The purpose of this study was to clarify the metabolic events that explain why supplemental glucose is detrimental during cardiac storage. Four solutions were used to flush and store porcine hearts: St. Thomas Hospital Solution (STHS), University of Wisconsin (UW) solution, and UW + 90 mM histidine, and UW + 90 mM histidine + 11 mM glucose. Despite equivalent increases in lactate in the two histidine-buffered groups throughout 10 h of storage, glycogen utilization was evident in the group without supplemental glucose. The presence of glucose resulted in a reduction in energy production, presumably mediated by direct inhibition of glycogenolysis. Furthermore, UW + histidine was the only group to show consistent improvements in ATP and total adenylates. It was concluded that inclusion of the buffering agent, histidine, to UW solution promotes anaerobic energy production as a result, in part, of preserved high levels of the regulatory control enzyme, phosphofructokinase.
Current clinical techniques for the preservation of cardiac allografts result in the stoppage of both oxygen and nutrient supplies; consequently, the cumulative injury that occurs is, at present, unavoidable. The practical result of the progressive damage to the graft is that hearts can be safely stored for only 4–6 h before the risk of primary nonfunction becomes prohibitive to a successful transplant operation (1,2).
Preservation solutions play a key role in minimizing the detrimental effects of ischemia and maximizing the safe storage time of the organ. This was highlighted by introduction of the University of Wisconsin (UW) solution in 1987. Developed by Belzer and Southard, this solution has more than doubled the safe storage times of liver, kidney, and pancreas (3). Although this solution was not widely adopted for cardiac preservation, as it was with intra-abdominal organs, it served to demonstrate that the key to successful development of preservation solutions is to understand the patho-physiological changes that occur during ischemia, and then to formulate a solution to counteract these changes. Although the problems associated with low-temperature organ preservation are multifaceted, one important issue is the exhaustion of high-energy adenylates during cold storage. Previous investigators have shown that depletion of tissue energy stores [ATP, PCr, total adenylates (TAs)] are associated with poor recovery of the graft upon reperfusion (2,4).
Levels of high-energy molecules in the cold-stored myocardium reflect a delicate balance between energy production and energy utilization. In general, hypothermia is globally applied to stored organs in the clinical setting in order to reduce the rate of energy-consuming reactions. The current study addresses the aspect of energy production in this delicate balance. In the anaerobic state, glycolysis becomes the primary source of energy production (5). In terms of energy yield, glucose metabolism through this pathway is inefficient, producing only 5–8% (depending upon the source of glucose) of the ATP produced by oxidative metabolism (6). Consequently, it is vital to ensure a brief storage period in order to avoid the potentially lethal complications of the low energy state that accompanies extended periods of ischemia.
A number of studies have attempted to improve cardiac preservation and recovery from prolonged storage by supplementing the preservation solution with glucose added as exogenous substrate for purposes of energy production (7–9). Unexpectedly, these metabolic and functional studies document a consistent paradoxical decrease of cardiac quality following treatment with relatively low concentrations of glucose (11 mM). The present study clarifies the underlying metabolic events that explain why glucose supplementation is detrimental to cardiac storage.
Materials and Methods
Biochemicals and animals
All chemicals were of analytical reagent (AR) grade and were purchased from Sigma Chemical Co. (Oakville, Ontario, Canada) or BDH Inc. (Toronto, Canada). Adult Landrace-Yorkshire pigs (35–40 kg) obtained from the University of Alberta swine farm were used as cardiac donors. All animals were treated humanely in accordance with the regulations set forth by the Canadian Council on Animal Care (Animal Protection Act).
Pigs used in these experiments were assigned to one of four experimental groups (n = 4–6 for each group). The groups differed only in the solution used to flush and store the hearts, with other factors being held constant. St. Thomas Hospital solution II (STHS) served as the control group because of its common use in cardiac surgery and because its composition is a simple saline formulation. The solutions used for each of the experimental groups were adjusted to a pH of 7.4 and were as follows:
Group 1 (control). St. Thomas Hospital solution II (STHS) containing (mM): sodium chloride, 110; potassium chloride, 16; magnesium chloride, 16; calcium chloride, 1.2; and sodium bicarbonate, 10.
Group 2 (UW). UW solution modified to contain (mM): lactobionate, 90; raffinose, 25; potassium hydroxide, 90; sodium hydroxide, 25; magnesium sulfate, 5; potassium phosphate, 25; adenosine, 5; and mannitol, 90.
Group 3 (HIS). Group 2 plus histidine, 90 mM.
