All study protocols had been reviewed and had received previous approval from the Mayo Foundation Institutional Animal Care and Use Committee.
Thirty male and female adult mongrel dogs that weighed twenty-three to twenty-eight kilograms were anesthetized with thirty milligrams of pentobarbital per kilogram of body weight and were killed. The tibiae and vessels were isolated as described previously13. Both bones were wrapped in towels that had been moistened with Krebs-Ringer solution and were kept in a refrigerator at 4 degrees Celsius. All procedures were performed under sterile conditions. Ex vivo perfusion was carried out with a previously described technique13.
Flow Rate
The rate of flow through the nutrient artery initially was set at 0.2 milliliter per minute for an hour and then was gradually increased to 0.5, 1.0, 1.5, and 2.0 milliliters per minute at thirty-minute intervals in the tibiae that had been stored at 4 degrees Celsius for one or three days. Because of the increased resistance to perfusion observed in the tibiae that had been stored for longer periods, an initial flow rate was set at 0.2 milliliter per minute for 1.5 hours in the tibiae that had been stored for five or seven days. The flow rate then was gradually increased to 0.5, 1.0, and 1.5 milliliters per minute at thirty-minute intervals.
Selection of the Drugs
The endothelium synthesizes a variety of substances that affect vascular smooth-muscle tone and the coagulation of blood. To evaluate this endothelial eccrine function after varying periods of ischemic injury, we assayed concentrations of endothelin in the perfusate effluent of the tibiae after perfusion with or without adrenomedullin, acetylcholine, NG-monomethyl-L-arginine acetate, indomethacin, norepinephrine, and tetraethylammonium. Endothelin-1,2[125I] was assayed as previously described13. Tetraethylammonium (a potassium-channel blocker) inhibits the activity of endothelium-derived hyperpolarized factor in endothelium-dependent vasodilation24. All drugs were prepared in a manner that has been previously described13.
Experimental Design
In phase 1 of the study, the right and left tibiae from six dogs (twelve paired tibiae) were used in each of four ischemic groups that differed according to the duration of storage (one, three, five, or seven days). The tibiae from an additional six dogs were utilized in phase 2 of the study. The contralateral tibia of each dog served as a control for each experiment. The control tibia received an infusion of Krebs-Ringer solution in place of the particular pharmacological agent that was infused into the experimental bone but was otherwise treated identically.
Phase 1
Vascular endothelial function after storage at 4 degrees Celsius for one, three, five, or seven days: After the baseline perfusion pressure had stabilized at a flow rate of 2.0 milliliters per minute in the tibiae that had been stored for one or three days and at a rate of 1.5 milliliters per minute in the tibiae that had been stored for five or seven days, a standard norepinephrine dose-response curve for doses of sixty-four, 128, and 256 picomoles was generated with use of 0.1-milliliter bolus injections into the perfusate (t1) at the same time in both the experimental and the control tibia (Fig. 1). Then, the experimental tibia was continuously perfused with 0.1 milliliter of 10-5-molar acetylcholine solution per minute and the control tibia, with Krebs-Ringer solution. Five minutes later, a second norepinephrine dose-response curve (T1) was generated with use of the same three doses of norepinephrine as those at t1 (Fig. 1). The perfusion of acetylcholine (and of Krebs-Ringer solution in the control tibia) was then discontinued. Twenty minutes later, a norepinephrine dose-response curve (t2) was generated with use of the same doses as those at t1 and T1. Then, the experimental tibia was continuously perfused with 2 x 10-4-molar NG-monomethyl-L-arginine acetate at a rate of 0.1 milliliter per minute and the control tibia, with Krebs-Ringer solution. Five minutes later, another norepinephrine dose-response curve (T2) was generated with use of the identical doses in both tibiae. This same sequence was then repeated a third time with 2 x 10-4-molar indomethacin solution used in place of the NG-monomethyl-L-arginine acetate in the experimental tibia (t3 and T3). Twenty minutes after generation of the T3 norepinephrine dose-response curve, the perfusate effluent from each tibia was collected for twenty minutes and then stored for later assay of endothelin-1. Finally, 0.1 milliliter of 10-5-molar adrenomedullin solution or Krebs-Ringer solution was infused into the perfusate of each tibia, and the subsequent change in the baseline perfusion pressure was observed during a thirty-minute period. For this step in the experiment, the selection of the experimental and control tibial specimens was altered in order to determine if previous infusion with acetylcholine, NG-monomethyl-L-arginine, or indomethacin had influenced the effect of adrenomedullin on baseline perfusion pressure. Therefore, three tibiae that had been experimental specimens in the previous steps and three that had been controls received an infusion of adrenomedullin. The contralateral tibiae were allocated to the control group and received an infusion of an identical volume of Krebs-Ringer solution.
