INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

IN RECENT YEARS, it has become increasingly clear that the importance of the endothelium is not restricted to the modulation of vascular tone and homeostasis. Several studies have established that both the coronary vascular and the endocardial endothelium play a paracrine role in regulating cardiac contractile function (for reviews, see Refs. 1, 17, 27). Cardiac endothelial cells release several diffusible agents, including nitric oxide, endothelin, prostanoids, and kinins, that modify cardiac myocyte function. Such paracrine effects of endothelial cells have been demonstrated both in vitro and in vivo (11), confirming their likely importance.

Besides these known agents, endothelial cells release other unidentified substances that have potent effects on myocardial contraction. Both the superfusates of pure cultures of endothelial cells and the coronary effluent of isolated buffer-perfused rat hearts are reported to alter the contraction of isolated cardiac myocytes and cardiac trabeculae (5, 6, 12, 15, 18, 19, 21). The stimuli that influence the release of these substances from endothelial cells remain poorly understood. Ramaciotti et al. (15) reported circumstantial evidence suggesting that coronary flow rate and ambient PO₂ were important regulatory factors. Indeed, endothelial cell function (e.g., the release of vasoactive mediators) is known to be highly sensitive to acute alterations in flow-induced shear stress as well as acute moderate hypoxia (reviewed in Ref. 25). Recently, we reported that cultured large-vessel endothelial cells superfused with moderately hypoxic buffer (PO₂ 40-50 mmHg) for 1-6 h released a stable, low-molecular-mass substance(s) that induced a potent, reversible inhibition of cardiac myocyte shortening not attributable to reduction in cytosolic Ca²⁺ transients (19). This inhibitory effect appeared to be independent of subcellular second messenger signaling pathways. Thus this substance markedly depressed the translocation of actin filaments over myosin molecules in an in vitro motility assay and reduced the rate of actin-activated cardiac myosin ATPase activity in solution (19). The inhibitory activity was not attributable to generally recognized cardioactive or vasoactive substances released by endothelial cells.

If a similar substance were released by coronary endothelial cells in the whole heart in response to hypoxia, this could have important implications for the cardiac adaptive response to hypoxia. In the present study, we therefore 1) investigated the effects of the coronary effluent of acutely hypoxic isolated hearts on cardiac myocyte contraction, 2) compared this with the effects of hypoxic cultured coronary microvascular endothelial cell (CMEC) superfusate, and 3) studied the subcellular mechanisms underlying the effects of hypoxic coronary effluent.

	METHODS

Top Abstract Introduction Methods Results Discussion References

All studies conformed with the United Kingdom Home Office Guidance on the Operation of the Animals (Scientific Procedures) Act 1986. For isolation of hearts, cardiac myocytes, or CMEC, adult Wistar rats of either sex weighing 250-350 g were terminally anesthetized with pentobarbital sodium (60 mg/kg ip). Hearts were rapidly excised and placed into ice-cold HEPES buffer of the following composition (in mM): 117 NaCl, 5.7 KCl, 4.4 NaHCO₃, 1.2 NaH₂PO₄, 1.25 CaCl₂, 1.7 MgCl, 20 HEPES, and 10 glucose, pH 7.4 at 37°C.

Collection of coronary effluent from isolated perfused rat hearts. Isolated hearts were mounted on a nonrecirculating Langendorff apparatus and retrogradely perfused with HEPES buffer (37°C, gassed with 100% O₂, PO₂ >680 mmHg). Indomethacin (10 µM) was included to inhibit cyclooxygenase, and acebutolol (1 µM) was included to inhibit any -adrenergic effects secondary to catecholamine release. A constant coronary flow that achieved a mean perfusion pressure of 70-80 mmHg was maintained. Hearts were paced at ~10% above intrinsic rate by a right atrial electrode at ~10% above threshold voltage. A water-filled latex balloon attached to a SensoNor 840 transducer was used to measure isovolumic left ventricular pressure. Left ventricular end-diastolic pressure was set at ~10 mmHg. Left ventricular and coronary perfusion pressures were recorded on a chart recorder. Hearts in which mean coronary perfusion pressure or left ventricular pressure varied by >5% or those that had significant arrhythmia during an equilibration period of 25 min were excluded from the study.

