Abstract
Thyroid hormones influence the function of many organs and mediate their diverse actions through two types of thyroid hormone receptors, TRα and TRβ. Little is known about effects of ligands that preferentially interact with the two different TR subtypes. In the current study the comparison of the effects of the novel synthetic TRβ-selective compound GC-1 with T3 at equimolar doses in hypothyroid mice revealed that GC-1 had better triglyceride-lowering and similar cholesterol-lowering effects than T3. T3, but not GC-1, increased heart rate and elevated messenger RNA levels coding for the If channel (HCN2), a cardiac pacemaker that was decreased in hypothyroid mice. T3 had a larger positive inotropic effect than GC-1. T3, but not GC-1, normalized heart and body weights and messenger RNAs of myosin heavy chain α and β and the sarcoplasmic reticulum adenosine triphosphatase (Serca2). Additional dose-response studies in hypercholesteremic rats confirmed the preferential effect of GC-1 on TRβ-mediated parameters by showing a much higher potency to influence cholesterol and TSH than heart rate. The preferred accumulation of GC-1 in the liver vs. the heart probably also contributes to its marked lipid-lowering effect vs. the absent effect on heart rate. These data indicate that GC-1 could represent a prototype for new drugs for the treatment of high lipid levels or obesity.
THE HEART IS an important target for thyroid hormone action (1). T3 increases the heart rate, speed, and force of systolic contraction (2), shortens the duration of diastolic relaxation (3), affects vascular tone (4), and lowers plasma lipid levels (5, 6). A few attempts have been made to develop thyroid hormone analogs that induce a differential thyroid hormone response and, for example, preferentially lower lipid levels but do not increase heart rate (7, 8). Such compounds might have medical utility. In the past, thyroid hormone analogs were designed without a detailed knowledge of specific cardiac effects of T3 receptor isoforms or of the ligand binding pockets of the thyroid hormone receptors (TRs) (7–9). Thyroid hormone actions are mediated by two different types of TRs, TRα and TRβ. Recently, the ligand-binding domains of both TRα and TRβ have been crystallized and structurally characterized in detail (10, 11). In addition, it has become apparent in animal models, in which either TRα or TRβ is deleted, that distinct and differential cardiac effects are exerted by these isoforms. For example, in TRα1-deleted animals marked bradycardia occurs (12, 13). In contrast, in mice with deletion of TRβ there is no lowering of heart rate, but the animals exhibit inner ear deafness (14).
Recently, the novel thyroid hormone analog GC-1 has been synthesized (15). In this analog the three iodines of T3 are replaced by methyl and isopropyl groups. A methylene linkage replaces the biaryl ether linkage between the two phenol groups, and the amino acid side-chain has been changed to an oxyacetic acid group (Fig. 1). GC-1 has approximately a 10 times higher binding affinity to TRβ1 compared with TRα1 and shows TRβ-selective actions in cells in culture. However, it is unclear what influence GC-1 has on T3mediated changes in cardiovascular action and lipid levels in vivo. To explore its in vivo actions, GC-1 was administered to hypothyroid mice and hypercholesteremic rats, and its effects were compared with those of equimolar doses of T3. The results show that significant and distinct differences in T3vs. GC-1 action occur.
Chemical structure of T3 and the thyromimetic compound GC-1.
Chemical structure of T3 and the thyromimetic compound GC-1.
Materials and Methods
GC-1 and T3 effects in hypothyroid mice
Experimental design. All procedures described were performed in accordance with the guidelines of the committee on animal research at the University of California, San Diego. The effects of GC-1 were studied in hypothyroid mice and compared with those of equimolar doses of T3. The amount of T3 (3.5 ng/g BW·day) that was found to return hypothyroid mice to the euthyroid state (16) is defined as 1 × T3. The equimolar amount of GC-1 (1 × GC-1) equals 1.8 ng/g BW·day, calculated from the mol wt of T3 (mol wt, 651) and GC-1 (mol wt, 328.4). Accordingly 4.5 × GC-1 equals 8.0 ng/g BW·day, 9 × GC-1 equals 16.2 ng/g BW·day, 4.5 × T3 equals 15.8 ng/g BW·day, and 9 × T3 equals 31.5 ng/g BW·day. The 1 × dose is equivalent to 5.38 nmol/kg, the 4.5 × dose to 24.67 nmol/kg, and the 9 × dose to 49.33 nmol/kg. The effect of each dose was studied after 4 weeks of daily ip injection in groups of six age-matched mice. One week before the start of the injection mice were made hypothyroid by administration of a special iodine-deficient mouse chow containing the thyroid hormone biosynthesis inhibitor 5-propyl-2-thiouracil (0.15%; Harlan Teklad, Madison, WI). Mice were kept on this hypothyroid regimen during the study. One hypothyroid group of seven mice remained untreated. A group of six untreated euthyroid mice fed standard mouse chow was used as the control group. Electrophysiological measurements were performed in all mice at the end of the treatment period. Hemodynamic measurements were performed in control and hypothyroid mice treated with 1 × GC-1 and 1 × T3. Serum for lipid measurement and cardiac tissue for determination of the messenger RNA (mRNA) levels of certain proteins were collected from all mice at the end of the study. Each group consisted of male and female mice, aged of 10–15 weeks.
In vivo hemodynamic measurements. Heart rate was calculated from electrophysiological tracings. Mice were lightly sedated with ketamine (0.033 mg/g BW) and pentobarbital (0.033 mg/g BW) and restrained (17). Four needles were placed sc on each limb close to the trunk, and 6 min after injection of sedatives an electrocardiogram (ECG) was obtained. Heart rate was calculated from the RR interval. For measurement of in vivo hemodynamics, mice were anesthetized with a mixture of ketamine (0.1 mg/g BW) and xylazine (0.007 mg/g BW) and placed under a dissecting microscope (18). ECG electrodes were connected to the limbs, and the body temperature was monitored with an anal temperature probe and kept at 37 C. The animals were intubated and ventilated with room air. A bilateral vagotomy was performed. The right jugular vein and right carotid artery were cannulated with polyethylene catheters, and the carotid line was connected to a pressure gauge. Arterial pressure was measured at this time point. Next, a left-sided thoracotomy was performed. The pericardium was opened, and a high fidelity pressure transducer (1.8 French, Millar Instruments, Houston, TX) was inserted into the left atrium and advanced into the left ventricle. Continuous left ventricular (LV) pressure, differential LV pressure, aortic pressure, and ECG were recorded on a personal computer. After recording the measurements, increasing doses of isoproterenol were infused through the venous line (5, 50, 500, and 1000 ng). The following parameters were analyzed: heart rate (HR), carotid arterial pressure, LV end-systolic pressure (LVESP), and the maximum values of LV pressure derivatives (dP/dtmax, dP/dtmin). Effects of isoproterenol on hemodynamic parameters were analyzed 45–60 sec after drug administration.
