INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

AMONG THE SEVERAL TYPES of voltage-dependent Ca²⁺ channels identified in various cells, only L-type and T-type channels are expressed in cardiac myocytes. Although L-type channels have been shown to play a central role in the excitation-contraction coupling of cardiac myocytes, the function of T-type channels in the heart remains poorly understood. T-type Ca²⁺ currents (I_Ca,T) are observed in pacemaker cells, in which they are shown to participate in the electrogenesis of impulse generation (5, 27), and are clearly described in embryonic or neonatal cultured atrial and ventricular myocytes, whereas they are rarely observed in adult ventricular myocytes, except in the guinea pig (15, 12), suggesting that T-type channels are associated with development and postnatal growth of cardiac cells.

Recently, full-length cDNAs encoding three homologous but distinct new ₁-subunits (_1G, _1H, _1I) have been identified (4, 9, 17). Their expression in oocytes and mammalian cells gives rise to currents with all the typical properties of native T-type channels. Northern blot analyses have indicated that messengers for both _1G- and _1H-subunits were expressed in adult human heart (17, 4), whereas no I_Ca,T has been observed in human cardiac cells isolated from either atrial appendage or ventricular tissue (16). It is important to investigate whether cardiac I_Ca,T is related to these ₁-isoforms.

In the present paper, we describe the postnatal evolution of T-type versus L-type Ca²⁺ currents in neonatal rat atrial and ventricular myocytes. Information on the molecular nature of the cardiac T-type channels is provided by comparison of the functional properties of neonatal cardiac I_Ca,T with those reported for recombinant _1E-, _1G-, _1H-, and _1I-subtype Ca²⁺ currents and by RT-PCR experiments. The results suggest a linkage between neonatal cardiac I_Ca,T and the _1G gene. The contribution of T-type currents to Ca²⁺ signaling in neonatal atrial cells was investigated by measuring the secretion of atrial natriuretic factor (ANF).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell isolation. Single rat atrial and ventricular myocytes were enzymatically isolated following standard procedures. Four-day-old to adult rats of either sex were anesthetized with pentobarbital sodium. The heart was quickly removed and rinsed in a warm (37°C) Tyrode solution (mM: 112 NaCl, 6 KCl, 2 MgCl₂, 4 NaHCO₃, 1.5 KH₂PO₄, 25 HEPES, 10 pyruvic acid, and 5.85 glucose with 17.7 mg/l phenol red, 60 mg/l penicillin G, and 100 mg/l streptomycin, pH 7.5 adjusted with NaOH) supplemented with 2 mM Ca²⁺. The heart was then perfused retrogradely by aortic cannulation on a Langendorff system with a Ca²⁺-free Tyrode solution for 8 min followed by a 0.7 mg/ml collagenase (type II, Worthington) solution made with the Tyrode solution and containing 10 µM Ca²⁺. Perfusion solutions were warmed to 37°C, and the duration of the enzyme solution perfusion was varied with the age of the animal from which the heart had been removed, from 15 min for the youngest animals to 30 min for the oldest. At the end of the perfusion, the atria were separated from the ventricles and mechanical dissociation of the myocytes was performed with a smooth-tip Pasteur pipette. Cells were then maintained up to 8 h in a high-K⁺ medium containing (mM) 100 potassium glutamate, 10 potassium aspartate, 25 KCl, 10 KH₂PO₄, 2 MgSO₄, 20 taurine, 5 creatine, 0.5 EGTA, 20 glucose, and 5 HEPES with 0.1% BSA (pH 7.2 adjusted with KOH) at 4°C.