Group 4 (HG). Group 3 plus glucose 11 mM. Insulin 100 IU/L was added to facilitate glucose uptake.
Surgical procedure for cardiectomy
Pigs were anesthetized with ketamine (87.5–100 mg/kg IM). The animal's heart rate and oxygen saturation were continuously monitored with a Nellcor N180 pulse-oximeter to ensure an adequate oxygen saturation (95–100%). A median sternotomy was performed and the animal was mechanically ventilated once the chest was opened. The inferior vena-cava (IVC) and superior vena-cava (SVC) were dissected circumferentially so that ligatures could be placed around them; the ligatures were not tied at this time. The pericardium was opened and the right atrial appendage was retracted to expose the aortic root. The adventitial layer was sharply removed and the aorta was cannulated. Once cannulation was complete the IVC and SVC were ligated with the previously placed ligatures, and the heart emptied by allowing it to beat for several cycles. The aorta was then cross-clamped distal to the cannula and 350 mL (0.01 L/kg) of experimental solution (previously cooled to 4 °C and oxygenated with 100% O2) was infused at a pressure head of approximately 100 cm H2O. The time needed to infuse this volume of preservation solution was approximately 1–2 min. The heart arrested within several seconds of the start of the infusion and very quickly there was a noticeable blanching of the myocardium. Topical ice slush was not employed during delivery of the flush solution; attempts were made to minimize ischemic time and to flush, excise and transfer the heart to the perfusion apparatus as quickly as possible. A small incision was made in the right atrium to allow drainage of the effluent. As the heart was being perfused the apex was sharply excised and immediately ‘snap frozen’ in liquid nitrogen using Wollenberger clamps. This sample was designated the initial t = 0 sample. When the infusion was complete, the heart was rapidly excised and was immediately transferred to a recirculating perfusion apparatus (Figure 1) in a 4 °C incubator. Hearts were continuously perfused (+O2) for 1 h at a pressure of 100 cm H2O with the same solution used to flush the hearts. At the end of perfusion, the hearts were removed from the apparatus and stored in their respective flush solutions at 4 °C. Additional transmural tissue samples from the left ventricle were taken at 1, 2, 4, and 10 h and were snap frozen. All samples were stored at −65 °C until processed.
Titration of preservation solutions
The buffering capacity and characteristics were assessed for each preservation solution via the sequential addition of small amounts of 1.0 M HCl to 100 mL of preservation solution. After the addition of each aliquot of HCl, the pH was measured following a 1–2-min period of equilibration. The unit volume of HCl was translated to mmoles H+ per liter of preservation solution.
Sample preparation and metabolite assays
Tissue samples homogenized in 6% perchloric acid (containing 1 mM EDTA); a volume of the well-mixed homogenate was then removed and added to 100 mM Na-acetate for subsequent determination of glycogen contents. Precipitated protein in the remaining homogenate was removed by centrifugation (15 min at 3000 g). Protein was assessed for determination of wet/dry weight ratios. The neutralized extracts were used immediately for assays of phosphocreatine (PCr), adenosine triphosphate (ATP), adenosine diphosphate (ADP), and adenosine monophosphate (AMP). Extraction procedure and metabolite assays were conducted as described previously (10,11). Metabolites were assayed enzymatically based on the absorbance of NADH at 340 nm.
Measurement of key enzymes, glycogen phosphorylase, phosphofructokinase and isocitrate dehydrogenase, were performed as described previously (11).
Data are reported as means ± SEM; n = 4–6. Metabolite levels were reported in terms of µmol/g wet weight. Changes in metabolite levels were assessed at each time point compared with the control group using an anova followed by Dunnett's post hoc comparison test; p < 0.05 was reported.
Tissue water content
Calculation of water contents revealed no significant changes over 10 h of cold storage within or between groups. Average water content for all groups was 65.3% ± 0.5.
Titration of experimental preservation solutions (Figure 2). The decline in solution pH values for STHS, UW and HIS/HG solutions followed a near-linear trend with a slight curve in all lines. The HIS and HG solutions differed only in the addition of glucose and insulin, hence buffering characteristics followed identical curves. Quantities of H+ added to each experimental solution facilitating a 1.0 pH unit drop from 7.4 to 6.4 were 3.3, 12.6 and 40.0 mmol/L.