Endothelial cell isolation and culture. CMEC were isolated and cultured as described previously (13). Briefly, hearts mounted on a Langendorff apparatus were perfused at 37°C with buffer 1 of the following composition (in mM): 118 NaCl, 4.7 KCl, 1.2 NaH₂PO₄, 1.2 MgSO₄, 25 NaHCO₃, and 11 glucose, pH 7.4 (gassed with 95% O₂-5% CO₂). Epicardial mesothelial cells were devitalized with 70% (vol/vol) ethanol. Buffer 1, with the addition of 0.25 µM CaCl₂ and 0.04% collagenase (type II, Sigma Chemical), was then perfused and recirculated for 30 min. Ventricles (excluding visible large vessels) were chopped into 15 ml of recirculating solution containing 200 mg of bovine serum albumin (fraction V, Sigma Chemical) and triturated gently at 37°C for 10 min. The suspension was filtered through nylon gauze and centrifuged (150 g, 3 min) to sediment myocytes, and the supernatant, with the addition of 100 mg of bovine serum albumin, 0.01% trypsin, and 50 µM CaCl₂, was incubated at 37°C for 15 min. The CMEC pellet was obtained by centrifugation (1,000 g, 10 min), washed twice in buffer 1 with 250 and 500 µM CaCl₂, respectively, and resuspended in 40 ml of prewarmed medium 199 (GIBCO) with 10% newborn calf serum, 10% fetal calf serum, 250 U/ml benzylpenicillin, 250 µg/ml streptomycin, 12.5 µg/ml amphotericin B, and 50 µg/ml gentamycin. Cell suspensions were plated in 75-cm² tissue culture flasks (37°C). After 1 h, unattached cells and debris were washed off. CMEC were cultured to confluence (5-7 days) in medium 199 (with added serum and antibiotics as above) and then subcultured in fresh flasks. CMEC were characterized as endothelial by their typical "cobblestone" morphology, uptake of fluorescently labeled acetylated low-density lipoprotein by >99% of cells, and negative staining for smooth muscle -actin or characterized as microvascular by the rapid formation of capillary-like tubes on the basement membrane preparation Matrigel (8). Porcine aortic endothelial cells were cultured exactly as described previously (19). At passages 2-4, rat CMEC and porcine aortic endothelial cells were transferred to siliconized stirrer vessels with 3-4 ml of microcarrier beads and maintained on beads for ~1-2 wk until use.

Endothelial cell superfusion. Paired identical aliquots of confluent cells on microcarrier beads (1.5-2 ml; i.e., 5-7 × 10⁶ cells) were washed with HEPES buffer and placed in cartridges with 0.8-µm filters (19). Cartridges were continuously superfused with HEPES buffer at 1 ml/min (37°C) for 6 h. Both cartridges were initially superfused with normoxic buffer (PO₂ 160 mmHg). After 1 h, the solution superfusing one cartridge was switched to hypoxic buffer (PO₂ 40-50 mmHg), whereas the other cartridge continued to be superfused with normoxic buffer. Matched single-pass superfusates of both cartridges were collected over consecutive 60-min intervals and stored at 70°C. The ionic composition (Na⁺, K⁺, Mg²⁺, Cl, and Ca²⁺), osmolality, and pH of superfusates were not significantly altered during the period of study. Endothelial cell viability was also unaltered during either normoxic or hypoxic superfusion, as confirmed by trypan blue exclusion and the absence of lactate dehydrogenase in the superfusates.

