Therapeutic Ultrasound Increases Myocardial Blood Flow in Ischemic Myocardium and Cardiac Endothelial Cells: Results of In-vivo and In-Vitro Experiments

Abstract

Background:

Therapeutic ultrasound (US) can reduce infarct size in a model of coronary thrombosis even when sonothrombolysis is ineffective. We hypothesized that US-induced cardioprotection is mediated by molecules released from the vascular endothelium that increase myocardial blood flow (MBF) and also have direct tissue salvaging effects.

Methods:

We performed in vivo and in vitro experiments using a 1.05 MHz transducer. For the in-vivo experiments we studied 10 control and 10 US treated dogs undergoing occlusion of the left anterior descending coronary artery (LAD). MBF was measured using myocardial contrast echocardiography. For the in-vitro experiments we exposed primary mouse cardiac endothelial cells (ECs) to US at baseline or following oxygen-glucose deprivation (OGD) and measured eNOS phosphorylation (p-eNOS) as well as adenosine and the eicosanoids: epoxyeicosatrienoic acids (EETs), dihydroxyeicosatrienoic acids (DHETs), and hydroxyl-eicosatetraenoic acids (HETEs).

Results:

In vivo: US treatment caused higher MBF (20±10 vs 10±8, p=0.03) and higher wall thickening (3±3% vs 1±0.4%, p=0.01) in the collateral-derived border zone compared to control. Epicardial MBF in LAD bed also tended to be higher (17±17 vs 5±4, p=0.05) in US treated versus control animals; however endocardial MBF in this region was similar to controls (13±14 vs 14±7). In-vitro: p-eNOS and adenosine increased (129±11% and 286±63%, respectively, p<0.01) with US compared to unstimulated cells. Similar results were obtained with EETs. After OGD, p-eNOS decreased and was restored with application of US. Similar changes were noted with EETs. Cell viability decreased with OGD and returned to near baseline with US.

Conclusions:

US increases MBF in ischemic tissue in-vivo. This effect is likely mediated by the release of a plethora of coronary vasodilators during US treatment that have direct tissue salvaging effects. Therapeutic US, therefore, has potential for treatment of acute and chronic myocardial ischemia independent of its effect on thrombolysis.

Introduction

More than six decades ago, it was observed that ultrasound (US) treatment increased blood flow in femoral arteries of dogs^{1, 2} and patients.^{3, 4} US also normalized the pH of ischemic tissue and reversed cyanosis of skeletal muscle during femoral artery ligation in a rabbit model.⁵ Additionally, tissue perfusion and myocardial capillary density increased and pH normalized with US in a model of total coronary occlusion. These effects of femoral and coronary artery ligations were attenuated with endothelial nitric oxide (NO) synthase inhibitor N-ω-nitro-L-arginine methyl ester (L-NAME) suggesting that they resulted from release of NO.^{6, 7}

Accordingly, the majority of the literature on the tissue beneficial effect of US implicates NO and its isoforms.^8–10 We have recently shown that brain endothelial cells (ECs) also release other vasodilator compounds like epoxyeicosatrienoic acids (EETs) and adenosine when exposed to US.¹¹ For the present study we hypothesized that US (1.05 MHz) increases myocardial blood flow (MBF) in ischemic tissue through the release of a plethora of coronary vasodilators from myocardial ECs that also have direct tissue salvaging effects. We chose this frequency because we had previously noted effective sonothrombolysis in-vitro with this frequency.¹²

Methods

The study protocols were approved by our institutional animal research committee and conformed to the American Heart Association guidelines for the use of animals in research.

InVivo Experiments

Animal Preparation:

Twenty anesthetized adult male mongrel dogs (30-35 kg) were studied. A 6F sheath was placed in the femoral artery and connected to a pressure transducer for arterial pressure measurements. Catheters were placed in the femoral vein for administration of drugs, fluids and microbubbles, which were used to measure MBF.

The heart was exposed by left lateral thoracotomy and suspended in a pericardial cradle. The proximal sections of the left anterior descending (LAD) and left circumflex (LCx) coronary arteries were dissected from surrounding tissue. Ultrasonic time-of-flight flow probes (Series SC, Transonics. Ithaca, NY) were placed on both coronary arteries and connected to a digital flow meter (Model T206, Transonics, Ithaca, NY). Mean pressures and epicardial coronary blood flow (CBF) were digitally recorded on a multichannel recorder (PowerLab, AD Instruments, Inc., Colorado Springs, CO) and displayed on-line on a computer system (iMac, Apple Inc., Cupertino, CA) throughout the experiment. Occlusion of the LAD was produced by placing an adjustable occluder around the LAD and confirmed by complete abolition of epicardial blood flow.