Electrophysiological recordings. All experiments were conducted at room temperature (20-22°C). Whole cell Ca²⁺ currents were recorded from single myocytes using the patch-clamp method with a Biologic RK 300 or Axopatch 200A amplifier interfaced to a PC computer. Fire-polished pipettes were made from borosilicate glass and had a resistance of 2.5-3.5 M. They were filled with an internal solution containing (mM) 130 CsCl, 10 EGTA, 25 HEPES, 3 ATP (Mg), and 0.4 GTP (Na) (pH at 7.2 with CsOH). Current recordings were performed in a bath solution containing (mM) 2 CaCl₂, 5 4-aminopyridine, 136 tetraethylammonium (TEA)-Cl, 1.1 MgCl₂, 25 HEPES, and 22 glucose with 17.7 mg/l phenol red (pH at 7.4 with TEA-OH). Data were filtered at 5 kHz. Mibefradil (Produits Roche) was prepared freshly and directly dissolved in the external recording solution. NiCl₂ was diluted into the recording solution from a stock solution (1 M) to the appropriate concentrations. Stimulation protocol was applied from a holding potential (HP) of 100 mV. Experiments addressing the effects of Ni²⁺, mibefradil, and SNX-482 (a gift from Dr. Robert Newcomb, Neurex) on Ca²⁺ currents were elicited by test voltage steps to 30 mV applied at 10-s intervals. pClamp6 software was used for data acquisition and analysis. Current-voltage curves were fitted using a modified Boltzmann equation
where I is the current for the test potential V, V_rev is the Ca²⁺ current reversal potential, G_LVA and G_HVA are the maximal low-voltage-activated (LVA) and high-voltage-activated (HVA) conductances, V_0.5LVA and V_0.5HVA are potentials for half-maximal activation, and k₁ and k₂ are related to the steepness of the voltage dependence of activation. For inactivation curves, Ca²⁺ currents were recorded after a 5-s conditioning prepulse, normalized, and fitted with a Boltzmann equation. Results are presented as means ± SE. Two 4-day-old rats, seven 8-day-old rats, two 3-wk-old rats, two 2-mo-old rats, and one 4-mo-old rat were used for the electrophysiological recordings.

Transient expression of recombinant Ca²+ channels. cDNAs encoding ₁-, ₂-, and -subunits and the reporter gene (CD8) were inserted in vertebrate expression vectors. Human _1G (GenBank accession number AF126965) was inserted in pBK-CMV vector (Statagene); rat brain _1E, _2a-, and ₂ were cloned in pMT2 vector (21) and CD8 cDNA in a cytomegalovirus-driven vector. Human embryonic kidney cells expressing the SV40 large antigen (tsA201 cells) were grown in DMEM medium supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin (vol/vol). For optimal transfection, cells were plated at 50-70% confluence. A Ca²⁺ phosphate transfection procedure was used with an _1G-CD8 or a _1E-₂-_2a-CD8 cDNA mix at respective molar ratios of 1:0.1 and 1:1:1:0.1. This cDNA ratio was already proven to allow the expression of all channel subunits together with the reporter gene (7). Three micrograms of mixed cDNA were used per 35-mm petri dish; after an overnight transfection, the cells were rinsed with fresh culture medium. Cells were plated at low density 24 h after transfection and used for patch-clamp studies 24 h later. Positively transfected cells were identified using anti-CD8 antibody-coated beads (Dynal). About 20% of the transfected cells were positive to the anti-CD8 antibody-coated beads and >90% expressed a Ca²⁺ current.

Secretion of ANF. The atrial tissue from 8-day-old rats was dissected and cut into small pieces of 2-mm² maximum size, transferred to the normal Locke buffer [in mM: 140 NaCl, 5 KCl, 1 MgCl₂, 10 HEPES-NaOH, 10 glucose, and 2 CaCl₂, pH 7.4, containing 0.01% bovine serum albumin (wt/vol)], and washed twice with the same buffer. The atrial tissue was placed onto filters (0.45-µm Acrodisc LCPVDF, Gelman Sciences) and perfused (Minipulse peristaltic pump, Gilson) for 45 min with normal Locke buffer at a flow rate of 50 µl/min. The flow rate was slowly increased during this period to 100 µl/min. Collection of the perfusate over 5-min periods started 60 min after the tissues were loaded onto the filter. The whole set-up was mounted in an incubator, and the perfusion experiments were performed at 37°C. ANF content in each fraction was then determined by a competitive RIA using a RIA kit purchased from Phoenix Pharmaceuticals (RK-005-24). Depolarization medium contained 50 or 20 mM K⁺ (final concentration), and a constant Na⁺ concentration was maintained by equimolar substitution of KCl for N-methyl-D-glucamine chloride. The results presented correspond to the average ANF content of standard aliquots taken from the collected fractions arising from at least three separate groups of tissues. The amount of hormone released during each period was calculated by subtracting the amount of hormone released under basal conditions from that observed during and directly after the stimulus. Student's t-test was used for statistical tests.

