Triiodothyronine Increases Contractility Independent of β-Adrenergic Receptors or Stimulation of Cyclic-3',5'-Adenosine Monophosphate

Male, Sprague-Dawley rats weighing between 250-350 g were used in all experiments. Animals received humane care under a protocol approved by the Animal Care and Use Committee of the Bowman Gray School of Medicine. Our protocol was performed in compliance with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health.* Animals were anesthetized with intraperitoneal pentobarbital (100 mg/kg). The hearts were removed while still beating and were rapidly arrested with iced Dulbecco's phosphate-buffered saline to protect against the obligatory ischemia during instrumentation. A 22-G catheter and a latex balloon were placed into the left ventricle via the left atrial appendage and sutured in place. A 20-G catheter was pushed into the aorta and sutured in place.

The heart was suspended using the Langendorff method, and perfused with Dulbecco's phosphate-buffered saline (37 degrees Celsius) containing CaCl²1.8 mM, MgCl²2.6 mM, and glucose 5.5 mM. A gas mixture of 95% Oxygen²and 5% CO²was bubbled continuously through the buffer solution. The mean ischemic time (from transection of the aorta to perfusion with oxygenated buffer) was 19 min. All hearts began beating immediately when perfusion resumed. After perfusion, the rate of change in pressure (dP/dT) stabilized over a 5-10-min period. Fifteen to 20 min of perfusion was allowed before administration of drugs. The aortic infusion pressure was maintained constant at 90 mmHg and the infusion rate at approximately 250 ml/h. The left-ventricular balloon was filled with 0.15 ml of 0.9% saline. The pressure transducer was zeroed after the balloon was filled with this volume. The end-diastolic pressure did not change during the course of the experiments (13 plus/minus 1 mmHg). All drugs were administered as a single bolus injection of 100 micro liter through the aortic perfusion catheter. When necessary, drugs were diluted in Dulbecco's phosphate-buffered saline. T³was supplied by SmithKline Beecham Pharmaceuticals (Philadelphia, PA), and thyroxine (T⁴) was purchased from Flint Laboratories (Deerfield, IL). The T³was provided in glass vials containing 10 micro gram T³base per milliliter diluent. The diluent also contained citric acid 0.175 mg, NH sub 3 2.19 mg, and benzyl alcohol 6.8% by volume. SmithKline Beecham was unable to supply us with T³-free diluent for study; nevertheless, T³demonstrated efficacy even when diluted to the point that concentrations of these "factory additives" became nominal. Isoproterenol and propranolol were purchased from Elkins-Sinn (Cherry Hill, NJ) and SoloPak (Franklin Park, IL), respectively. All other chemicals were reagent grade and purchased from commercial sources.

Left-ventricular pressure was recorded on paper strips. Maximum dP/dT and maximum -dP/dT were determined using tangents to the pressure-time curve. Drug effects (T³, T⁴, isoproterenol, and propranolol) were compared using the change in maximum dP/dT and maximum -dP/dT from baseline to minimize the potentially confounding effects of variability in baseline contractility between heart preparations (control values are presented in Table 1).

Beta-Adrenergic receptor blockade was induced with propranolol 1 micro mol. After the hearts returned to baseline contractility (5-8 min after propranolol), 0.8 nmol isoproterenol was administered to confirm receptor blockade. T³, 0.74 nmol, was then administered to assess its effect on contractility in the presence of beta blockade. Finally 0.8 nmol isoproterenol was again administered to confirm persistent beta blockade.

Rat hearts (n = 10) were isolated in the same manner as for the contractility studies. Ventricular tissue was minced and suspended in 10 ml Dulbecco's medium with addition of CaCl²0.8 mM and 10 mg collagenase (Sigma Chemical, St. Louis, MO). The tissue was vigorously stirred for 20 min at 37 degrees Celsius. The suspension was filtered through gauze, then centrifuged at 2,000g for 10 min, yielding a pellet of cells. The pellet was resuspended in 2.5 ml Dulbecco's medium, separated into 100-micro liter aliquots, and added to tubes containing varying concentrations of isoproterenol or T³. Fifty microliters isobutylmethylxanthine (10 mM) was added to all reaction tubes to inhibit cAMP hydrolysis. Baseline cAMP was determined in nonstimulated cell suspensions. The cells were incubated with agonists at 37 degrees Celsius for 10 min. Forskolin (10 sup -6 M) was used as a positive control to confirm that the cells were viable. One hundred microliters 3% perchloric acid was added to the tubes to lyse the cells. Five hundred microliters 30% KHCO³was added to neutralize the acid. The tubes were centrifuged for 10 min at 2,000g at 4 degrees Celsius. cAMP was measured in the supernatant using radioimmunoassay from Diagnostic Products (Los Angeles, CA).