Isolation of ventricular myocytes and assessment of myocyte function. Ventricular myocytes were isolated by Ca²⁺-free collagenase digestion and were loaded with the Ca²⁺-sensitive fluorescent probes fura 2-AM or indo 1-AM (6 and 10 µM, respectively) as described previously (19). At least 45 min were allowed for deesterification of the indicators. Cells were studied within 8 h of isolation. A drop of myocyte suspension was placed in a chamber on the stage of a Nikon Diaphot inverted fluorescence microscope and superfused with HEPES buffer or test solutions (PO₂ 160 mmHg) at 1 ml/min. Single adherent cells were studied according to previously established criteria, i.e., they were rod-shaped and free of membrane blebs or granulation, with <1 spontaneous contractile wave per minute and a stable contraction pattern. Experiments were performed at room temperature (23°C) to minimize cell leakage of fluorescent probes (23). Myocytes were field stimulated at 0.5 Hz. Cell length was monitored by either a custom-designed photodiode array system or a video edge-detection system (Crescent Electronics). The amplitude of unloaded twitch contraction is reported as the percent decrease in diastolic length. Fura 2 fluorescence was excited at 340 and 380 nm by a rotor-mounted system, and emission was measured at 510 nm. The ratio of fluorescence after excitation at these wavelengths (340/380 ratio) was used as an index of intracellular Ca²⁺. Indo 1 fluorescence was excited at 360 nm, and the 410/480 nm emission ratio was recorded as an index of intracellular Ca²⁺. No attempt was made to calibrate cytosolic Ca²⁺ because of the uncertain subcellular compartmentation of the probes (23). Data files of 6-10 consecutive steady-state beats (fluorescence ratio and cell length) recorded at intervals were averaged for analyses.

Effect of enzymatic hydrolysis on hypoxic coronary effluents. To investigate the chemical nature of the substance(s) in coronary effluent that exerted biological activity on cardiac myocytes, effluents were subjected to attempted enzymatic hydrolysis. The effects of acid phosphatase, 3'-nucleotidase, 5'-nucleotidase, trypsin, sulfatase, glutamic-pyruvic transaminase, and glutaminase were studied. Samples of hypoxic coronary effluent were split into two identical aliquots (~10-20 ml each), and one of the enzymes listed was added to one aliquot (0.01 U/ml). Both aliquots were then incubated at 37°C for 30 min, heated at 90°C to denature the enzyme, and centrifuged to remove denatured products. The paired aliquots were then tested in random order on individual cardiac myocytes, and the effect of the enzyme-treated aliquot was expressed as a percentage of the effect obtained with the untreated aliquot.

Materials. Fura 2-AM and indo 1-AM were purchased from Calbiochem, medium 199 with L-glutamine was from GIBCO, and all other chemicals and reagents were from Sigma Chemical.

Statistics. Data are given as means ± SE. Comparisons were made by an unpaired or paired Student's t-test, as appropriate, on absolute values, and differences were considered significant if P < 0.05.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Effect of coronary effluent collected during hypoxia. On exposure to coronary effluents collected during hypoxia, reductions in isolated myocyte twitch shortening ranging from ~10 to 100% depression were observed. The majority of this variability was at the level of the effluents and not the myocytes, i.e., any individual effluent had broadly similar effects on several cardiac myocytes.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Example of potent effect of hypoxic coronary effluent on isolated rat cardiac myocyte contraction. A: slow time-base recording of cell twitch shortening during exposure to normoxic (i), hypoxic (ii; heavy horizontal bar), and posthypoxic coronary effluents (iii) collected from the same heart. All effluents were reequilibrated for temperature, PO2 (~160 mmHg), and pH before they were tested on isolated cardiac myocytes. B: fast time-base recordings of cell shortening and associated intracellular Ca2+ transients during exposure to effluents indicated.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Example of less potent effect of hypoxic coronary effluent on isolated rat myocyte contraction. Effluents were reequilibrated for temperature, PO2 (~160 mmHg), and pH before they were tested on isolated cardiac myocytes. Top: fast time-base recordings of intracellular Ca2+ transients. Bottom: associated twitches during exposure to matched normoxic or hypoxic coronary effluent.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Percent changes (relative to control HEPES buffer) in cell twitch shortening (TS; left) and amplitude of Ca2+ transients (fluorescence ratio as % of control) on exposure to coronary effluents. Numbers in parentheses indicate number of myocytes studied. * P < 0.05; *** P <0.0001 vs. control.

