vein as described above. This dose was validated to induce coronary hyperemia without influencing systemic hemodynamics, as validated in C57BL/6DoseResponse mice. This dose seems to induce maximum coronary hyperemia, because an additional dose of 320 ␮g/kg/min in C57BL/ 6Adenosine mice did not further increase the hyperemic flow velocity. Maximal hyperemia was typically obtained within 2 to 3 min. Flow velocity measurements were made at baseline (normoxia) and during hypoxia-induced hyperemia. Moderate hypoxia was induced during anesthesia by changing the mixture of air and N2 into the gas mask. The gas flow was held constant, and the oxygen tension was continuously HYPOXIA AND CFVR.

Figure 2. Typical Doppler flow velocity signals during (A) rest and (B) hypoxia-induced hyperemia, as measured in the mid left coronary artery. The continuous white line indicates the diastolic phase of the coronary Doppler flow signals.

Wikström et al. Imaging of Coronary Atherosclerosis in Mice

JACC Vol. 46, No. 4, 2005 August 16, 2005:720–7

measured with an oximeter (Taylor Servomax Oxygen Meter, Servomex Group Ltd., Crowborough, United Kingdom). Baseline oxygen tension was 21% in air with no addition of N2 gas. Hypoxic gas mixture with an oxygen tension of 12.5%, corresponding to an altitude of 4,500 m (13), was obtained by quick reduction of air and addition of N2 gas. Coronary flow velocity measurements were made continuously during hypoxia for a maximum of 3 min. This provocation does not induce any echocardiographically evident myocardial ischemia or infarctions (data not shown), whereas prolonged exposure to lower oxygen tension has been shown to be harmful to the myocardium (10). Application of the hypoxia-induced CFVR protocol in atherosclerotic mice. BASIC ECHOCARDIOGRAPHY AND HYPOXIA-INDUCED CFVR. The basic echocardiography, calculations, and the hypoxia-induced CFVR protocol were performed as described above. The proximal LCA in the LDLR⫺/⫺ mice was visualized using a UBM with a transducer frequency of 40 MHz (Vevo 600, Visualsonics, Toronto, Canada). This set-up provides imaging with a theoretical resolution of 40 ␮m at a frame rate of 32 Hz. A similar imaging window was used as during CDE scanning, with careful probe adjustments to obtain maximal lumen diameter, thus avoiding off-axis-related underestimations of the MLD. A cine-loop sequence of at least 20 cardiac cycles was recorded and measured off-line (Imagepro plus 5.0, Media Cybernetics, Silver Spring, Maryland). Measurement of MLD was made in diastole, in a frame where it was best visualized. The proximal LCA was defined as a segment reaching from the level of the coronary ostium to approximately 500 ␮m into the proximal LCA (Figs. 1C and 1D). One operator, who was blinded to the animal identities and other cardiac parameters, analyzed all images. UBM.

The formaldehyde-fixed hearts were dehydrated and paraffin-embedded for microtome sectioning. Sections of 5 ␮m were collected every 50 ␮m from the level of the mitral valves and up to the aorta. The sections were stained with Picro-sirius red for light-microscope inspection (BX-60, Olympus Optical Co., Tokyo, Japan).

HISTOLOGY.

STATISTICS. The Student’s paired t test with adjustments for four comparisons over time using a Bonferroni correction (p values ⬍ 0.0125 [0.05 of 4]) are considered statistically significant) was used to detect changes in BP and HR

723

in C57BL/6DoseResponse mice. Student’s two-sample t test was performed between C57BL/6Adenosine and C57BL/ 6Hypoxia mice in: 1) echocardiography measurements, and 2) baseline and hyperemic coronary flow velocity and the resulting CFVR. Pearson’s test was used to study correlation (CFVR vs. MLD and HR vs. CFVR) (Prism 3.0, Graphpad Inc., San Diego, California). The coefficient of variation for interobserver variability was calculated using the following formula: standard deviation (x-y)/mean (x,y) ⫻ 100. A p value ⬍0.05 was considered statistically significant except when a Bonferroni correction was used. All data are expressed as mean ⫾ standard deviation.

RESULTS Establishment of the CFVR protocol in wild-type mice. HEMODYNAMIC EFFECTS OF INTRAVENOUS ADENOSINE. In the C57BL/6DoseResponse mice, only the maximal dose of adenosine (640 ␮g/kg/min) altered BP (p ⬍ 0.001), whereas HR remained unchanged during all doses (Table 1). BASIC ECHOCARDIOGRAPHY MEASUREMENTS.

