| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 1998;66:311-312
© 1998 The Society of Thoracic Surgeons
a Section of Cardiovascular Surgery, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA
Address reprint requests to Dr Schaff, Section of Cardiovascular Surgery, Mayo Clinic, 200 First St SW, Rochester, MN 55905
In vitro experimentation has a vital but focused role in cardiac surgical research. By allowing exacting control of many variables, in vitro studies permit exploration of pathways and mechanisms by which cellular or subcellular processes produce local or systemic responses in vivo. For example, protamine infusion in patients decreases peripheral vascular resistance with resultant hypotension [1], and in vitro studies demonstrate that protamine causes the release of the endogenous nitrovasodilator nitric oxide for the arterial endothelium [2].
By their very nature, in vitro experiments allow for many controls and extensive characterization of a measured observation. The Achilles heel of such studies is the possibility of documenting exciting findings that have little or no clinical relevance outside of the organ chamber or perfusion circuit. Separation of important versus irrelevant observations can be difficult, particularly if the strengths of in vitro experimentation (ie, multiple controls and many observations) are not exploited.
The article in this issue by Chardigny and associates [3] highlights some of these issues. Chardigny and associates seek to determine if exposure to cardioplegia causes functional intimal damage and constriction of internal mammary artery grafts, but the design of the experiments does not allow a definitive answer to this question. The basis of their conclusions is that after exposure to cardioplegic solutions, vascular segments have varying impairment in relaxation to a single concentration of acetylcholine (10-7 mol/L). But is this an indication of global endothelial injury, or is it simply a pharmacologic artifact? Chardigny and associates implicate impaired nitric oxide production by the endothelium, but we are not given results of parallel studies of rings with and without endothelium. Additionally, experiments were not performed with an inhibitor of cyclooxygenase, such as indomethacin. Thus, in vitro prostanoid production by the endothelium and vascular smooth muscle could confound the results. Further, response to the agonist acetylcholine is the only measure of endothelial function, and only one concentration of the compound is used, but concentration-response curves were used for norepinephrine.
Concentration-response curves are very useful because they demonstrate the sensitivity of the tissue for the drug and the maximal response that the tissue can express to the drug. This allows more definitive and complete inferences about the altered relaxation observed after intimal injury. It is possible that the arteries in the study by Chardigny and associates [3] relaxed completely to acetylcholine, but at a higher concentration (a shift in the concentration-response curve to the right). In our initial studies [4] in the human internal mammary artery, the threshold for endothelium-dependent vasodilation to acetylcholine was between 10-8 and 10-7 mol/L. It is not unreasonable to conclude that the internal mammary arteries in Chardigny and associates’ experiment could have relaxed completely at a concentration of 10-6 mol/L.
Because only a single agonist was used to examine endothelium-dependent vasodilation, one cannot be sure that the impaired relaxation was caused by an idiosyncratic change in the cell receptor affinity to acetylcholine (with no real physiologic significance) or true impairment in the ability of the endothelial cells to synthesize and release endothelium-derived relaxing factor (EDRF). Two additional experiments would answer this question. First, perform concentration-response curves to another receptor-dependent mediator of endothelium-dependent vasodilation, such as adenosine diphosphate, a physiologically important platelet-derived compound. Second, perform similar experiments using a receptor-independent mediator of EDRF release such as the calcium ionophore A23187. These additional data will clarify whether there is a significant and general impairment in receptor-mediated release of EDRF and whether the endothelial cell retains the ability to generate EDRF independent of receptor dysfunction.
Chardigny and associates also suggest that vascular smooth muscle relaxes normally after exposure to cardioplegia exposure, but again, only a single, very high (10-5 mol/L) concentration of one drug (sodium nitroprusside) was employed. Indeed, in our initial experiments on the human internal mammary artery, the onset of vasodilation to sodium nitroprusside occurred at a concentration between 10-8 and 10-7 mol/L, a concentration 100 to 1,000 times less than that used by Chardigny and associates [3]. Thus, it is possible that there is impaired vascular smooth muscle function (ie, a shift in the concentration-response curve to the right) in the preparation used by Chardigny and associates, but it is masked because of the high concentration of the drug used. For completeness, another smooth muscle vasodilator, such as isoproteronol, should be employed to determine if smooth muscle vasodilation was indeed unaltered, and from these data, Chardigny and associates could tell us if cyclic guanosine monophosphate-mediated relaxation (sodium nitroprusside) and cyclic adenosine monophosphate-mediated relaxation (isoproterenol) were maintained or impaired.