Myocardial Contrast Echocardiography (MCE):

MCE was used to determine the perfusion bed size of the LAD¹³, the spatial extent of the risk area during coronary occlusion ¹⁴, the collateral-dependent ischemic zone¹⁵, and MBF at each stage of the experiment¹⁶ as previously described.

For contrast, lipid-shelled decafluorobutane microbubbles were prepared by sonication of an aqueous lipid dispersion of polyoxyethylene-40-stearate and distearoyl phosphatidylcholine saturated with decafluorobutane gas. Microbubble concentration was measured by electrozone sensing (Multisizer III, Beckman Coulter, Brea, CA).

Intermittent harmonic imaging was performed using a phased array system (Sonos 5500, Philips) with ultrasound transmitted at 1.3 MHz and received at 3.6 MHz. The microbubble infusion rate (1-2 mL·min⁻¹, at 1×10⁷ microbubbles·mL⁻¹) was adjusted for each dog at baseline to provide optimal myocardial opacification with minimal left ventricular (LV) posterior wall shadowing. The transducer was fixed in position distal to the occluder and a water bath placed over the heart served as an acoustic interface. The dynamic range, overall gain, depth and focus were optimized at the beginning of each experiment for high mechanical index (MI) intermittent imaging (MI of 1.0) and were held constant throughout.

For measuring MBF, background-subtracted pulsing interval (PI) versus acoustic density plots were generated separately from the central portion of the risk area as well as the collateral dependent zones. The resultant plots were fitted to the function: y =A(1−e^−βt), where y is the acoustic density at PI t, A is the plateau acoustic density representing myocardial blood volume (MBV), and β is the rate of rise of acoustic density denoting the mean myocardial microbubble velocity. The product A·β represents MBF.¹⁶ For deriving epicardial and endocardial MBF, regions-of-interest were placed over the outer and inner one-thirds of the myocardium, respectively within the risk area.

Wall thickening Measurements:

Wall thickening (WT) measurements were made from real-time harmonic images acquired at the same mid-papillary muscle plane as the MCE images¹⁷ in regions-of-interest derived from MCE images where the LAD perfusion bed, the LCx perfusion bed, and the collateral-defined border zones are defined. Plots of %WT over the entire systolic contraction sequence in the central 75% of these regions are then generated¹⁷.

Study Protocol:

Figure 1 panel A depicts the study protocol and Table 1 depicts the US parameters used. After acquisition of baseline data (stage A), coronary occlusion was performed (stage B). Animals were then randomized to receive either US for 30 min, n=10) or control (n=10) administered 45 min after coronary occlusion to simulate the clinical setting.

Figure 1:

Experimental Protocols for the in vivo (A) and in vitro (B) experiments. MCE=myocardial contrast echocardiography; TTC=triphenyl tetarzolium chloride; EC=primary endothelial cells; OGD=oxygen-glucose deprivation; PRAPA=peak rarefactional acoustic pressure amplitude; EET=Epoxyeicosatrienoic acids; DHET=Dihydroxyeicosatrienoic acid; HETE=Hydroxyeicosatetraenoate acid. See text for details.

Table 1.

Ultrasound Parameters Used for the Study

Parameter	In-vivo	In-vitro
Frequency	1.10 MHz	1.05 MHz
Peak negative pressure	1 MPa	02, 0.3, 0.4 and, 0.5 MPa
Mode	Pulsed	Pulsed
# cycles	50	50
Pulse Repetition Frequency	50	50
Duty cycle (%)	0.25	0.25
# active elements per transducer	3	1
Outer element power	30%
Middle element power	50%	100%
Inner element power	100%
Transducer diameter	81.8 mm	81.8 mm
Central hole	none	20 mm
Transducer height	19 mm	19 mm
Focussed/unfocussed	Unfocussed	Focused
Beam width at focus	60 mm (natural focus)	1.34 mm
Insonification area	28.3 cm2	1.4 mm2

Immediately following US treatment (Stage C) MCE was performed to measure immediate effects of US. Another 45 minutes were allowed with continued coronary occlusion (Stage D) at which time MCE was repeated in order to determine prolonged effects of US treatment. Finally, MCE was performed 90 min after reperfusion and before animal sacrifice.

