Daily lecture: Low-O2 affinity erythrocytes improve performance of ischemic myocardium

Interesting article about shifting the oxyhemoglobin dissociation curve.

Gösta Berlin, Keith E. Challoner, and Robert D. Woodson

ABSTRACT

O₂ transport and O₂ diffusion interact in providing O₂ to tissue, but the extent to which diffusion may be critical in the heart is unclear. If O₂ diffusion limits mitochondrial oxygenation, a change in blood O₂ affinity at constant total O₂ alter cardiac O₂ consumption ( transport shouldO₂) and function. To test this hypothesis, we perfused isolated isovolumically working rabbit hearts with erythrocytes at physiological blood-gas values and P₅₀ (PO₂ required to half-saturate hemoglobin) values at pH of 7.4 of 17 ± 1 Torr (2,3-bisphosphoglycerate depletion) and 33 ± 5 Torr (inositol hexaphosphate incorporation). When perfused at 40 and 20% of normal coronary flow, mean O₂ decreased from the control value by 37 and 46% (P <> as cardiac work, decreased by 38 and 52%, respectively (P <> Perfusion at higher P₅₀ during low-flow ischemia improved O₂ by 20% (P <>P <> modest improvement at basal flow (P <>P <> The improvement in O₂ and function due to the P₅₀ increase demonstrates the importance of O₂ diffusion in this cardiac ischemia model.

blood oxygen affinity; oxygen dissociation curve; inositol hexaphosphate; isolated heart; rabbit

	INTRODUCTION

THE ROLE OF O₂ diffusion in O₂ delivery remains a controversial and difficult area. It is well known that O₂ flux from erythrocytes to cells of an organ depends on diffusion. Because the O₂ pressure in cells, including cardiac myocytes, is only a few Torr (15), the O₂ diffusion gradient depends heavily on the O₂ pressure in the microvasculature at the point of its release from hemoglobin, a variable determined in part by the position and shape of the blood O₂ dissociation curve (ODC). That changes in ODC position might enhance or limit O₂ flow to cells in certain settings seems intuitively evident, given the existence of the Bohr phenomenon, the relationship between blood O₂ affinity and hemoglobin concentration in mutant hemoglobins, the presence of higher O₂ affinity in fetuses, the relative left ODC shift of animals native to high altitude, and the rise in 2,3-bisphosphoglycerate (BPG) and P₅₀ (PO₂ required to half-saturate hemoglobin) in anemia and low cardiac output states (7, 49). These observations are also consistent with the notion that the O₂ pressure head is regulated in a range that does not greatly exceed what is needed for O₂ flux. Indeed, in the case of myocardium, the fact that blood flow varies inversely with P₅₀ (47) provides further support for this idea, as does the tight relationship between cardiac work and coronary flow. Nevertheless, experiments that provide unambiguous evidence of modulation of in vivo O₂ off-loading by ODC shifts are comparatively sparse, and many experiments have shown only modest or no effect. Apart from its physiological significance, this is a matter of some importance in clinical medicine, given the changes in ODC position that are known to occur with cardiac disease, storage of red blood cells (RBCs), disturbances of acid-base balance and the like (40), as well as the possibility of therapeutic manipulation of the ODC (46).

A specific setting in which the O₂ diffusion gradient could be of considerable importance is myocardial O₂ delivery (50). Myocardial blood flow is characterized by a major degree of microheterogeneity, with flow rates in millimeter-range tissue volumes varying 6- to 10-fold under basal conditions (3, 5, 11, 12, 24, 41). This heterogeneity rises as tissue volume falls (12) and is greatest as the tissue volume analyzed approaches the domain of a single capillary (27). Local myocardial substrate uptake and O₂ consumption (O₂) are also heterogeneous (24) and only somewhat matched to flow. When a major coronary vessel is constricted, downstream local flow also decreases but is initially random with respect to original local flow (9, 24). Anaerobic metabolism appears in foci with the greatest relative reductions in flow (24), a phenomenon believed to account for the patchiness of myocardial infarction after insults that reduce cardiac perfusion (3). Given that basal myocardial O₂ extraction is normally high and locally variable (43), this could be simply because limited O₂ extraction reserve caps O₂ sooner in local areas with higher extraction. Alternatively, if O₂ diffusion between capillary units is of importance (50), one might expect dysoxia in loci that are most dependent on diffusion from adjacent regions. Accordingly, induced shifts of the ODC with other O₂ transport variables held constant furnish a useful method to test the importance of local O₂ diffusion in myocardial ischemia.