The basic

echocardiography data are shown in Table 2. CFVR MEASUREMENTS IN C57BL/6 MICE. The LCA flow velocity at baseline and in the hyperemic condition was successfully measured in all C57BL/6 mice (Table 3). Both adenosine and hypoxia induced a profound increase in blood flow velocity within 3 min. Adenosine at 320 ␮g/kg/min did not further increase flow velocity in any animals compared with the lower dose of 160 ␮g/kg/min (data not shown). Application of the hypoxia-induced CFVR protocol in atherosclerotic mice. BASIC ECHOCARDIOGRAPHY MEASUREMENTS AND HYPOXIA-INDUCED CFVR. The basic echocardiography data (Table 2), as well as coronary flow velocity measurements and CFVR (Table 3), are tabulated.

The LDLR⫺/⫺ mice had a total cholesterol level of 10.3 ⫾ 0.5 mmol/l, triglycerides 1.4 ⫾ 0.1 mmol/l, low-density lipoprotein 5.7 ⫾ 0.5 mmol/l, and high-density lipoprotein 3.8 ⫾ 0.1 mmol/l.

PLASMA LIPID PROFILE.

Qualitative evaluation of the histologic sections from the LDLR⫺/⫺ mice showed lumen-narrowing atherosclerotic plaque in the proximal but not in the mid LCA segment (Fig. 3).

HISTOLOGY.

Table 1. Changes in Blood Pressure and Heart Rate During Intravenous Adenosine in C57BL/6DoseResponse Mice

State

Average Mean Arterial Pressure (mm Hg)*

⌬Mean Arterial Pressure Dose-Baseline (mm Hg)

Baseline Dose 80 Dose 160 Dose 320 Dose 640

74 ⫾ 5 74 ⫾ 5 68 ⫾ 11 65 ⫾ 8 56 ⫾ 5

— ⫺0.4 ⫾ 2.2 ⫺6.5 ⫾ 8.8 ⫺18.9 ⫾ 24.6 ⫺26.7 ⫾ 20.2

*Mean ⫾ standard deviation.

p

Average Heart Rate (beats/min)

⌬ Heart Rate Dose-Baseline (beats/min)

p

— NS NS NS ⬍0.0001

390 ⫾ 22 408 ⫾ 36 429 ⫾ 46 427 ⫾ 44 434 ⫾ 45

— 18 ⫾ 30 38 ⫾ 41 39 ⫾ 34 45 ⫾ 38

— NS NS NS NS

724

Wikström et al. Imaging of Coronary Atherosclerosis in Mice

JACC Vol. 46, No. 4, 2005 August 16, 2005:720–7

Table 2. Baseline Echocardiography Data

Left ventricular mass (mg) End diastolic volume (␮l) Ejection fraction (%) Fraction shortening (%) Stroke volume (␮l) CO (ml/min) Heart rate (beats/min)

should be taken when comparing absolute magnitudes of the CFVR with those in humans.

C57BL/6 Hypoxia

C57BL/6 Adenosine

LDLRⴚ/ⴚ

69 ⫾ 5 85 ⫾ 12 70 ⫾ 11 40 ⫾ 9 60 ⫾ 15 22 ⫾ 3 368 ⫾ 53

68 ⫾ 18 87 ⫾ 30 58 ⫾ 9* 31 ⫾ 6* 49 ⫾ 14 25 ⫾ 7 508 ⫾ 58*

81 ⫾ 7 58 ⫾ 9 64 ⫾ 7 35 ⫾ 6 37 ⫾ 4 15 ⫾ 2 390 ⫾ 15

*Difference (p ⱕ 0.05) between C57BL/6Hypoxia and C57BL/6Adenosine, mean ⫾ standard deviation are displayed. CO ⫽ cardiac output; LDLR ⫽ low-density lipoprotein receptor.

Good visualization of the proximal LCA was obtained in 9 of 10 LDLR⫺/⫺ mice. The MLD ranged from 190 to 370 ␮m and averaged 274 ⫾ 70 ␮m. UBM MINIMAL LUMEN DIAMETER MEASUREMENTS.