At the end of the experiment, a needle was passed through the heart at the site of the imaging plane and heart frozen. Post-mortem triphenyl tetrazolium chloride (TTC staining) of the frozen heart section corresponding to the 2DE imaging plane was performed to measure infarct size (IS)¹⁸ that was then normalized to the total LV cross myocardial sectional area and also expressed as a percent of MCE-derived risk area.

Statistical Methods:

Average measurements between groups (US versus control) were compared at stages C and D via linear mixed models adjusting for initial value recorded at stage B. A random intercept was included to account for within-subject correlation. The factors group and stage and the interaction of group and stage were included in the models. Overall group effect (across both stages C and D), and the differences between US and control individually at stage C and at stage D were tested with F-tests under the linear mixed model framework. Including data from stage E in the mixed models did not meaningfully change results.

In-vitro experiments

Primary Mouse Cardiac Endothelial Cells Isolation and Culture:

We isolated primary mouse vascular endothelial cells (ECs) from 8-week old male C57BL6 mice (Charles River Laboratory, Wilmington, MA) as previously described.¹¹ ECs were then treated with US either at baseline (no oxygen-glucose deprivation (OGD) or after OGD. For the latter an anaerobic chamber filled with an anoxic gas mixture (5% CO2, 5% H2, and 90% N2) was used. The O² concentration was maintained at 0 parts per million using a palladium catalyst. ECs were subjected to OGD for 2 h at 37°C.

Ultrasound Experimental Setup:

Cells were placed in a tank filled with 0.2 μm filtered and degassed water at 37.4°C. The US parameters used are depicted in Table 1. The treatment details are described previously.¹¹ Four different peak rarefactional acoustic pressure amplitudes (PRAPAs) were used: 0.2, 0.3, 0.4, and 0.5 MPa.¹¹

Post US Treatment Sample Collection:

Medium was removed and frozen on dry ice for adenosine analysis. ECs were scraped in ice-cold phosphate buffered saline (PBS) containing antioxidant solution (1:100), and frozen on dry ice. A portion of the cell suspension was used to determine protein concentration for normalization of eicosanoid concentrations; the remaining cell suspension was analyzed for eicosanoid profile. ECs in duplicate wells were lysed and frozen on dry ice for phospho-eNOS protein analysis by Western blot. Preparation of samples and calibration for adenosine and eicosanoids metabolites using LC-MS/MS analysis as well as Western blotting were previously described in details.¹¹ All Mass Spec analyses were performed at the OHSU Metabolomics Core.

Cell Viability:

EC viability following US stimulation was quantified in separate plates. The medium was replaced with endothelial culture medium and ECs were incubated for 24 h. Cell viability was then assessed by the reduction of MTT (3-4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) (Sigma, St. Louis, MO), which is converted to a formazan, by viable cells that is measured spectrophotometrically at 540 nm.

Statistical Methods:

Data are presented as mean ± SEM. Groups were compared by one-way ANOVA with the post hoc Bonferroni multiple comparisons test for eicosanoid analysis (each eicosanoid separately), or two-way ANOVA with the post hoc Tukey pairwise multiple comparison test for adenosine analysis and two-way ANOVA for analysis of eNOS phosphorylation. Differences were considered significant at p < 0.05.

Results

In Vivo Experiments

Hemodynamic Data:

Table 2 lists blood pressures, heart rate, and LCx CBF between US treated and control groups at all stages. While the blood pressure in the US treated group tended to decline over time, no significant differences were found between the control and US groups at any of the stages.

Table 2.

Hemodynamic Results (mean±1SD)

Variables	Stage A	Stage B	Stage C	Stage D	Stage E
Systolic Blood Pressure (mmHg)-Control	110±30	108±26	104±24	102±29	102±30
Systolic Blood Pressure (mm Hg)-US	113±28	103±14	96±22	90±20	88±22
Mean Blood Pressure (mm Hg)-Control	94±29	95±27	90±24	87±29	80±28
Mean Blood Pressure (mm Hg)-US	99±26	91±14	85±22	79±20	76±23
Heart rate (beats·min−1)-Control	111±26	110±22	109±22	115±27	108±28
Heart rate (beats·min−1)-US	120±11	113±16	109±18	114±31	105±20
LCx CBF (mL·min−1)-Control	29±10	32±12	31±6	28±12	35±14
LCx CBF (mL·min−1)-US	46±14	39±13	32±11	40±23	36±23

Stage A=baseline

Stage B=LAD occlusion prior to treatment

Stage C=LAD occlusion with or without treatment

Stage D=LAD occlusion 45 min after treatment completion

Stage E=90 min after reperfusion

US=ultrasound; LCx= left circumflex coronary artery

MCE Data:

Figures 2 and 3 depict examples from control and US treated groups, where panels A to C depict LAD perfusion bed size, risk area, and ROI’s placed over the myocardium at baseline MCE image (stage A) from where time versus acoustic intensity plots were derived. Panels D and E depict the actual plots derived from the risk area and border zone during coronary occlusion (stage B) and 45 min after treatment period (stage D). Videos 1 to 4 depict the MCE aligned end-systolic images at different PI’s from where the plots were derived. It is apparent that transmural MBF in the risk area and border zone were significantly higher at stage D in the US treated animal compared to control.

Figure 2:

MCE images from a control dog. Panels A to C depict LAD perfusion bed size, risk area, and ROI’s placed over the myocardium at baseline MCE image (stage A) from where time versus acoustic intensity plots were derived. Panels D and E depict the actual plots derived from the risk area and border zone during coronary occlusion (stage B) and 45 min after treatment period (stage D). Videos 1 and 2 depict the MCE aligned end-systolic images at different PI’s from where the plots were derived at stages B and D, respectively. See text for details.

Figure 3:

MCE images from a dog undergoing US treatment. Panels A to C depict LAD perfusion bed size, risk area, and ROI’s placed over the myocardium at baseline MCE image (stage A) from where time versus acoustic intensity plots were derived. Panels D and E depict the actual plots derived from the risk area and border zone during coronary occlusion (stage B) and 45 min after treatment period (stage D). Videos 3 and 4 depict the MCE aligned end-systolic images at different PI’s from where the plots were derived at stages B and D, respectively. See text for details.

MBF velocity (β) and total MBF (A·β) within the MCE-defined LAD risk area were significantly different between US and control groups at stage D and demonstrated borderline significance at Stage C. No differences were noted at stage E (Table 3). To further define the differences in these parameters between the 2 groups, we examined endocardial and epicardial MBF by defining regions-of-interest over the outer and inner one-third, respectively, of the LAD risk area in stages B to D. While US increased epicardial MBF velocity (β) significantly with a trend towards increase in total MBF (A·β) compared to control at stage C (end of treatment period), it did not change endocardial values significantly. A similar trend was noted at Stage D, which did not reach significance because of the large standard deviation in the values.

Table 3.

MCE Results (mean±1SD) from the LAD Risk Area

Variables	Stage A	Stage B	Stage C	Stage D	Stage E
Myocardial Blood Volume (A)-Control	49±19	35±28	30±15	27±12	50±17
Myocardial Blood Volume (A)-US	67±12	44±16	43±20	37±19	50±14
MBF Velocity (β)-Control	0.47±0.23	0.35±0.31	0.22±0.16	0.18±0.10	0.36±0.35
MBF Velocity (β)-US	0.45±0.18	0.28±0.15	0.30±0.11	0.45±0.35‡	0.37±0.16
Total MBF (A•β)-Control	24±18	14±16	6±4	5±4	22±32
Total MBF (A•β)-US	31±16	13±8	13±8	17±17§	19±9
Epicardial MBF velocity (β)-control		0.30±0.26	0.15±0.07	0.19±0.16
Epicardial MBF velocity (β)-US		0.38±0.35	0.31±0.21*	0.27±0.18
Endocardial MBF Velocity (β)-Control		0.26±0.13	0.34±0.34	0.37±0.17
Endocardial MBF Velocity (β)-US		0.46±0.39	0.41±0.20	0.33±0.12
Epicardial MBF (A•β)-Control		11±9	7±3	7±7
Epicardial MBF (A•β)-US		16±14	16±15†	11±12
Endocardial MBF (A•β)-Control		11±7	13±14	9±6
Endocardial MBF (A•β)-US		19±10	14±7	9±7

Stage A=baseline

Stage B=LAD occlusion prior to treatment

Stage C=LAD occlusion with or without treatment

Stage D=LAD occlusion 45 min after treatment completion

Stage E=90 min after reperfusion

US=ultrasound; MBF=myocardial blood flow;

^‡p=0.01 compared to control

^§p=0.02 compared to control

^*p=0.04 compared to control

^†p=0.05 compared to control

In the collateral-defined border zone, MBF (A·β) was also significantly higher during US exposure at stage C (end of treatment period, Table 4) and showed a similar trend at Stage D, but not at Stage E. Thus, the US induced increase in MBF in the LAD bed during coronary occlusion resulted from maintenance rather than decline of epicardial MBF compared to control, with no effect on endocardial MBF. Similarly, US resulted in maintenance of transmural MBF in the border zone compared to control.