Several recent studies of ODC shifts on oxygenation of the heart and other tissues have been performed with the compound 2-[4-[[(3,5-dimethylanilino)carbonyl]methyl]phenoxyl]-2-methylproprionic acid (RSR13). This molecule crosses the RBC membrane and interacts reversibly with hemoglobin, producing appreciable reductions in blood-O₂ affinity (1). Results indicate that this drug may improve oxygenation when flow is blood decreased, particularly in models of ischemic heart disease and stroke (21, 29, 44, 45), implying that an increase in the O₂ diffusion gradient may increase O₂ flux. However, there is at least some evidence that RSR13 has effects on vascular tone other than those expected from the rightward ODC shift (Ref. 32 and Woodson, unpublished observations), although other observations have shown no such effect (29, 44). This could be a confounding variable, especially because vascular tone appears to mediate the microheterogeneity of blood flow (3, 4). In any case, it would be desirable to establish whether comparable effects of ODC shifts can be demonstrated when the ODC is shifted in other ways, particularly given the paucity of positive results in the literature.

These considerations prompted us to examine the role of O₂ diffusion in cardiac ischemia, in which we tested the hypothesis that a shift in the ODC due to the presence of intraerythrocytic inositol hexaphosphate (IHP) would improve O₂ diffusion and O₂ when the latter is limited by reduced O₂ transport. We employed an isolated, isometrically contracting rabbit heart in these studies, a preparation widely used in studies of cardiac physiology and metabolism. The rabbit heart is known to display the same microheterogeneity of blood flow and metabolism observed in larger animals and humans (27, 35). This model allowed us to evaluate myocardial function and O₂ at a normal coronary flow rate and during ischemia when the erythrocyte (RBC) P₅₀ was increased from a subnormal value to a supranormal one. Although other investigators have studied effects of altered RBC O₂ affinity in the isolated heart, they employed quite different models and/or did not examine effects of altering P₅₀ during ischemia.

MATERIALS AND METHODS

Preparation of RBCs

Krebs-Henseleit buffer. Krebs-Henseleit buffer (KHB) was prepared as follows. The basic solution (in mM: 118 NaCl, 4.7 KCl, 2.75 CaCl₂, 1.2 MgSO₄, 1.2 KH₂PO₄, 0.52 Na₂EDTA, 25 NaHCO₃, and 11 dextrose and 1,000 U sodium heparin per liter) was equilibrated by bubbling with 95% O₂-5% CO₂ at room temperature. Bovine serum albumin (1.5%) was then added, and the solution was filtered (0.22 µm).

High-affinity RBCs (control cells). Human packed RBCs stored for 6-14 days in standard CPDA-1 solution (citrate-phosphate-dextrose-adenine) were washed three times in an isotonic saline solution (1,350 g, 5 min); the supernatant and the buffy coat were carefully removed. RBCs were then diluted with KHB. At this stage, the RBC solution was stored in a refrigerator overnight at a hematocrit of 30-40%. The following day, the RBCs were further washed twice in saline containing 10 mM CaCl₂, 10 mM MgCl₂, and 2 mM glucose. Base excess was corrected to ~0 meq/l (pH of 7.4 at PCO₂ of 40 Torr) with addition of NaHCO₃. The cells were diluted with KHB to give a hematocrit of 25%. The diluted RBC suspension was passed through a leukocyte removal filter (PALL RC100).

