INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A SMALL PERCENTAGE of the aged population fulfills the criteria for so-called "latent" hyperthyroidism; i.e., serum levels of thyroid-stimulating hormone (TSH) are decreased, but the circulating hormones thyroxine and triiodothyronine are within the normal range (26, 31). The therapeutic relevance of this finding is disputed (28-30), but these patients are known to have an approximately threefold increased risk to develop atrial fibrillation over time (26). Until now, it has also been unclear whether latent hyperthyroidism constitutes nothing but a subclinical risk factor or whether it represents a pathophysiological entity by itself, as suggested by studies showing that latent hyperthyroidism is more common in patients already presenting with atrial fibrillation (2, 5, 22). Therefore, it seems straightforward to use cardiac atrial tissues from such patients to look for a molecular pathophysiological basis of latent hyperthyroidism.

Animal experiments have been frequently used to investigate the molecular sequelae of manifest hyperthyroidism. Among other factors, Ca²⁺-handling proteins are altered. For the sarcoplasmic reticular proteins phospholamban (PLB) and sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA, the Ca²⁺ pump), an increased SERCA-to-PLB ratio has been found. This result was based on reduced PLB expression [dog ventricle (17) and rat atrium (13)], increased SERCA expression [rat ventricle (3) and baboon ventricle (14)], or both [cultured rat myocytes (16)]. In any case, improved Ca²⁺-sequestering function of the sarcoplasmic reticulum should result from these changes in experimental hyperthyroidism. At the sarcolemmal level, Ca²⁺ handling is altered because of increased expression of L-type Ca²⁺ channels, the major source of activator Ca²⁺ in heart. Action potentials (1), whole cell Ca²⁺ currents (1, 25), and the density of dihydropyridine-binding sites (15) consistently indicate a marked increase of cardiac Ca²⁺ channels in experimental hyperthyroidism. In addition, Ca²⁺ channels are well-known targets of -adrenergic stimulation, which is indeed augmented in hyperthyroidism (7).

On the basis of these clinical and experimental results, we decided to study Ca²⁺ channel function and expression in human atria. The atrium is the functionally relevant tissue (regarding atrial fibrillation). The heart is a well-characterized target organ of thyroid hormones. We took advantage of the fact that a small atrial biopsy is routinely taken at every cardiac surgery in which cardiopulmonary bypass is used. This allowed us to screen a large number of patients within a reasonable time. Conceptually, this strategy allows for a comparison between affected and "control" patients, because the reason for surgery (mostly coronary bypass surgery) was independent of the variable of interest, i.e., thyroid state (i.e., normal or latent hyperthyroidism).

We deliberately chose the more complicated single-channel recording mode: this technique, in contrast to the whole cell mode, not only allows detection of the direction of change of current but also displays the mechanism of regulation in more detail. Indeed, our own single-channel studies (27) of human ventricular myocytes (in heart failure patients) revealed results undisclosed by whole cell experiments. Together with an estimate of the number of total channels, single-channel behavior is the main determinant of the Ca²⁺ current density. Therefore, we measured Ca²⁺ channel expression at the protein level. To obtain a more complete picture of Ca²⁺ homeostasis, the sarcoplasmic reticular proteins PLB and SERCA were also quantified by Western blot analysis.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient selection. Over a period of 18 mo, TSH screening was performed routinely on patients before elective cardiac surgery. A TSH value 0.1 µU/ml was required for enrolling a patient in the latent hyperthyroid group; a value 1.0 µU/ml was required for controls (normal range 0.5-3.5 µU/ml). In every case, the free serum triiodothyronine and the free serum thyroxine levels had to be within the normal range (2.0-5.0 pg/ml and 0.8-2.3 ng/ml, respectively). The entry criteria for both groups was as follows: 1) no evidence of previous or actual thyroid disease, 2) no iodine exposure 4 wk before surgery (except for purposes of cardiac diagnosis, which was carried out within this time frame for 4 patients classified as latent hyperthyroid and 6 controls), 3) no actual therapy with iodine-containing drugs, thyroid hormones, or thyreostatic agents, and 4) full documentation of actual medication. All patients enrolled in the latent hyperthyroid group were followed up. TSH suppression was confirmed, and thyroidal autonomy was found scintigraphically in every case. A similar number of control cases were randomly (i.e., taken by chance from the cohort presenting with normal TSH values) selected from the patients fulfilling the same entry criteria. Controls were enrolled over the same time period, and no attempt was made to match the two groups regarding demographic or anamnestic parameters. The control patients received no follow-up. The protocol was approved by the local Ethics Committee.