Phase 2
Effect of tetraethylammonium after storage at 4 degrees Celsius for seven days: In the group of tibiae that had been stored for seven days in phase 1 of the study, perfusion of acetylcholine had attenuated the norepinephrine-induced pressor response, whereas perfusion of NG-monomethyl-L-arginine acetate and indomethacin had had no effect. This observation suggested that acetylcholine induced vascular smooth-muscle relaxation that was unrelated to both nitric oxide and a prostaglandin. Therefore, phase 2 of the study was designed to determine if this observation could be explained by acetylcholine-induced release of endothelium-derived hyperpolarized factor in the tibiae that had been stored for seven days. After the baseline perfusion pressure had stabilized at the maximum flow rate of 1.5 milliliters per minute (as in phase 1 of the study), a standard norepinephrine dose-response curve (t1) was generated for doses of sixty-four, 128, and 256 picomoles with use of 0.1-milliliter bolus injections into the perfusate. Next, one tibia was continuously perfused with 0.1 milliliter of 10-5-molar acetylcholine per minute and the control tibia, with Krebs-Ringer solution at the same rate. Five minutes later, a norepinephrine dose-response curve (T1) was generated with use of the same norepinephrine doses as had been previously used (Fig. 1). The perfusion of acetylcholine was then discontinued, and twenty minutes later a norepinephrine dose-response curve (t2) was generated as previously described (Fig. 1). Then, the experimental bone was perfused continuously with a combination of 10-5-molar acetylcholine solution and 3 x 10-3-molar tetraethylammonium solution at 0.1 milliliter per minute and the control bone, with Krebs-Ringer solution. Five minutes later, a final norepinephrine dose-response curve (T2) was generated with use of the same doses as previously described.
Analysis of the Data
A variety of quantitative parameters of endothelial eccrine function were evaluated.
1. The baseline perfusion pressures were assessed at each flow rate for the four groups.
2. The concentrations of endothelin-1 were determined in phase 1 of the study.
3. The change in each norepinephrine dose-response curve after injection of acetylcholine, NG-monomethyl-L-arginine acetate, or indomethacin (T1, T2, and T3) in the experimental tibiae and after injection of Krebs-Ringer solution in the control tibiae was expressed as a percentage of each initial dose-response curve (t1, t2, and t3) for each tibia. The initial norepinephrine dose-response curves (t1, t2, and t3) were defined as the maximum (100 per cent) norepinephrine-induced smooth-muscle contractile response, and each T1, T2, or T3 norepinephrine dose-response curve was expressed as the percentage that the curve differed from its t1, t2, or t3 counterpart (T1/t1 x 100, T2/t2 x 100, and T3/t3 x 100). The pressor response for each norepinephrine dose-response curve was calculated by adding the pressor responses for the three doses that had been used.
4. The change in baseline perfusion pressure resulting from an injection of 0.1 milliliter of 10-5-molar adrenomedullin solution was compared with that produced by the injection of Krebs-Ringer solution in phase 1 of the study.
5. Norepinephrine dose-response curves (T1 and T2) were expressed as a percentage of the initial dose-response curves (t1 and t2) in phase 2 of the study.
The experimental design involved consistent use of paired tibiae from the same animal for comparison of experimental and control values, and statistical evaluation of the data was performed with use of the paired two-tailed Student t test. P values of less than 0.05 were considered significant.
Baseline Perfusion Pressure
Baseline perfusion pressures at each flow rate were compared among the four groups of tibiae (storage at 4 degrees Celsius for one, three, five, or seven days) (Fig. 2). No pharmacological agents were administered in these studies. The baseline perfusion pressure increased as the duration of storage increased. The differences among the groups were all significant (p < 0.05) at all flow rates except for that between the groups that had been stored for five and seven days when the perfusion pressure was measured at a flow rate of 0.5 milliliter per minute.
Concentrations of Endothelin-1 in the Venous Effluents of the Perfused Tibiae
Any samples in which the level of endothelin-1 was less than 0.5 picogram per milliliter were assigned the value of 0.5 picogram per milliliter for the purpose of statistical analysis because values of less than 0.5 picogram per milliliter cannot be accurately quantitated by the assay technique that was used in this study.
The concentrations of endothelin-1 were a mean (and standard error of the mean) of 0.53 ± 0.08, 0.88 ± 0.07, 8.7 ± 3.2, and 10.8 ± 4.3 picograms per milliliter in the tibiae that had been stored for one, three, five, and seven days, respectively (Fig. 3).