Northern blot analysis. Messenger RNA levels for sarcoplasmic reticulum calcium adenosine triphosphatase (Serca2), myosin heavy chain α (MHCα), MHCβ, and the hyperpolarization-activated cyclic nucleotide-gated channel 2 (HCN2) were determined. RNA was isolated from myocardial tissue by a guanidinium thiocyanate method (19), and polyadenylated RNA was obtained using an oligo(deoxythymidine) kit (QIAGEN, Chatsworth, CA). mRNA was separated on a 1% agarose gel and transferred to a nylon membrane. A 1.6-kb fragment of the 5′-end of the rat Serca2 complementary DNA, a 224-bp fragment of the second exon of mouse HCN2 complementary DNA, and oligonucleotides specific for the rat MHCα and rat MHCβ were used to make 32P-labeled probes (Multiprime DNA labeling systems, Amersham Pharmacia Biotech, Aylesbury, UK). Radioactivity was assessed on a blue film (Eastman Kodak Co., Rochester, NY), and the resulting image was quantified with Image 1.61 software.
Measurements of cholesterol and triglycerides in mice. Blood was collected from all mice at the end of the study; the serum was separated by centrifugation and immediately frozen in liquid nitrogen. Triglycerides and cholesterol were measured by enzymatic methods (Roche, Basel, Switzerland).
Statistical analysis. Parameters in control, hypothyroid mice and in hypothyroid mice treated with different doses of GC-1 and T3 were analyzed by ANOVA and Fisher’s post-hoc tests. P < 0.05 was assumed significant. Data are given as the mean ± sem.
Dose-response studies in hypercholesteremic rats
Male Sprague Dawley rats (Harlan Sprague Dawley, Inc.), weighing 300–400 g, were used for this study. The rats were fed a diet containing 1.5% cholesterol and 0.5% cholic acid (Harlan Teklad Rodent Chow, Madison, WI) for 2 weeks before drug treatment and for the entire course of the study. The rats were given via oral gavage either vehicle (n = 6; a separate vehicle group was used for the T3-treated group and the GC-1 treated group, also n = 6) or the following drug treatment groups: 1.54–924 nmol/kg·day T3 (n = 5/group) or 46.2–2920 nmol/kg·day GC-1 (n = 5/group). The doses correspond to 1.0–601.5 ng/g·day T3 and 15.2–958.9 ng/g·day GC-1. The vehicle or drugs were administered once per day for 7 days. Vehicle consisted of 5% cremaphor, 5% ethanol, 10% m-pyrol, and 80% water. The animals were weighed each day before drug administration. After 7 days, the rats were anesthetized using 40 mg/kg pentobarbital sodium, ip, and the heart rates were measured by ECG (Gould Instruments, Valley View, OH). A 3-ml blood sample was taken via the inferior vena cava for each animal. The blood samples were centrifuged at 2000 rpm for 20 min, and the plasma was used for cholesterol, triglyceride, and TSH measurements. Plasma cholesterol measurement was performed using a Cobas Mira S analyzer (Roche, Indianapolis, IN). Plasma TSH was measured using a RIA kit designed for rat TSH (Amersham Pharmacia Biotech, Arlington Heights, IL).
The doses chosen were based on the lowest dose of T3 (1 μg/kg·day or 1.54 nmol/kg·day). For the purpose of comparison between GC-1 and T3, all doses are denoted as nanomoles per kg/day. Increasing doses of T3 or GC-1 were then given in half-log increments.
Determination of T3, T4, and GC-1 levels in plasma and tissues
Male Sprague Dawley rats were anesthetized with 30 mg/kg pentobarbital, ip. Either T3 or GC-1 (10μ mol/kg) was injected directly into the right jugular vein (n = 3/group). Plasma and tissue samples were collected 1 h after treatment. For various tissues, samples were weighed and homogenized with 3 vol deionized water. For T3 and T4 assays, 100-μl samples were treated with 250μ l acetonitrile containing 100 ng/ml internal standard (IS-I). For GC-1 assays, 25-μl samples were treated with 50 μl acetonitrile containing 100 ng/ml internal standard (IS-II). After centrifugation to remove precipitated proteins, the clear supernatant was dried with nitrogen in silanized glass vials and then reconstituted with 50 μl acetonitrile-water (50:50). Due to endogenous production of T3 and T4 hormones, standard curve and quality control samples for T3 and T4 measurements were prepared using acetonitrile-water (50:50) mixture in the same fashion as for the tissue treatment. A 10-μl aliquot of the reconstituted solution was analyzed by liquid chromatography/tandem mass spectrometry.
Sample analysis was performed using a Shimadzu LC system (Shimadzu Scientific, Kyoto, Japan) with an Inertsil (MetaChem Technologies, Torrance, CA) ODS-2 HPLC column (2 × 50 mm, 5μ m) interfaced to a Micromass Quattro (Micromass UK Limited, Manchester, UK) tandem mass spectrometer. The mobile phases consisted of 0.1% formic acid in water (A) and 0.1% formic acid in methanol (B). With a flow rate of 0.3 ml/min, a linear gradient was started from 95% A/5% B to 5%A/95% B over 2.5 min and was held at 5% A/95% B for an additional 1.5 min. The mobile phase was then returned to initial conditions, and the column was reequilibrated for 2 min. The total analysis time was 6 min. MS/MS conditions were as followings: electrospray ionization; ultra high purity nitrogen as nebulizing gas (100 liters/h) and desolvation gas (900 liters/h); desolvation temperature, 350 C; and source temperature, 150 C. (M+H)+ species for T3, T4, GC-1, and internal standards were selected in MS1 and collisionally dissociated with argon at a pressure of 1.5 × 10−3 T to form specific product ions that were subsequently monitored by MS2. Instrument parameters for each compound are shown in Table 1.
