INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

ISCHEMIC HEART DISEASE is responsible for nearly 500,000 deaths annually in the United States (1). Because the adult myocardium has no process to repair damaged myocytes or restore function to infarcted regions, a significant effort has been invested in therapies designed to augment the performance of remaining cardiac tissue. In addition to pharmacological support, current therapeutic options for patients with end-stage heart disease include cardiac transplantation, mechanical circulatory assist devices, and dynamic cardiomyoplasty. However, because each of these suffers limitations, investigators continue to explore methods to improve the performance of regions containing irreversibly damaged muscle.

In the currently evolving era of molecular cardiology, proposed methods for regenerating function in the damaged heart have included cellular cardiomyoplasty (transplantation of regenerative cells into the heart) and cardiac gene therapy (15, 17, 24). However, demonstrating the safety and efficacy of new therapies requires a reproducible model of injury in which the discrete effects of these technologies can be measured in vivo. This study is an attempt to create such a model.

In creating such a model, small animals offer several advantages over larger animals. Rabbits, in particular, provide several unique advantages over other animal models of myocardial damage. Unlike the mouse, rat, and canine heart, the rabbit myocardium expresses contractile protein isoforms similar to the human heart (14); and it lacks silent collateral vessel growth often present in the dog (22). Furthermore, rabbit myocardium has many properties in common with the human heart, including prominent postextrasystolic potentiation and a positive force-frequency relationship, whereas rodent myocardium does not (21). Yet, current methods for evaluating myocardial performance in small animals (including rabbits) or in isolated heart preparations are limited to gross functional measurements such as the first derivative of pressure (dP/dt) (13). Although improved function has been reported in rabbits (13) as well as in transgenic mice (15, 17, 24) or in rats (19, 20) using dP/dt as an index of systolic and diastolic performance, dP/dt is limited by an inability to detect small changes in the contractile state independent of changing hemodynamic conditions, which can be significant in small animals. Other investigators have recently described in vivo functional assessment of transgenic mice, but those assessments were limited to normal myocardium. Although those investigators did monitor function postinjury, those data were obtained in vitro (25). Further in vitro regional assessments of function obtained through sonomicrometry have been reported in mice (13), but these functional analyses suffer from the limitations associated with all in vitro assessments, including a variable degree of ischemia, baseline drift, and lack of correlation with function in the intact animal (2).

Assessing in vivo regional and global myocardial performance in a small animal after myocardial injury has been difficult because 1) the available implantable hardware was formerly too large relative to the heart, 2) small hearts are difficult to instrument and subsequently injure in vivo, or 3) small animals have a rapid heart rate not easily captured by ultrasound or other noninvasive techniques. Although M-mode echocardiography has been used to monitor myocardial function in the mouse, this technology is expensive and is technically limited by its extreme dependence on probe placement. Furthermore, echocardiography primarily yields global assessments of myocardial performance, which are not applicable to acute or relatively small regions of injury. In the few instances where regional assessments have been attempted, the measurements appear to be heart rate dependent. Left ventricular (LV) function has also been assessed through the use of conductance catheters, but because these studies are generally terminal, no serial measurements can be obtained in vivo.

Monitoring regional LV function in vivo by sonomicrometry and micromanometry is a well-established technique in larger animals (7). For preload recruitable stroke work (PRSW) comparison, the linear relationship between stroke work (SW) and end-diastolic segment length (also known as the Frank-Starling relationship) has been employed in a variety of animal and human studies to assess cardiac contractility in a load-insensitive fashion. Recently, we adapted this technology acutely to rabbits (23). The present study extends those findings to an in vivo setting of myocardial injury.

