INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ISOLATED CARDIOMYOCYTE MODELS of simulated ischemia (SIsc) have lent much insight into the pathophysiology of myocardial ischemic injury. Free from neurohumoral influences and potentially confounding contributions from nonmyocyte cell types, these models have contributed to our understanding of both the ionic dyshomeostasis and the altered signal transduction accompanying ischemia (1, 7). In these studies (1, 7, 14), manipulations to protect the ischemic heart have largely focused on reducing necrotic cell death, which has typically been assessed by using indicators of sarcolemmal membrane permeability, such as trypan blue and enzyme release.

Results from several studies (6, 8, 21), however, indicate that reperfusion after myocardial ischemia can result in exacerbation of injury and apoptosis (programmed cell death). In fact, several studies (6, 8, 13, 21, 35) suggest that apoptosis is a reperfusion-related phenomenon. Because of reports (25, 34) indicating evidence of apoptosis in the peri-infarct and remote nonischemic zones of infarcted human myocardium, the mechanisms governing this death program have come under intense investigation. The majority of studies (5, 8, 21) examining myocardial reperfusion-related apoptosis have used techniques such as DNA fragmentation and evidence of DNA laddering. However, these techniques detect apoptosis at a very late stage, and cardiomyocytes with fragmented DNA have been reported (10, 17) to exhibit extensive membrane blebbing, mitochondrial and nuclear margination, and a condensed or rounded morphology. In contrast, data from numerous cell types have indicated that one of the earliest events in the apoptotic cascade is externalization of the phosphoaminolipid phosphatidylserine from the inner face of the plasma membrane to the outer cell surface (33). Fluorophore-labeled annexin V (a protein that exhibits nanomolar affinity for phosphatidylserine) binding to externalized phosphatidylserine has been extensively employed as a reliable marker of apoptosis (3, 13, 27).

If apoptosis is a physiologically relevant aspect of reperfusion injury and is truly unique from necrosis, as suggested by several noncardiac studies (20), then merely measuring cell death during ischemia based on membrane permeability may underestimate the extent of myocyte death during ischemia-reperfusion. In the present study, we examined the effects of SIsc and simulated reperfusion (SRP) in isolated rat ventricular myocytes by using annexin V-propidium iodide, markers of apoptosis and necrosis, respectively, morphology, mitochondrial free calcium concentration ([Ca²⁺]_m), and twitch amplitude.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Isolation of ventricular myocytes. Ventricular myocytes were enzymatically dissociated from male Sprague-Dawley rats (350-400 g) as previously described (24). Cardiomyocytes were suspended in standard HEPES buffer composed of (in mM) 130 NaCl, 4.7 KCl, 1.2 MgSO₄, 1.2 KH₂PO₄, 10 HEPES, and 11 glucose (containing 200 µM CaCl₂ at a pH of 7.4). This procedure routinely yielded >70% rod-shaped myocytes. For the experimental protocols, the same buffer with 1 mM CaCl₂ was used. All data were collected within 6 h of myocyte isolation.

Myocyte imaging and analysis system. To visualize and quantify annexin V-propidium iodide staining, measurement of [Ca²⁺]_m, and cardiomyocyte length, width, and twitch amplitude, we used an inverted microscope (model IX70, Olympus America; Melville, NY) equipped with a 75-W xenon arc lamp. Excitation and emission wavelengths were selected with the use of a variety of filter combinations. For functional measurements, cardiomyocytes were placed in a 0.3-ml recording chamber (Warner Instruments; Warrendale, PA), the floor of which was a 22 mm × 22 mm glass coverslip coated with laminin to enhance cell adherence. Both epifluorescence and the cardiomyocyte image (×200 magnification) were collected by the use of a charge-coupled device camera that was attached to a side port of the microscope. Image analysis was done with the use of a Pentium-based computer and custom-made software (Coyote Bay; Manchester, NH).

Annexin V-propidium iodide staining in cardiomyocytes. Myocytes were stained with fluorescein isothiocyanate (FITC), annexin V, and propidium iodide, as per the manufacturer's directions (Vybrant Apoptosis Assay Kit 2, Molecular Probes) and visualized under a fluorescence microscope. Cells were counted regardless of morphology. Cells not binding FITC-annexin V and excluding propidium iodide were classified as annexin V-negative (ANve) (26). Myocytes that bound FITC-annexin V [excitation wavelength (_ex) = 488 nm and emission wavelength (_em) = 520 nm] but excluded propidium iodide (_ex = 540 nm and _em = 630 nm) were termed annexin V-positive (AN+ve) and myocytes permeant to propidium iodide (regardless of whether or not they bound FITC-annexin V) were deemed necrotic (31).