Data are presented as means plus/minus SEM. Contractility data are reported as the measured increase in maximum dP/dT (compared with baseline conditions) after drug administration. Lusitropic data are reported as the measured increase in maximum -dP/dT (compared with baseline condition) after drug administration. cAMP is reported as percent increase above baseline after incubation with isoproterenol or T sub 3. Data were analyzed by Wilcoxon's rank-sum test, Fisher's sign test, or doubly repeated-measures analysis of variance, as appropriate. Holms' correction was used if multiple comparisons were made. Except when otherwise stated, P < 0.05 was considered significant.

All drug doses are given in moles, not in concentrations or infusion rates.

The effects of T³, T⁴, and isoproterenol on dP/dT are shown in Figure 1. The mean peak increase in maximum dP/dT with the highest dose of T³(0.74 nmol, n = 6) was 335 plus/minus 38 mmHg/s; whereas the mean peak effect of the highest dose of isoproterenol (0.8 nmol, n = 6) was an increase of 530 plus/minus 110 mmHg/s. Peak effects of T³were measured after a median delay of 15 s; the corresponding median delay for isoproterenol was 60 s; these differences were significant. T³effects on contractility rapidly dissipated over a 60-90-s period after drug administration; the effects of isoproterenol were more sustained, dissipating by 240 s after drug administration. T⁴even as much as 12.9 nmol, produced no significant change in dP/dT (n = 4).

The effects of these drugs on maximum -dP/dT are shown in Figure 2. Maximum -dP/dT was affected similarly to maximum dP/dT. The peak effect with the highest dose of T³was - 227 plus/minus 42 mmHg, whereas the peak effect of the highest dose of isoproterenol was -393 plus/minus 52 mmHg. Time course of these peak effects corresponded to the same time course as for maximum dP/dT. There was no significant effect of T⁴on maximum - dP/dT. No significant change in heart rate was produced by any of these drugs over the dose ranges studied (Figure 3).

The effects of isoproterenol and T³were assessed after pretreatment with 1 micro mol propranolol. Propranolol caused a transient (30-90 s) asystole, with return to baseline contractility after 5-8 min. Ten minutes after propranolol treatment, when the heart had returned to baseline contractility and heart rate, isoproterenol (0.8 nmol, n = 6) was administered to verify the adequacy of beta-adrenergic receptor blockade (the peak increase in maximum dP/dT when isoproterenol was given after propranolol was only 40 plus/minus 20 mmHg/s) (Figure 4). Five minutes after isoproterenol, T³was administered. The peak increases in maximum dP/dT after T³were 390 plus/minus 68, 155 plus/minus 44, and 132 plus/minus 27 mmHg/s. respectively, for 0.74, 0.30, and 0.15 nmol, increases comparable to those observed in the absence of beta-adrenergic receptor blockade, which were 393 plus/minus 29, 207 plus/minus 55, and 103 plus/minus 36 mmHg/s, respectively (Figure 1and Figure 4). After dissipation of the T³effect and return to baseline, isoproterenol (0.8 nmol, n = 6) was again administered to verify the persistence of beta-adrenergic receptor blockade (Figure 4). After the second exposure to isoproterenol, contractility increased maximally by only 40 plus/minus 20 mmHg/s. To summarize, after treatment with propranolol, isoproterenol did not increase dP/dT; however, propranolol had no effect on the contractile response to T³(Figure 4).

Production of cAMP was measured after incubation with concentrations of T³or isoproterenol that increased contractility (Figure 5) (n = 10 separate heart preparations). Isoproterenol significantly increased cAMP generation at does greater than or equal to 10 sup -7 M. However, T³had no effect on cAMP production even at doses as high as 10 sup -6 M. The differences between isoproterenol and T³were statistically significant.

Our results confirm that T³, but not T⁴, rapidly increases the contractility of isolated rat hearts. This T³effect, unlike that of isoproterenol, does not appear to require interaction with the beta-adrenergic receptor or stimulation of cAMP production. The mechanism by which T³produces these effects remains unclear.

