Introduction

Top Summary Introduction Procedures Results Discussion References

Repolarization from the plateau phase of the AP in ventricular myocytes is controlled by a delicate balance between inward and outward currents in the setting of a high membrane resistance. Important outward currents that determine repolarization are I_Kr and I_Ks (Sanguinetti and Jurkiewicz, 1990). Several class III antiarrhythmic agents block I_Kr and thereby prolong APD and the QT interval on the electrocardiogram. Excessive APD prolongation by these drugs causes LQT, which is associated with torsades de pointes, a ventricular tachyarrhythmia that can degenerate into ventricular fibrillation and cause sudden death.

LQT can also be inherited. The finding that mutations in HERG, the gene that encodes I_Kr channels, cause inherited LQT (Curran et al., 1995; Sanguinetti et al., 1995, 1996a) provided a mechanistic link between acquired LQT and one form of inherited LQT. The most common form of LQT is caused by mutations in KvLQT1, a novel K⁺ channel gene (Wang et al., 1996). Expression of KvLQT1 in either Xenopus laevis oocytes or mammalian cell lines induced a K⁺ current with biophysical properties unlike any known cardiac K⁺ current. Coexpression of KvLQT1 with minK induced a current that was essentially identical to cardiac I_Ks, indicating that KvLQT1 and minK proteins coassemble to form I_Ks channels (Barhanin et al., 1996; Sanguinetti et al., 1996b). Thus, dysfunction of either I_Kr or I_Ks can increase the risk of cardiac arrhythmia and sudden death.

An activator of I_Kr or I_Ks channels might be useful for the treatment of LQT that results from excessive pharmacological block of these channels or from mutations in the genes that encode the channel proteins. We previously described the properties of L-735,821, a benzodiazepine that is a potent and selective stereospecific blocker of cardiac I_Ks (Salata et al., 1996). In this study and one preliminary report (Salata et al., 1997), we describe another 1,4 benzodiazepine, R-L3 (L-364,373), that is a stereospecific activator of cardiac I_Ks.

	Experimental Procedures

Top Summary Introduction Procedures Results Discussion References

Isolation of guinea pig ventricular myocytes. Guinea pig ventricular myocytes were isolated as described previously (Salata et al., 1995). After isolation, the cells were stored in HBS containing 132 mM NaCl, 4 mM KCl, 1.8 mM CaCl₂, 1.2 mM MgCl₂, 10 mM HEPES, and 10 mM glucose, pH 7.2, at 24-26° until studied, usually within 8 hr after isolation.

Action potential studies. Transmembrane potentials were recorded using conventional microelectrodes filled with 3 M KCl (tip resistances, 30-50 M) using an Axoclamp 2B amplifier (Axon Instruments, Foster City, CA) as described previously (Fermini et al., 1995). Cells were superfused with HBS maintained at 37° at a rate of 2 ml/min APs were evoked with brief current pulses (1 msec, 1.2 times threshold) delivered at a frequency of 1 Hz through the recording electrode using an active bridge circuit. Only cells showing normal AP configurations and resting membrane potentials equal to or more negative than 85 mV were used in this study. Cells were studied after a 5-min control period and after 5 min of superfusion with each concentration of drug. After reaching a steady state effect under each condition, 20 individual APs were sampled and digitally averaged. The APA_50ms and APD₉₀ were measured from the digitally averaged records. Concentration-response relationships were determined by measuring APs or currents in each cell under control conditions and during superfusion with successively increasing concentrations of a given drug.