Assay for caspase-3-like activity. The protease caspase 3 is a cysteine protease that cleaves substrates with a Asp-Glu-Val-Asp (DEVD) motif at the terminal aspartate (30). Myocytes were incubated with the FITC-conjugated cell permeant aminopeptide, GDEVDGI, and propidium iodide as per manufacturer's directions (OncoImmunin; Gaithersburg, MD). Uncleaved GDEVDGI is not fluorescent, but after cleavage by caspase-3-like proteases, the cleaved fractions fluoresce (_ex = 488 nm and _em = 520 nm).

Measurement of [Ca²⁺]_m: use of rhod 2 and Mn²⁺ quenching. Myocytes were loaded with rhod 2-acetoxymethyl ester (10 µM; Molecular Probes; Eugene, OR) for 20 min, followed by washout of the dye. Aliquots of the rhod-2-labeled cell suspension were placed in the temperature-controlled recording chamber and suffused with HEPES buffer containing MnCl₂ (100 µM, 20 min). Because some of the deesterified rhod 2 is distributed in the cytosolic compartment, it is necessary to eliminate the cytosolic component of the Ca²⁺ signal (23). Mn²⁺, which binds rhod 2 with greater affinity than Ca²⁺, quenches the cytosolic Ca²⁺ signal but does not enter mitochondrial compartment (at least over 120 min) (23). After 20-min exposure to MnCl₂, suffusion with standard HEPES buffer was resumed and fluorescence measurements (_ex = 540 nm and _em = 580 nm) were made. Conversion of the fluorescence signal to [Ca²⁺] was done by the method of Delcamp et al. (4). [Ca²⁺]_m (expressed in nM) was calculated using the equation
where R is the measured fluorescence intensity (corrected for background fluorescence), R_min is the fluorescence at zero calcium, and R_max is the fluorescence under saturating [Ca²⁺] (2.5 mM). R_min was derived by exposing myocytes to 25 µM digitonin in a Ca²⁺-free HEPES buffer containing 10 mM EGTA, and R_max was derived by exposing the cardiomyocytes to HEPES buffer containing 2.5 mM CaCl₂ (no EGTA). A dissociation constant (K_d) of 570 nM was used (Molecular Probes).

Measurement of cell length, cell width, and twitch amplitude. Maximal cell length parallel to the longitudinal axis (cardiomyocyte length) and maximal cell width perpendicular to the longitudinal axis (cardiomyocyte width) was measured using a micrometer grid. Amplitude of cell shortening (in response to electrical stimulation at 0.5 Hz, 37°C), expressed as a percentage of diastolic cell length, was computed as described previously by our laboratory (24).

SIsc and SRP protocol. After a 60-min postisolation equilibration period, aliquots of cell suspensions were placed in 1.5-ml Eppendorff tubes and pelleted at 100 g for 10 s, resulting in a pellet occupying a volume of ~100 µl. Excess supernatant was removed, leaving a thin fluid layer above the pellet. Mineral oil (200 µl) was then layered on top of the pellet to limit gaseous diffusion, and the Eppendorff tubes were placed in a nonshaking water bath at 37°C to simulate normothermic ischemia. At the end of ischemia, the mineral oil was aspirated and the cardiomyocytes resuspended either in phosphate-buffered saline or in standard HEPES buffer (gassed with 100% O₂, 37°C).

Data analysis. For quantification of ANve, AN+ve, and necrotic cells, ~300 cardiomyocytes were counted per heart per group and expressed as a percentage of all cardiomyocytes counted. Data are expressed as means ± SE. Within-group differences were determined by one-way analysis of variance (ANOVA), followed by Newman-Keuls post hoc analysis. A P value <0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Under baseline (nonischemic) conditions, the percentage of ANve, AN+ve, and necrotic cardiomyocytes, as determined by FITC-annexin V-propidium iodide staining, was 77.0 ± 1.5, 0.8 ± 0.2, and 22.0 ± 1.7, respectively (n = 12 hearts). Myocytes suspended under normoxic conditions retained baseline viability over the 4- to 5-h experimental period. Twenty or sixty minutes of SIsc alone did not result in an increase in AN + ve cells (~0.9% in each group, n = 5 hearts/group). Twenty minutes of SIsc did not result in additional necrosis from baseline (24.0 ± 2.1%), but sixty minutes of SIsc was associated with significant cardiomyocyte necrosis (47.7 ± 3.0%, P < 0.05 vs. baseline).