T³has both direct and indirect effects on the cardiovascular system. [12]We used as isolated heart preparation to eliminate indirect effects of T³on myocardial performance, T³has both chronic and acute categories. Chronic effects of T³may relate to protein synthesis and interaction of T³with nuclear proteins; acute effects may result from direct actions of T³, on integral membrane proteins, calcium uptake or calcium release from the sarcoplasmic reticulum, effects on mitochondrial function, or some other interaction with a cellular enzyme. In our study and in others, cardiac function improved acutely after T³administration, long before protein synthesis and changes in contractile proteins could occur and even faster than the effects of beta-adrenergic stimulation occurred. [13].

Alterations in beta-adrenergic receptor density on lymphocytes (which is commonly used as a marker for changes in the myocardial beta-adrenergic receptor complement) parallel T³concentrations during hypothyroidism and thyroid hormone replacement. [8]However, these alterations require days to develop. Production of new beta-adrenergic receptors on myocardial cells is unlikely to explain the acute inotropic effects of T³that we observed in our study or the acute improvement in myocardial function after cardiopulmonary bypass, [5,7]although increased beta-adrenergic receptor numbers or improved receptor coupling may play a role after more prolonged exposure. [13-18].

Seppet et al. have demonstrated an effect of thyroid hormone on the activities of sarcolemmal Calcium²+ -stimulated adenosine triphosphatase, ATP-dependent Calcium²+ uptake, ouabain-sensitive Sodium sup + -Potassium sup + adenosine triphosphatase, and Magnesium sup 2+ adenosine triphosphatase. [19]These studies were done in cells exposed to prolonged periods of thyroid hormone concentration changes, but the effects on these enzymatic functions may occur more rapidly and may be responsible for the acute alteration in contractility seen in this model. Dudley and Baumgarten have reported acute effects of thyroid hormone on Sodium sup + channel function. [20]Further study needs to be done to determine the acute effect of thyroid hormone on these channel functions because these changes may be responsible for the acute contractile changes demonstrated in our model.

Direct effects of T³on mitochondrial function, demonstrated by others, may have contributed to the acute increase in contractility demonstrated in our model. [21-23]Increased high-energy phosphate production may be responsible for the acute increase in contractility. Sterling has demonstrated direct effects of T sub 3 on adenine nucleotide translocase on the inner mitochondrial membrane. [23]T³effects on mitochondria, adenine nucleotide translocase, and energy substrate availability may have particular importance after ischemic insults such as aortic cross-clamping or, in our model, from the ischemic insult incurred during excision of the hearts. During ischemia, depletion of high-energy substrates may result in impaired contractility. Administration of T³and reperfusion with metabolic subtrates may accelerate recovery of myocardial contractility and decrease tissue lactate concentrations. [24]These effects in mitochondria occur rapidly and do not require protein synthesis, similar to the effects seen in our experiments. [22]Thus, T³actions on mitochondria may play a role in the acute improvement in contractile function of myocardial tissue alone or in conjunction with other mechanisms.

Maximum dP/dT may be influenced by changes in preload or left-ventricular end-diastolic volume in vivo. [25]We controlled preload using a constant-volume, left-ventricular balloon. The baseline or diastolic pressure from this volume did not change during the experiments. Changes in left-ventricular pressure correlate directly with changes in maximum dP/dT, and therefore, are not reported separately. Because a constant-volume left-ventricular balloon catheter was used, afterload was also constant. Thus, changes in maximum dP/dT measured by this model directly relate to changes in inotropy.

Heart rate did not change significantly after any drug treatment in our study. A lack of heart rate change with thyroid hormone has been reported previously [26]and may be related to the dual blood supply of the rat heart [27]: the atria are perfused by cardiomediastinal arteries branching from the internal mammary artery, rather than by the coronary arteries. When we perfuse isolated rat hearts with an aortic cannula, the atria may not receive adequate perfusion or complete exposure to the drugs. Inadequate perfusion leading to ischemia of the sinoatrial node could make the sinus node rate less sensitive to drug effects. Finally, these isolated hearts receive no basal stimulation from circulating catecholamines or sympathetic nerve endings, possibly also contributing to the lack of effect on heart rate of T³, T⁴, and isoproterenol.

The lack of response to T⁴in our experiments is consistent with T⁴being an inactive precursor of T³. [12]Moreover, the heart lacks a membrane transport system for T⁴. Thus, T³appears to mediate the cardiac actions of thyroid hormone. During hypothermic cardiopulmonary bypass, there is decreased conversion of T⁴to T³in tissue, and as a consequence, a relative decrease in intracellular T³. This may account for the decreased contractility commonly observed in patients separating from cardiopulmonary bypass. [2-7].