Finally, Chardigny and associates appear to have made a cardinal error when examining endothelium-dependent vasodilation. Before exposure to acetylcholine, rings were precontracted with norepinephrine, which, by their own data, has an impaired ability to induce contraction in two of the experimental groups. Indeed, after exposure to University of Wisconsin solution or the Broussais solution, vascular segments contract to only 10% and 50% of their previous maximal tension, respectively (see their Figure 5). In addition, the dose of norepinephrine used (10-7 mol/L) is on the steep portion of the concentration-response curve. Thus, vascular relaxation was compared between groups that had different initial vasoconstriction. This problem may have been prevented by precontracting with compounds such as prostaglandin F2
or endothelin, which do not have altered vasoreactivity after cardioplegic exposure.
One must also be careful of the "spin" one places on data. Chardigny and associates, who have an interest in endothelial cell physiology, suggest that high-potassium cardioplegia is harmful because it impairs endothelium-mediated vasodilation. However, a research group interested in adrenergic receptors and vascular smooth muscle might prefer to entitle the report "High-Potassium Cardioplegia Attenuates Norepinephrine-Induced Vasoconstriction—Protection Against Postoperative Catecholamine-Mediated Vasospasm."
In the coronary artery, either hypothermic or normothermic exposure to commercially available preservation solutions with a potassium ion concentration from 45 mEq/L [5] to 140 mEq/L [6] does not irreversibly impair release of EDRF or alter function of vascular smooth muscle. Indeed, after treatment with either normothermic (37°C) or hypothermic (7°C) hyperkalemic cardioplegia, endothelium-dependent vasodilation to acetylcholine or the platelet-derived compound adenosine diphosphate is maintained [5, 6]. Also, endothelium-dependent vasodilation to sodium fluoride, which promotes EDRF release by acting on a pertussis toxin-sensitive G protein [7], is maintained. The ability of vascular smooth muscle to contract (in response to voltage-dependent potassium ions or to receptor-dependent prostaglandin F2a) or relax (in response to sodium nitroprusside, which is mediated by cyclic guanosine monophosphate, or isoproteronol, which is mediated by cyclic adenosine monophosphate) is unaffected by either hypothermia or cardioplegic exposure [5, 6].
The worry that hyperkalemia causes vascular injury is an interesting one. When EDRF was first discovered, Furchgott and Zawadzki [8] demonstrated that high-potassium solutions impair endothelium-dependent relaxation. However, this impairment is caused by potassium-mediated endothelial cell depolarization and is rapidly reversed with washout of the hyperkalemic solution. The fact that hyperkalemia depolarizes cells is the exact characteristic that allows surgeons to use potassium ions to produce diastolic myocardial arrest during cardiac operations. Potassium-mediated depolarization also induces constriction of vascular smooth muscle. In their present study, Chardigny and associates demonstrate vasoconstriction (140 mEq/L potassium ions) of the internal mammary artery to University of Wisconsin solution, a finding previously observed in the coronary artery [6]. Such hyperkalemia-mediated vasoconstriction has important implications in the model one uses to study preservation solutions. Some investigators have used constant-flow perfusion systems to evaluate the vascular effect of preservation solutions [9, 10]. With such techniques, an isolated heart is perfused at a constant flow, after which vascular reactivity is studied. As the present study [3] and previous reports [6] indicate, constant infusion of hyperkalemic cardioplegia in such a model could increase vascular tone (vasoconstriction) and secondarily increase pressure and shear stress in the constant-flow circuit. Both shear stress and excessive pressure cause endothelial cell dysfunction [11–14]. Thus, one could erroneously conclude hyperkalemic cardioplegia is toxic to the endothelium, whereas the mechanism is that of barotrauma in the constant perfusion circuit.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