mRNA preparation and RT-PCR analysis. Total RNA preparations from cerebellum, atrial, and ventricular tissues of 8-day-old animals and from kidney, skeletal muscle, cerebellum, atrial, and ventricular tissues of adult rats were done using TRIzol (Life Technologies) according to the manufacturer's protocol.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ca²+ currents in neonatal cardiac myocytes. Figure 1A illustrates representative Ca²⁺ currents recorded from freshly dissociated 8-day-old rat atrial myocytes in response to increasing depolarizations. At a HP of 100 mV, I_Ca,T is activated by a depolarization around 50 mV and peak current is observed around 30 mV. For test potentials above 30 mV, an L-type current is activated that displays a maximum amplitude near +10 mV. At a HP of 50 mV, I_Ca,T is mostly inactivated whereas the L-type current remains almost unaffected. The difference between current traces at HP of 50 mV and 100 mV for each potential reflects the I_Ca,T. Figure 1B represents the mixed current-voltage curve of T-type and L-type currents present at a HP of 100 mV as well as the current-voltage curve of the L-type current recorded at a HP of 50 mV; the difference between the two curves indicates the theoretical I_Ca,T-voltage relationship. Both T-type and L-type currents were observed in atrial and ventricular cells at this age. I_Ca,T was always present with a larger amplitude than L-type current in atrial myocytes. In contrast, I_Ca,T was not systematically observed in ventricular cells, and its amplitude was always smaller than that of L-type current. Application of the -adrenergic agonist isoproterenol (2 µM) during a double test pulse protocol activating either T-type current or mostly L-type current had no effect on the T-type current, whereas a large current increase (from 70 to 160 pA) for the higher depolarization was related to enhancement of the L-type current amplitude (Fig. 1C). This result mainly indicates that there is no significant contribution of L-type current at a depolarization of 30 mV, where I_Ca,T is maximally activated.

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Typical Ca2+ currents from atrial myocytes of 8-day-old rat. A: whole cell Ca2+ currents recorded in a freshly dissociated atrial myocyte. Representative traces are shown for step depolarizations to the indicated command potentials from holding potential (HP) of 100 mV (left) and 50 mV (middle); traces shown at right represent the difference current attributable to T-type Ca2+ current (ICa,T). B: current-voltage relationships for Ca2+ current densities from the representative cell used in A: mixed current-voltage curve of T-type and L-type currents from HP 100 mV (), current-voltage curve of L-type current from HP 50 mV (), and current-voltage curve of ICa,T obtained from the difference of those 2 current-voltage curves (). C: effect of the application of 2 µM isoproterenol (Iso) during a double test pulse from HP 100 mV, activating the ICa,T at 30 mV and mostly the L-type current at +10 mV. Note that ICa,T was not affected by isoproterenol (n = 5).

Postnatal evolution of I_Ca,T density. T-type and L-type Ca²⁺ current density were measured at the peak current (30 and +10 mV, respectively) in atrial and ventricular myocytes from 4-day-old to adult rat hearts (Fig. 2). I_Ca,T in atrial cells is already predominant in 4-day-old myocytes, with a density of 4.52 ± 0.63 pA/pF (n = 14). However, I_Ca,T is maximal in 8-day-old rat atrial myocytes, with a density of 5.78 ± 0.38 pA/pF (n = 51). The density of the I_Ca,T had decreased to 2.44 ± 0.27 pA/pF after 3 wk (n = 11) and was low in adult rat atrial myocytes (n = 11). The situation is different in ventricular cells because I_Ca,T is also observed but has a smaller amplitude (vs. L-type) in 8-day-old rats (1.47 ± 0.26 pA/pF, n = 9) and then disappears in ventricular cells from 3-wk-old rats. We did not detect any major changes for I_Ca,T activation and inactivation properties in the different stages investigated (see Table 2).

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Evolution of Ca2+ current density. A: current-voltage relationship recorded from HP 100 mV on freshly isolated atrial myocytes at 8 days (n = 51) and 3 wk (n = 11) and in adult (4 mo; n = 11) and on freshly isolated ventricular myocytes at 8 days (n = 9) and 3 wk (n = 6) and in adult (2 mo; n = 6). B: histograms of peak current density amplitudes for freshly isolated atrial myocytes at 4 days (n = 14), 8 days (n = 51), and 3 wk (n = 11) and in adult of 4 mo (n = 11) and for ventricular myocytes at 8 days (n = 9) and 3 wk (n = 6) and in adult of 2 mo (n = 6). Maximum amplitudes were measured at 30 mV for the ICa,T (filled bars) and at +10 mV for the L-type current (open bars). Data are shown as means ± SE.