Instrument parameters fro compounds
| Compound | Electrkospray | Reaction | Cone (V) | Collision (V) |
|---|---|---|---|---|
| T3 | + | 652→605 | 55 | 20 |
| T4 | + | 778→731 | 5 | 30 |
| IS-I | + | 492→155 | 35 | 40 |
| GC-1 | − | 327→269 | 40 | 20 |
| IS-II | − | 354→69 | 15 | 10 |
| Compound | Electrkospray | Reaction | Cone (V) | Collision (V) |
|---|---|---|---|---|
| T3 | + | 652→605 | 55 | 20 |
| T4 | + | 778→731 | 5 | 30 |
| IS-I | + | 492→155 | 35 | 40 |
| GC-1 | − | 327→269 | 40 | 20 |
| IS-II | − | 354→69 | 15 | 10 |
Results
Body and heart weights
Hypothyroid mice showed typical evidence of severe thyroid hormone deficiency (Table 2); body weight, heart weight, and heart weight to body weight ratio (HW/BW ratio) were markedly decreased compared with those in euthyroid age-matched mice. Treatment of hypothyroid mice with 1 × T3, 4.5 × T3, and 9 × T3 increased body weight, heart weight, and HW/BW ratio. Treatment with 1 × GC-1 and 4.5 × GC-1 did not increase body weight or HW/BW ratio. At the 9 × GC-1 dose, heart weight was significantly increased, and there was a trend (albeit insignificant) for an increase in body weight. No toxic effects were observed in mice treated with GC-1. The appearance and behavior of GC-1-treated mice were similar to those of T3-treated mice.
Body and heart weights
| BW (g) | Heart wt (mg) | HW/BW ratio | |
|---|---|---|---|
| Control | 22.5 ± 0.9a | 120 ± 3a | 5.3 ± 0.2a |
| Hypothyroid | 18.8 ± 0.6 | 79 ± 3 | 4.2 ± 0.1 |
| 1 × T3 | 20.8 ± 0.4b | 110 ± 1a | 5.3 ± 0.1a |
| 4.5 × T3 | 21.6 ± 0.6b | 120 ± 2a | 5.6 ± 0.1a |
| 9 × T3 | 21.3 ± 0.7b | 125 ± 3a | 5.9 ± 0.2a |
| 1 × GC-1 | 18.4 ± 0.5 | 81 ± 3 | 4.4 ± 0.1 |
| 4.5 × GC-1 | 19.8 ± 0.7 | 84 ± 2 | 4.3 ± 0.1 |
| 9 × GC-1 | 20.5 ± 0.7 | 92 ± 2a | 4.5 ± 0.1 |
| BW (g) | Heart wt (mg) | HW/BW ratio | |
|---|---|---|---|
| Control | 22.5 ± 0.9a | 120 ± 3a | 5.3 ± 0.2a |
| Hypothyroid | 18.8 ± 0.6 | 79 ± 3 | 4.2 ± 0.1 |
| 1 × T3 | 20.8 ± 0.4b | 110 ± 1a | 5.3 ± 0.1a |
| 4.5 × T3 | 21.6 ± 0.6b | 120 ± 2a | 5.6 ± 0.1a |
| 9 × T3 | 21.3 ± 0.7b | 125 ± 3a | 5.9 ± 0.2a |
| 1 × GC-1 | 18.4 ± 0.5 | 81 ± 3 | 4.4 ± 0.1 |
| 4.5 × GC-1 | 19.8 ± 0.7 | 84 ± 2 | 4.3 ± 0.1 |
| 9 × GC-1 | 20.5 ± 0.7 | 92 ± 2a | 4.5 ± 0.1 |
Hypothyroid mice were treated with 3.5 ng/g BW T3 (1 × T3), which was a dose assumed to restore euthyroidism, with the equimolar amount of 1.8 ng/g BW GC-1 (1 × GC-1), and with higher doses of both compounds (4.5 × T3, 9 × T3, 4.5 × GC-1, 9 × GC-1). T3-treated mice showed increased body weight, heart weight, and heart weight to body weight ratios (HW/BW ratio) at all doses. 1 × GC-1 and 4.5 × GC-1 did not increase body weight, heart weight, or heart weight to body weight ratio. Only 9 × GC-1 treatment led to an increase in heart weight.
P < 0.0001 vs. hypothyroid.
P < 0.05 vs. hypothyroid.
Cardiac effects of GC-1 and T3 in mice
Heart rate in hypothyroid mice (Table 3) was markedly decreased to 335 ± 21 beats/min from a heart rate in control mice of 472 ± 26 beats/min (P < 0.001 control vs. hypothyroid). Whereas treatment with 1 × T3 lead to an increase to 502 ± 21 beats/min (P < 0.001 hypothyroid vs. 1 × T3-treated), mice treated with the equimolar amount of 1.8 ng/g BW GC-1 (1 × GC-1) retained a low heart rate of 358 ± 28 beats/min (P < 0.01 control vs. 1 × GC-1-treated). Higher equimolar doses of the compounds let to similar increases in heart rate; 4.5 × T3 and 4.5 × GC-1 increased heart rates to 536 ± 24 and 527 ± 35 beats/min, respectively, and treatment with 9 × T3 and 9 × GC-1 resulted in heart rates of 584 ± 28 and 570 ± 31 beats/min, respectively. All other hemodynamic parameters measured, systolic and diastolic arterial pressure (Pa sys and Pa dias), end-systolic left ventricular pressure (LVESP), and maximum speeds of contraction (dP/dtmax) and relaxation (dP/dtmin), were also severely decreased in hypothyroid mice (Table 3). 1 × GC-1 showed a positive inotropic effect (Figs. 2 and 5) by increasing dP/dtmax to 4424 ± 510 from 2945 ± 273 mm Hg/sec in hypothyroid mice (P < 0.05 hypothyroid vs. 1 × GC-1-treated), although this was less than the effect of 1 × T3 that resulted in a dP/dtmax of 6224 ± 888 mm Hg/sec (P < 0.05 1 × T3-treated vs. 1 × GC-1-treated). Treatment with 1 × T3, but not with 1 × GC-1, also increased Pa sys, Pa dias, LVESP, and dP/dtmin (P < 0.05 hypothyroid vs. 1 × T3-treated). The positive inotropic response was assessed by the increase in arterial pressure and the maximal speed of contraction after administration of isoproterenol. Administration of increasing doses of isoproterenol iv resulted in positive inotropic effects in all groups. The hemodynamic effects of 500 ng isoproterenol are presented as the percent increase from the basal level; arterial pressure increased from basal levels in control, hypothyroid, 1 × T3-treated, and 1 × GC-1-treated mice by 25 ± 5%, 23 ± 2%, 25 ± 5%, and 36 ± 9% respectively, dP/dtmax increased by 16 ± 6%, 25 ± 11%, 14 ± 5%, and 17 ± 9% respectively. These data show that the sympathetic responses of the hypothyroid, 1 × GC-1, and 1 × T3 groups were not different from that of the control group.