Effect of superfusates of hypoxic CMEC. Superfusates obtained from two batches of rat cultured CMEC were studied. Figure 4 shows an example of a potent effect of hypoxic CMEC superfusate on myocyte shortening. In this cell, addition of hypoxic superfusate resulted in a rapid total inhibition of shortening that was not accompanied by a corresponding reduction in amplitude of the intracellular Ca²⁺ transient. Myocyte diastolic length was also significantly decreased. Replacement of hypoxic superfusate by posthypoxic superfusate resulted in a rapid recovery of cell shortening and diastolic cell length. These effects were similar to those observed with hypoxic coronary effluent (Fig. 1) and those previously reported with pig aortic endothelial cells (19).

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Example of potent effect of hypoxic superfusate of rat cultured coronary microvascular endothelial cells (CMEC) on isolated rat cardiac myocyte contraction. A: slow time-base recording of cell twitch shortening during exposure to normoxic (i), hypoxic (ii; heavy horizontal bar) and posthypoxic CMEC superfusates (iii). All endothelial superfusates were carefully reequilibrated for temperature, PO2 (~160 mmHg), and pH before they were tested on isolated cardiac myocytes. B: fast time-base recordings of cell shortening and associated intracellular Ca2+ transients during exposure to superfusates indicated.

View larger version (12K): [in this window] [in a new window]   Fig. 5.   Percent changes in cell TS, diastolic length (DL), and amplitude of fluorescence ratio transients (FR) on exposure to hypoxic rat CMEC and pig aortic endothelial superfusates (relative to matched normoxic superfusates). Numbers in parentheses indicate number of myocytes studied. * P < 0.05; ** P < 0.01; *** P < 0.001 vs. control.

Assessment of calcium-myofilament interaction. Myocyte contractile amplitude is closely influenced by the peak intracellular Ca²⁺ and the myofilament responsiveness to Ca²⁺. The relative contribution of these factors to altered contraction may be assessed from the relationship between peak twitch shortening and peak intracellular Ca²⁺ (fluorescence ratio). Figure 6 (A and B) shows this relationship for the twitches recorded from three myocytes that, after a period of quiescence, were electrically stimulated (0.5 Hz). With the use of linear regression as a simple and convenient analysis, the sequential reductions in fluorescence transient and twitch shortening during the "negative staircases" were found to be highly correlated. In contrast, there was a very weak correlation between peak fluorescence ratio and peak twitch shortening for contractions recorded during exposure of myocytes to hypoxic coronary effluents (expressed as a percentage of the respective baseline values before addition of effluents) (Fig. 6C). In other words, the changes in cell shortening were, to a large extent, unrelated to alterations in intracellular Ca²⁺.

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Relationship between peak intracellular Ca2+ and peak TS of isolated rat cardiac myocytes. A: "negative staircase" of Ca2+ transients and twitch contractions in a cell stimulated at 0.5 Hz after a quiescent period of >3 min. B: peak fluorescence ratio vs. peak TS values during negative staircases in 3 myocytes. All values are expressed as a percentage of initial (largest) twitch. C: peak fluorescence ratio vs. peak TS during exposure of 15 myocytes to hypoxic coronary effluent.