Hypoxiainduced CFVR was found to correlate with UBM measurements of the proximal MLD (p ⬍ 0.005, R2 ⫽ 0.8707) (Fig. 4). CORRELATION BETWEEN CFVR AND MLD.

No correlation was found between baseline HR and CFVR, neither in the whole study population nor in any subgroups.

CORRELATION BETWEEN CFVR AND HR.

DISCUSSION In summary, we have shown that CDE and UBM are highly feasible techniques for assessing in vivo mouse coronary artery function and morphology. Coronary hyperemia can be induced by either adenosine-infusion or moderate hypoxia. An adenosine dose of 160 ␮g/kg/min causes significant coronary hyperemia without any apparent effects on the systemic hemodynamics. Moderate acute hypoxia induced significant coronary flow increase, comparable with that of adenosine. Proximal left coronary artery atherosclerosis develops in the LDLR⫺/⫺ mice, which can be imaged in vivo using UBM. Hypoxiainduced CFVR correlates well with MLD in the mouse left coronary artery. Establishment of the CFVR protocol in wild-type mice. CFVR AND ANESTHESIA IN MICE. Several echocardiographic studies performed in trained, conscious mice showed improved cardiac function compared with measurements during anesthesia (14). Anesthesia is, however, a highly necessary approach in the current study because of its semi-invasive nature. Also, various anesthetic agents have been shown to have different influences on coronary vasodilator reserve in rats as measured by positron emission tomography (15). In the current study, isoflurane-assisted anesthesia is used because it is a well-tolerated, non-invasive approach that is widely used during echocardiography. Nevertheless, although mouse CFVR values from the current study are comparable with previously published data from, for example, ex vivo set-ups (16 –19), caution still

ADENOSINE-INDUCED CORONARY HYPEREMIA IN C57BL/6

The CFVR averaged at 2.0 in C57BL/6Adenosine mice. The dose used in the present study (160 ␮g/kg/min) would theoretically render a blood concentration between 4.5 and 6.5 ␮M/min, calculated from a blood volume of 2 to 3 ml. This dose was found to induce maximal coronary hyperemia without significant systemic hemodynamic impact. However, the resting values of HR and LCA flow velocity were slightly higher in the adenosine group, which could probably be caused by the relative invasiveness of tail vein catheterization. Although an elevated HR is known to influence basal coronary flow in man through increased cardiac workload, in anesthetized mice with decreased baseline HR no correlation was found between HR and CFVR, which may suggest a minor role of HR for CFVR in this experimental setting. Adenosine is commonly used in clinical practice to induce maximal coronary hyperemia. Doses of 160 ␮g/ kg/min result in CFVR values of 3 to 5 in healthy subjects and ⬍2 in subjects with significant coronary stenosis (8). The mechanism for the adenosine-mediated vasodilatation is not entirely understood, but is most probably mediated through both NO-independent and NO-dependent pathways (20). In mice, adenosine mainly acts through the adenosine A2A receptor (17). Ex vivo perfusion studies in mice hearts have shown that adenosine usually induces resistance lowering and flow increases that correspond to a coronary flow reserve (CFR) between 1.4 and 4.4 (16 –19). The differences in flow responses reported in these studies are most probably attributable to differences in adenosine concentrations (0.1 to 740 ␮M) and other methodologic differences, such as basal metabolism and oxygen saturation of the perfusion medium. However, these high adenosine concentrations are not possible to use in living animals without considerable influence on systemic hemodynamics. In addition, many ex vivo perfusion set-ups lack the important parameter of afterload, which is a factor known to influence basal metabolism and thus basal coronary flow and CFR. Based on these major methodologic discrepancies, values from in vivo and ex vivo approaches may differ substantially. MICE.

Table 3. Coronary Flow Velocity Measurements and CFVR C57BL/6 hypoxia C57BL/6 adenosine LDLR⫺/⫺

Baseline

Hyperemia

CFVR

20 ⫾ 2 26 ⫾ 5* 19 ⫾ 7

40 ⫾ 10 50 ⫾ 11 34 ⫾ 7

2.03 ⫾ 0.48 1.94 ⫾ 0.45 1.93 ⫾ 0.60

*Difference (p ⱕ 0.05) between C57BL/6Hypoxia and C57BL/6Adenosine, mean ⫾ standard deviation are shown. CFVR ⫽ coronary flow velocity reserve; LDLR ⫽ low-density lipoprotein receptor.