Investigation of molecular nature of cardiac I_Ca,T. RT-PCR analysis was performed using reverse-transcribed products from mRNA collected from rat neonatal and adult cardiac tissues as well as from other tissues (skeletal muscle, cerebellum, kidney) used as positive or negative controls. Specific primers (see METHODS) were designed to detect the presence of _1E, _1G, _1H, _1I, and _1C in heart and other selected tissues. Figure 3 shows that _1I mRNA is not present in the heart but is found in the cerebellum as expected (9), whereas _1G and _1H mRNAs are both present in the atrial and ventricular tissues of young and adult rats. We also detected _1E mRNA in atrial tissue from young rats. As expected from previous studies, our control experiments demonstrated the presence of _1C in both cerebellum and cardiac tissues, the presence of _1H in the kidney but not in the cerebellum, and the presence of _1G in the cerebellum but not in the kidney (4).

View larger version (62K): [in this window] [in a new window]   Fig. 3.   RT-PCR analysis of 1I, 1G, 1H, 1E and 1C mRNAs from different tissues of young and adult rats. 1I mRNA is not present in the heart but is present in the cerebellum. 1G and 1H mRNAs are both present in the atrial and ventricular tissues of young and adult rats. 1E mRNA is present in atrial tissue from young rats. Control experiments show the presence of 1C in both cerebellum and cardiac tissues, the presence of 1H in the kidney but not in the cerebellum, and the presence of 1G in the cerebellum but not in the kidney.

Biophysical properties of cardiac I_Ca,T. I_Ca,T properties were studied in atrial myocytes from 8-day-old rats at a depolarization of 30 mV, at which I_Ca,T is maximally activated without a significant contribution of L-type current. The current activates more rapidly with increasing depolarizations, the time to peak values ranging around 25 ms for a potential of 50 mV to <10 ms for a test potential above 50 mV (Fig. 4A). As previously reported in various cell types, I_Ca,T also inactivates faster with depolarization, with a time constant of 12.9 ± 0.4 ms at 30 mV. Measurements were not performed for higher depolarizations because of contamination by L-type current. The steady-state inactivation relationship obtained using a classical double-pulse protocol indicates a half-inactivation potential of 68.4 mV with a slope of 4.09 mV. The normalized values of steady-state activation (obtained from the current-voltage curve) and inactivation (Fig. 4B) reveal a window current between 65 and 45 mV. Membrane repolarization at the time to peak current displays slow deactivating tail currents, a hallmark of T-type currents, because the time constant is 1.3 ms for repolarization at 100 mV and 4.1 ms at 60 mV (Fig. 4C).

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Biophysical characteristics of ICa,T from 8-day-old atrial myocytes. A: time to peak of Ca2+ currents in atrial cells (n = 12). B: voltage-dependent activation and inactivation curves of ICa,T. The steady-state inactivation curve (n = 32, smooth curve) is fitted using a simple Boltzmann function [V0.5 = 68.4 mV, slope factor (k) = 4.09 mV], whereas activation curve is constructed from values of half-activation potential [V0.5(T) = 41 mV] and with the slope factor (kT = 5.16 mV) obtained from the fit of the current-density voltage curve. C: typical recordings illustrating deactivation kinetics of a ICa,T. Activation at 30 mV is followed by repolarization to 100 mV and 60 mV. Time constants of deactivation are 1.3 and 4.1 ms, respectively; n = 5.

Pharmacology of cardiac I_Ca,T. Several ions and molecules have been used to distinguish between L-type and T-type Ca²⁺ channels such as Ni²⁺ or mibefradil, a nondihydropyridine compound considered the most selective T-type versus L-type channel blocker (11). We found that the cardiac T-type channel is poorly sensitive to Ni²⁺ because the apparent K_d for inhibition [as shown by 50% inhibitory concentration (IC₅₀)] is ~160 µM (Fig. 5A). Moreover, Ni²⁺ could not be used to separate the L-type from the T-type current because the IC₅₀ for L-type channel inhibition was a similar concentration (192 µM; not shown). The inhibition of I_Ca,T by mibefradil occurred with an IC₅₀ of 0.1 µM (Fig. 5B). From experiments using an antisense strategy, it was previously suggested that the neonatal cardiac I_Ca,T might be related to the _1E-subunit (18). We also tested the effect of SNX-482, a recently described potent specific blocker of recombinant _1E currents (14). No effect on cardiac I_Ca,T or expressed human _1G T-type current was observed in the presence of 100 nM SNX-482, whereas the same concentration of toxin, as expected, strongly inhibited _1E-generated current in HEK-293 cells (Fig. 5C).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Pharmacological properties of ICa,T from 8-day-old atrial myocytes. A: dose-response curve of Ni2+ on cardiac ICa,T; 50% inhibitory concentration (IC50) = 160 µM (n = 8). B: dose-response curve of mibefradil on cardiac ICa,T; IC50 = 0.11 µM (n = 11). C: comparison of the effect of SNX-482 between cardiac ICa,T and recombinant 1G and 1E currents expressed in HEK-293 cells. Representative traces before () and after () application of 100 nM SNX-482 on cardiac and recombinant 1G and 1E currents are shown. Note the absence of effect on the cardiac T-type and 1G-related currents, whereas 1E-related currents are strongly inhibited. Test depolarizations were elicited from HP 100 mV to the maximum peak current (30 mV for the cardiac T-type and 1G currents and 0 mV for 1E current). The rate of stimulation was 0.1 Hz.