Low-affinity RBCs (IHP-loaded cells). Packed RBC units were stored for 6-14 days at 4°C. The cells were washed once in isotonic saline and then passed through a leukocyte removal filter (PALL RC100). After two more washes in isotonic saline, IHP was incorporated into the cells by the continuous-flow hypotonic dialysis technique, similar to that described by Teisseire et al. (38). The method was modified by reducing the flow rate of the RBCs through the hemodialyzer (Lundia 1C plate dialyzer) to 10 ml/min and by diluting the IHP solution with 0.15 M NaCl (1:1 vol/vol for the first 5 experiments and 1:1.5 for the subsequent experiments) to reduce the degree of P₅₀ shift. After they were resealed, the cells were washed once in isotonic saline, once in hypotonic saline (240 mosmol/kgH₂O) to lyse the most fragile cells, and two times in isotonic saline containing 10 mM CaCl₂, 10 mM MgCl₂, and 2 mM glucose. The RBCs were then diluted with KHB containing albumin (1.5%) and stored in a refrigerator overnight. On the day of perfusion, the cells were washed once in isotonic saline, once in hypotonic saline, and finally three times in saline with CaCl₂, MgCl₂, and glucose. The cells were then diluted with KHB with 1.5% albumin and NaHCO₃ to achieve a hematocrit of 25% and pH 7.4. The IHP incorporation resulted in a P₅₀ of 25-42 Torr (mean shift of 16.0 ± 5.1 Torr, range of 9-26 Torr). Mean recovery of RBCs was 61%. Supernatant hemoglobin concentration during perfusion was consistently below 0.1 g/dl, and the concentrations of ionized calcium, sodium, and potassium were within the normal range.

Isolated Heart Preparation

Experimental procedures were approved by the Animal Care Committee of the University of Wisconsin and were conducted in accord with the Guiding Principles in the Care and Use of Animals of the American Physiological Society and the Guide for the Care and Use of Laboratory Animals [DHSS Publication No. (NIH) 85-23]. Our method paralleled those used in other laboratories (25, 42). Male New Zealand White rabbits weighing between 1.5 and 2 kg were anesthetized with an 8:1 mixture of ketamine-xylazine administered intramuscularly and then were given 1,000 U of sodium heparin intravenously. The heart was quickly removed after an intravenous bolus injection of pentobarbital sodium (25-30 mg/kg). The heart was placed in a heated cabinet, the ascending aorta was immediately cannulated, and retrograde perfusion was started at once with either KHB solution (series A) or human RBCs suspended in KHB solution (series B). The time from sternal incision to cardiac perfusion was well under 1 min. A drain was created in the apex of the left ventricle (LV) by puncture with an 18-gauge needle to allow egress of blood from the Thebesian vessels. A cannulated, fluid-filled balloon connected to a pressure transducer was placed in the LV via a left atriotomy for measurement of LV pressure during isovolumic contraction. A second catheter was placed in the pulmonary artery to collect myocardial venous effluent. Aortic pressure was monitored by a pressure transducer connected to a stopcock inserted into the line just above the aortic cannula.

Perfusion Setup

A schematic diagram of the perfusion setup is shown in Fig. 1. The suspended RBCs were brought to physiological blood-gas concentration and temperature in a primary circuit. From a continuously stirred, covered reservoir, suspended RBCs were pumped at a relatively high rate (about 25 ml/l) through a membrane oxygenator (SciMed Life Systems, Minneapolis, MN) and a transfusion filter (PALL Ultipor) to a second similar overflow reservoir, from which they returned by gravity to the main reservoir. Red blood cells were then propelled by a second pump at the desired flow rate from the overflow reservoir, which also served as a bubble trap, to the heart cannula. Perfusate temperature was recorded by a needle probe in the aortic line just above the heart. Blood passing through the heart was not recirculated, which avoided influence of metabolites. The reservoirs were water-jacketed to maintain a perfusate temperature close to 37°C, and the entire apparatus was enclosed in a thermostated cabinet. The system was designed so as to avoid settling of RBCs, with the possibility of altered perfusate hematocrit, at any point in the circuit.