In vivo hemodynamic measurements. Hypothyroid mice showed markedly decreased heart rate and maximum speed of contraction (dP/dtmax) compared with control mice. Heart rate increased in mice treated with 1 × T3 (3.5 ng/g BW), but not in mice treated with equimolar amounts of GC-1 (1.8 ng/g BW). Both compounds had a positive inotropic effect on dP/dtmax. This positive inotropic effect was higher in 1 × T3-treated mice, which had increased dP/dtmax, than in hypothyroid mice and 1 × GC-1-treated mice.
In vivo hemodynamic measurements. Hypothyroid mice showed markedly decreased heart rate and maximum speed of contraction (dP/dtmax) compared with control mice. Heart rate increased in mice treated with 1 × T3 (3.5 ng/g BW), but not in mice treated with equimolar amounts of GC-1 (1.8 ng/g BW). Both compounds had a positive inotropic effect on dP/dtmax. This positive inotropic effect was higher in 1 × T3-treated mice, which had increased dP/dtmax, than in hypothyroid mice and 1 × GC-1-treated mice.
Hemodynamic measurements from control and hypothyroid mice, and from hypothyroid mice treated with 1 × T3 (3.5 ng/g BW) and 1 × GC-1 (1.8 ng/g BW)
| Heart rate (beats/min) | Arterial pressure (mm Hg) | LVESP (mm Hg) | DP/dtmax (mm Hg/s) | DP/dtmin (mm Hg/s) | ||
|---|---|---|---|---|---|---|
| Systolic | Diastolic | |||||
| Control | 472 ± 26 | 96 ± 9 | 78 ± 8 | 69 ± 3 | 5065 ± 397 | 3826 ± 372 |
| Hypothyroid | 335 ± 21a | 73 ± 3a | 61 ± 5a | 49 ± 2a | 2945 ± 273a | 2128 ± 203a |
| 1 × T3 | 502 ± 21b | 102 ± 9b | 85 ± 7b | 74 ± 8b | 6224 ± 888b | 4427 ± 1013b |
| 1 × GC-1 | 358 ± 28 | 76 ± 4 | 62 ± 3 | 57 ± 3 | 4424 ± 520c | 2385 ± 184 |
| Heart rate (beats/min) | Arterial pressure (mm Hg) | LVESP (mm Hg) | DP/dtmax (mm Hg/s) | DP/dtmin (mm Hg/s) | ||
|---|---|---|---|---|---|---|
| Systolic | Diastolic | |||||
| Control | 472 ± 26 | 96 ± 9 | 78 ± 8 | 69 ± 3 | 5065 ± 397 | 3826 ± 372 |
| Hypothyroid | 335 ± 21a | 73 ± 3a | 61 ± 5a | 49 ± 2a | 2945 ± 273a | 2128 ± 203a |
| 1 × T3 | 502 ± 21b | 102 ± 9b | 85 ± 7b | 74 ± 8b | 6224 ± 888b | 4427 ± 1013b |
| 1 × GC-1 | 358 ± 28 | 76 ± 4 | 62 ± 3 | 57 ± 3 | 4424 ± 520c | 2385 ± 184 |
All measured hemodynamic parameters were severely decreased in hypothyroid mice. This decrease was completely reversed by 1 × T3, which normalized heart rate, arterial pressure, LVESP, dP/dtmax, and dP/dtmin. 1 × GC-1 had a differential effect on the hemodynamic system; it led to an increase in dP/dtmax, but did not change heart rate.
P < 0.05, hypothyroid vs. control mice.
P < 0.05, 1 × T3-treated vs. hypothyroid mice.
P < 0.05, 1 × GC-1 treated vs. hypothyroid mice.
Levels of mRNAs coding for cardiac proteins
Cardiac tissue of hypothyroid mice showed changes in MHC isoform expression typically seen in hypothyroidism (20), with high mRNA levels of MHCβ and almost no MHCα mRNA. Treatment of hypothyroid mice with 3.5 ng/g BW T3 (1 × T3) completely reversed the pattern of MHC isoform expression; control MHCα mRNA levels were restored, and MHCβ mRNA was no longer detected. In contrast, treatment with 1 × GC-1 did not increase MHCα expression, and the decrease in MHCβ mRNA was smaller than that in 1 × T3-treated mice. Treatment with 9 × GC-1 increased MHCα mRNA to euthyroid levels, but MHCβ mRNA was still detectable (Figs. 3 and 5).
Northern blots of messenger RNA of MHCα, MHCβ, sarcoplasmic calcium ATPase (Serca2), and hyperpolarization-activated cyclic nucleotide-dependent channel (HCN2) and 28S in cardiac tissue. The decrease in MHCα and the increase in MHCβ in hypothyroid mice were fully compensated in mice treated with 1 × T3 (3.5 ng/g BW), 4.5 × T3, and 9 × T3. Treatment of hypothyroid mice with 1 × GC-1 did not decrease MHCβ or increase MHCα compared with hypothyroid levels. Higher doses of GC-1 reversed MHCα mRNA to control levels, but MHCβ mRNA was undetectable even with the highest dose of GC-1 (16.2 ng/g BW, 9 × GC-1). Serca2 mRNA, which was decreased in hypothyroid mice, was fully restored by treatment with any T3 dose, whereas GC-1 only increased Serca2 mRNA at 16.2 ng/g BW. HCN2 mRNA levels, which were severely decreased in hypothyroid mice, were restored by T3 and increased further with higher T3 doses. In contrast, treatment with 1 × GC-1 did not increase levels of HCN2 mRNA, and only 9 × GC-1 reconstituted normal HCN2 levels. 28S was used to control for equal loading.