I_Ca,T contribution to ANF secretion. To determine whether T-type channels might contribute to a physiological function such as hormonal secretion, ANF release was measured with peptide-specific RIA in perfused atrial tissue of 8-day-old rats. Figure 6A shows that membrane depolarization evoked by a high concentration of K⁺ (50 mM) induced a large increase in ANF secretion from 4.1 ± 0.2 to 56.4 ± 7 pg/15 min. The same protocol applied in the presence of 5 µM nitrendipine, which at this concentration blocks L-type without affecting T-type Ca²⁺ current (not shown), leads to 21.78 ± 2.5 pg/15 min of ANF secretion, which corresponds to a large inhibition (61%). Application of 1 µM mibefradil, which preferentially blocks I_Ca,T at this concentration, almost abolished ANF secretion (6.03 ± 0.8 pg/15 min, equivalent to 89% inhibition). Another set of experiments was performed using 20 mM K⁺ to induce a weaker depolarization activating mostly T-type channels (19). A significant increase of ANF release was induced by application of the 20 mM K⁺ solution from 4.9 ± 0.4 to 14.4 ± 1.8 pg/15 min. This release is 25% lower than the value obtained with 50 mM K⁺ (Fig. 6B). In these experimental conditions, nitrendipine did not significantly block the evoked ANF release (12.03 ± 0.5 pg/15 min) whereas ANF release was completely abolished by the application of mibefradil (1 µM). It is also interesting to note the existence of a basal level of ANF secretion (Fig. 6C). The application of 1 µM mibefradil also reduced basal ANF release by 33.5% (from 4.9 ± 0.4 to 3.26 ± 0.1 pg/15 min).

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Effects of Ca2+ channel blockers on the atrial natriuretic factor (ANF) release from 8-day-old rat atrial tissues. ANF release was stimulated by 50 (A) or 20 (B) mM K+ either in absence of channel blockers or in presence of 5 µM nitrendipine or 1 µM mibefradil. C: basal ANF release in absence or presence of 1 µM mibefradil. Each of the bars in A represents a mean of 3 measurements, and each of the bars in B and C represents a mean of 4 measurements; 3-4 rats were used for each series of experiments. *P < 0.02, ***P < 0.001; NS, not significant.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We report here the postnatal evolution of I_Ca,T density in freshly dissociated atrial and ventricular myocytes from rat heart. This current exhibits the typical properties of T-type currents; in particular, slowly deactivating tail currents are observed during membrane repolarization. An overlap between the steady-state activation and inactivation relationships, i.e., a window current, indicates that a fraction of the T-type channel remains open and might generate a maintained Ca²⁺ influx in the voltage range between 65 and 45 mV. Our results show that I_Ca,T density in atrial cells peaks at postnatal day 8. It must be noted that in all cases, when present, I_Ca,T had similar biophysical properties. Our results differ from those of Xu and Best (26), who found a maximum I_Ca,T density in 5-wk-old rat atrial myocytes. They did not record Ca²⁺ currents before postnatal week 3, and the LVA Ca²⁺ current density reported in their experiments is four times lower than the density observed here 1 wk after birth.