JACC Vol. 46, No. 4, 2005 August 16, 2005:720–7

Wikström et al. Imaging of Coronary Atherosclerosis in Mice

725

Figure 3. Typical histology of the proximal (A, B) and mid (C) left coronary artery in low-density lipoprotein receptor (LDLR)⫺/⫺ mice. Coronary artery lesions were found in the proximal but not in the mid left coronary artery. Arrowheads indicate coronary lesions. Scale bar ⫽ 200 ␮m. HYPOXIA-INDUCED CORONARY HYPEREMIA IN C57BL/6

Moderate acute hypoxia induced profound changes in flow velocity in all animals without increasing HR. The normal physiologic response to hypoxia is an increase in respiration and HR, which is mediated through neural and metabolic pathways. The following increase in cardiac workload, estimated by the rate-pressure product, also increases coronary flow. Nevertheless, Kaufmann et al. (21) showed that an increase in myocardial flow remains after compensation for the increase in rate-pressure product, which indicates a direct hypoxia-induced dilation. Further, in the same study, the investigators showed an inverse correlation between coronary flow and arteriolar PO2. Although hypoxic stimuli seem to induce a slight increase of HR in humans, anesthetized mice did not show any HR changes during moderate hypoxia, a phenomenon also known from studies in newborn lambs, which also showed a CFR value around 2 (22). Whether this is attributable to the species difference or to effects of anesthesia is currently not clear. Nevertheless, even moderate hypoxia seems to be a potent trigger of coronary hyperemia, which provides a non-invasive method of measuring CFVR in mice. Taken together, intravenous adenosine infusion and acute moderate hypoxia induced comparable coronary hyperemia. These two protocols provide us with relatively safe and non-invasive techniques for assessing mouse CFVR in vivo. Application of the hypoxia-induced CFVR protocol in atherosclerotic mice. HYPOXIA AND CFVR IN LDLRⴚ/ⴚ MICE. Moderate reduction of oxygen tension seems to induce substantial coronary hyperemia in LDLR⫺/⫺ mice. MICE.

Figure 4. The correlation between coronary flow velocity reserve (CFVR) and minimal lumen diameter (MLD) in low-density lipoprotein receptor (LDLR)⫺/⫺ mice (p ⬍ 0.005, R2 ⫽ 0.8707).

Further, the correlation found between CFVR and MLD may indicate that 12.5% O2 indeed causes adequate vasodilatation to override the auto-regulatory mechanism of the coronary resistance vessels, thus making the proximal lumen diameter the flow-limiting factor. It is also possible that the rapid baseline HR and short diastolic phase make mice more susceptible to hypoxia. The CFVR ranged from 1.4 to 2.8 in these aged LDLR⫺/⫺ mice, compared with the CFVR range of 1.5 to 5.0 in humans. The narrower range of CFVR in these mice could make it difficult to find a cut-off value for significant coronary artery stenosis, like the diagnostic concept in the human setting. CFVR versus MLD. In the present study, UBM was used to assess the minimal proximal LCA lumen diameter in atherosclerotic mice. The atherosclerotic lesions in the proximal LCA were found to intrude more or less into the lumen, resulting in a relatively heterogenic lumen diameter (200 to 400 ␮m) in the group. Subsequent studies of histologic sections of the mid and proximal segments of the LCA showed advanced lumen narrowing lesions in the proximal but not the mid segments. The findings are in line with those of previous studies in apoE⫺/⫺ mice, which showed that the proximal parts of the coronary arteries were the first and most prominent sites of coronary lesion formation (10). Recently, by using a 30-MHz UBM system, Zhou et al. (7) showed that the proximal segments of mouse LCA could be visualized. We showed previously that by using CDE it is possible to visualize the entire LCA and to make flow measurements in the mid and distal segments, which enables CFVR measurement downstream of coronary lesions. Further, when using the UBM system, there are difficulties in visualizing the proximal LCA in alternative imaging planes than the one that is tangential to the epicardial surface. This may cause incorrect measurement of coronary artery lumen diameter in cases of asymmetric plaque growth. Thus, despite the good correlation between CFVR and MLD, it is conceivable that CFVR may better reflect asymmetric lumen narrowing, as well as the longitudinal extent of the coronary atherosclerotic lesion. Theoretically, MLD should impact flow velocity by the second power, according to the Poiseuille law. Interestingly, when applying a non-linear exponential regression analysis, an improved correlation was found between CFVR and