Northern blots of messenger RNA of MHCα, MHCβ, sarcoplasmic calcium ATPase (Serca2), and hyperpolarization-activated cyclic nucleotide-dependent channel (HCN2) and 28S in cardiac tissue. The decrease in MHCα and the increase in MHCβ in hypothyroid mice were fully compensated in mice treated with 1 × T3 (3.5 ng/g BW), 4.5 × T3, and 9 × T3. Treatment of hypothyroid mice with 1 × GC-1 did not decrease MHCβ or increase MHCα compared with hypothyroid levels. Higher doses of GC-1 reversed MHCα mRNA to control levels, but MHCβ mRNA was undetectable even with the highest dose of GC-1 (16.2 ng/g BW, 9 × GC-1). Serca2 mRNA, which was decreased in hypothyroid mice, was fully restored by treatment with any T3 dose, whereas GC-1 only increased Serca2 mRNA at 16.2 ng/g BW. HCN2 mRNA levels, which were severely decreased in hypothyroid mice, were restored by T3 and increased further with higher T3 doses. In contrast, treatment with 1 × GC-1 did not increase levels of HCN2 mRNA, and only 9 × GC-1 reconstituted normal HCN2 levels. 28S was used to control for equal loading.
Effects of T3 and GC-1 on blood lipid levels. Cholesterol and triglyceride levels were increased in hypothyroid mice. Treatment of hypothyroid mice with T3 and GC-1 decreased cholesterol to euthyroid levels at all doses. GC-1 also lowered triglycerides to control levels, whereas T3 had no effect on triglyceride levels at any dose. GC-1 was more effective at lowering triglyceride levels than treatment with the equimolar dose of T3.
Effects of T3 and GC-1 on blood lipid levels. Cholesterol and triglyceride levels were increased in hypothyroid mice. Treatment of hypothyroid mice with T3 and GC-1 decreased cholesterol to euthyroid levels at all doses. GC-1 also lowered triglycerides to control levels, whereas T3 had no effect on triglyceride levels at any dose. GC-1 was more effective at lowering triglyceride levels than treatment with the equimolar dose of T3.
Differential effects of 1 × GC-1 on cardiac activity and lipid levels. To compare the effects of treatment with 1 × T3 (3.5 ng/g BW) and equimolar amounts of GC-1 (1.8 ng/g BW, 1 × GC-1) on different parameters in hypothyroid mice, the effect was expressed as a percentage of the parameter’s change from the hypothyroid to the euthyroid level, which was defined as 100%. This comparison shows that 1 × T3 treatment normalized all cardiac parameters to the euthyroid level, whereas treatment with 1 × GC-1 changed only dP/dtmax significantly and did not increase any of the other cardiac parameters. In contrast, treatment of hypothyroid mice with 1 × GC-1 lowered elevated cholesterol and triglyceride levels, whereas 1 × T3 treatment normalized only cholesterol levels, but not triglyceride levels.
Differential effects of 1 × GC-1 on cardiac activity and lipid levels. To compare the effects of treatment with 1 × T3 (3.5 ng/g BW) and equimolar amounts of GC-1 (1.8 ng/g BW, 1 × GC-1) on different parameters in hypothyroid mice, the effect was expressed as a percentage of the parameter’s change from the hypothyroid to the euthyroid level, which was defined as 100%. This comparison shows that 1 × T3 treatment normalized all cardiac parameters to the euthyroid level, whereas treatment with 1 × GC-1 changed only dP/dtmax significantly and did not increase any of the other cardiac parameters. In contrast, treatment of hypothyroid mice with 1 × GC-1 lowered elevated cholesterol and triglyceride levels, whereas 1 × T3 treatment normalized only cholesterol levels, but not triglyceride levels.
Hypothyroidism resulted in a 70% decrease in mRNA for the sarcoplasmic calcium pump protein (Serca2) and almost completely abolished mRNA for the cardiac pacemaker channel HCN2. Changes in Serca2 were fully reversed in mice treated with T3 at all doses (P < 0.05, T3 treated vs. hypothyroid; T3 treated not different from control), whereas treatment with GC-1 restored Serca2 mRNA levels only at the highest dose of 9 × GC-1. mRNA levels for HCN2 were fully restored by treatment with 1 × T3, and further increased over euthyroid levels with treatment with 4.5 × T3 and 9 × T3 respectively. GC-1-treated mice had decreased HCN2 levels with 1 × GC-1 and 4.5 × GC-1 treatments, and only treatment with 9 × GC-1 restored normal HCN2 levels (Figs. 3 and 5).
Lipid-lowering effects in mice
Beneficial effects on the lipid profile were equal or greater in GC-1-treated mice than in T3-treated mice (Figs. 4 and 5); cholesterol levels were elevated in hypothyroid mice to 165 mg/dl compared with 74 ± 3 mg/dl in euthyroid mice (P< 0.0001). Both T3 and GC-1 treatment at all doses decreased cholesterol to normal levels. Triglycerides were elevated in hypothyroid mice to 126 ± 15 mg/dl compared with 70 ± 15 mg/dl in euthyroid mice (P < 0.05). GC-1 reestablished normal triglyceride levels at all doses, whereas T3 treatment failed to decrease triglyceride levels from the elevated hypothyroid levels.
Dose-responses to GC-1 and T3 in rats
The effects of T3 or GC-1 on plasma cholesterol, heart rate, and plasma TSH are shown in Fig. 6. These data are shown as the percent change for each of these parameters relative to their respective vehicle-treated group values. Plasma cholesterol was reduced by T3 and GC-1 in a dose-dependent manner. The vehicle-treated group values for cholesterol were 187 ± 25 and 180 ± 24 mg/dl for the T3 and GC-1 studies, respectively. For the purpose of comparison, ED50 values for cholesterol reduction are shown in Table 4. ED50 was defined as the dose of drug causing a 50% reduction in cholesterol compared with the vehicle-treated group value, and this was chosen because it was approximately half of the maximal effect (80–90% reduction). As shown in this table, the ED50 for cholesterol lowering was 12-fold lower with T3 than with GC-1. For the T3-treated group, cholesterol was significantly reduced (compared with vehicle, P < 0.05) starting at the 15.4 nmol/kg·day dose. For GC-1, significant cholesterol lowering vs. that in vehicle-treated animals was seen at the 154 nmol/kg·day dose and higher doses (P < 0.05).