Voltage-clamp of guinea pig ventricular myocytes. Whole-cell voltage-clamp studies were performed using a List EPC-7 (Medical Systems, Greenvale, NY) or Axopatch 200A (Axon Instruments) amplifier as described previously (Fermini et al., 1995). Pipettes were made from square bore (1.0 mm outer diameter) borosilicate capillary tubing (Glass Company of America, Bargaintown, NJ). Pipettes were filled with 0.5 M K⁺ gluconate, 25 mM KCl, and 5 mM K₂ATP to minimize "rundown" (Giles and Shibata, 1985) and had resistances of 3-7 M (average, 5.5 ± 0.3 M). Series resistance was compensated 40-70%. Currents were low-pass filtered (3 dB at 0.2 kHz) before digitization at 1 kHz. I_Ks was measured during superfusion of the cells at a rate of 2-3 ml/min with normal HBS (35°) containing 0.4-1 µM nisoldipine to block L-type Ca²⁺ current and standard selective I_Kr blockers in 100-fold excess of the IC₅₀ (e.g., 3µM MK-499) to completely block I_Kr (Sanguinetti and Salata, 1996). Cells were voltage clamped at a V_h of 50 mV to inactivate I_Na. Time-dependent I_Ks amplitude was measured as the difference from the initial instantaneous current, after the settling of the capacity transient, to the final current level at the end of a depolarizing pulse. I_Ktail was measured as the difference from the holding current level to the peak tail current amplitude on return to V_h. I_Ktail was normalized to the maximum measured amplitude (I_Ktail-max) after 7.5-sec pulses. Averaged data were fit to a Boltzmann distribution of the form: I_Ktail/I_Ktail-max = 1/(1 + exp[(V_1/2  V_t)/k]) with a nonlinear least-squares fitting routine (Origin; Microcal Software, Northampton, MA) to estimate the half-point (V_1/2) and slope factor (k) for this relationship. The time courses of I_Ks activation and deactivation were fit with a double exponential relationship of the form: I_t = A₀ + A₁e^t/1 + A₂e^t/2 using a Chebeshev noniterative fitting technique (pCLAMP; Axon Instruments).

Culture and voltage clamp of AT-1 cells. The original mouse atrial tumor (AT-1) mouse colonies were established by Dr. Loren Field (Krannert Institute of Cardiology, Indianapolis, IN) (Field, 1988), and the lineage at Merck Research Laboratories was established from tumor cells that were provided by Dr. Dan Roden (Vanderbilt University, Nashville, TN). AT-1 cells were propagated in vivo, and their isolation and culturing were conducted as described previously (Delcarpio et al., 1991; Yang et al., 1994; Jurkiewicz et al., 1996).

cRNA injection and voltage-clamp of oocytes. The isolation and maintenance of X. laevis oocytes, in vitro transcription of KvLQT1 and hminK cRNA, and its injection into oocytes were performed as described previously (Sanguinetti et al., 1995, 1996b). Stage V and VI oocytes were injected with 11.5 ng of KvLQT1 cRNA (46 nl of a 250 ng/µl solution) alone or coinjected with 11.5 ng of KvLQT1 plus 1.25 or 0.1 ng of hminK cRNA. Currents were recorded 2-4 days later using standard two-microelectrode voltage-clamp techniques and a Dagan TEV-200 amplifier. Oocytes were bathed at room temperature (22-25°) in a solution containing 94 mM NaCl, 2 mM KCl, 2 mM MgCl₂, 0.1 mM CaCl₂, and 5 mM HEPES, pH 7.6.

Materials. R-L3 (Fig. 1) was prepared as described previously by Evans et al. (1987). Its enantiomer, S-L3, was prepared by the same procedure described for R-L3, with L-tryptophan acid chloride hydrochloride used in place of the D-isomer: ¹H NMR (CDCl₃) identical to that of R-L3. The chemical and chiral purity of R-L3 and S-L3 were determined to be >99%. S-L3: high performance liquid chromatography (Vydac C-18, 15 × 0.46 cm, 16 min gradient 95:5 to 5:95 0.1% H₃PO₄/H₂O:CH₃CN, 1.5 ml/min, 215 and 254 nM) retention time (rt) = 11.0 min, >96%, coelutes with R-L3. high performance liquid chromatography (Chiralcel absorbance 25 × 0.46 cm, 90/10 hexane/EtOH, 1.5 ml/min, 280 nM) rt = 9.95 min, 98.2%, contains <1% of R-L3 (R-L3: rt = 10.94 min, 99.6%, contains <0.5% of S-L3). TLC (silica, 10% Et₂O in CH₂Cl₂): single component, Rf = 0.43, coelutes with R-L3. Calc. for C₂₅H₂₀FN₃O: C 75.55, H 5.07, N 10.57; found: C 75.57, H 5.17, N 10.46. 