View larger version (27K): [in this window] [in a new window]   Fig. 7.   Representative phase-plane plots of instantaneous cell length vs. fluorescence ratio for isolated myocyte twitch contractions. A: effect of matched normoxic, hypoxic, and posthypoxic coronary effluents. B: effects of matched normoxic, hypoxic, and posthypoxic CMEC superfusates. C: effect of 2,3-butanedione monoxime (BDM). D: effect of pyrophosphate (Pyro). Con, control. Phase-plane loops proceed counterclockwise and start from points indicated by arrow. Fluorescence ratio signals were treated by Savitzky-Golay least-squares quintic smoothing (Fig P software; Biosoft, Cambridge, UK).

View larger version (14K): [in this window] [in a new window]   Fig. 8.   Steady-state relationship between cell shortening and intracellular Ca2+ during tetanic contraction of intact single cardiac myocyte. Effect of hypoxic coronary effluent is shown relative to that of matched normoxic effluent (normoxic + hypoxic). Top: tetanic elevation of Ca2+. Bottom: tetanic shortening.

Agent(s) responsible for activity of hypoxic coronary effluent. It was previously reported that the activity of hypoxic endothelial cell superfusate 1) was not accounted for by known cardioactive factors released by living cells, 2) probably involved a low-molecular-mass, stable factor acting directly on the cardiac myofilaments independent of known subcellular signaling pathways, and 3) was specific for cardiac myosin and did not affect smooth muscle myosin (19). In view of these features, we considered the possibility that the inhibitory activity may involve a nucleotide or nucleoside analog or product that specifically inhibited cardiac but not smooth muscle myosin or that some intermediate product of high-energy metabolic pathways in the endothelial cell may be responsible. Accordingly, we surveyed the effects of several pharmacological agents on myocyte twitch shortening and Ca²⁺ transients (Table 2). However, none of these compounds reproduced the effects of the hypoxic coronary effluents or endothelial superfusates. A negative inotropic effect associated with a reduction in the intracellular Ca²⁺ transient was observed with four compounds. The changes in twitch shortening and fluorescence ratio with these compounds were as follows: adenosine 2'-monophosphate, 17.6 and 12.5%; adenosine 3'-monophosphate, 24.3 and 16.7%; adenosine 5'-O-(3-thiotriphosphate), 28.2 and 15.3%; and sodium pyrophosphate, 69.7 and 48.6%, respectively.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The main finding of the present study is that isolated rat hearts respond rapidly and reversibly to acute moderate hypoxia by releasing an unidentified stable substance that inhibits isolated cardiac myocyte shortening and reduces diastolic cell length independent of changes in intracellular Ca²⁺. These effects are similar to those of the superfusate of hypoxic rat CMEC in culture (present study) and to those previously reported for hypoxic large vessel endothelial cells in culture (19). The action of hypoxic coronary effluent appears to result predominantly from a Ca²⁺-independent modification of myofilament function as assessed both during twitch and tetanic contraction.

The present study significantly extends the previous data obtained in cultured large-vessel endothelial cells (19) in that it 1) establishes that the release during hypoxia of substances that inhibit myocyte contraction is not simply a cultured cell phenomenon but is applicable to the whole heart; 2) demonstrates that in the intact heart, this response can be evoked by just 5 min of moderate hypoxia and can also be rapidly "switched off" within 5 min of reoxygenation; and 3) suggests that the cardiodepressant factor(s) in coronary effluent of hypoxic rat hearts may derive from coronary endothelial cells in situ. The latter would be consistent with previous reports suggesting the presence of endothelial-derived cardioactive factors in the coronary effluent of isolated perfused rat hearts (5, 12, 15). However, the results of the present study do not exclude the possibility that other cell types within the heart may also play a role. The rapidity of release of the cardioactive factor is consistent with previous studies showing that acute moderate luminal hypoxia (PO₂ 30-60 mmHg in buffer-perfused systems) is an effective stimulus for rapid changes in several aspects of endothelial function (25), including the release of vasodilators such as nitric oxide and prostanoids in isolated arteries and the intact circulation (9, 14). Because endothelial cells have a low "metabolic" sensitivity to hypoxia, with no change in high-energy phosphate content or cell viability for several hours even during severe hypoxia (PO₂ <10 torr) (7), these rapid responses to moderate hypoxia may represent an undefined hypoxia "sensing" process unrelated to changes in energy production (14).