An increase in left-ventricular function after T³administration has been reported in stunned, but not in normal hearts. [28]The hearts in our model more closely resemble stunned than normal hearts, because a period of ischemia is inevitable during instrumentation. Isolated, perfused hearts have no available endogenous thyroid hormone. The response of intact hearts (in vivo) to administered T³(e,g., in the setting of reduced blood concentrations of thyroid hormone after cardiopulmonary bypass), may be less than what we measured in isolated hearts (not exposed to thyroid hormone at baseline). Furthermore, thyroid hormone may be a less effective inotrope when administered to patients with normal circulating thyroid hormone concentrations.

Diastolic function is also altered by T³, as demonstrated by the changes in maximum -dP/dT we report. In hypothyroidism the diastolic relaxation time is prolonged [29]; in hyperthyroidism the isovolumic diastolic relaxation time is significantly shorter. [30]Nevertheless, because the T³effects we observed appeared and abated so quickly, they may not have a common mechanism with the chronic actions of thyroid hormone. Because a constant-volume balloon was placed in the left ventricle and diastolic pressure was unchanged over time, alterations in left ventricular end-diastolic pressure could not he used as a measure of diastolic function. However, in our study, maximum -dP/dT, an indirect measure of diastolic relaxation, demonstrated acute alterations similar in time course to maximum positive dP/dT changes. This suggests diastolic relaxation function, like systolic function, may improve after T³without requirement for protein synthesis. This data also suggests that an independent thyroid action on diastolic relaxation does not occur.

We recognize that our study inevitably has limitations, chief among which may be the differences between our model and the clinical settings in which T³and beta-adrenergic agonists are likely to be given. As we have noted previously, there are species related differences between humans and rats in cardiac blood supply. The relative distributions of beta¹- and beta²-adrenergic receptors in heart tissue varies among species and in different clinical conditions [31,32]. Despite these cooncerns, we doubt that our conclusions regarding the relative efficacies of T³and T⁴, and regarding T³'s mode of action are appropriate only for isolated rat hearts.

It is possible that the physical properties of the left ventricle may have been altered by the drugs we administered. Drug-mediated coronary vasodilation could have modified the dP/dT we measured. We assessed drug contractile effects at only one left-ventricular end-diastolic pressure rather than over a range of pressures (i.e., we did not obtain a true Frank-Starling curve). It is also possible, but unlikely, that the drugs we studied could manifest different effects at doses other than the ones we studied. It is also problematic to relate the drug dose we used to concentrations in other studies because, in each case, we administered bolus doses, not infusions of the agents, so that we could observe the time course of the drug effects.

Additional human and animal studies of T³may better define the uses of T³as a positive inotropic agent. The positive inotropic efficacy of T³, independent of any effect on beta-adrenergic receptors or cAMP, separates T³from many other intravenous inotropes, and may provide clinicians the opportunity to combine T³with other drugs (with different mechanisms) to produce a greater aggregate response. In addition, we have not overlooked the potential use of T³for the treatment of beta-adrenergic antagonist overdose.

In normal patients, blood concentrations of total and free T sub 3 range from 1.5-3 nM and 15-40 pM, respectively. [33]The T³concentrations we used (up to 10 sup -6 M) should be sufficient to activate cAMP production in the isolated myocyte, if such an action were present; however, no effect was seen. Moreover, our concentrations were far in excess of those required (1.5-3 nM in an oxygenated, protein-free perfusate) to increase contractility in isolated hearts above that seen in the absence of T³. [14].

Studies in patients are also needed to define the role of T sub 3 after cardiopulmonary bypass, and whether T³may interact with other inotropes to produce additive, or even synergistic effects. Studies must also determine whether T³has major actions on blood flow through normal, diseased, or bypassed coronary arteries before the drug can be used with confidence in patients with coronary artery disease. A better understanding of the mechanism of T³'s acute inotropic effects may lead to a better understanding of normal regulation of myocardial contractility and to more effective resuscitative efforts in patients.

The authors thank Kimberly Ward Black for her diligence in performing many of the assays reported in this work.

*Guide for Care and Use of Laboratory Animals. Publications 85-23. Bethesda, MD, Public Health Services. National Institutes of Health, revised 1985.