Figure 1, A-C, show the effects of 20 min of SIsc and 20 min of SRP on cardiomyocytes. As seen in Fig. 1A, under normal lighting conditions, there are two rod-shaped myocytes and one rounded myocyte. Only the rounded myocyte was permeable to propidium iodide indicating necrosis (Fig. 1B). As shown in Fig. 1C, of the two apparently normal, rod-shaped cardiomyocytes, one fluoresced green due to FITC-annexin V binding and was therefore classified as AN+ve.

View larger version (56K): [in this window] [in a new window]   Fig. 1.   After 20 min of simulated ischemia (SIsc) and 20 min of simulated reperfusion (SRP) cardiomyocytes were stained with fluorescein isothiocyanate (FITC)-annexin (AN) V/propidium iodide and viewed under several lighting conditions. A: bright field, showing two rod-shaped and one rounded myocyte. B: propidium iodide fluorescence [excitation wavelength (ex) = 540 nm, emission wavelength (em) = 630 nm], where the nucleii of rounded myocyte is visible, indicating that this cell is necrotic. C: FITC-annexin V fluorescence (ex = 490 nm, em = 520 nm), rod-shaped cardiomyocyte, excluding propidium iodide (in B) fluoresces green and is classified as annexin V-positive (AN+ve).

Effects of 20-min SIsc, followed by 20- or 60-min SRP on the percentage of cardiomyocytes staining positive for annexin V are shown in Fig. 2. Compared with baseline, 20 min SRP was associated with an ~12-fold increase in AN+ve cells. Extending reperfusion time from 20 to 60 min did not result in a significant increase in this value. The percentages of necrotic cells at 20 and 60 min SRP were 28.4 ± 2.2 and 34.0 ± 3.0, respectively (both values, P < 0.05 vs. baseline). When myocytes (n = 3 hearts) were reperfused (20 min) in the absence of extracellular Ca²⁺, the percentage of AN+ve cells was 1.5 ± 0.3, a value significantly lower compared with myocytes reperfused in the presence of extracellular Ca²⁺ (7.5 ± 1.5%). In myocytes (n = 3 hearts) subjected to 20-min SIsc and reperfused (20 min) with buffer saturated with 100% N₂, 4.1 ± 0.5% of cells stained AN+ve (P < 0.05 vs. control HEPES reperfusion). Finally, exposure of normoxic cardiomyocytes to H₂O₂ (n = 3 hearts) resulted in 14.6 ± 2.1% AN+ve cells.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Effects of 20-min SIsc (ISC), followed by 20- or 60-min SRP (RP) on annexin V staining in ventricular myocytes. *P < 0.05 vs. baseline.

Although AN+ve cardiomyocytes exhibited rod-shaped morphology (Fig. 1C), closer examination revealed that these cells exhibited decreased cell width. Figure 3A shows cell length and width in baseline (nonischemic), reperfused ANve and AN+ve cardiomyocytes (n = 35 cells / group). Whereas cell length was not different between groups, cell width in the AN+ve cells was about one-half that of the ANve myocytes. Consequently, cell length-to-width ratio was 6.8 ± 0.4 in AN+ve cardiomyocytes, twice the ratio (3.4 ± 0.1) observed in ANve cells (Fig. 3B). Despite significantly different cell length-to-width ratio, the twitch amplitude in both groups (n = 15 cells/group) was comparable (ANve, 5.3 ± 0.4%; AN+ve, 4.5 ± 0.5%) but considerably less than that observed in cardiomyocytes subjected to normoxic, normothermic incubation (9.3 ± 1.0%).

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Diastolic length and width (A) and length-to-width ratio (B) in baseline (nonischemic) and reperfused annexin V negative (ANve) and AN+ve myocytes. *P < 0.05 vs. baseline; #P < 0.05 vs. ANve.