Creating an effective model in which novel therapeutic regimens can be compared requires a myocardial lesion of reproducible size, depth, and location in which regional in vivo functional assessments are possible. Because of biological variability and differences in coronary vascularization, the size, depth, and location of ischemia-induced myocardial lesions are unpredictable. For example, in a rabbit coronary ligation model of ischemic injury, despite the presence of an area at risk comprising up to 41% of the LV, infarct size comprises only 68.7 ± 20% of the area at risk, and myocardial performance is statistically unchanged (4). This extent of variability makes functional assessment of the damaged region and precise delivery of cells or genes to the damaged area unlikely. Thus to create a reproducibly sized and positioned lesion, with little or no hibernating myocardium, we have adapted a method of cryoinjury previously described in canines and used clinically to treat both supraventricular and ventricular tachyarrhythmias (5, 6, 9, 11). The current study was designed to determine whether the cryoinjured rabbit left ventricle provides 1) a histological model of myocardial injury similar to postischemic lesions; and 2) a model in which reproducible changes in the slope and intercept of the linear PRSW relationship between SW and end-diastolic length can be detected in vivo.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Experimental preparation. All animals were studied under guidelines published in the NIH Guide for the Care and Use of Laboratory Animals (DHHS publication No. NIH 85-23, Revised 1985) and approved by the Institutional Animal Care and Use Committee at Duke University Medical Center. Research subjects were evaluated to elucidate the effects of local cryoinjury on regional myocardial performance.

Using techniques developed in this laboratory, we anesthetized 11 New Zealand White rabbits (4-5 kg) with ketamine (60 mg/kg im), xylazine (6 mg/kg im), fentanyl citrate (10 µg/kg iv), and succinylcholine (1 mg/kg iv), and we then endotracheally intubated and mechanically ventilated (Bird Products, Palm Springs, CA) the rabbits. A sterile left thoracotomy through the fifth intercostal space was performed, and the pericardium was divided to expose the heart. Ultrasonic dimension transducers (Physiologic Monitoring Systems, Duke University, Durham, NC) were positioned along the minor axis of the heart between the left circumflex coronary artery and the posterior interventricular groove, ~1.5 cm apart (Fig. 1A). The hardware was exited through a separate rib space, tunneled through the chest wall, and exteriorized through a dorsal Silastic skin button. The thoracotomy was repaired in layers.

View larger version (40K): [in this window] [in a new window]   View larger version (159K): [in this window] [in a new window]   View larger version (137K): [in this window] [in a new window]   Fig. 1.   Surgical preparation used to measure myocardial segment length and left ventricular (LV) pressure within cryoinjured myocardium along the posterolateral border of the left ventricle. A: miniature cylindrical ultrasonic transducers measure minor axis midwall segmental length within region of cryoinjury (shaded area). During data collection, a Fogarty embolectomy catheter was placed in inferior vena cava, and a micromanometer was placed into the left ventricle via a separate carotid cutdown. B: application of 1-cm cryoprobe (arrow) to epicardial surface of rabbit heart. C: frozen left ventricle immediately after removal of cryoprobe from epicardial surface.

Each animal was allowed to recover 5-7 days before baseline myocardial performance data were obtained. After the control study, each animal was allowed to recover for an additional 5-7 days. Each was then reanesthetized and mechanically ventilated as cited above, and the initial thoracotomy incision was reopened. A 1-cm diameter cryoprobe (Frigitronics, Shelton, CT) was cooled to 70°C by continuously circulating nitrous oxide and placed on the epicardial surface between the ultrasonic dimension transducers for ~1-3 min at the location indicated by the shaded area in Fig. 1A. During the creation of the cryoinjury, electrocardiographic tracings were continuously monitored for detection of arrhythmias and for evidence of S-T segment elevation. Persistent S-T elevation (beyond thawing) was used as an index of transmural damage (10). After cryoablation, the frozen area was allowed to completely thaw for 3-5 min. The thoracotomy was again repaired in layers. The animals were allowed to recover for 2 or 6 wk before being returned to the laboratory for final assessment of regional myocardial performance in vivo.

Data acquisition. Animals were mildly sedated with ketamine (30 mg/kg) and acepromazine (1 mg/kg). During the initial study (control before cryoinjury), a right neck cutdown was performed to expose the right external jugular vein and the right internal carotid artery. A 2.5-Fr micromanometer (PC500, Millar Instruments, Houston, TX) was advanced retrogradely into the left ventricle through the carotid artery. To induce vena caval occlusion (VCO) on inflation, a 4-Fr Fogarty embolectomy catheter (model 120804F, Baxter Healthcare, Irvine, CA) was passed into the proximal inferior vena cava (IVC) via the jugular vein. The appropriate position of both catheters was confirmed fluoroscopically.

The previously implanted ultrasonic transducer pairs were connected to a sonomicrometer (James W. Davis Consultant, Durham, NC) with a sampling rate of 1 kHz and a resolution of 0.05 mm. Digitized data of regional minor axis dimensions and LV intracavitary pressure were recorded under steady-state conditions and during VCO.