Effect of 7 days of treatment with GC-1 or T3 on plasma cholesterol, TSH, and heart rate in cholesterol-fed rats. Cholesterol and TSH were reduced by both compounds in a dose-dependent manner. T3 caused a dose-dependent increase in heart rate within the cholesterol-lowering dose range, whereas GC-1 had no effect.
Effect of 7 days of treatment with GC-1 or T3 on plasma cholesterol, TSH, and heart rate in cholesterol-fed rats. Cholesterol and TSH were reduced by both compounds in a dose-dependent manner. T3 caused a dose-dependent increase in heart rate within the cholesterol-lowering dose range, whereas GC-1 had no effect.
Percent changes in cholesterol (Colest.), heart rate (HR), and TSH in T3- or GC-1-treated hypercholesteremic rats were calculated from the respective vehicle control group
| Compound | ED50 (Colest.) | ED15 (HR) | ED30 (TSH) | ED15/ED50 (HR/Colest.) | ED30/ED50 (TSH/Colest.) |
|---|---|---|---|---|---|
| T3 | 17.2 | 22.8 | 11.7 | 1.24 | 0.68 |
| GC-1 | 201 | >2920 | 251 | >14.5 | 1.19 |
| Compound | ED50 (Colest.) | ED15 (HR) | ED30 (TSH) | ED15/ED50 (HR/Colest.) | ED30/ED50 (TSH/Colest.) |
|---|---|---|---|---|---|
| T3 | 17.2 | 22.8 | 11.7 | 1.24 | 0.68 |
| GC-1 | 201 | >2920 | 251 | >14.5 | 1.19 |
EDs are expressed as the dose lowering cholesterol by 50%, that lowering TSH by 30%, or that increasing heart rate by 15%. Values for the GC-1 or T3 dose are expressed as nanomoles per kg/day.
T3 increased heart rate in a dose-dependent manner, whereas GC-1 had little effect on heart rate within the dose range tested (Fig. 6). Heart rates were 364 ± 9 and 378 ± 8 beats/min for the vehicle-treated groups for the T3 and GC-1 studies, respectively. Heart rate was increased by T3, and significance was noted starting at the 46.2 nmol/kg·day dose (P < 0.05 vs. vehicle). No significant changes in heart rate were observed for GC-1 with the doses used in this study. Potencies for changes in heart rate were expressed as ED15, which is the dose causing a 15% increase in heart rate relative to the time-matched vehicle group value. The ED15 was chosen because a 10–15% increase in heart rate is viewed as being the lower end of clinically relevant changes in heart rate. In addition, the potency ratios for ED15 (heart rate)/ED50 (cholesterol) are shown in Table 4. The ratio was approximately 10-fold higher for GC-1 compared with T3, suggesting that GC-1 was significantly more selective for lowering cholesterol than it is for causing tachycardia.
TSH was reduced by both GC-1 and T3 in a dose-dependent manner, although T3 was approximately 20-fold more potent than GC-1. T3 significantly reduced TSH starting at the 4.62 nmol/kg·day dose, whereas GC-1 did so starting at the 154 nmol/kg·day dose (P < 0.05 compared with their respective vehicle group values). The potency for TSH suppression was expressed as ED30, because the maximal percent suppression was approximately 60% in these animals. As TSH suppression and cholesterol lowering are both presumably TRβ-mediated effects, it would be expected that the potency ratios for TSH vs. cholesterol lowering should be similar for T3 and GC-1. GC-1 was slightly more selective for lowering cholesterol than was TSH, although this difference was only 2-fold, which is not pharmacologically significant.
Tissue distribution of GC-1 and T3
Measurements of GC-1 and T3 in plasma and tissues revealed distinct differences in their organ distributions (Table 5). The tissue to plasma ratios in the liver were similar for GC-1 and T3 (2.09± 0.53 and 5.86 ± 1.93, respectively). In contrast the tissue/plasma ratio in the heart was 30 times higher for T3 (3.61 ± 0.69 ng/g) than for GC-1 (0.12 ± 0.002 ng/g). Absolute values for plasma levels were 75 times different (3188 ± 445 ng/ml for T3 and 42 ± 8 ng/ml for GC-1).
Tissue levels (nanograms per ml) and tissue/plasma ratios (T/P ratio; nanograms per g) of T3 and GC-1 in rats
| T3 | T/P ratio, T3 | GC-1 | T/P ratio, GC-1 | (T/P ratio GC-1)/(T/P ratio T3) | |
|---|---|---|---|---|---|
| Liver | 18,690 ± 6,148 | 5.86 ± 1.93 | 90.4 ± 23.2 | 2.09 ± 0.53 | 0.357 |
| Heart | 11,511 ± 1,100 | 3.61 ± 0.69 | 5.6 ± 1.2 | 0.12 ± 0.002 | 0.033 |
| Plasma | 3,188 ± 445 | 1 | 42.5 ± 8.2 | 1 | 1 |
| Muscle | 3,131 ± 1,253 | 0.98 ± 0.39 | 11.6 ± 3.6 | 0.27 ± 0.12 | 0.276 |
| Adipose | 18 ± 24 | 0.01 ± 0.01 | <LLOQa | ||
| Brain | 1,798 ± 137 | 0.56 ± 0.04 | 5.2 ± 0.4 | 0.13 ± 0.03 | 0.23 |
| T3 | T/P ratio, T3 | GC-1 | T/P ratio, GC-1 | (T/P ratio GC-1)/(T/P ratio T3) | |
|---|---|---|---|---|---|
| Liver | 18,690 ± 6,148 | 5.86 ± 1.93 | 90.4 ± 23.2 | 2.09 ± 0.53 | 0.357 |
| Heart | 11,511 ± 1,100 | 3.61 ± 0.69 | 5.6 ± 1.2 | 0.12 ± 0.002 | 0.033 |
| Plasma | 3,188 ± 445 | 1 | 42.5 ± 8.2 | 1 | 1 |
| Muscle | 3,131 ± 1,253 | 0.98 ± 0.39 | 11.6 ± 3.6 | 0.27 ± 0.12 | 0.276 |
| Adipose | 18 ± 24 | 0.01 ± 0.01 | <LLOQa | ||
| Brain | 1,798 ± 137 | 0.56 ± 0.04 | 5.2 ± 0.4 | 0.13 ± 0.03 | 0.23 |
The relation of the T/P ratio of GC-1 to the T/P ratio of T3 was 10 times lower in the heart than in the liver. Tissue was harvested 1 h after iv injection of 10 μmol/kg T3 or GC-1.