Effect of Acetylcholine, NG-Monomethyl-L-Arginine Acetate, or Indomethacin on the Pressor Response to Norepinephrine
Acetylcholine significantly attenuated the pressor response to norepinephrine in each group compared with the response in the control tibiae (p < 0.05) (Fig. 4-A). Perfusion of NG-monomethyl-L-arginine acetate significantly increased the pressor response to norepinephrine only in the tibiae that had been stored at 4 degrees Celsius for one day (p < 0.05) (Fig. 4-B). Perfusion of indomethacin significantly decreased the pressor response to norepinephrine in the tibiae that had been stored for one, three, or five days (p < 0.05) (Fig. 4-C). The attenuation effect of acetylcholine on the norepinephrine-induced pressor response in the tibiae that had been stored for three days was significantly greater than that of the tibiae that had been stored for five days (p < 0.05). The maximum attenuation effect of indomethacin was observed in the tibiae that had been stored for three days. The data suggest that nitric oxide was produced only in the tibiae that had been stored at 4 degrees Celsius for one day and that prostaglandins produced in the tibiae that had been stored at 4 degrees Celsius in the present study acted predominantly as vascular smooth-muscle constrictors.
Effect of Adrenomedullin on Perfusion Pressure
Thirty minutes after a bolus injection of 0.1 milliliter of 10-5-molar adrenomedullin solution, the mean baseline perfusion pressure decreased by 10.1 ± 1.6, 8.9 ± 1.4, 9.2 ± 3.5, and 9.8 ± 1.8 millimeters of mercury (1.35 ± 0.21, 1.19 ± 0.19, 1.23 ± 0.47, and 1.31 ± 0.24 kilopascals) in the tibiae that had been stored for one, three, five, or seven days, respectively (Fig. 5). These results were significantly different from those in the control specimens for a duration of thirty minutes after the injection of the adrenomedullin (p < 0.05).
Endothelium-Derived Hyperpolarized Factor in the Phase-2 Tibiae That Had Been Stored for Seven Days
Perfusion of 0.1 milliliter of 10-5-molar acetylcholine solution per minute decreased the mean pressor response to norepinephrine to 54.6 ± 3.7 per cent (control value, 96.5 ± 5.6 per cent) (p < 0.05). The addition of 0.1 milliliter of 3 x 10-3-molar tetraethylammonium solution per minute to the perfusate increased the mean pressor response to norepinephrine to 168 ± 20 per cent (control value, 83.6 ± 5.7 per cent) (p < 0.05).
Vascularized bone transfer is a well accepted technique for the reconstruction of large defects in bone after resection of a tumor, débridement for the treatment of osteomyelitis, or traumatic bone loss6,22,25. With improved immunosuppression or immunomodulation techniques, the future use of vascularized whole-bone allografts or xenografts may be a reality for clinical applications. However, for these procedures to be successful, techniques must be developed to maximize the time that these grafts can be stored; preferably, it should be possible to store them for several days. Many methods are currently used for the preservation of organs for transplantation. The most common techniqe is hypothermic storage of tissue that has been perfused with special solutions such as the University of Wisconsin solution10,20 and the Euro-Collins solution3. The benefits of cold storage for prolonging tissue preservation are well known. In the present study, emphasis was placed on understanding the effects of increasing durations of hypothermic storage of bone on vascular smooth-muscle contractility and on endothelial eccrine products, which are modulators of vascular smooth-muscle contractility.
The results of the present study suggest that the storage of tibiae at 4 degrees Celsius for one or three days is associated with low-level production of endothelin-1 (less than one picogram per milliliter). However, storage at 4 degrees Celsius for five days produced a mean (and standard error of the mean) of 8.7 ± 3.2 picograms of endothelin-1 per milliliter, which is approximately a ten-fold increase compared with that produced after three days of storage. There was also a marked increase in the baseline perfusion pressure as the duration of storage increased.
The vascular smooth-muscle relaxing effect of acetylcholine and the vascular smooth-muscle contractile response to norepinephrine were demonstrated in each group in the present study. However, the acetylcholine-induced vascular smooth-muscle relaxation decreased as the duration of storage increased. There was also a significant decrease in the effect of acetylcholine on the norepinephrine-induced pressor response in the tibiae that had been stored at 4 degrees Celsius for five days compared with those that had been stored for three days (p < 0.05). Dysfunction of the endothelial cells is typically associated with an increase in vascular smooth-muscle contractility2.