726

Wikström et al. Imaging of Coronary Atherosclerosis in Mice

MLD compared with the linear relationship. However, because of the relatively small study population, this issue needs to be addressed in future studies. CFVR in various mouse strains. The CFVR values induced either by adenosine or hypoxia averaged 2 in mice, whereas corresponding values in healthy humans usually exceed 3. This discrepancy may be explained by several factors. First, the housing of mice in cages without voluntary daily physical activity may indeed result in reduced cardiac performance. Second, one may speculate that the small body size and the high HR could lead to an elevated basal metabolism with corresponding lower cardiac reserve capacity compared with the human counterpart. Third, it is also now evident that mice are only capable of increasing their oxygen uptake by a factor of two when they are forced to run on a treadmill (23), compared with the nearly 20-fold increase in humans. However, whether this major difference in CFVR between humans and mice is attributable to constitutional or environmental etiology still remains unclear. In non-atherosclerotic wild-type mice (C57BL/6) with a proximal LCA lumen diameter of approximately 500 ␮m, CFVR seems to be affected by certain factors, i.e., coronary endothelial and smooth muscle cell function, other than the coronary artery morphology. In a parallel study using non-atherosclerotic rats, we recently found comparably low pre-interventional CFVR that was markedly improved after voluntary exercise with concomitant improved endothelial function (data to be published). Thus, it is conceivable that the similar low CFVR values in wild-type and atherosclerotic mice, which are also supported in a previous ex vivo study (16), could be a consequence of sedentary cage life. This issue is currently under investigation in our laboratory. Moreover, by removing genes, such as the LDL-receptor gene, the cardiovascular phenotype is most likely altered in more ways than the intended. Thus, in the current work we did not intend to address this complicated issue regarding strain differences; instead, the C57BL/6 mice were used for development of the CFVR protocol in terms of adenosine tolerance and feasibility. The ability of our methods to predict coronary artery lumen diameter is shown in the present work by using aged LDLR⫺/⫺ mice with various degrees of proximal coronary artery atherosclerosis as evident in histology. Methodologic consideration and implications. The maximal theoretical resolution of the 15-MHz probe is 150 ␮m. We have tested various gate sizes ranging from 100 ␮m using UBM to 1.5 mm using CDE in our initial studies and found similar flow velocity values. Instead of being a disadvantage, we have found that a larger gate size (1 to 1.5 mm) seems to provide more distinct Doppler signals, because the coronary arteries will move within the sample volume despite the cardiac motion. Because the tissue motion velocity is substantially lower than the coronary flow velocity in mice, a good signal-to-noise ratio can be ob-

JACC Vol. 46, No. 4, 2005 August 16, 2005:720–7

tained, resulting in more distinct Doppler signals from the coronary arteries despite the larger sample volume. In clinical studies, Doppler-assisted CFVR has been validated to predict coronary artery stenosis (8). Recently, using quantitative angiography and invasive CFVR measurement, Sanjay et al. (9) showed good correlation between CFVR and MLD. In mice, coronary circulation has typically been studied in ex vivo perfusion set-ups with wellcontrolled hemodynamic parameters, which may facilitate detailed mechanistic studies (17). Although high-resolution magnetic resonance imaging has been used successfully to study mouse cardiac function as well as tissue damage after temporal LAD ligation, single-photon-emission computer tomography has been used recently to evaluate mouse myocardial perfusion (24,25). However, no previous methods have been capable of assessing in vivo coronary artery function and morphology in a quantitative manner. We showed here that UBM and CFVR could be used to visualize the naturally occurring coronary atherosclerosis in living mice. Indeed, atherosclerotic mice showed highly variable coronary lesion phenotypes, and correlations between peripheral and coronary artery atherosclerosis were poor (data on file). These facts may further emphasize the necessity of using powerful imaging methods to select animal models with pathogenic phenotypes relevant to human coronary artery diseases. The finding that the average CFVR overlapped between young healthy and old atherosclerotic mice may suggest that the current CFVR values could be strain specific. However, the relationship between CFVR and MLD is probably related to flow-limiting morphologic changes in coronary arteries, and should also be valid in other atherosclerotic mouse strains with advanced coronary artery lesions. Nevertheless, caution should still be taken when comparing absolute values of CFVR values of the current mouse strains with others. These potential strain-dependent coronary artery characteristics indeed justify further investigation. Study limitations. The major limitation of the study is the lack of gold standards to assess mouse coronary flow reserve and coronary artery lumen diameter in vivo. Microsphere technique, a gold standard for CFR measurement from humans to rats, is however hampered in mice because of the extremely small blood volume (26). Assessment of MLD, on the other hand, is complicated by substantial methodologic difficulties regarding perfusion fixation and sectioning in the histologic preparations. In vivo MLD of the coronary arteries is influenced on the one hand by morphology of the vessel, and on the other hand by physiologic parameters, such as BP, as well as by intramyocardial pressure. Further, there are considerable variations of the coronary artery diameter during the cardiac cycle. Using UBM, a 40% diameter increase is evident during diastole compared with systole in wild-type mice (data on file). Obviously, this issue can only be addressed properly in vivo. Thus, histology should preferentially be used in a qualitative way to address