The chemical identity of the substance released in response to hypoxia is currently unknown. The activity was not attributable to any changes in measured ionic composition, osmolality, pH, or PO₂ of the coronary effluent. We have previously reported (19) that the biological actions of a similar or identical substance released by cultured large vessel endothelial cells are not attributable to known cardioactive factors released by these cells (e.g., nitric oxide, adenosine, prostanoids, endothelin-1) and are found in low-molecular-mass fractions (relative molecular weight <500). Activity of the factor was maintained for several hours at room temperature and for several weeks at 70°C and was not significantly diminished by heating at 90°C. In the present study, we also investigated the possibility that the inhibitory activity may involve a nucleotide or nucleoside analog (or product) that specifically inhibited myofilament function or that some other intermediate product of high-energy metabolic pathways in the endothelial cell was involved. However, a survey of several potential candidate compounds failed to identify any that reproduced the typical actions of the factor released during hypoxia (Table 2). A significant negative inotropic effect of pyrophosphate (1 mM) was observed, but this was associated with a concomitant reduction in the Ca²⁺ transient. A number of the other compounds tested [adenosine 5'-O-(3-thiotriphosphate), adenosine 2'-monophosphate, and adenosine 3'-monophosphate] exerted small effects on myocyte shortening at high (100 µM) doses, but these were also attributable to changes in the Ca²⁺ transient. We also found that activity was not significantly reduced by attempted enzymatic hydrolysis.

Inhibition of cardiac myocyte contraction by the endothelial factor without corresponding reduction in the intracellular Ca²⁺ transient suggests that this effect occurred predominantly via an interaction with the myofilaments, rather than any excitation-contraction coupling process capable of modulating cytosolic Ca²⁺ levels. In keeping with this, there was a poor correlation between changes in peak fluorescence ratio and alterations in twitch shortening on exposure to hypoxic coronary effluent, in contrast to the good correlation between these parameters during negative staircases (Fig. 6). Analysis of the instantaneous relation between intracellular Ca²⁺ and cell length during electrically stimulated twitch contractions indicated that there was a marked insensitivity of cell length to changes in intracellular Ca²⁺ (Fig. 7). This was manifested as a loss of hysteresis in the "phase-plane loop." This effect was quite different from that observed with BDM, which reduces myofilament Ca²⁺ sensitivity (2). In the case of BDM, the hysteresis in the phase-plane loop was to a large extent maintained, but the position of the loop was shifted right and down, i.e., toward reduced myofilament responsiveness to Ca²⁺ (22). A further point of note was the reduction in cell length induced by the endothelial factor, even at low (diastolic) Ca²⁺. The data obtained in experiments in which myocytes were electrically tetanized after inhibition of the sarcoplasmic reticulum Ca²⁺ ATPase by thapsigargin were consistent with these findings. Myocyte cell length was reduced by the endothelial factor, whereas the amplitude of cell shortening on steady elevation of intracellular Ca²⁺ was significantly depressed (Fig. 8). The underlying molecular mechanism(s) responsible for these changes remains to be defined. One possibility that we have suggested is that, in the presence of the endothelial factor, cross bridges are able to attach and to generate force but that their cycling rate is reduced (19). This could occur, for example, because of an inhibition of the transition of cross bridges out of high force states, somewhat analogous to the latch-state cycling of smooth muscle in which force is maintained while cross bridges turn over more slowly (3). Such a mechanism would be, to the best of our knowledge, unique for any paracrine agent released by living cells.