Cell isolation. Immediately after surgical excision, the piece (~300 mg) of right atrial tissue was placed in a preoxygenated solution of 4°C (solution A) composed of (in mM) 100 NaCl, 10 KCl, 5 MgSO₄, 20 dextrose, 50 taurine, and 5 MOPS (pH 7.4) and taken to the laboratory within 15 min. After removal of fat and connective tissue, part of the tissue was dissected into <1 × 1 × 1-mm pieces in oxygenated solution A at room temperature. (The remaining tissue was deep-frozen in liquid nitrogen and stored at 80°C for molecular biological analysis.) Samples were digested in 10 ml of solution A in the presence of collagenase (1.5 mg/ml; type CLS 1, Worthington Biochemical), trypsin (1 mg/ml; type III, Sigma Chemical), and BSA (10 mg/ml, Sigma Chemical) at 37°C for 40 min under continuous stirring and shaking. After the supernatant was decanted and the tissue was washed with solution A once, it was incubated again (20-65 min, terminated after appearance of single viable myocytes) in the presence of collagenase (0.5 mg/ml) and BSA (1 mg/ml). The digestion was stopped by dilution of the cell suspension with solution A (1:5, 37°C). After centrifugation (10 min, 700 U/min; Heraeus Labofuge 400) the pellet was resuspended and stored in a solution (solution B) containing (in mM) 50 potassium glutamate, 40 KCl, 20 KH₂PO₄, 20 taurine, 20 KOH, 3 MgCl₂, 10 HEPES, 5 EGTA, and 10 dextrose (pH 7.4).

Electrophysiological measurements. Cells were placed in petri dishes containing (in mM) 120 potassium glutamate, 25 KCl, 2 MgCl₂, 10 HEPES, 2 EGTA, 1 CaCl₂, 1 Na-ATP, and 10 dextrose (pH 7.4 with NaOH, 21-23°C). Pipettes (borosilicate glass, 7-10 M) were filled with (in mM) 70 BaCl₂, 110 sucrose, and 10 HEPES (pH 7.4 with tetraethylammonium hydroxide). Single Ca²⁺ channels were recorded in the cell-attached configuration (depolarizing test pulses of 150 ms at 1.67 Hz). Pulse generation, data acquisition (10 kHz), and filtering (2 kHz, 3 dB, 4-pole Bessel) were done using an Axopatch 200A and pClamp 6.0 software (both from Axon Instruments, Foster City, CA).

Drugs. In some experiments, FPL-64176 {methyl 2,5-dimethyl-4-[2-(phenylmethyl)benzoyl]-1H-pyrrole-3-carboxylate; RBI, Natick, MA; 1 mM dissolved in DMSO} or isoproterenol (Sigma Chemical, Deisenhofen, Germany; 0.1 mM dissolved in water, stabilized by ascorbic acid) was added to the bath as a 30-µl bolus. The actual bath volume (~3 ml) was measured after the experiment. The calculated final drug concentrations were 6.3-10.7 µM FPL-64176 and 0.8-1.1 µM isoproterenol. Unless noted, channels were recorded in the absence of agonists.

Single-channel data analysis. Linear leak and capacity currents were digitally subtracted using the average currents of nonactive sweeps. Openings and closures were identified by the half-height criterion by the use of the pClamp software. The availability (percentage of depolarizing test pulses eliciting 1 opening) and the peak current of the ensemble average were corrected for the number of channels in the patch (N), where necessary: N was derived from the maximum current amplitude observed divided by the unitary current amplitude. Peak current was normalized by division through N. The availability was corrected by the square-root method: (1  availability_corrected) is the Nth root of (1  availability_uncorrected). The open probability (open state density during active sweeps) and the mean closed time were evaluated for patches containing one channel only (>60% of all patches). Patches containing more than two channels were entirely discarded. Activation and inactivation data (obtained from 60 sweeps for each voltage protocol) were fitted to a Boltzmann-type function (8) according to the following equation: y = y_max · 1/{1 + exp [(V_0.5  V)/k]}, where y (y_max) is the (maximum) availability of channels, V is the membrane potential, V_0.5 is the potential at which activation or inactivation is half-maximal, and k is the slope factor describing the steepness of the curve at V_0.5.