It has been reported (23) that although cytosolic free Ca²⁺ levels may be normal in reperfused rod-shaped myocytes, [Ca²⁺]_m is elevated. Because increased [Ca²⁺]_m is thought to play a role in initiating apoptosis (28), we measured [Ca²⁺]_m in nonischemic and reperfused myocytes. As shown in Fig. 4, nonischemic cells exhibited a baseline [Ca²⁺]_m of 111 ± 14 nM. After SRP, [Ca²⁺]_m in ANve cells was 214 ± 22 nM and further elevated in AN+ve myocytes 382 ± 9 nM.

View larger version (10K): [in this window] [in a new window]   Fig. 4.   Mitochondrial free Ca2+ concentration ([Ca2+]m) in nonischemic cardiomyocytes and myocytes subjected to 20 min of SIsc and 20 min of SRP. Reperfused myocytes were subdivided into ANve and AN+ve groups. [Ca2+]m was measured using rhod 2 and the Mn2+ quenching technique described in METHODS. *P < 0.05 vs. baseline; #P < 0.05 vs. ANve.

Activation of the executioner enzyme caspase 3 is a hallmark feature of the apoptotic cascade (5, 17, 30) and caspase-3-like activity was examined in cardiomyocytes subjected to 20 SIsc, followed by 20- or 60-min SRP. Caspase-3-like activity appeared in 0.03 ± 0.02% cells after 20-min SRP and in 2.85 ± 0.30% of cardiomyocytes after 60-min SRP. Myocytes staining positive for caspase 3 exhibited membrane blebs and a rounded morphology.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study, rat ventricular myocytes subjected to brief SIsc and SRP, but not SIsc alone, exhibited a significant increase in annexin V staining. Rod-shaped myocytes staining positive for annexin V excluded propidium iodide, indicating maintenance of sarcolemmal integrity. However, these same cells exhibited significantly decreased cell width and increased [Ca²⁺]_m levels. During SRP, caspase-3-like activity was observed in cells with a rounded morphology and membrane blebbing. These results suggest that use of annexin V in the setting of SIsc-SRP uncovers a population of cardiomyocytes whose characteristics appear to be consistent with cells undergoing apoptosis.

The use of isolated adult mammalian cardiomyocyte models of SIsc-SRP can serve to distinguish between events occurring during ischemia versus those resulting from reperfusion and also provide a controlled environment for studying mechanisms underlying ischemia-reperfusion injury. However, the majority of single-cell studies (1, 7, 14) have typically addressed necrotic death during ischemia using markers of membrane permeability such as trypan blue staining and enzyme release. Although reperfusion can be associated with cardiomyocyte apoptosis (8, 21, 35), there have been few, if any, studies examining the deleterious effects of reperfusion (12). In fact, unlike necrosis, which occurs during ischemia-reperfusion, the majority of studies (6, 8, 21) suggest that myocardial apoptosis is a reperfusion-related phenomenon.

In the present study, necrotic cell death was assessed with propidium iodide, which stains the nuclei of cells with permeant membranes (15, 30). Our data indicated that only the long (60 min) simulated ischemic interval was associated with significant cardiomyocyte necrosis. On the other hand, SRP (but not SIsc alone) was associated with significant annexin V binding in cells excluding propidium iodide. Annexin V, a protein originally isolated from human placenta, but since found in other tissue, binds phosphatidylserine, a phosphoaminolipid that is externalized onto the outer surface of the plasma membrane during apoptosis or programmed cell death (33). When myocytes were reperfused with HEPES buffer lacking Ca²⁺ or saturated with N₂ (PO₂ ~30 mmHg), annexin V staining was significantly reduced. The exposure of nonischemic cardiomyocytes to the ROS generator H₂O₂ also resulted in a significant increase in annexin V-positive cells. These observations are consistent with the hypothesis that apoptosis occurs primarily during reperfusion (6, 8, 21) and is associated with intracellular oxidative stress and altered Ca²⁺ handling (15, 35, 29).

In addition to excluding propidium iodide, which is indicative of a noncompromised sarcolemmal membrane, AN+ve cells also maintained rod-shaped morphology. Hypoxia-reoxygenation studies (9, 12) in isolated cardiomyocytes have indicated that retention of rod-shaped morphology is critically dependent on maintenance of intracellular ATP levels. Thus our data suggested the presence of some intracellular ATP in AN+ve cells. Despite the appearance of rod-shaped morphology, detailed examination of annexin V myocytes revealed that these cells exhibited decreased cell width, indicative of cell shrinkage. This is in contrast to reperfused ANve cells, which exhibited the increased cell width indicative of cell swelling. Measurement of unloaded shortening after sRP was not different between AN+ve and ANve myocytes, although shortening was reduced compared with nonischemic cells.