There was a significant variability in the magnitude of effect of the hypoxic coronary effluent on isolated cardiac myocyte shortening, which was largely due to variability of effluents rather than of myocyte responses. This variability could not be accounted for by differences in baseline contractile function or coronary perfusion of the hearts or by the level of contractile depression during hypoxia (Table 1). All the animals used for these studies were similar in age, weight, source, and strain. Furthermore, great care was taken to ensure that conditions such as temperature, pH, ionic composition, and PO₂ were precisely controlled. A more likely possibility that may account for the observed variability in response is that the endothelial factor could be released abluminally and that what was detected in the coronary effluent might represent "spillover" into the coronary circulation. Such a mechanism is known to exist for other endothelial cell-derived factors, e.g., endothelin-1 (26). It is also feasible that the endothelial factor might be degraded in the coronary circulation and that this process may be variable. The precise reasons for the variability remain to be defined; in preliminary studies, we have found that the proportion of heart effluents that exert significant reduction in myocyte twitch shortening can be increased to a maximum of ~70% by recirculating the effluent collected during hypoxia and concomitantly reducing coronary flow rate to ~66% of the baseline level during hypoxia (unpublished data).

Because the present study was performed using isolated buffer-perfused hearts, extrapolation of the findings to the situation in vivo involves many untested assumptions. Nevertheless, our results could be of relevance to certain cardiac pathological conditions. The level of isolated heart hypoxia studied (PO₂ ~35 Torr) was within a pathophysiologically relevant range that might occur during partial reduction in coronary blood flow (10). The release of a factor that inhibits myocyte shortening in response to hypoxia could serve to reduce energy turnover and myocardial O₂ consumption. Such a mechanism may therefore function to facilitate the maintenance of myocardial oxygen supply-demand balance during hypoxia or partial coronary flow restriction. Indeed, it is recognized that oxygen supply-demand balance during acute hypoxia or a temporary partial reduction in coronary flow may be facilitated by a stable decrease in oxygen demand associated with a proportional depression of contraction (4, 16). Restoration of oxygen supply or coronary flow usually leads to prompt recovery of function. The mechanism(s) responsible for this adaptive downregulation of oxygen demand remains unclear, but a sustained decrease in high-energy phosphates or phosphorylation potential is not thought to be responsible. Our findings raise the possibility that coronary microvascular endothelial cells, sited at the interface between vascular O₂ supply and the O₂-utilizing tissue, may "sense" hypoxia and trigger adaptive responses such as the release of the contraction-inhibiting substance described in the present study. Alternatively, it is also feasible that such a factor could "protect" against the deleterious effects of Ca²⁺ overload during prolonged hypoxia and/or immediately on reoxygenation. By causing the cardiac myofilaments to become relatively insensitive to rises in intracellular Ca²⁺, such a factor could minimize cellular hypercontracture at reoxygenation. (Although the "hypoxic factor" reduces the resting length of externally unloaded cells, the extent of myocyte shortening is still lower than that during normal twitch contraction.) Whereas release of the factor is "turned off" rapidly on cessation of brief hypoxia, the effects on cardiac myocytes in situ would presumably take some minutes to be reversed. A recent study by Stephan and colleagues (24) reported the release of a cardiodepressant mediator by isolated hearts subjected to ischemia, although the effects and subcellular mechanisms of action of this agent in myocardium were not explored.

In conclusion, we have shown that the reoxygenated coronary effluent of isolated hypoxic rat hearts releases an unidentified substance that inhibits myocyte shortening independent of changes in intracellular Ca²⁺. These effects are similar to those of the reoxygenated superfusate of hypoxic cultured rat CMEC. On the basis of an assessment of Ca²⁺-myofilament interaction in intact cardiac myocytes, it appears that this substance exerts Ca²⁺-independent effects on cross bridge function. Such effects could act as a "protective" mechanism during hypoxia-reoxygenation (or ischemia-reperfusion), for example, by reducing myocardial energy turnover and O₂ demand or by minimizing some of the effects of Ca²⁺ overload.