During the terminal study (postcryoinjury), a left neck cutdown was performed to access the left internal carotid artery, and a right groin cutdown was done to access the right femoral vein. A micromanometer pressure transducer was passed via the carotid artery into the LV cavity, and a Fogarty embolectomy catheter was passed into the proximal IVC via the femoral vein. Digitized data were collected in a fashion similar to the control studies.

Histological preparation. During the completion of the final study, each animal was anticoagulated with intravenous heparin (200 U/kg) and subjected to euthanasia as outlined by the Duke University Medical Center animal care guidelines. The heart was excised, rinsed in saline, and placed in a 30% sucrose solution at 4°C for 24 h. Each heart was then divided into 5-mm cross-sectional slices from apex to base and frozen in liquid nitrogen. Thin (8 mm) sections from each of the frozen slices were stained with hematoxylin and eosin (H+E) for visualization of the muscle and with Masson-Trichrome (MT) to delineate fibrous tissue. Each specimen was closely examined grossly and microscopically to verify that all animals included in the comparison of functional data demonstrated transmural injury.

Data analysis. Analog pressure and dimension data were processed through a 50-Hz low-pass filter, digitized in real time at 200 Hz (ADAC model 1012, Woburn, MA), and collected on a PC-based system. Data analysis was performed on a microprocessor (VAX Station II/GPX, Digital Equipment, Maynard, MA) using interactive software developed in this laboratory. The cardiac cycle was defined using the LV dP/dt as described previously (8). Diastole was defined as beginning 15 ms after the first zero crossing of dP/dt, following peak negative dP/dt and ending 15 ms before the systolic upstroke of LV intracavitary pressure. Beginning ejection and end ejection were placed at peak positive and peak negative dP/dt, respectively. Mean LV pressure was defined as the mean LV intracavitary pressure during ejection.

Percent systolic shortening (%SS) was calculated as

(1)

where L_ed and L_es were end-diastolic and end-systolic segment lengths, respectively. An analog of regional myocardial stroke work (SW_R) was calculated as the integral of LV intracavitary pressure (P) and segment length (L) over each cardiac cycle

(2)

For each VCO, the relationship between SW_R and L_ed was fitted to the linear regression

(3)

where M_w is the slope, and L_w is the x-intercept for the relationship.

Regional ventricular performance in the context of regional ischemia is more accurately assessed as the area beneath each SW_R versus L_ed relationship during VCO (16). This area, the preload recruitable work area (PRWA), is determined by

(4)

where L_{w max} was the maximal L_w recorded throughout the entire study for each animal.

Statistical analysis. All Frank-Starling relationship indexes from the above equations were derived using least-squares regression techniques on a personal computer. Student's paired t-tests were used to compare indexes before and after cryoinjury. All data are presented as means ± SE, unless otherwise noted. Statistical significance was considered to occur at P < 0.05.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Myocardial injury. Cryoprobe (Fig. 1B, arrow) application for up to 3 min to the posterior LV surface of the rabbit heart was associated with S-T segment wave elevation in the electrocardiogram, significant bradycardia, and extended freezing (Fig. 1C) of the LV. Left atrial congestion occurred acutely but diminished with thawing and the return to a normal heart rate. On thawing, a discrete border of injury on the epicardial surface could be seen. Within the damaged region, distension and thrombosis of subepicardial arterial branches and veins could be visualized acutely. Although no obvious acute differences could be seen after a 1- or 3-min lesion was thawed, the extent of chronic injury correlated with the duration of cryoprobe application. One to three percent of the rabbits subjected to 1-min cryoinjury died acutely as did ~30% of the animals subjected to 3 min of cryoinjury, despite the use of antiarrhythmic and inotropic agents. In these animals that died acutely, the extent of myocardial injury was often much >30% and appeared to correlate with distension and thrombosis of a major diagonal branch of the left anterior descending coronary artery. Subepicardial hemorrhagic necrosis and thrombosis were evident. In two animals that survived the 3-min cryolesion but died within 2 wk (before the completion of the studies), we found evidence of organized left atrial thromboses.

Although electrocardiogram changes were obvious during the bradycardia associated with both 1 and 3 min of cryoinjury, only S-T segment elevation that persisted beyond thawing was associated with transmural defects at 2-6 wk. However, by 2 wk after cryoinjury, assessment of S-T segment elevation was unreliable in anticipating transmural injury primarily due to the effects of respiratory variation and the normal rapid heart rate in unanesthetized rabbits.