Fig. 1. Schematic diagram of perfusion setup. Suspended red blood cells (RBC) are continuously recirculated through oxygenators at a rapid rate in the 2 primary circuits (see text), from which they are pumped at the desired flow rate to the isolated heart. Perfusate temperature is registered by a needle probe (not shown) in the aortic inflow line just above the heart. The lower catheter, syringe, and transducer are for balloon inflation and left ventricular (LV) pressure measurement. The upper catheter and syringe are for anaerobic sampling of myocardial venous return. Except during sampling, venous return flows freely from the severed pulmonary artery and is discarded. A small LV apical stab wound allows drainage from the Thebesian circulation. The entire system is enclosed in a heated cabinet, and all reservoirs are water jacketed at 37°C (not shown). P50, PO2 required to half-saturate hemoglobin.

Series A. In this series, we used normal stored human RBCs ("control cells") to evaluate the reproducibility and sensitivity of the isolated heart model and to study the effects of ischemia on LV physiological parameters. Hearts (n = 9) were paced at a rate of 160-180/min (4-8 V, 10-ms pulse duration). They were initially perfused with KHB by gravity at a constant aortic pressure of ~90 mmHg. The intraventricular balloon volume was set to produce an end-diastolic pressure of 10 mmHg (2). The balloon volume was held constant during the experiment so that developed LV pressure [peak LV systolic pressure minus peak LV diastolic pressure (LVS-LVD)] reflected the contractile state of the myocardium. Hearts were allowed to stabilize for ~15 min under these conditions. Hearts that did not generate an LVS pressure of at least 60 mmHg or whose function declined during the stabilization period were discarded (2). About 20% of hearts were rejected for these reasons.

Perfusion by pump was then started with oxygenated RBCs at a constant flow rate of 9 ml/min. This corresponds to a perfusion rate of 2.1 ± 0.2 ml · min¹ · g ventricular wet weight¹ (mean ± SD), which is similar to the rate used by others in RBC-perfused isolated hearts (2, 20, 25) and close to the means reported for awake rabbits (28) and anesthetized, open-chest rabbits (17, 43). This flow rate produced a mean aortic pressure of 95 ± 22 mmHg. Ischemia was then induced by reducing the flow rate to 3.5 ml/min and then to 2 ml/min for at least 5 min. Hearts were allowed to recover for at least 5 min at a flow rate of 9 ml/min after each level of ischemia. Finally, flow was interrupted completely for 2 min (total ischemia), after which the flow rate was returned to 9 ml/min.

Series B. In this series, hearts (n = 12) were perfused with suspended control RBCs immediately after isolation at a flow rate of 9 ml/min, paced (130-180/min), and allowed to stabilize for ~15 min. Each heart was then perfused with control (high-affinity) and with IHP-loaded (low-affinity) RBCs at flow rates of 9.0 ml/min, 3.5 ml/min, and back to 9.0 ml/min. Arterial and venous samples were obtained in duplicate after at least 5 min of perfusion, and the results were averaged. The order of perfusion with control and IHP-loaded RBCs was randomly varied such that the order for half of the hearts was C9-IHP9-IHP3.5-C3.5-C9-IHP9, whereas the order for the other half was IHP9-C9-C3.5-IHP3.5-IHP9-C9, where C indicates perfusion with control cells, numbers indicate rates of perfusion (in ml/min), and IHP indicates perfusion with IHP-loaded cells. In most experiments, hearts were then exposed to total ischemia (no perfusion) for 2 min once (n = 10) or twice (n = 3), after which the flow rate was returned to 9.0 ml/min. Total experimental time including the stabilization period was 60-90 min.