ATP (Figure 3). Initial values ranged from 10.8 to 12.5 µmol/g in all groups. Out of all groups, HIS showed the greatest difference from the control (STHS); values were significantly higher at 2, 4, and 10 h (p < 0.05). By the end of the 10-h period, only the HIS group showed a significant increase in ATP levels compared with initial values; a final tissue concentration of 18.1 µmol/g corresponded to a 63% increase relative to initial ATP levels (p < 0.01). In contrast, the UW and HG groups were unable to demonstrate superior levels of ATP past 2 h. In summary, beyond 2 h of ischemic cold storage, the HIS solution was the only experimental solution demonstrated to be clearly superior to STHS in maintaining tissue ATP.
Energy charge (Figure 3). Atkinson (12) described another useful measure of tissue energetics, ‘Energy Charge’ as:
In illustrating the significance of this measurement, Pegg uses the analogy of a battery in describing the Energy Charge as being a measure of the ‘charged up’ state of the adenylate pool (13). As ATP levels decline relative to ADP and AMP levels, EC also drops, indicating that less of the TA pool exists in a form that is immediately available for cellular work. Regulatory enzyme control of the glycolytic pathway also responds to levels of EC, as all three terms in the equation are both positive and negative allosteric effectors; hence in any study of glycolytic energy production, changes in EC values need to be examined. Energy charge values immediately following the myocardial flush (t = 0) ranged from 0.90–0.95 for all groups. Only the HIS group was significantly higher than STHS at t = 0 h. After 1 h of perfusion, the experimental groups had EC values between 0.91 (UW) and 0.93 (HIS and HG groups), and were significantly greater than STHS (0.87, p < 0.05). By the end of the storage period, EC was significantly higher in all experimental groups (0.91–0.93) compared with the control group (0.85, p < 0.05).
Phosphocreatine and TAs (data not shown)
Immediately following the flush (t = 0), phosphocreatine levels in all groups ranged from 7.6 to 8.6 µmol/g. By 4 h, all groups showed a rapid decline ranging from 45% (HIS) to 60% (UW) decreases from initial values (p < 0.05). Over the remaining 6 h of storage, all groups exhibited a further progressive decline, although at a much slower rate than was observed over the first 4 h. Final PCr tissue concentrations at 10 h were between 39% and 48% of initial values in all experimental groups. Mean levels were 2.2, 3.8, and 3.4 µmol/g in the UW, HIS, and HG groups, respectively.
Total adenylates for all groups at t = 0 h ranged from 12.4 to 14.3 µmol/g. After 2 h, the control group exhibited a 33% decline and was significantly lower than all experimental groups at this time point (p < 0.05). The only group that exhibited a consistent elevation above the STHS group was HIS; TA levels were greater than the control at 2, 4, and 10 h of storage (p < 0.05). After 10 h, levels in the HIS group increased to 18.1 µmol/g, representing a 63% increase from t = 0 and a 76% increase over the control group (p < 0. 05).
ADP and AMP (data not shown)
Adenosine diphosphate levels were 2.3 µmol/g in the STHS (control) group and remained unchanged throughout the storage time course. Initial ADP levels in the three experimental groups (UW, HIS, HG) ranged from 1.4 to 1.6 µmol/g. After 10 h, levels in the UW, HIS, and HG groups had risen to a range of 2.0–2.3 µmol/g; not significantly different from the STHS group. Initial levels ranged between 0.11 and 0.18 µmol/g. Within 4 h, levels of this low-energy adenylate rose to a range of 0.25–0.34 µmol/g and remained elevated throughout the rest of the 10-h period of storage.
Glycogen Phosphorylase. Initial values of glycogen phosphorylase ranged between 2.0 (UW) and 2.5 (HIS) for all the groups, with no significant difference from the control (STHS). After 4 h only the HIS group dropped significantly to 1.7 from the initial value. By 10 h, all groups except UW underwent a significant drop from initial values, with the HIS group (1.0) also being statistically different from the control.
Phosphofructokinase (PFK). Initially, upon entry into storage, PFK ranged from 4.0 (STHS) to 15.9 (HIS), with all groups being significantly higher than the control. After 4 h the HIS group dropped significantly from the control to 10.7 and experienced an increase significantly different from the control at 10 h. In contrast, the HG group experienced a significant increase from the control and initial values at both 4 h and 10 h. There was no notable change in the STHS and UW groups during the 10-h period.
Isocitrate dehydrogenase (ICDH). Similarly, at 0 h, levels of ICDH ranged from 12.3 to 14.7 with no significant difference between the control and all the experimental groups. Treatment with HIS resulted in a minor yet gradual decline over 10 h from 14. 7 to 10.7 IU/g. STHS, UW, and HG groups did not exhibit a significant change from the control or initial values throughout the 10 h.