Western blot techniques. For expression of L-type Ca²⁺ channel _1C-subunits, total cell lysates (40 µg/lane) of each sample were subjected to SDS-PAGE on 7.5% gels for resolution of L-type _1C-subunit expression. Protein was transferred electrophoretically to a nitrocellulose membrane. Equal transfer among lanes was verified by reversible staining with Ponceau red. Immunoblotting was performed with polyclonal antibodies directed against the L-type channel _1C-subunit (1:1,000; Alomone Labs, Jerusalem, Israel) and calsequestrin (1:1,000; SWANT, Bellinzona, Switzerland). Detection was performed with the enhanced chemiluminescence technique (Amersham, Braunschweig, Germany). Densitometric analysis of immunoblots was performed on an Epson GT 8000 scanner with the analysis software ScanPak (Biometra, Göttingen, Germany). For this series of experiments, biopsies were taken from seven patients included in the single-channel analysis, six additional control patients (3 men and 3 women, mean age 62.8 ± 4.6 yr), and one additional patient with latent hyperthyroidism (1 woman, 77 yr of age). For SERCA and PLB expression, frozen tissue (from the same patients as in the single-channel study) was homogenized in 100 µl of 10 mM NaHCO₃ with use of a microdismembrator (Braun Melsungen, Melsungen, Germany), then 200 µl of 20% SDS were added and protein was extracted at 25°C for 30 min. The samples were pelleted at 14,000 g at 4°C for 20 min, and the protein concentration in the supernatant was assayed (20). SDS extracts made as described above were thawed, and SDS buffer (19) was added. The samples were heat treated for 10 min at 30°C. Thirty micrograms of homogenate sample protein were loaded per lane. These amounts were in the linear range for PLB, SERCA2a, and calsequestrin (data not shown). Gels were run (23) using 8% polyacrylamide separating gels. Electrophoresis was initially run at 40 mA per gel for 30 min, and then the current was increased to 60 mA. Proteins were electrophoretically transferred to nitrocellulose membranes (45 µm pore size; TA 85, Schleicher & Schuell, Dassel, Germany) in 50 mM sodium phosphate buffer (pH 7.4) for 180 min at 1.5 A at 4°C, as reported elsewhere (24). Then membranes were blocked in Tris-buffered saline containing 2.0% BSA for 30 min and incubated overnight at 4°C with antibodies directed against PLB (monoclonal A₁, PhosphoProtein Research, Bardsey, UK) (4), SERCA (2A7-A1)(21), and calsequestrin (12). PLB antibody was detected using ¹²⁵I-labeled goat anti-mouse IgG (ICN Biomedicals, Eschwege, Germany), and SERCA and calsequestrin antibody were detected using ¹²⁵I-labeled protein A (ICN Biomedicals, Eschwege, Germany). Bound radioactivity was visualized and quantified in a PhosphorImager with use of ImageQuant software (version 3.30, Molecular Dynamics, Sunnyvale, CA).

Statistics. Values are means ± SE. Significance was checked in such cases by two-tailed Student's t-test (P < 0.05) for paired or unpaired observations, as appropriate. For comparison of patient characteristics, Fisher's exact test was used (2-tailed, P <0.05).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient characteristics. The baseline characteristics of all patients included in the single-channel analysis are compiled in Table 1. In most cases, the diagnosis leading to cardiac surgery was coronary heart disease. Accordingly, medication was not significantly different between the two groups: in patients with latent hyperthyroidism, NO donors were taken in 11 cases (11 controls), -blockers in 8 cases (8 controls), angiotensin-converting enzyme inhibitors in 9 cases (6 controls), and diuretics in 6 cases (4 controls). The patients with latent hyperthyroidism were older, and this group consisted of more female patients (P < 0.05). This was expected from our random sampling strategy and from the higher prevalence of coronary heart disease in older women. The free levels of triiodothyronine were significantly higher in latent hyperthyroidism but were still within the normal range for every patient. Atrial fibrillation (actual or in history) was threefold higher in latent hyperthyroidism (not significant).