Cardiac T-type channels in 8-day-old cardiomyocytes are blocked by a micromolar concentration of mibefradil (IC₅₀ = 0.1 µM), as expected from previous studies on cardiovascular cells (11). Moreover, we found that neonatal I_Ca,T was poorly sensitive to Ni²⁺ (IC₅₀ = 160 µM). Several genes coding for a T-type channel pore-forming subunit (_1G, _1H, _1I) have been identified in rat, mouse, and human (4, 8, 9, 13, 17). Among these genes, _1H mRNA was found in adult human heart (4) and _1G was also detected in rat and human heart (17). However, _1H was initially considered to be the cardiac T-type isoform. It is interesting to note that _1H- and _1G-related currents in expression systems strongly differ in their Ni²⁺ sensitivity (IC₅₀ = 13 and >150 µM, respectively; Refs. 10, 13), a property that may be considered as a signature to establish correlations between native and recombinant channels. Thus the Ni²⁺ sensitivity of the neonatal atrial I_Ca,T reported here is much closer to the value reported for _1G-related currents. This result suggests that _1G is related to the cardiac I_Ca,T expressed in rat neonatal cells. Several studies have reported that in adult myocytes of rabbit, frog, cat, and guinea pig, cardiac I_Ca,T are completely blocked by low Ni²⁺ concentration (40 µM) (2, 5, 12, 15, 27). We cannot exclude the possibility of different T-type isoform expression among various species. A developmental switch between isoforms is likely and is consistent with the recent report by Monteil et al. (13) demonstrating developmental regulation of _1G transcript in human heart.

A previous study supported the idea that cardiac I_Ca,T may be related to the _1E gene (18) although the experimental conditions differed in the sense that I_Ca,T expression was enhanced by hormone treatment. Our experiments do not support this hypothesis, because we show here the absence of an effect of SNX-482 on I_Ca,T. A recent study reported the existence of a SNX-482-resistant R-type current (23), but the permeation and conductance properties of the channel underlying this current are clearly different from any genuine T-type channel. RT-PCR analysis was performed on neonatal and adult rat atrial and ventricular tissues to determine the molecular make-up of T-type channels. Although the presence of _1E mRNA is detected in 8-day-old cardiomyocytes, this may indicate that the protein is not targeted to the cell surface or is not functional. Moreover, atrial and ventricular tissues used for mRNA preparation could contain intracardiac neurons known to express several Ca²⁺ channels including the R-type channel assumed to be encoded by the _1E-subunit (6). As expected (9), the _1I isoform is not expressed in cardiac tissue from either young or adult rats. Although I_Ca,T are not detected in adult ventricular myocytes, both _1G and _1H transcripts are present in neonatal and adult atrial and ventricular tissues. This might indicate that unidentified factors or unidentified subunits may control T-type gene expression or function during development. This is consistent with the report by Xu and Best (25) of the enhancement of the cardiac I_Ca,T by growth hormone. We cannot exclude that such a factor interferes with the targeting of the protein to the membrane or interacts with the T-type channel protein to inhibit its activity. Altogether, our data on the molecular and pharmacological characteristics of the neonatal cardiac T-type channel strongly suggest that it is encoded by the _1G gene. This is in agreement with the recently published work of Satin and Cribbs (20), who reported the _1G isoform as supporting the T-type current in a cell line derived from mouse atrial tissue.

The physiological role of the T-type channel in atrial cells remains unclear. It is often proposed to be involved in pacemaking activity in the heart, more specifically, in sinoatrial node cells (5). In other cell types, T-type channels have been shown to contribute to Ca²⁺-dependent hormone secretion such as aldosterone in adrenal glomerulosa cells (3, 19) or insulin in a pancreatic -cell line (1). Cardiac tissue is known to be the main source of ANF release, a key regulator in the homeostasis of salt and water and in the maintenance of blood pressure, which in the normal adult heart is mainly restricted to both atria. Substantial changes in ANF gene expression take place at the time of birth, the ANF gene being highly expressed during fetal life both in atria and ventricle. A peak of ANF mRNA is observed in atrial tissue the first day after birth followed by a progressive decrease over 2 wk to reach near-adult levels (24). ANF secretion was shown to be stimulated by several factors, including atrial stretch and vasoactive agents such as angiotensin II, endothelin, or vasopressin, and most studies found that their action occurred via an elevation of the intracellular Ca²⁺ concentration (22). Our results show that ANF secretion is sensitive to dihydropyridine, confirming that L-type Ca²⁺ channels participate in ANF secretion. However, ANF release is more sensitive to mibefradil, especially during a weak depolarization activating mostly T-type channels, indicating their substantial contribution to ANF secretion. In addition, mibefradil inhibits a basal component of ANF secretion, which might be explained by the existence of the window current reflecting a population of T-type channels open within the range of the cells' resting potential. Because of the predominance of T-type versus L-type Ca²⁺ channel expression in the early postnatal period, our data suggest an important contribution of cardiac T-type channels in Ca²⁺ signaling and in physiological functions such as hormone release during the neonatal period.