Lactate accumulation provided a quantitative assessment of anaerobic metabolism during the static period of cold storage. Lactate increased in all groups by 4 h of storage, indicating glycolytic activity. By 10 h, lactate accumulation was significantly greater in the HIS and HG groups than in the control; final values were approximately double that of STHS, 22 and 19 µmol/g (for HIS and HG respectively) vs. 10.5 µmol/g (STHS).
Initial myocardial glycogen levels ranged from 34.0 to 39.8 µmol/g. Glycogen levels declined a linear fashion and were statistically significant following 10 h of storage in all groups (p < 0.05). Total glycogen decrease was 5.9 µmol/g in STHS; values for the HG and UW groups were not significantly different from the control (6.2 and 8.4 µmol/g). However, glycogen utilization in HIS was the only experimental group that demonstrated statistically greater glycogen depletion than STHS; values in the HIS group dropped by 13.3 µmol/g (p < 0.05).
The current study investigated several key aspects of anaerobic energy production. These included the effects of solution-buffering agents on energy production, as well as the possible benefit of including glucose as substrate within the flush-storage solution. With respect to buffering of the preservation solution, it was hypothesized that enhanced buffering would augment glycolytic energy production resulting in improved tissue energetics during cold storage. The data presented are consistent with this premise and are supported by a previous study from our lab examining the effects of various buffers on porcine cardiac metabolism (13). The STHS solution utilizes HCO3– as its primary buffer (14), and has a much lower buffering capacity than the UW or histidine-containing solutions (HIS and HG) employed in this study. The UW solution was not found to be superior to STHS in maintaining glycolytic energy production over the 10-h cold ischemic period, as evidenced by a lack of significant differences in lactate accumulation, tissue ATP, and energy charge values. However, when the buffering capacity of the UW solution was increased via the addition of 90 mM histidine, lactate production, ATP, TA, and energy charge were all found to be significantly greater than in hearts flushed and stored with STHS. Previous investigators have shown that a decline in cellular ATP levels is associated with poor myocardial performance upon reperfusion (2,11). The mechanism behind this finding is that energy is needed by the cardiomyocyte for excitation-contraction coupling, and for fueling the membrane ion pumps that regulate normal intracellular ionic homeostasis (12). This suggests that the injury sustained by the myocardium is at least in part related to the depletion of high-energy molecules (ATP, PCr) during cold storage (15). It follows therefore that an important aspect of any flush-storage solution should be its ability to preserve myocardial energetics during ischemia. It must be pointed out that the model used in this study does not directly parallel that of the clinic. Times of flushing, topical cooling with ice-slush (which was not employed in this study), and additional oxygenated perfusion for 1 h are all confounding variables that make direct application to clinical practice impossible. However, the model employed in this study was designed/chosen for its relative simplicity, and the brief perfusion period was necessary in order to prove a biochemical point.
The rationale behind the hypothesis that increased preservation solution buffering is beneficial in maintaining tissue energy levels comes from understanding the regulatory control mechanisms of anaerobic metabolism. Three key regulatory enzymes control flux through the glycolytic pathway: glycogen phosphorylase, phosphofructokinase (PFK), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (6). Of these enzymes, PFK is particularly sensitive to pH changes (16,17). Under conditions of low oxygen (hypoxia or anoxia), proton accumulation can be quite rapid, as demonstrated by Anderson et al. (18) who found that intracellular pH fell to 6.6 within 5 min of the onset of warm ischemia in rabbit myocardium. Under conditions of hypothermic storage, the drop in intracellular pH occurs at a slower rate however, the magnitude of the decline is still sufficient to facilitate a reduction in PFK activity (19). Glycolysis is therefore self-limiting in the setting of cold organ storage as a direct consequence of proton accumulation; low pH induces a reversible hysteretic loss of PFK activity (17). By providing intracellular buffers, as proposed by Bretschneider in 1975 (20), the means are available to ‘soak up’ excess protons and lessen the pH decline that results in an inhibition of glycolytic energy production. Not only were ‘energy charge’ values elevated immediately following the initial flush, demonstrating an immediate effect of histidine, activities of the primary regulatory control enzyme, PFK, were significantly elevated in these groups as well. This data clearly support the importance of the role of improving tissue energetics via the inclusion of supplemental buffering agents (13).