Identification of Ca²⁺ channel activity. In about one-third of all patches where gigohm seals were obtained, single-channel activity could be elicited. To check whether these channels are L-type Ca²⁺ channels, pharmacological and biophysical tests were applied. Figure 1 shows the voltage dependence of activity and the effect of the L-type-specific Ca²⁺ channel activator FPL-64176 (18). Test pulses between 20 and +30 mV elicited openings of decreasing amplitude, whereas the number of openings increased with positive potentials. FPL-64176 (n = 5) caused a typical prolongation of openings, allowing calculation of a single channel conductance (Fig. 1C) of 22.3 ± 1.1 pS. All these features, together with the long-lasting channel activity throughout the 150-ms test pulse (see below), indicate that the channels recorded are typical for human cardiac L-type channels (8). Only on two occasions were rapidly inactivating channels with a lower conductance (probably representing T-type channels) observed (not shown).

View larger version (1610K): [in this window] [in a new window]   Fig. 1.   Single-channel properties of barium currents in human atrial myocytes: effect of FPL-64176. A: representative current response to test pulses (0, +10, and +20 mV). Scale bars, 20 ms and 0.5 pA. B: channel shown in A after addition of FPL-64176. Scale bars as in A. C: unitary current amplitude (i) vs. test pulse voltage in 5 channels exposed to ~10 µM FPL-64176. Conductance (see RESULTS) was calculated by linear regression of individual channels.

Voltage dependence of channel activity. To characterize the channels in more detail, a series of experiments was conducted where holding potentials and test potentials were varied over a wide range. Channel availability was measured and analyzed as a Boltzmann function of these holding and test potentials (Fig. 2). It can be seen, first, that holding potentials of 70 mV or less and test potentials of at least +10 mV were sufficient for maximum availability. Second, the shape and position of these "activation" and "inactivation" curves were similar between the two groups. Boltzmann functions of availability yielded the following values for activation (k = 5.7 and 4.6 mV and V_0.5 = 3.3 and 4.2 mV) and inactivation (k = 9.8 and 8.9 mV and V_0.5 = 33.1 and 40.9 mV) for latent hyperthyroidism and controls, respectively. This allowed for an unbiased comparison of the groups by use of a voltage protocol with holding and test potentials of 100 and +20 mV, respectively (see below). Third, the only apparent difference (see also below) was the maximum availability (y_max), which was higher for latent hyperthyroidism for activation (57 vs. 45%) and inactivation curves (60 vs. 42%).

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Dependence of channel availability on test potential (A and C) or holding potential (B and D). Channels were from controls (A and B) or patients with latent hyperthyroidism (C and D). Data are from 10 (A) or 6 (B-D) patches. Lines are best-fit Boltzmann-type functions.

Single-channel gating in latent hyperthyroidism. Channels were now stimulated using voltage pulses from 100 to +20 mV (see above) over prolonged periods. Experiments with 120 sweeps (240 in most cases) recorded under these conditions were included in the analysis. Figure 3 shows representative examples from a control and a patient with latent hyperthyroidism. In the latter case the ensemble average current trace (bottom trace) is larger. This is due to an increased availability (fraction of sweeps containing channel activity) and an increased open probability within such active sweeps. These effects were more pronounced in patients not receiving -blocker treatment (Fig. 3). Unexpectedly, this treatment had an increasing effect by itself on the same gating parameters. The details of channel gating are therefore given separately in Table 2. However, in patients treated with -blockers, we found the same qualitative influence of latent hyperthyroidism, although it was less pronounced than in the group of nontreated patients. In summary, latent hyperthyroidism leads to increased single-channel activity, caused by an increase in the availability (cf. Fig. 2) and open probability (mainly due to shorter closed times; Table 2). Open times were not significantly different. The (chronic) influence of latent hyperthyroidism on single-channel behavior (increase of peak current, availability, and open probability due to shorter closed times) resembles the acute effect of -adrenergic stimulation as known from animal experiments (10, 11). To determine whether -adrenergic stimulation exerts similar effects in human atrial L-type channels, we exposed eight patches (6 from controls and 2 from patients with latent hyperthyroidism) to isoproterenol. Regardless of thyroid state and pretreatment with -blockers (n = 4), peak current and availability were enhanced in every individual case. In summary, peak current rose from 31.4 ± 6.8 to 58.8 ± 14.8 fA (P < 0.05), and availability increased from 43.0 ± 6.4 to 64.2 ± 6.2% (P < 0.05). In the four one-channel patches, mean closed time dropped from 7.6 ± 1.8 to 3.3 ± 0.75 ms [not significant (NS)]. Open times were not markedly affected (from 0.64 ± 0.12 to 0.82 ± 0.29, NS). The data set is too small for a meaningful comparison between the subgroups (thyroid state and -blocker pretreatment) but shows that acute isoproterenol effects resemble qualitatively the chronic influence of latent hyperthyroidism.