Table 4.

MCE Results (mean±1SD) from the LAD Border Zone

Variables	Stage A	Stage B	Stage C	Stage D	Stage E
Myocardial Blood Volume (A)-Control	68±26	62±12	50±25	32±21	64±26
Myocardial Blood Volume (A)-US	82±13	61±24	61±15	58±26	64±19
MBF Velocity (β)-Control	0.44±0.19	0.27±0.20	0.22±0.12	0.29±0.19	0.30±0.14
MBF Velocity (β)-US	0.47±0.18	0.28±0.13	0.32±0.15	0.29±0.18	0.32±0.20
Total MBF (A·β)-Control	30±20	16±13	10±8	9±6	18±8
Total MBF (A·β)-US	37±12	17±10	20±10*	15±8	19±13

Stage A=baseline

Stage B=LAD occlusion prior to treatment

Stage C=LAD occlusion with or without treatment

Stage D=LAD occlusion 45 min after treatment completion

Stage E=90 min after reperfusion

US=ultrasound; LAD=left anterior descending artery; MBF=myocardial blood flow;

^*p=0.03

No changes were noted in any of the MCE parameters derived from the LCx bed between US and control groups at or between any of the stages. It must be noted, however, that a large portion of the LCx bed (the posterolateral wall) was not exposed to US.

Wall thickening data:

WT was normal in the LAD and LCx beds at baseline (Table 5). It remained unchanged in the LCx bed at all stages. There was marked reduction in WT in the LAD risk area after coronary occlusion (Stage B) that remained unchanged throughout Stages C and D in both groups. In contradistinction, while the control animals showed progressive deterioration of WT in the collateral-supplied border zone to akinesia after coronary occlusion, the US treated bed showed no further decline (Figure 4B), maintaining hypokinesia throughout Stages D and E, so that there was a significant difference in WT between control and US treated group during these stages. WT improved equally between US treated and control group after reperfusion.

Table 5.

Regional Wall Thickening Results (mean±1SD)

Variables	Stage A	Stage B	Stage C	Stage D	Stage E
% Wall Thickening (LAD risk area)-Control	35±1	3±5	0.3±0.3	1±2	5±5
% Wall Thickening (LAD risk area)-US	35±1	1±1	1±1	1±2	4±4
% Wall Thickening (LAD border zone)-Control	35±1	4±7	1±0.4	1±1	7±7
% Wall Thickening (LAD border zone)-US	36±1	3±3	3±3*	3±2*	8±5
% Wall Thickening (LCx bed)-Control	35±1	35±1	35±1	35±1	35±1
% Wall Thickening (LCx bed)-US	36±1	35±1	36±1	36±1	36±1

Stage A=baseline

Stage B=LAD occlusion prior to treatment

Stage C=LAD occlusion with or without treatment

Stage D=LAD occlusion 45 min after treatment completion

Stage E=90 min after reperfusion

US=ultrasound; LAD=left anterior descending coronary artery; LCx=left circumflex coronary artery

^*p=0.01 compared to control

Figure 4:

Panel A illustrates normalized intracellular eNOS phosphorylation in endothelial cells as a function of PRAPA (peak rarefactional acoustic pressure amplitude) at 15 and 45 min after ultrasound exposure (see text for details). Panel B depicts normalized adenosine levels as a function of PRAPA at 15 and 45 min after US exposure (See text for details).

LAD Perfusion Bed Size, Risk Area, and Infarct Size:

Table 6 shows the results of the perfusion bed size, risk area, and IS (both in absolute terms and as percent of risk area) from the 2 groups of dogs. There is no difference in any of these parameters between the two groups.

Table 6.

LAD Perfusion Bed Size, Risk Area, and Infarct Size (mean±1SD)

Variable Measured	Control	US Treated
Perfusion Bed Size (%LV myocardium)	44.6±11.8	41.5±12.3
Risk Area (%LV myocardium)	14.9±7.9	15.2±12.2
Border Zone (% Perfusion Bed Size - % Risk Area)	27.5±11.2	28.9±12
Infarct Size (% LV myocardium)	8.5±9.9	4.5±5.4
Infarct Size/Risk Area Ratio	53.5±20.4	57.6±26.8

US=ultrasound; LAD=left anterior descending coronary artery; LV=left ventricular

In Vitro Experiments

Baseline conditions

eNOS Phosphorylation:

Intracellular levels of phosphorylated eNOS demonstrated a dose dependent increase with increasing PRAPAs at 15 min following US exposure (Figure 4A). Statistical significance was reached for PRAPAs of 0.3 MPa or greater (121 ± 6%, 122 ± 9%, and 129 ± 11% of control for 0.3, 0.4, and 0.5 MPa, respectively, n=6, p < 0.05). By 45 min after US exposure eNOS phosphorylation levels returned to baseline.