The detection limit for GC-1 was approximately 1 ng/ml.
Discussion
Recently accumulated evidence firmly indicates that binding of T3 to products of the c-erbAα gene (TRα1) vs. those of the c-erbAβ gene (TRβ1) or its splice variant TRβ2 leads to distinct thyroid hormone effects (21, 22). It is therefore of interest to develop new T3 agonists and antagonists with preferred binding and action through the various TR isoforms and explore influences of these analogs on specific thyroid hormone-responsive physiological functions.
Described in this report are the effects of the T3 analog GC-1, which exhibits similar binding affinity and actions in cell culture as T3 for the TRβ1 subtype, but approximately 10 times weaker binding and actions in cell culture for the TRα1 subtype (15). Upon administration of T3 at a dose that largely restored a euthyroid status (3.5 ng/g BW, defined as 1 × T3) and equimolar amounts of GC-1 (1.8 ng GC-1/g BW, defined as 1 × GC-1), distinct differences in the responses of specific thyroid hormone-responsive parameters were noted. Thus, 1 × GC-1 lowered cholesterol levels to the euthyroid range, as did administration of 1 × T3. Triglyceride levels were lowered significantly with GC-1 compared with T3, which exerted no significant effect on triglyceride levels. 1 × GC-1 did not increase heart rate from the hypothyroid level, but T3 at this dose significantly increased heart rate and restored it to the normal euthyroid level. The mRNA coding for the pacemaker channel HCN2, which was severely decreased in hypothyroid mice, was restored by 1 × T3 to control levels and further increased with higher T3 doses, whereas 1 × GC-1 did not affect HCN2, and only higher GC-1 doses restored mRNA HCN2 levels. These effects of T3 and GC-1 on HCN2 paralleled their effects on heart rate. Restoration to control values by 1 × T3, but not by 1 × GC-1, was also seen for mRNAs of MHCα, MHCβ, and Serca2 and for body weight, heart weight, and HW/BW ratio. Cardiac contractility was increased by 1 × GC-1 and 1 × T3, although this effect was larger and included more parameters with 1 × T3.
The potential mechanism underlying this and other differential responses between GC-1 and T3 could be linked to differences in TRβ isoform distribution in specific organs. Findings by other investigators have shown that TRβ receptors are the predominant isoform in the liver, accounting for 80% of T3 receptor binding (23). In addition, the T3-mediated lowering of cholesterol levels is most likely due to an elevated clearance rate of cholesterol mediated by increased expression of hepatic low density lipoprotein receptors (24) and an increase in specific lipid-lowering liver enzymes (6, 25, 26). The strong effect of GC-1 on lowering cholesterol, which is equal to that of T3, and the even more enhanced effect of GC-1 on lowering triglyceride levels, which exceeds the effect of T3, is therefore compatible with a TRβ-preferred action of GC-1. In contrast to the liver, close to half of the T3 receptors in the heart are of the TRα1 subtype (23). The target genes of thyroid hormone effects on the heart rate are not known, although HCN2 and HCN4 are possible candidates. As mentioned previously, mice with deletion of the TRα (12), but not the TRβ (14), receptor have marked bradycardia. The absence of GC-1 effects on increasing heart rate from the hypothyroid value is compatible with the previous finding that the EC50 on TRα1 for GC-1 is approximately 10 times greater than that of T3, and T3 action on heart rate is mediated through TRα1. In contrast, a T3 dose of 3.5 ng/g BW returned heart rate from a decreased hypothyroid rate to the normal euthyroid range. These differential effects of T3vs. GC-1 on heart rate are compatible with selective TRβ actions of GC-1. However, in this context we cannot exclude tissue-selective differences in the uptake of the various compounds as contributors to the differential responses.
Recently, the gene coding for the cardiac pacemaker channel If, which appears to play an essential role in setting the heart rate, has been cloned (27, 28). This hyperpolarization-activated inward rectifier channel conducts the monovalent cations potassium and sodium at a 4:1 ratio. This channel has been termed hyperpolarization-activated cyclic nucleotide-regulated (HCN) channel. Four isoforms of it occur in the brain (HCN1–HCN4) (27, 28). Two of these, HCN2 and HCN4, are expressed in the heart, with a marked predominance of HCN2 (Gloss, B., and W. H. Dillmann, unpublished observations). As the expression of this gene might be involved in thyroid hormone effects on heart rate, we quantitated the response of HCN2 to T3 and GC-1 and found a response in line with that of heart rate. HCN2 mRNA does not significantly increase upon treatment with 1 × GC-1, in contrast to treatment with 1 × T3, which results in elevated HCN2 mRNA levels compared with euthyroid levels. It appears therefore quite likely that the discrepant heart rate and HCN2 responses to T3 and GC-1 may indicate a cause-effect relationship between the heart rate and the level of expression of the HCN2 channel. Thus, HCN2 may be a major target for thyroid hormone regulation of the heart rate. In line with the effects on heart rate are other markedly smaller effects of GC-1 vs. T3 on other cardiac parameters. This includes the finding that 1 × GC-1 has no significant effect on heart weight and mRNAs coding for proteins related to cardiac contraction, such as MHCα, MHCβ, and Serca2. It is of interest to note that the first derivative of the rise in systolic pressure, dP/dtmax, significantly increased by treatment with 1 × GC-1, although this effect was smaller than that with 1 × T3 treatment. Some specific cardiac functions, such as dP/dtmax, show a partial response to a TRβ-preferred agent.