View larger version (36K): [in this window] [in a new window]   Fig. 3.   Single-channel effect of latent hyperthyroidism. A: consecutive traces from a 1-channel patch of a control patient. Top trace: voltage protocol; bottom trace: average current of whole ensemble; middle traces: consecutive sweeps. Scale bars, 15 ms and 2 pA (unitary current) or 15 fA (ensemble average). B: plot shown in A, but from a patient with latent hyperthyroidism. C: summary of results from controls (n = 11) and from patients with latent hyperthyroidism (Lat hyp, n = 7). There was no -blocker treatment in either group. * Statistical significance (P < 0.05).

Ca²⁺ channel expression. We checked whether latent hyperthyroidism regulates the expression level of the L-type channel in human atrial cardiomyocytes. Immunoblot analyses of lysates from cardiomyocytes identified a 210-kDa band, corresponding to the expected size of L-type channel _1C-subunit (Fig. 4). In samples obtained from patients without latent hyperthyroidism, only a weak signal was detected. Latent hyperthyroidism led to a marked, significant (P < 0.01) increase in the abundance of protein (3.23 ± 0.45-fold; Fig. 4B). All data were normalized against the expression of calsequestrin, inasmuch as this protein has been shown not to be altered in heart disease. Taken together, these data show that in patients with latent hyperthyroidism the expression of the L-type channel _1C-subunit is upregulated in cardiac atrial myocytes compared with samples obtained from patients with normal thyroid function.

View larger version (7011K): [in this window] [in a new window]   Fig. 4.   Cardiac expression of L-type channel 1-subunit in latent hyperthyroidism. Cellular lysates of cardiomyocytes were subjected to SDS-PAGE (40 µg protein/lane), immunoblotted with anti-L-type 1c-antibody primary antibody, and visualized by a chemiluminescence technique. A: in a representative immunoblot, a weak signal at expected size of 210 kDa could be detected in euthyroid patients (control, C). In samples obtained from patients with latent hyperthyroidism (HT), expression level increased >3-fold. Protein from rat ventricle was used as a standard in all blots. B: densitometric analysis of 3 independent experiments. * P < 0.01 vs. control.

Expression of SERCA and phospholamban. As shown in Fig. 5, SERCA and PLB expression were not significantly altered in patients with latent hyperthyroidism. SERCA expression tended to be slightly increased (P = 0.085). The SERCA-to-PLB ratio was likewise moderately higher, and this difference also failed statistical significance (P = 0.11). Sarcoplasmic reticular Ca²⁺-handling proteins, in contrast to the L-type Ca²⁺ channel, are therefore largely unaffected by thyroid state in these patients (Table 3).

View larger version (67K): [in this window] [in a new window]   Fig. 5.   Autoradiograms of nitrocellulose strips: protein expression of sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA), calsequestrin (CSQ), and phospholamban (PLB) in atria of control patients (Ctr) and patients with latent hyperthyroidism (LH). Atrial homogenates were prepared as described in METHODS, subjected to gel electrophoresis, and transferred to nitrocellulose. Nitrocellulose strips were treated with antibodies against PLB, SERCA, and CSQ. PLB was detected using 125I-labeled protein A. Bound radioactivity was visualized in a PhosphorImager. Left: molecular weight standards.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our study shows a marked increase of expression of Ca²⁺ channels in atria from patients with latent hyperthyroidism. At the single-channel level, the channels are more active. The pattern of activity, i.e., increased peak current, availability, and shorter closed times, resembles acute isoproterenol action. This may be speculatively explained by a higher baseline level of -adrenergic stimulation and increased cAMP-dependent phosphorylation. With the assumption that all channels detected by Western blot are functionally active, atrial cells from patients should reveal a dramatically higher whole cell Ca²⁺ current. It is tempting to speculate that this represents the cellular electrophysiological basis of atrial fibrillation found in such patients in epidemiological studies (26). We also noted a higher incidence of chronic atrial fibrillation in patient history (3 of 14 vs. 1 of 15), but the absolute numbers are much too small for a meaningful statistical analysis. It is worthwhile to further address these issues by comparing whole cell currents, which can be obtained more easily in larger numbers of patients, with thyroid state, medication, and their clinical outcome. Such data would help firmly elucidate the functional role of Ca²⁺ channel regulation in latent hyperthyroidism.