JACC Vol. 46, No. 4, 2005 August 16, 2005:720–7

occurrence and extension of coronary artery atherosclerosis, rather than the quantitative approach for assessing in vivo MLD. However, transthoracic ultrasound has been used by several investigators to assess coronary artery morphology in humans with good correlation to quantitative angiography (27). Using UBM with its excellent resolution, coronary artery diameter can be obtained in similar imaging windows. To further confirm the current methods, it is conceivable that the invasive synchrotron radiation microangiographic technique could be of value (28). Conclusions. The CFVR and UBM techniques can be used to measure atherosclerosis-related lumen narrowing of the left coronary artery in living mice. These non-invasive techniques may provide us with novel tools for following up disease status in mouse coronary arteries in a quantitative manner. Acknowledgments The authors thank Mrs. Jia Jing for excellent technical assistance and Dr. Göran Långström for statistical advice. Reprint requests and correspondence: Dr. Li-ming Gan, Göteborg University, Institute of Physiology and Pharmacology, Department of Physiology, Box 432, 405 30 Göteborg, Sweden. E-mail: [email protected].

REFERENCES 1. Daugherty A. Mouse models of atherosclerosis. Am J Med Sci 2002;323:3–10. 2. Breslow J. Transgenic mouse models of lipoprotein metabolism and atherosclerosis. Proc Natl Acad Sci U S A 1993;90:8314 – 8. 3. Christensen G, Wang Y, Chien KR. Physiological assessment of complex cardiac phenotypes in genetically engineered mice. Am J Physiol 1997;272:H2513–24. 4. Baumgart D, Schmermund A, Goerge G, et al. Comparison of electron beam computed tomography with intracoronary ultrasound and coronary angiography for detection of coronary atherosclerosis. J Am Coll Cardiol 1997;30:57– 64. 5. Sakuma H, Kawada N, Takeda K, Higgins CB. MR measurement of coronary blood flow. J Magn Reson Imaging 1999;10:728 –33. 6. Gan LM, Wikstrom J, Bergstrom G, Wandt B. Non-invasive imaging of coronary arteries in living mice using high-resolution echocardiography. Scand Cardiovasc J 2004;38:121– 6. 7. Zhou Y-Q, Foster FS, Nieman BJ, Davidson L, Chen XJ, Henkelman RM. Comprehensive transthoracic cardiac imaging in mice using ultrasound biomicroscopy with anatomical confirmation by magnetic resonance imaging. Physiol Genomics 2004;18:232– 44. 8. Hozumi T, Yoshida K, Ogata Y, et al. Noninvasive assessment of significant left anterior descending coronary artery stenosis by coronary flow velocity reserve with transthoracic color Doppler echocardiography. Circulation 1998;97:1557– 62. 9. Chugh SK, Koppel J, Scott M, et al. Coronary flow velocity reserve does not correlate with TIMI frame count in patients undergoing non-emergency percutaneous coronary intervention. J Am Coll Cardiol 2004;44:778 – 82.