Whereas the GC-1-induced lowering of cholesterol shows a similar response to T3 at doses of 1.8, 8.1, and 16.2 ng/g BW, other responses differ. Thus, GC-1-induced lowering of triglyceride levels is more efficient at all doses compared with T3. In contrast, other parameters, such as body weight or HW/BW ratio, showed no change from the hypothyroid level at any GC-1 dose, in contrast to T3, which resulted in marked and progressive increases in these parameters. The mechanisms underlying these discrepant T3vs. GC-1 responses are currently unclear. The GC-1 effect on heart weight, which only occurs at high GC-1 doses, is in line with a significantly diminished GC-1 effect on parameters such as heart rate and reversal of MHCα RNA levels and MHCβ mRNA levels compared with T3. In addition, the T3-induced increase in body weight was already maximal at doses of 3.5 ng/g BW and did not further increase at higher T3 doses. It is unlikely that the decreased response of heart weight or MHCα mRNA levels of GC-1-treated mice is due simply to a much weaker thyromimetic action than T3, as this would not explain the equivalent effects of GC-1 and T3 on cholesterol levels and the greater effects of GC-1 than T3 on triglyceride levels. Results from mice with deletion of TRα vs. TRβ clearly indicate that for some effects T3 can exert a similar influence, being bound to either TRα1 or TRβ1, and the different TR isoforms can exert functional cross-coverage for each other. In contrast, other effects can only be fully mediated by T3 bound to either the TRα or the TRβ. Another mechanism that has to be considered for differential effects of GC-1 is the difference in tissue distribution; tissue/plasma ratios of T3 are similar for heart and liver, whereas the tissue/plasma ratio of GC-1 for the liver is 17 times higher than that for the heart. This difference in tissue distribution ratio could contribute to a higher stimulation of hepatic effects such as lipid lowering by GC-1 without increasing the heart rate. This observation implies that organ-specific distribution contributes to the effect of this new thyroid hormone analog. Much lower absolute levels of GC-1 than of T3 in plasma and tissues also indicate that GC-1 may be more rapidly metabolized than T3. Details of T3vs. GC-1 pharmacokinetics need to be clarified by an in-depth study; however, it may be that the tissue-selective thyromimetic effects of GC-1 arise from a combination of selective tissue uptake and selective activation of TR 127.
The comparative effects of GC-1 and T3 on rats were determined to show the relationship between heart rate vs. cholesterol and TSH vs. cholesterol. This was performed in rats because of the greater reliability of heart rate and TSH measurements in this species. As TSH and cholesterol lowering are both thought to be primarily due to TRβ activation (12, 14), the potency ratios for these two plasma parameters should be similar for T3 and GC-1. The ED30 (TSH)/ED50 (cholesterol) potency ratios were similar for GC-1 and T3, which is consistent with them being mediated by the same receptor subtype. GC-1 was significantly more selective for cholesterol lowering compared with its tachycardic potency. The selectivity was more than 15-fold, and this may be conservative, as we could not observe significant tachycardia even at high doses. These data are consistent with the TRβ binding selectivity for this compound (15). Tachycardia is thought to be primarily due to TRα stimulation, and our results are consistent with this hypothesis; however, selective tissue uptake may also play a role, as mentioned previously.
Although GC-1 is relatively selective for TRβ, it has a slightly weaker binding affinity (2-fold less potent) compared with T3. The differences in potency for cholesterol and TSH suppression were on the order of 10- to 20-fold less potent for GC-1. These differences in potency of TRβ-mediated effects of GC-1 and T3 were not observed in mice. Cholesterol levels were similarly lowered by equivalent T3 and GC-1 doses, and GC-1 was even more effective in lowering triglycerides. The rats were euthyroid and had only elevated cholesterol levels, so an influence on triglyceride lowering could not be assessed. The effects of GC-1 and T3 on cholesterol levels in mice were maximal at the 1× doses; thus, a difference of potency for this parameter might have become apparent at lower GC-1 and T3 doses. T3 and GC-1 were administered orally to rats to determine whether the in vitro selectivity could be observed when administered in the manner in which patients will be treated. The selectivity should not be affected by differences in oral absorption, although the apparent potency could certainly be affected. The results clearly showed that therapeutic cholesterol-lowering activity was observed for GC-1 without concomitant tachycardia, unlike T3; therefore the in vitro studies correctly predicted the selectivity observed in vivo.
Previous reports (7, 8) of thyroid hormone analogs have indicated that some thyroid hormone analogs can lead to lipid lowering without significant cardiac effects. A differential effect on the heart was observed for the compound 3,5-diiodothyropropionic acid, which increased cardiac contractility without increasing heart rate and MHC isoform expression (29). The actual mechanism of action of this compound is also not clear, as 3,5-diiodothyropropionic acid differs in several aspects from the chemical structure of thyroid hormones, and its binding to thyroid receptors was not studied.
There has been extensive interest in using thyroid hormones for treating a variety of indications, including weight loss, lowering plasma lipid levels, and treatment of hypothyroid states in the elderly. Unfortunately, the use of thyroid hormones for these indications has been limited, predominantly by the fact that the compounds have adverse effects on the heart, particularly on heart rate and rhythm. The findings in the current study and in some previous reports (7, 8) indicate that lowering of cholesterol without a significant increase in heart rate is possible. Our report shows that a compound such as GC-1 has TRβ-selective actions, with over 14 times greater potency on lowering cholesterol and TSH levels than on heart rate, with similar or greater effects than T3 on triglyceride levels and diminished effects relative to T3 on weight gain and heart rate. These findings provide further support for the idea that compounds with such selectivity can be generated. Whereas the mechanism for these in vivo actions is not fully understood and may involve a combination of selective tissue uptake and selective TR subtype activation, the results suggest that compounds with properties similar to GC-1 may be useful therapeutic agents for the treatment of a variety of thyroid hormone-related metabolic disorders.
Acknowledgments
We are thankful to Brian Lin and Van Lee for the help with the animal work. We also thank William G. Humphreys, Shu Y. Chang, and Rocco J. Georg from Bristol-Myers Squibb Co. (Princeton, NJ) for assistance with the assessment of the GC-1 distribution using mass spectrometry.