Adenosine:

US caused a step-wise increase in adenosine release with increasing PRAPAs at both 15 and 45 min (Figure 4B). However, only the 15 min levels reached statistical significance at the 2 highest PRAPAs (286 ± 63% and 245 ± 45% of control at 0.4 and 0.5 MPa, n=6, p < 0.05).

Eicosanoids:

US resulted in increased intracellular levels of all EET regioisomers at 15 min for two PRAPAs (0.2 and 0.5 MPa) (Figure 5). At 0.2 MPa, levels of 8,9-, 11,12-, and 14,15-EETs increased to 151 ± 13%, 157 ± 15%, and 158 ± 15% of control, respectively (n=6, p < 0.05), and at 0.5 MPa they increased to 152 ± 16%, 158 ± 17%, and 156 ± 17% of control, respectively (n=6, p < 0.05). The DHET regioisomers (5,6-, 8,9-, 11,12-, and 14,15-DHETs) were below the threshold of detection (<0.15 pg eicosanoid/μg protein) at both 15 and 45 min after US exposure.

Figure 5:

Normalized intracellular 8,9-, 11,12-, and 14,15-EET levels as a function of PRAPA (peak rarefactional acoustic pressure amplitude) in the same order as in Figure 4 at 15 min after ultrasound exposure (see text for details).

US exposure did not result in significant changes in the levels of HETEs compared to control (n=6, p > 0.05) at any time-point or PRAPA.

Oxygen-Glucose Deprivation

ECs underwent OGD for 2 hr prior to US exposure. For these experiments we used the PRAPA that consistently caused a cellular response (0.5 MPa) at baseline. Measurements were made 15 min after US exposure.

eNOS Phosphorylation:

ECs exposed to OGD alone resulted in a significant decrease in intracellular eNOS phosphorylation (69 ± 7% of control, n=8, p < 0.05). On US exposure, the phosphorylation normalized to control levels (Figure 6A).

Figure 6:

Panel A shows normalized intracellular eNOS phosphorylation at baseline, OGD (oxygen-glucose deprivation) and OGD+US (ultrasound) conditions. Panel B illustrates normalized adenosine levels at baseline, OGD and OGD+US conditions (see text for details).

Adenosine:

ECs exposed to OGD increased adenosine levels (193 ± 33% of control, n=11, p < 0.05). US did not cause any further increase in adenosine levels (Figure 6B).

Eicosanoids:

When ECs were exposed to OGD, no significant effect on EETs regioisomers was seen. US increased the levels of all regioisomers significantly (n=7, p < 0.05, Figure 7A). No significant effect on DHET regioisomers was seen on OGD. US exposure caused a significant increase in all DHET regioisomers (n=7, p < 0.05).

Figure 7:

Panel A illustrates normalized 8,9-, 11,12-, and 14,15-EET (epoxyeicosatrienoic acid) levels at baseline, OGD (oxygen-glucose deprivation) and OGD+US (ultrasound) conditions. Panel B depicts normalized cell viability in baseline and OGD with and without US (see text for details).

Similarly, when ECs were exposed to OGD, no significant effect on HETEs regioisomers was seen whereas US exposure resulted in a significant increase in 18- and 20- HETE but not 5-, 11-, 12-, and 15-HETE (data not shown, n=7, p < 0.05).

Cell Viability:

At baseline, the presence or absence of US did not significantly alter cell viability. OGD, as expected, reduced cell viability significantly compared to baseline. US exposure resulted in an increase in cell viability to near baseline levels. (n=7, p < 0.05, Figure 7B).

Discussion

In Vivo Experiments

The new finding from this study is that while in control animals epicardial MBF in the occluded LAD perfusion bed and transmural MBF in the collateral-derived border zone continued to decline over time, it was maintained at the post-occlusion level with a 30 min application of 1.05 MHz US. The maintenance of MBF in the collateral-derived border zone was also associated with maintenance of %WT in this zone compared to control where WT deteriorated over time. However, endocardial MBF in the LAD bed was not favorably affected by US, nor was IS.