Histology and ultrastructure. At 2 wk postinjury, gross examination of hearts that were cryoinjured for a duration of 1 min demonstrated a hemorrhagic area ~1.6 cm in diameter on the epicardial surface of the left ventricle, which represented ~30% of LV surface area. As depicted grossly in a short-axis view in Fig. 2A, at 2 wk this 1-min lesion extended into the myocardium resulting in hemorrhagic necrosis and a fibrous yellow nontransmural scar very similar to that seen after an acute myocardial infarction in patients. However, in a reciprocal relation to the typical human postischemic lesion, the cryoinjury arose from the epicardial surface and contained intact subendocardial muscle tissue. H+E and MT staining of the injured myocardium demonstrated necrotic myocytes at the periphery of the scar, fibrous tissue throughout the scar, and subepicardial hemorrhagic necrosis.

View larger version (87K): [in this window] [in a new window]   Fig. 2.   Representative micrographs of a typical cryoinjured heart showing gross pathology at 2 wk after a 1-min cryoinjury (A) demonstrating a nontransmural lesion. B: hematoxylin and eosin (H+E) staining (×4 magnification) of a transmural lesion occurring 2 wk after a 3-min cryoinjury, where M is intact myotubes at periphery of lesion and sc is fibrotic scar. Boxed region at interface of scar and lesion is depicted in higher magnification in D. C: Masson-Trichrome staining of a transmural lesion shown in B. D: H+E staining (×10 magnification) of boxed region from B depicting transmural lesion interface with normal myocardium.

Application of the cryoprobe to the epicardial surface for 3 min resulted in a similarly sized but transmural defect as indicated in Fig. 2, B-D. H+E staining of the 3-min lesion at 2 wk (Fig. 2, B and D) demonstrated a complete loss of myocytes within the region of injury despite the presence of thin connective tissue strands, scattered cellular debris, and lymphocytic infiltration. MT staining (Fig. 2C) revealed enormous elaboration of connective tissue, lack of intact muscle, and establishment of a scar (labeled as sc in Fig. 2) coincident with the accompanying pale region within the H+E-stained sections. In contrast to a 1-min cryoablation, no intact subendocardium was seen with 3-min cryoablation. Furthermore, both H+E and MT staining delineated a discrete border zone (see box in Fig. 2B) where intact myocardium gave way to scar (Fig. 2D); however, no hemorrhagic necrosis was obvious. Light micrographs of the injured region at a higher magnification demonstrated an inflammatory infiltrate and collagen deposition but a lack of cardiocytes within the lesion. Blood flow data obtained after technicium-99 infusion by single photon emission computerized tomograph imaging (12) 24 h to 2 wk after cryoinjury (data not shown) demonstrated that flow to the damaged area was sharply interrupted. Ultrastructural analysis (Fig. 3, A-F) at 2-6 wk indicated both intact (Fig. 3C) and thrombosed coronary vessels. Transmission electron microscopy of a normal zone (Fig. 3A) distinct from the scar demonstrated ordered sarcomeres with intact z lines (labeled as z) and mitochondria (labeled as m) in close apposition to a patent vessel containing red blood cells. Transmission electron microscopy of a typical scar at 2 wk (Fig. 3, B-F) revealed a discrete demarcation of injury, with abrupt disruption of myofiber integrity at the border lesion, such that regions of normal cardiocytes (labeled as M in Fig. 3B) could be visualized directly adjacent to areas of collagen (labeled as C) and chronic inflammatory cells (labeled as *) within the scar. As demonstrated in Fig. 3C a patent vessel containing red blood cells could also be identified adjacent to normal sarcomeres (M) at the edge of the lesion. However, surrounding the vessel were numerous areas of packed threadlike collagen fibers (labeled as Col) consistent with developing scar. In addition, at the border of the cryoinjury, severely damaged, dying myocytes could be detected (M in Fig. 3D) adjacent to fibroblasts (fb) and collagen (Col). At 6 wk postinjury, gross examination of the region demonstrated a thinned area of pale yellow fibrotic tissue ~1.6 cm in diameter, covering ~30% of LV surface area and transmural in depth. Ultrastructurally, densely packed collagen similar to that seen in the cross section in Fig. 3E, few fibroblastic cells and patent microvessels similar to those seen in Fig. 3F were obvious at 6 wk.