Wikström et al. Imaging of Coronary Atherosclerosis in Mice

727

10. Caligiuri G, Levy B, Pernow J, Thoren P, Hansson GK. Myocardial infarction mediated by endothelin receptor signaling in hypercholesterolemic mice. Proc Natl Acad Sci U S A 1999;96:6920 – 4. 11. Heaps CL, Tharp DL, Bowles DK. Hypercholesterolemia abolishes voltage-dependent K⫹ channel contribution to adenosine-mediated relaxation in porcine coronary arterioles. Am J Physiol Heart Circ Physiol 2005;288:H568 –76. 12. Johansson ME, Hagg U, Wikstrom J, Wickman A, Bergstrom G, Gan LM. Hemodynamically significant plaque formation and regional endothelial dysfunction in cholesterol-fed ApoE ⫺/⫺ mice. Clin Sci (Lond) 2005;108:531– 8. 13. Wyss CA, Koepfli P, Fretz G, Seebauer M, Schirlo C, Kaufmann PA. Influence of altitude exposure on coronary flow reserve. Circulation 2003;108:1202–7. 14. Yang X-P, Liu Y-H, Rhaleb N-E, Kurihara N, Kim HE, Carretero OA. Echocardiographic assessment of cardiac function in conscious and anesthetized mice. Am J Physiol Heart Circ Physiol 1999;277: H1967–74. 15. Croteau E, Benard F, Bentourkia Mh, Rousseau J, Paquette M, Lecomte R. Quantitative myocardial perfusion and coronary reserve in rats with 13N-ammonia and small animal PET: impact of anesthesia and pharmacologic stress agents. J Nucl Med 2004;45:1924 –30. 16. Godecke A, Ziegler M, Ding Z, Schrader J. Endothelial dysfunction of coronary resistance vessels in apoE⫺/⫺ mice involves NO but not prostacyclin-dependent mechanisms. Cardiovasc Res 2002;53:253– 62. 17. Flood A, Headrick JP. Functional characterization of coronary vascular adenosine receptors in the mouse. Br J Pharmacol 2001;133:1063–72. 18. Bratkovsky S, Aasum E, Birkeland CH, Riemersma RA, Myhre ESP, Larsen TS. Measurement of coronary flow reserve in isolated hearts from mice. Acta Physiol Scand 2004;181:167–72. 19. Talukder MAH, Morrison RR, Mustafa SJ. Comparison of the vascular effects of adenosine in isolated mouse heart and aorta. Am J Physiol Heart Circ Physiol 2002;282:H49 –57. 20. Buus NH, Bottcher M, Hermansen F, Sander M, Nielsen TT, Mulvany MJ. Influence of nitric oxide synthase and adrenergic inhibition on adenosine-induced myocardial hyperemia. Circulation 2001;104:2305–10. 21. Kaufmann PA, Schirlo C, Pavlicek V, et al. Increased myocardial blood flow during acute exposure to simulated altitudes. J Nucl Cardiol 2001;8:158 – 64. 22. Oskarsson G, Pesonen E, Gudmundsson S, Ingimarsson J, Sandstrom S, Werner O. Coronary flow reserve in the newborn lamb: an intracoronary Doppler guide wire study. Pediatr Res 2004;55: 205–10. 23. Faldt J, Wernstedt I, Fitzgerald SM, Wallenius K, Bergstrom G, Jansson JO. Reduced exercise endurance in interleukin-6-deficient mice. Endocrinology 2004;145:2680 – 6. 24. Wu MC, Gao DW, Sievers RE, Lee RJ, Hasegawa BH, Dae MW. Pinhole single-photon emission computed tomography for myocardial perfusion imaging of mice. J Am Coll Cardiol 2003;42:576 – 82. 25. Yang Z, Berr SS, Gilson WD, Toufektsian MC, French BA. Simultaneous evaluation of infarct size and cardiac function in intact mice by contrast-enhanced cardiac magnetic resonance imaging reveals contractile dysfunction in noninfarcted regions early after myocardial infarction. Circulation 2004;109:1161–7. 26. Richer C, Domergue V, Gervais M, Bruneval P, Giudicelli JF. Fluospheres for cardiovascular phenotyping genetically modified mice. J Cardiovasc Pharmacol 2000;36:396 – 404. 27. Kiviniemi TO, Saraste M, Koskenvuo JW, et al. Coronary artery diameter can be assessed reliably with transthoracic echocardiography. Am J Physiol Heart Circ Physiol 2004;286:H1515–20. 28. Yamashita T, Kawashima S, Ozaki M, et al. In vivo angiographic detection of vascular lesions in apolipoprotein E-knockout mice using a synchrotron radiation microangiography system. Circ J 2002;66:1057–9.