Measurements

Heart rate and LV and aortic pressures were recorded continuously (Gould 481 strip-chart recorder). Duplicate arterial (oxygenated blood in the reservoir) and venous (pulmonary artery catheter) blood samples were taken after ~5 min at each flow rate for measurement of pH, PO₂, PCO₂ (Radiometer ABL 30, Copenhagen, Denmark), O₂ content and saturation, and hemoglobin concentration (CO-oximeter, model 282, Instrumentation Laboratory, Lexington, MA). O₂ content was determined from O₂ saturation and hemoglobin concentration with allowance for dissolved O₂. O₂ extraction was expressed as follows: (arterial O₂ content venous O₂ content)/arterial O₂ content. O₂ was calculated as the product of perfusion rate (calibrated) and arteriovenous O₂ content difference. LV-developed pressure was expressed as LVS-LVD. Cardiac work was expressed as the double product (LVS-LVD) × heart rate. ODCs were determined with either a Hemox Analyzer (TCS Medical Products) or with a Hem-O-Scan (Aminco) at 37°C and expressed at pH 7.4.

Histology

Three hearts from series A were examined histologically after perfusion with control RBCs. Muscle fiber structure was intact with normal striations and no visible edema at 1.5 h, the maximal time of any experiment. Compared with normal hearts, perfused hearts showed minimal, spotty hemorrhage in the LV myocardium, with a tendency of more hemorrhage with increasing perfusion time. These hemorrhages involved <5%> punctate hemorrhages could be seen grossly. These changes are not surprising in light of absence of platelets and coagulation proteins in the perfusate and compare favorably with what others have observed grossly (M. Vogel, personal communication, and Ref. 42). By contrast, there was considerably more hemorrhage in the right ventricular wall. Because our study dealt only with LV function, we believe this did not affect our conclusions. The behavior and gross appearance of experimental hearts were similar to those of the histologically examined hearts. We found no other studies in which histopathology in this preparation was described.

Statistics

Duplicate values obtained for each parameter during each perfusion condition were first averaged. Differences in parameters with changes in flow rate at constant P₅₀, and with changes in O₂ affinity at constant flow rate, were examined by paired t-test. Differences as a function of P₅₀ in series B following total ischemia were examined by unpaired t-test.

RESULTSSeries A

Table 1 shows that arterial blood gases were close to the physiological range. Temperature averaged 35.2 ± 1.1°C. Figure 2 displays the relationship of LV-developed pressure, positive change in pressure over time (+dP/dt), and LV work as a function of flow rate (n = 9). At a normal flow rate of 9 ml/min (2.1 ± 0.2 ml · min¹ · g ventricular wet wt¹), mean value ± SD for LVS-LVD averaged 58 ± 6 mmHg and LV work was 9,463 ± 1,027 mmHg · beats · min¹. These values are in agreement with those of Apstein et al. (2), whose methodology closely paralleled ours. When total O₂ transport was decreased to simulate ischemia by the reduction of flow rate to ~40% of the initial value (3.5 ml/min = 0.8 ± 0.1 ml · min¹ · g¹), to 20% (2 ml/min ± 0.4 ml · min¹ · g¹), and to 0% (total ischemia), there were progressive decreases in LV function. Thus LVS-LVD decreased from the starting value by 39, 53, and 77% with the three flow decrements, respectively (Fig. 2A; P <>t-test). Peak +dP/dt decreased by similar amounts (Fig. 2B; P <> better, with changes in LV relaxation rate, as judged by negative peak dP/dt, paralleling changes in +dP/dt. Cardiac work decreased by 39, 52, and 77%, respectively (Fig. 2C; P <> Myocardial O₂ extraction increased by 73 and 163% at 40 and 20% of the initial perfusion rate, respectively (Fig. 2D; P <> whereas O₂ decreased by 37 and 46% (P <>

Table 1. Arterial blood gas parameters and temperature

Fig. 2. Effect of various perfusion rates with control red blood cells and total ischemia on peak LV systolic pressure minus peak LV diastolic pressure (LVS-LVD; A), peak positive change in pressure over time (+dP/dt; B), LV work (C), and O2 extraction (D).

Hearts recovered completely with restoration of perfusion to the basal level after the two levels of partial ischemia; there were no statistically significant changes postischemia in any parameter from basal values. With restoration of perfusion to the basal level after total ischemia, which occurred at the end of the protocol, there was a 15% decrease from starting value in mean LVS-LVD (P <>P <> arteriovenous O₂ extraction and O₂ were unchanged.