View larger version (179K): [in this window] [in a new window]   Fig. 3.   Transmission electron micrographs depicting normal myocardium at a site distant to cryolesion containing both intact sarcomeres, including z lines (z), mitochondria (m), and a patent vessel composed of an endothelial cell nucleus (E Nu) and containing red blood cells (RBC) (A). B: myocardium at myocyte/scar interface depicting intact myocytes (M) with z lines as well as damaged myocytes lacking z lines. In addition, a mast cell nucleus (nu) and granules (*) can be seen adjacent to collagen (c). C: in a different region of the scar/myocyte interface, intact myocytes (M) can be seen adjacent to a patent vessel containing RBC and to fibrosis evidenced by collagen deposition (Col). D: severely damaged myofibers (M) are seen in close apposition to dense regions of collagen (Col) and fibroblasts (fb). E: densely packed regions of collagen can be seen in cross section of scar interior. No myocytes are present within this region. F: damaged myocytes (M) on periphery of injured region surrounded by microvessels (E Nu) and collagen (C).

Regional function. Six animals underwent regional functional assessment before cryoinjury and 2 wk after cryoinjury. Three animals were further analyzed at 6 wk. Representative dynamic raw data for both the control and the 2-wk cryoinjured states during a typical VCO are shown in Fig. 4. No obvious difference in LV pressure or heart rate can be seen. Examination of the hemodynamic data for control versus cryoinjury in each animal showing no change in heart rate and mean ejection are depicted in Table 1 and Fig. 5A. Similarly, LV end-diastolic pressure, albeit more variable, was unchanged after cryoinjury (Fig. 5A; Table 1). In contrast, myocardial work loops constructed between LV pressure and segment length differed before and after cryoablation in all animals studied. Typical loops through the entire cardiac cycle in one animal during VCO for the control and cryoinjured states are shown in Fig. 5B. After cryoinjury, narrowed work loops were observed indicating less work per cycle primarily due to decreased systolic shortening. Furthermore, transmural cryoablation was associated with a 55 to 77% decrease in SW resulting in a decrease in the mean from 14.6 ± 2.7 to 4.9 ± 0.9 kerg postcryoinjury (Table 1; Fig. 5A). Similarly, the percent systolic shortening was decreased 80-92% resulting in a mean decrease from 18 ± 2.3 to 3 ± 0.04% (Fig. 5A; Table 1). The three animals examined also exhibited a similar or greater significant (P < 0.05) decrease in systolic shortening within the cryoinjured region at 6 wk after insult (data not shown).

View larger version (59K): [in this window] [in a new window]   Fig. 4.   Representative dynamic digitized raw data acquired during in vivo physiological study, showing LV intraventricular pressure, segment length, and dP/dt from one animal before and after cryoinjury.

View larger version (20K): [in this window] [in a new window]   View larger version (20K): [in this window] [in a new window]   Fig. 5.   A: mean hemodynamic data (±SE; n = 6) during steady-state conditions, including heart rate (HR), mean ejection pressure (MEP), and LV end-diastolic pressure (LVEDP), do not change after cryoinjury. However, a significant decrement in stroke work (SW) and percent systolic shortening (%SS) is seen in cryoinjured state. B: dynamic digitized LV intraventricular pressure-segment length work loops from one study subject. Segmental SW for each beat is calculated as the area within each work loop and is regressed against end-diastolic segment length to obtain the regional preload recruitable SW (PRSW) relationship.

The linear PRSW relationship between regional SW and end-diastolic segment length is used to quantify systolic performance. A decrease in the slope of this relationship reflects decreased contractility, whereas an increase in slope, as is seen after inotropic intervention, reflects increased contractility. Similarly, changes in the x-intercept of the PRSW relationship can be used as an index of ventricular dilation. Construction of the PRSW relationships obtained before and after cryoinjury (Fig. 6A) yielded highly linear relationships in all six animals studied with r² values >0.95. Cryoinjury was accompanied by a 22 to 44% decrease in M_w of this Frank-Starling relationship with a mean decrease from 9.0 ± 0.7 to 3.1 ± 0.5 kerg/mm (Table 1; Fig. 6B) consistent with reduced contractility (7). Although a trend rightward was seen in the x-intercept (L_w) of the PRSW relationship, no significant change in the mean value (7.0 ± 1.1 vs. 7.6 ± 1.0) was observed. Mean data (n = 6) demonstrating a significant decrement from 23 ± 4.6 to 6.2 ± 1.0 kerg · mm² (P = 0.006) in the area under the PRSW curve (PRWA) at 2 wk after cryoinjury is depicted in Table 1 and Fig. 6B. This reflects the decreased regional work done by the injured myocardium. Three animals studied 6 wk after cryoinjury demonstrated a continued or worsened decrement in M_w with no recovery of the trend toward a rightward shift of the L_w of the PRSW relationship.