On the basis of these observations, it is clear that strategies are needed to prevent the deterioration of the normal function of the vascular endothelium during cold ischemic periods exceeding three days. In the tibiae that had been stored for seven days, perfusion of acetylcholine attenuated the norepinephrine-induced pressor response, but perfusion of NG-monomethyl-L-arginine acetate or indomethacin was not found to have a significant effect. This observation suggests that, at this late period, acetylcholine induced the release of a vascular smooth-muscle relaxing agent, but this agent was neither nitric oxide nor a prostaglandin. Another acetylcholine-induced vascular smooth-muscle relaxing agent that has been recently identified is endothelium-derived hyperpolarized factor24. This agent is known to be selectively antagonized by tetraethylammonium24. Tetraethylammonium was therefore perfused, in the presence of acetylcholine, into the tibiae that had been stored for seven days. Perfusion with this combination resulted in a marked increase in the mean pressor response to norepinephrine to 168 ± 20 per cent (control value, 83.6 ± 5.7 per cent), whereas the mean pressor response to norepinephrine in the presence of acetylcholine alone was decreased to 54.6 ± 3.7 per cent (control value, 96.5 ± 5.6 per cent). These results suggest that endothelium-derived hyperpolarized factor is still produced by the vascular endothelium in bones that have been stored for seven days. It is plausible that endothelial synthesis of endothelium-derived hyperpolarized factor may be less susceptible to the effects of ischemia than are other known vascular smooth-muscle relaxing agents of endothelial origin.
Adrenomedullin is known to be one of the most potent vasodilator drugs in vivo15. Adrenomedullin causes a decrease in total peripheral resistance and blood pressure without affecting cardiac output and heart rate7,11. In the present study, a 0.1-milliliter bolus injection of 10-5-molar adrenomedullin solution significantly relaxed the vascular smooth muscle in all four groups of tibiae (p < 0.05). In previous studies, we found that adrenomedullin has a long-lasting vasodilator effect in bone14, even in the absence of the vascular endothelium12, and that it is capable of decreasing the baseline perfusion pressure in bones that have been stored at room temperature for four, eight, or twenty-four hours13. As cardiac output and heart rate are not altered during the marked systemic vasodepressor response to adrenomedullin and because adrenomedullin has a direct mechanism of action on vascular smooth muscle, activation of the adrenomedullin vasodilator mechanism may offer a therapeutic alternative in the clinical management of so-called no-reflow phenomenon, which is known to occur in some clinical states of excessive ischemia.
Endothelin has three known isoforms (endothelin-1, endothelin-2, and endothelin-3). Endothelin-1 is produced by the vascular endothelium and is one of the most potent known vasoconstrictors27. It is also the most important member of the endothelin family found in the peripheral vascular bed. The circulating concentration of endothelin has been demonstrated to be increased in patients who have essential hypertension and atherosclerotic vascular disease as well as in the presence of reperfusion injury9,17. Endothelin is approximately 100 times more potent than noradrenaline on a molar basis in the microvasculature system of the rabbit1, and it has a prolonged mechanism of action26. Endothelin has been reported to be an important factor in the pathogenesis of cerebral vasospasm16. Moreover, the potent vascular smooth-muscle relaxing agent nitric oxide is known to be functionally impaired with prolonged ischemia and in certain situations of reperfusion injury8. In the present study, production of nitric oxide was not demonstrated after three days of cold storage. Moreover, prostaglandins produced in tibiae that had been stored for one, three, or five days demonstrated predominantly vasoconstrictor activity. Thus, the increased production of the potent vasoconstrictor endothelin-1, the loss of the vascular smooth-muscle relaxing activity of prostaglandins, and the decreased production or effectiveness of nitric oxide strongly favor increased vascular tone and vessel spasm16,22.
In conclusion, the results of the present study suggest that the eccrine function of the vascular endothelium in bone is altered in the interval from three to five days of hypothermic storage. When practical, tactics to minimize the ischemic injury of the vascular endothelium should be employed. When ischemia-induced injury of the vascular endothelium is established but not complete, adrenomedullin may be of some value in minimizing the factors that promote sustained vascular smooth-muscle constriction and the resulting elevated peripheral vascular resistance. Production of endothelium-derived hyperpolarized factor was apparently preserved even in the tibiae that had been stored for seven days, suggesting that it is the most ischemia-resistant component among the known vascular smooth-muscle relaxing agents that are produced in the vascular endothelium. Although a variety of vascular smooth-muscle relaxing agents are known to have a relationship to nitric oxide and prostacyclin, the importance of endothelium-derived hyperpolarized factor as a relaxing agent, particularly after ischemia-induced endothelial injury, may merit special attention.