The higher epicardial MBF in the LAD perfusion bed can also explain the findings of the clinical study where despite lack of differences in coronary artery patency between US treated and control groups, ST segment elevation was less in the US treated group.¹⁹ ST segment elevation is thought to be a marker of epicardial injury²⁰ and any amelioration of that injury by increased MBF could result in reduced ST elevation or even its resolution. The maintenance rather than further deterioration of post coronary occlusion MBF and %WT in the collateral-derived border zone after US treatment could also contribute to reduction in the number of electrocardiographic leads demonstrating ST segment elevation. Higher epicardial MBF has been shown by MCE in patients with chronic CAD.²¹

Previous studies have exposed tissue to US from 1 to 120 minutes, andsimilar to our study, the effects were measured for 60-120 minutes following US exposure. Our aim was to simulate the clinical situation where US would be delivered between time of diagnosis and attempted infarct related artery recanalization. This would unlikely be more than 30 min. The beneficial effects of US in this instance would be required only until reperfusion was established. In our study, the effect of US lasted for at least 60 min after exposure, which would be acceptable in the clinical setting.

In vitro experiments

The in vitro experiments were performed determine the effects of US on release of coronary vasodilators that could explain the increase in MBF seen in vivo. The new findings of this study are that normal primary mouse cardiac ECs show increased eNOS phosphorylation coupled with release of adenosine and increased intracellular EETs at 15 min after <1 min of US exposure that subsides by 45 min. Following OGD, decreased EC viability is associated with a decrease in eNOS phosphorylation. Application of US during OGD results in normalization of EC viability and eNOS phosphorylation, which are coupled with increased levels of EETs, DHETs and HETEs. Our findings suggest that US protects EC against OGD-induced cell injury, in part, from enhanced synthesis of P450 eicosanoids.

Role of eNOS and NO

Prior studies have reported NO release and eNOS activation with US.²² US was found to normalize the pH, and to reverse cyanosis of skeletal muscle during femoral artery ligation in a rabbit model.⁵ Tissue perfusion and myocardial pH normalization were also found with US application in a model of total coronary occlusion.⁶ These effects of femoral and coronary artery ligations were attenuated with L-NAME.⁷

In another study, gracilis muscles of rabbits were exposed to US after femoral artery ligation. Compared to control, US increased calcium–dependent and calcium–independent NOS activity.⁸ US during ischemia improved functional capillary density 24 hours after reperfusion. In a separate study, US after ischemia not only improved capillary density but also arteriolar diameter and arteriolar and venular flow velocity and flow. L-NAME attenuated these effects.⁹ Another report showed real-time NO concentrations increased with increasing US energies without thermal effect that was partially reversed by L-N^G-monomethyl Arginine citrate (a nitric oxide synthase inhibitor).²³ Our eNOS phosphorylation results, though not new, served as an internal validation of the efficacy of the US approach used in our study for its direct tissue effect.

Role of Adenosine

This is the first study to show that adenosine is released from primary cardiac ECs on US exposure. Adenosine exerts both cardioprotective and vasodilators effects, through its actions on myocardial Ai and vascular A_2A receptors, respectively.^24–26 Stimulation of the A_2A receptor has also been implicated in ischemic post-conditioning²⁷ and in protection from reperfusion injury when administered immediately after the event.^28,29 On the other hand, the A₃ receptor has low density on myocytes, but it is believed to be beneficial in the presence of high adenosine levels, such as in myocardial ischemia.^25,26 Similarly, the A_2B receptor could participate in coronary vasodilation when adenosine levels are high.³⁰ Finally, there is evidence to suggest that there is interaction between the different adenosine receptors that potentiates cardioprotection.³¹

Role of Eicosanoids

This is the first study demonstrating increased levels of EETs when primary cardiac ECs are exposed to US. Free intracellular EETs can be increased either by de novo synthesis from arachidonic acid via cytochrome P450 expoxygenase or by release from preexisting pools stored in membrane phospholipids.² In our study US increases only EETs at baseline. Therefore, it is unlikely that this effect is mediated by non-selective arachidonic acid conversion to all its metabolites. The selectivity to EET could occur because US specifically targets stored EETs or selectively activates P450 epoxygenase.