View larger version (13K): [in this window] [in a new window]   View larger version (9K): [in this window] [in a new window]   Fig. 6.   A: PRSW relationship between segment SW and end-diastolic segment length before and after cryoinjury from one rabbit demonstrating the type of linear data used to generate means ± SE values (n = 6) presented in B. Characteristic, significant decreased slope, and trend toward a rightward shift of linear relationship are evident for cryoinjury-induced regression when compared with control. B: PRSW indexes (means ± SE; n = 6) demonstrating significant decrement in slope (Mw) and lack of significant change in x-intercept (Lw) 2 wk after cryoinjury.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Developing a reproducible clinically relevant, yet cost-efficient model, in which discrete changes in myocardial performance can be measured in vivo is essential for evaluating new treatments for myocardial infarction and congestive heart failure. Both canine and porcine models of occlusive myocardial injury have been described as well as rodent and rabbit models (14, 28). However, these models suffer from a major physiological limitation. It is extremely difficult to obtain an accurate assessment of regional myocardial performance within the injured zone in the aforementioned models because vessel occlusion does not create a reproducible, distinct area of injury. The placement and extension of the injury is extremely dependent on the anatomy of the given animal as well as the presence of collateral vessels. Also, in these models, areas of ischemic or stunned myocardium bordering the nonviable injury may contribute variably to the overall myocardial performance in the region (27). Thus when functional data obtained during the initial study are compared with later data in an ischemic model, biological variability compromises statistical assessment of regional function. Although this variability in infarct size and location does not alter histological outcome, it could have a dramatic impact on regional physiology and performance affecting evaluation of the intervention.

Because the ability to reliably control the size and position of an injured area is a prerequisite for evaluating regional myocardial performance chronically and because cryoinjury appears to meet these criteria, we have developed a cryoinjury model in rabbits. The effects of cryoablation on viable myocardium have been well documented (5, 9, 16, 26). By applying a cryoprobe to the epicardial surface of the rabbit heart for at least 3 min (16), we have created a transmural lesion lacking intact cardiocytes, which develops into a fibrotic scar. By histological analysis, this transmural lesion is similar to that seen after an acute myocardial infarction except that it initially arises from the epicardial rather than endocardial surface. The temporal development of the fibrotic transmural scar is more rapid in rabbits (2 wk) than in humans (7 wk), but it follows a similar pattern ranging from initial hemorrhagic necrosis and myocyte dropout to neutrophil and fibroblast infiltration and subsequent fibrovascular response leading to the development of fibrotic scar. Ultrastructural analysis of the lesion demonstrated a distinct demarcation between myocytes and mature scar consistent with the visual observation of white fibrosis adjacent to intact myocardium at the epicardial surface. The ability to choose the site of injury and reproducibly obtain a 1.6-cm transmural lesion similar histologically to that seen after a myocardial infarction in humans is an attractive feature of this model.

In our experience, shorter periods of cryoinjury (1 min) yield a lesion arising from the epicardial surface but containing cardiocytes in the subendocardial region. This difference depending on the time of initial freezing may account for the disparity of our data with previous reports, indicating that cryoinjury does not yield a myocyte-free transmural lesion (20). Our histological data indicate that if persistent S-T elevation and decreased heart rate after thawing are obtained at the time of creation of the lesion (which appears to indicate transmural freezing), this transmural defect is maintained. Of course, infliction of this degree of injury presumes a high mortality rate among study animals. Approximately 30% of the animals died acutely (within minutes of cryoinjury), which is a number that closely resembles the mortality figures for acute myocardial infarction. In the group of animals that died within minutes, the extent of myocardial injury was often much >30% (as indicated by darkening of the myocardial surface and histologically by hemorrhage) and correlated with cryolesion and thrombosis of a major branch of the left anterior descending coronary artery. Histological analysis of those hearts revealed hemorrhagic necrosis and thrombosed vessels within the lesion as well as a lesion extending beyond the normal radius of 0.8 cm.