It has been reported (23) that while cytosolic free Ca²⁺ levels may be normal in reperfused rod-shaped myocytes, [Ca²⁺]_m is elevated. Furthermore, myocardial reperfusion injury has been intimately linked to [Ca²⁺]_m overload (22, 23). Data from neuronal studies have suggested that increased [Ca²⁺]_m is associated with release of cytochrome c from the mitochondria (28), an early step in the apoptotic cascade (5). In the present study, use of the [Ca²⁺]_m indicator rhod 2 revealed that reperfused myocytes exhibited increased [Ca²⁺]_m compared with nonischemic cells. However, the salient finding was that [Ca²⁺]_m in AN+ve cells was significantly greater [Ca²⁺]_m than ANve cells. Validation of our [Ca²⁺]_m measurements was made by application of the uncoupler protonophore FCCP to normoxic cells, loaded with either the [Ca²⁺]_m indicator rhod 2 (in the presence of Mn²⁺ quenching) or the cytosolic free Ca²⁺ indicator fluo 3. This protocol resulted in an FCCP-induced decrease in [Ca²⁺]_m and an increase in cytosolic free Ca²⁺ (data not shown). Thus, whereas increased [Ca²⁺]_m has been implicated in the initiation of apoptosis, to our knowledge, this is the first report of elevated [Ca²⁺]_m during cardiomyocyte apoptosis.

Activation of executioner caspases, such as caspase 3, is a hallmark feature of apoptosis and is thought to be associated with irreversible morphological and nuclear damage (17, 26). Caspase-3-like activity appeared in cardiomyocytes after 20 min of SIsc and 60 min of SRP. However, the incidence of annexin V staining after 60 min of SRP was ~3-fold higher than the incidence of caspase-3-like activity. Furthermore, cells exhibiting caspase-3-like activity had a rounded morphology with membrane blebs. These findings are suggestive of an ongoing cardiomyocyte apoptotic program before caspase activation and loss of membrane integrity.

All of the characteristics that distinguished AN+ve myocytes from ANve cells and necrotic cells in the present study support the hypothesis that these cells are in the early stages of apoptosis. Requirement of reperfusion for the apoptotic program is well documented (6, 8, 21). Intracellular ATP is critical for execution of apoptosis (5). The presence of an intact cell membrane, cell shrinkage, and increased [Ca²⁺]_m is a hallmark characteristic of apoptotic cells (5, 28). Induction of the apoptotic program in cultured cardiomyocytes in response to exogenous ROS has also been reported (15, 31). Given these observations, the well-accepted use of annexin V as a reliable indicator of ongoing apoptosis (27, 30) and the appearance of caspase-3-like activity during SRP, our results suggest that in addition to necrosis, SRP after SIsc is accompanied by a cardiomyocyte apoptosis program.

There were several limitations in the present study. First, the relatively short SRP times (20 and 60 min) most likely underestimated the true extent of cardiomyocyte apoptosis and necrosis. A second limitation pertains to the caspase detection assay, which may not be entirely specific for caspase-3 activity, and accordingly, we have chosen to use the terminology "caspase-3-like" activity. Furthermore, the true sensitivity of this assay for detection of caspase-3-like activity is unknown. Nevertheless, this assay has been used for detection of caspase-3-like activity in individual cardiomyocytes (30) and in other nonmyocyte studies (11, 36). Third, the use of an isolated cardiomyocyte model may have underestimated the true extent of injury compared with an in vivo preparation. However, use of an isolated myocyte preparation permits both mechanistic determinations and morphological characterization of cardiomyocyte injury. Fourth, the present study did not examine the mechanism for phosphatidylserine externalization during SRP. However, data from other cell types indicate that both oxidative stress and calcium overload can result in externalization of this phosphaminolipid (2, 16).

In conclusion, our findings indicate that annexin V staining detects a subpopulation of myocytes that exhibit characteristics similar to cells associated with apoptosis. The use of indicators of apoptosis, such as annexin V, enables identification of myocytes in the early stages of apoptosis, when they still retain rod-shaped morphology and contractile capacity. Early detection of apoptosis will not only facilitate mechanistic studies relating to this death program but may also aid in the development of therapeutic interventions that can potentially arrest this program.