Evidence that the differences found in tissue energetics between the histidine-based solution and control were secondary to enhanced anaerobic metabolism was provided by the accumulation of lactate. Although, direct correlation between lactate and intracellular pH cannot be determined, it is clear that there is a positive correlation between buffering capacity and lactate accumulation throughout storage. Following the perfusion period, lactate increased linearly in all groups over time, consistent with the transition from aerobic to anaerobic metabolism. Total lactate accumulation was significantly higher in both histidine-buffered groups (HIS and HG groups) compared with STHS at the conclusion of the storage period, thus providing confirmation that glycolytic activity can be augmented by improved buffering of flush/storage solutions. Although there is a clear benefit to supplementing UW solution with an effective buffering agent during storage, whether or not there would be any maintenance of these benefits upon reperfusion remains speculative. The clearance of end products via lactate/H+ specific pumps can be influenced by alterations in intracellular or extracellular pH, hence removal of these potentially toxic compounds maybe facilitated by a similar buffering agent for a brief period of reperfusion prior to transplantation.
Interestingly, the benefit of solution buffering on maintaining ATP levels was lost with the addition of glucose (11 mmol/L). At the conclusion of 10 h of storage no difference in ATP was found between the HG and STHS groups despite the fact that the HG solution differed from the histidine solution only by the presence of glucose (plus insulin to facilitate glucose uptake). This finding is consistent with that of Hearse (7) who found that adding glucose to STHS resulted in a dose-dependent reduction in cardio-protection, and that insulin exacerbated this detrimental effect. Not only is glucose an allosteric inhibitor of glycogen phosphorylase, but also glucose forms a complex with phosphorylase ‘a’ (the active form of the enzyme), which provides a preferred substrate for phosphorylase phosphatase. The net result is an inactivation of glycogen phosphorylase in the presence of high glucose. Hence, even though there was no significant reduction in the phosphorylated (active) form of glycogen phosphorylase over the time frame of the current study (and at low temperatures), allosteric inhibition from glucose was clearly responsible for the reduction in glycogenolytic rate. This type of feedback inhibition is typical of this enzyme locus.
Of particular importance with respect to the ineffectiveness of supplemental glucose in promoting energy maintenance during prolonged static storage lies in the relative imbalance in glycolytic ATP production when exogenous glucose vs. endogenous glycogen are utilized as metabolic substrate. The yield of ATP from glycolysis when glucose is metabolized is 50% less than when endogenous glycogen is catabolized; phosphorylation at the hexokinase/glucokinase step requires one additional ATP for entry of glucose into the cell, and consequently only 2 mol ATP/mol glucose are generated from glucose catabolism compared with three ATP/glucosyl unit from glycogen. Thus, the presence of glucose in the HG solution likely resulted in an overall reduction in total glycolytic ATP production through decreased glycogenolysis and a greater reliance on exogenous glucose as substrate. This explanation is supported by the fact that no significant decrease in glycogen was observed in the group supplemented with glucose despite equivalent increases in anaerobic end-product in both histidine-buffered solutions (HIS and HG groups). In contrast, hearts that relied exclusively on endogenous glycogen for metabolic substrate (HIS group) were the only organs that exhibited a significant decline in glycogen stores over the entire period of static storage (1–10 h post flush) and a superior maintenance of ATP and TA levels.
In summary, the data presented in this communication support the hypothesis that buffering against pH changes during hypothermic ischemia improves myocardial tissue energetics through increased anaerobic ATP production. Enhanced glycolytic activity was demonstrated to occur through significant and equivalent increases in lactate production in myocardium flushed with histidine-buffered solutions. However, there was no metabolic benefit to supplementing cardiac preservation solutions with glucose as metabolic substrate throughout a prolonged period of preservation (up to 10 h). Of even greater importance is negative impact on cardiac energetics that glucose has on redirecting the substrate supply away from endogenous glycogen to exogenous substrate, which is less energetically favorable and has direct implications for energy maintenance throughout cold static storage.
Finally, it must be pointed out that although UW may not have a superior buffering capacity, this is not reason enough to suggest replacing UW solution with a HIS-based preservation solution. The UW solution is an effective preservation solution for other reasons: the inclusion of compounds targeted at maintaining purine carbon stores and limiting oxygen-free radical (OFR) generation during ischemia and perhaps more importantly upon reperfusion. We suggest however, that supplementing the UW solution with an effective (and nontoxic) buffering agent may provide additional benefits to cardiac allografts during static storage. These benefits to improving the energetic status throughout cold ischemia may even be more apparent upon reperfusion, a time when resumption of metabolism is of utmost importance. We anticipate that this will be borne out in future studies that compare UW with a histidine-supplemented UW solution in an experimental model with direct clinical relevance.
This research was supported by operating funds contributed by the Edmonton Civic Employees Charitable Assistance Fund and the Department of Surgery.