It may be argued that the findings are confounded by the higher age and higher proportion of female patients enrolled in the latent hyperthyroidism group. This was the outcome of our random sampling strategy for patient recruitment and may be circumvented in larger studies by use of controls matching all relevant parameters. However, at first glance, a significant bias of our data is unlikely: When the relevant electrophysiological parameters (peak current and availability) are plotted against age (both groups) or gender (latent hyperthyroidism), no correlation was revealed. Furthermore, in the expression study, specimens from additional control patients matching the other group (older, more women) were included, such that this particular result is not as much influenced by gender or age. Finally, a subgroup of patients (controls and patients with latent hyperthyroidism, respectively, n = 5, no -blockers) of the single-channel study was reanalyzed in an age (range 68-72 yr, mean 69.7 and 69.5 yr)- and gender (all men)-matched manner: Values similar to those in the whole data set (Table 2) were recorded for availability (24.6 ± 4.4 and 54.0 ± 9.4%, n = 6 and 4 patches, P < 0.05), peak current (15.6 ± 3.2 and 58.8 ± 23.4 fA, P = 0.05), and open probability (5.9 ± 2.9 and 9.0 ± 4.1%, NS) in these controls and hyperthyroid patients, respectively.

The unexpected finding was the effect of chronic -blockade on single-channel behavior. This treatment affected single-channel parameters in a manner resembling latent hyperthyroidism, and it partly occluded the effect related to thyroid function. We have no explanation to offer at this stage, but we believe that the presence of these drugs under our recording conditions is very unlikely, given the fact that cells are washed and subjected to medium changes >10 times between removal of the biopsy and patch-clamp measurement. Interestingly, -blocker treatment had no discernible effect on channel expression (not shown), such that a chronic adaptation of the channels seems to take place at a posttranslational level only.

This study extends a known cardiac effect of hyperthyroidism, i.e., increased number (15) and activity (1) of L-type Ca²⁺ channels, in three important and novel ways: 1) it is shown for the first time in native human tissue, 2) the effect is resolved at the single-channel level, and 3) it is seen already in the "latent" stage of hyperthyroidism, where free hormones are still "normal," yet the pronounced TSH suppression clearly indicates that thyroid hormone action is exceeding physiological boundaries. Latent hyperthyroidism is not considered a contraindication to surgery, thereby offering access to many sources of human tissue to be studied ex vivo by taking advantage of a clinically mild, but valid, endocrine "model" disorder.

Interestingly, SERCA levels only tended to increase very slightly, qualitatively similar but quantitatively behind expectations from some animal studies with manifest hyperthyroidism (3, 14, 16) and far behind the extent of Ca²⁺ channel upregulation. It would be helpful to identify genes with known promotor regions and thyroid hormone response elements (such as SERCA) (9) that are regulated in latent hyperthyroidism. Unfortunately, the promotor region of _1C-subunits of the L-type Ca²⁺ channel has not been cloned. It has been shown recently that 17-estradiol also regulates the expression pattern of the L-type channel _1C-subunit in cardiac tissue (6), demonstrating that the expression of the L-type Ca²⁺ channel is modulated by various endocrine factors. These issues should be further addressed in molecular biological studies.

In conclusion, latent hyperthyroidism has a clear-cut molecular manifestation in human right atrial tissue, i.e., increased expression and activity of L-type Ca²⁺ channels. This alteration may lead to a known clinical complication, atrial fibrillation. It is worthwhile to use this human model of endocrine disease to further study cardiac tissue from such patients to understand thyroid pathophysiology.