Series B

Because this model responded to stepwise decreases in total O₂ transport with parallel decrements in function, we evaluated the effect on function of ODC shifts in combination with changes in total O₂ transport. Table 1 shows that arterial blood gases and temperature series in the circuits with control and IHP-loaded cells were virtually identical and close to the physiological range. Although an arterial PO₂ slightly above the physiological level was employed, saturation of IHP-loaded cells, as expected, averaged 90 ± 2% (mean ± SD). P₅₀ averaged 17 ± 1 and 33 ± 5 Torr, respectively (P <>

When perfused at 9 ml/min with control cells, LVS-LVD averaged 84 ± 21 mmHg, work (double product) was 13,173 ± 3,047 mmHg · beats · min¹, mean aortic pressure was 82 ± 28 mmHg, coronary vascular resistance was 42 ± 16 mmHg · ml¹ · min · g, O₂ extraction was 3.1 ± 0.9 ml O₂/dl, and O₂ was 0.062 ± 0.022 ml · min¹ · g¹. These values are shown as 100% in Fig. 3. When perfused with IHP-loaded cells, there were small but significant increases in LVS-LVD (P <>P <>P <>2 extraction (P < src="http://jap.physiology.org/math/12pt/normal/Vdot.gif" alt="V" align="bottom">O₂P <> changes in mean aortic pressure or coronary vascular resistance. (

Fig. 3. Interaction of hemoglobin-O2 affinity and flow rate on LV-developed pressure (A), peak positive change in pressure over time (+dP/dt) (B), LV work (C), arteriovenous (A-V) O2 extraction (D), O2 consumption (O2; E), and coronary vascular resistance (CVR; F) during control red blood cell perfusion (open bars) and high P50 [inositol hexaphosphate (IHP)-loaded] RBC perfusion (solid bars) at various coronary perfusion rates. Differences between control and high P50 perfusion at the same flow rate are as follows: *P <>P <>P <>

When the flow rate was reduced to 3.5 ml/min, simulating ischemia, mean LVS-LVD, +dP/dt, work, aortic pressure, O₂, and coronary vascular resistance decreased significantly, as in series A, whereas O₂ extraction increased; this was true for control and for IHP-loaded cells in relation to their respective controls. Importantly, parameters of function and O₂ delivery improved significantly (P <> vs. control cells (Fig. 3). The increase in LV work and O₂ was sufficient to restore 17 and 20%, respectively, of the decrements due to this degree of ischemia.

Complete ischemia caused further significant decreases in functional parameters. Function during complete ischemia was independent of the type of RBC perfusion (control vs. IHP loaded) preceding the period of ischemia. Upon reperfusion after ischemia, function and O₂ improved significantly. These parameters were somewhat better when reperfusion was carried out with IHP-loaded cells, but the differences from reperfusion with control cells did not attain statistical significance.

Figure 4 depicts in vivo ODCs obtained by plotting arterial and venous O₂ saturation and pressure on perfusate samples obtained under the various conditions described above. These curves redemonstrate the right shift measured in vitro for IHP-containing cells and show that the ODC is less steep. At 9 ml/min, mean venous O₂ saturation was appreciably lower and mean venous PO₂ appreciably higher with the IHP-loaded cells (Table 2, Fig. 4). This same pattern was observed at 3.5 ml/min. Accordingly, arteriovenous O₂ content difference (Fig. 3) was significantly greater during perfusion with IHP-loaded RBC under both conditions and accounted for the significantly greater O₂ observed.

Fig. 4. In vivo O2 dissociation curves during control red blood cell perfusion () and IHP-loaded RBC perfusion (gray squares). Each point is the mean of duplicate arterial or venous samples obtained from individual hearts (series B) during basal or ischemic perfusion. Mean venous O2 saturation/venous PO2 values (venous point) for control cells at the basal perfusion rate (top ) and at the ischemic perfusion rate (bottom ) are shown. , Corresponding values for perfusion with IHP-containing cells.

Table 2. Venous P O2 and O2 saturation