The early pronounced histological changes evident after cryolesion appear to provide a model of myocyte clearing and inflammatory response similar to those seen after an acute myocardial infarction in humans. Subsequent scar formation with minimal ventricular dilation appears to develop within 2 wk and to be maintained for the duration of the experiments. This physiological progression suggests that the cryoinjured rabbit left ventricle may be provided as a small animal model in which the effects of specific therapeutic interventions on scar formation, ventricular remodeling, and ventricular dilation can be evaluated.

Even though smaller animals are more commonly used for cell transplantation and genetic manipulation studies, few investigators have obtained chronic regional myocardial function measurements in these models (17, 27). Most techniques used to quantify LV function in small animals rely on Langendorff-isolated heart preparations or invasive hemodynamic monitoring. Unfortunately, each of these techniques suffers from several limitations that can significantly affect myocardial performance. For example, isolated heart preparations are susceptible to variable degrees of ischemia, they obviously eliminate the possibility of serial measurements of LV performance, and they render in vivo function impossible to assess. Furthermore, invasive hemodynamic parameters obtained such as LV dP/dt_max, cardiac index, and LV mean ejection pressure that can be obtained chronically in vivo, are all heart rate dependent and load sensitive, so that they are dramatically altered with the state of autonomic stimulation (14, 18, 27). More recently, conductance catheters have been used to monitor regional myocardial performance in vivo, but these techniques are usually terminal, precluding the possibility of serial assessments of function.

The noninvasive measurements of myocardial performance have relied on techniques such as M-mode or two-dimensional echocardiography. Unfortunately, M-mode echocardiography on small animals requires a 12-MHz probe, which is cost prohibitive for many laboratories and is very sensitive to probe placement making serial assessments subject to error. Two-dimensional echocardiography, on the other hand, yields primarily global assessments of myocardial function. The model developed in this paper allows serial regional assessment of both normal and injured myocardium.

Until recently, no load- or heart rate-independent technique existed for monitoring regional myocardial performance. Glower et al. (8) first introduced the concept of PRSW as a method whereby assessments of regional myocardial contractility and geometry are obtained using the linear regression of regional SW versus regional end-diastolic segment length. This model has repeatedly been shown to be superior to other models when evaluating regional myocardial function due to its highly linear nature, independence of preload and hemodynamic alterations, and insensitivity to changes in afterload.

Although developed in large animal models, PRSW was applied by Silvestry et al. (23) to normal rabbit myocardium in a closed-chest, acute preparation. Our data extend this model of PRSW to the chronic model of cryoinjured rabbit myocardium. Although several authors have described the use of cryoinjury to simulate myocardial infarction or injury (3, 20), to our knowledge no one has described a load-independent regional assessment of myocardial performance within an area of cryoinjury.

The current study demonstrates that cryoinjury has no demonstrable effect on hemodynamics, including heart rate, LV end-diastolic pressure, and mean ejection pressure. Yet, by 2 wk postcryoinjury, regional SW and systolic shortening are significantly decreased, and they remain depressed or worsen over a subsequent 4-wk period. This decrease in SW and in systolic shortening within the region is presumably due to the lack of contractile cells (cardiocytes) within the region. Importantly, regional systolic performance does not fall to zero even within the transmural scar, presumably because the crystals are embedded within the myocardial wall, which is subjected to passive movement secondary to filling and contraction of the remaining normal left ventricle. This results in minimal movement of the crystals. Yet, the persistent decrement in regional function, including the decreased slope of the PRSW relationship to ~30% of control seen after a transmural cryolesion, provides a baseline against which interventions designed to improve function within the lesion (such as myoblast transfer and angiogenic gene delivery) can be measured locally in vivo.

Thus we have designed a myocardial cryoinjury model in rabbits that yields a reproducible transmural lesion, and when combined with available techniques for chronically assessing regional performance, the model ultimately allows for precise measurements of regional function before and after the imposed lesion. This model also provides a basal assessment of performance against which the effects of cellular cardiomyoplasty and other therapeutic interventions may eventually be evaluated.