After OGD, US increases P450 eicosanoids (all EETs and their DHET metabolites, in addition to 18-, and 20-HETEs), but does not increase lipoxygenase metabolites (5-, 11-, 12- and 15-HETEs). Again, the effect seems to be selective for P450 enzymes. Interestingly, unlike phosphorylated eNOS and adenosine, OGD had no effects on EETs formation. Application of US after OGD, however caused a marked increase in EETs production. Not surprisingly, the metabolites of EETs (DHETs) followed a similar course as EETs. It is interesting, however, that we only detect DHETs after OGD, but not at baseline. This suggests that the combination of US and OGD may activate the enzymes responsible for EETs conversion to DHETs (e.g., soluble epoxide hydrolase). EETs protection of the heart following ischemia has been extensively investigated.^32–35

Unlike EETs, however, none of the HETEs increased on US exposure under basal conditions. Under OGD conditions, however, levels of −18 and −20 HETE increased dramatically. Arachidonic acid is metabolized to 18-HETE by CYP enzymes.³² The cardiovascular effects of 18-HETE have not been well studied and those of 20-HETE are controversial.³⁶

Limitations of our study

Several limitations of the study are worth mentioning. First, we used a single US frequency-1.05 MHz. We found this frequency to cause effective sonothrombolysis in vitro when combined with microbubbles.¹² This frequency could also be used for imaging at 2.5 MHz (ultraharmonics), thus potentially allowing a single probe to be designed for therapy (both for sonothrombolysis as well as direct myocardial effect) and imaging. Therefore, it appeared to be an attractive choice. Whether other frequencies studied by us for sonothrombolysis (ranging from 40 KHz to 2 MHz)¹² would provide different results is not known.

Although our results are from a model of total coronary occlusion, as seen from the epicardial and border-zone results, US could be beneficial in instances of acute myocardial ischemia not associated with complete coronary occlusion, such as unstable angina, where maintenance of MBF above a critical level with US application could result in myocardial salvage. Similarly, US application may also have beneficial effects in chronic myocardial ischemia, resulting in improvement in regional and global function.

Our in vitro experiments clearly demonstrated release of a plethora of coronary vasodilators with US that could explain the increases in MBF in vivo although the in vitro results cannot be directly connected to the in vivo results. These mediators also have direct tissue salvaging effects. However, our first measurement of these metabolites was performed 15 min after US exposure. Whether these mediators were released earlier is not known. Additionally, the temporal sequence of the rise and fall of these mediators was not defined by our study.

The mechanism whereby ECs detect ultrasound and translate that into downstream cell-signaling is also not known. US can cause sonoporation of cells. If this were the mechanism of mediator release we would expect both vasodilators and vasoconstrictors to be released. Selective release of only vasodilators implies stimulation of specific pathways that are activated for cardioprotection, such as P450. Thus, significant additional basic science research needs to follow in order to better understand the beneficial US effects on tissue.

Conclusions

US at 1.05 MHz increases MBF in ischemic tissue in vivo. This effect is likely mediated by the release of different coronary vasodilators during US treatment. Primary mouse cardiac ECs when exposed to US at optimal PRAPA produce a series of compounds that not only increase MBF but by themselves are also cardioprotective. Potentially harmful compounds such as some of the HETEs are not produced. US reverses the reduction in eNOS phosphorylation and markedly increases EETs during OGD. Further studies are required to determine the upstream signaling events that lead to release of cardioprotective substances by US. Our results indicate that clinical studies should be designed whereby therapeutic ultrasound should not only be tested for sonothrombolysis in acute myocardial infarction and acute coronary syndrome but also for its direct myocardial protection effect.

Supplementary Material

1

Video 1: Myocardial replenishment end-systolic images from a control dog during coronary occlusion (Stage B) from where the time versus acoustic intensity plots shown in red in Panels D and E in Figure 2 were generated.

2

Video 2: Myocardial replenishment end-systolic images from a control dog after 45 min of no treatment during coronary occlusion (Stage D) from where the time versus acoustic intensity plots shown in blue in Panels D and E in Figure 2 were generated.

3

Video 3: Myocardial replenishment end-systolic images from a US treated dog during coronary occlusion before treatment (Stage B) from where the time versus acoustic intensity plots shown in red in Panels D and E in Figure 3 were generated.

4

Video 4: Myocardial replenishment end-systolic images from a dog 45 min after US treatment during coronary occlusion (Stage D) from where the time versus acoustic intensity plots shown in blue in Panels D and E in Figure 2 were generated.

Highlights.

• Ultrasound has direct effects on tissue that are cardioprotective

• Cardioprotection is offered, in part, by increased tissue blood flow induced by ultrasound

• A number of metabolites are released from endothelial cells that offer cardioprotection by increasing blood flow and by their direct tissue salvaging effects