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Chemical structure of R-L3.

Statistics. Data are expressed as mean ± standard error. Concentration-dependent changes in AP parameters and individual ionic currents were assessed by repeated-measures analysis of variance. Post hoc comparison of the treatment with the control mean values were made with Dunnett's t test to determine significant changes between the control and test group mean values. Statistical comparisons for the time constants of I_Ks activation and deactivation were made using a paired t test. A one-tailed probability (p < 0.05) was considered significant.

	Results

Top Summary Introduction Procedures Results Discussion References

R-L3 decreases APD of cardiac myocytes. R-L3 (0.1-1.0 µM) caused a concentration-dependent shortening of APD. Fig. 2A shows a representative example of APs recorded at a stimulus frequency of 1 Hz. R-L3 significantly decreased APD₅₀ and APD₉₀ without significantly affecting other AP parameters (Table 1). Shortening of APD was maximal at 1 µM; APD₉₀ was decreased at concentrations of 1 and 10 µM R-L3 by 14.2 ± 1.6% and 13.8 ± 4.0%, respectively.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Effects of R-L3 on action potentials of guinea pig isolated ventricular myocytes. A-C, APs were recorded during stimulation at 1 Hz during control (A) and after 10 min superfusion with R-L3 at 0.1, 1, and 10 µM in normal HBS; after 30 nM Iso alone and after the addition of 1 µM R-L3 (B); and after 100 nM timolol alone and after the addition of 1 µM R-L3 (C). Bar graphs, percentage changes in APD90 and APA50ms. Data are mean ± standard error.

R-L3 increases I_Ks of guinea pig myocytes in a concentration-dependent and stereospecific manner. R-L3 is structurally related to L-735,821, a benzodiazepine that selectively blocks I_Ks and prolongs APD of guinea pig ventricular myocytes (Salata et al., 1996). Therefore, we reasoned that R-L3 might shorten APD of guinea pig myocytes by activating I_Ks.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Modulation of IKs by R-L3 is concentration dependent in guinea pig isolated ventricular myocytes. A, Superimposed currents from a single cell before and after the addition of 0.03 and 0.3 µM R-L3 during a 3-sec voltage step from 50 to 10 mV. B, Percent increase in IKs at a Vt of 10 mV by R-L3 (six or more cells). C, Superimposed currents before and after the addition of 10 µM R-L3 during a 7.5-sec voltage step from 50 to +50 mV. D, I-V relationship for the time-dependent IKs measured at the end of 7.5-sec pulses (six cells).

R-L3 shifts the voltage dependence of activation and slows deactivation of I_Ks. The voltage dependence of I_Ks activation was estimated using 7.5-sec depolarizing steps from a V_h of 50 mV (Fig. 4A). The amplitude of the tail currents was normalized relative to the maximum amplitude and fit to a Boltzmann function (Fig. 4B). Because I_Ks does not achieve a steady state, even during extremely long pulses (Hice et al., 1994), the activation curves are isochronal. In control, the V_1/2 was 19.2 ± 1.6 mV, and k for this relationship was 11.0 ± 1.2 mV (five cells). In these same cells, R-L3 shifted V_1/2 to 3.0 ± 0.8 mV at 0.1 µM and 4.9 ± 3.4 mV at 1 µM but had no effect on k. The maximally activated I_Ks measured at a V_t of +60 mV (896 ± 196 versus 953 ± 183 pA, 1 µM) was slightly but not significantly increased by R-L3. Likewise, after pretreatment with 100 nM timolol, 1 µM R-L3 shifted the V_1/2 by 19 mV without affecting k, indicating that its effect was independent of -adrenergic stimulation. These results suggest that the primary mechanism of the increase in I_Ks by R-L3 is an effect on channel gating.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   R-L3 shifts the voltage-dependence of IKs activation in guinea pig isolated ventricular myocytes. A, Currents recorded at the indicated Vt before (control) and after the addition of 1 µM R-L3. B, Isochronal activation curves were determined from the normalized amplitudes of tail currents after 7.5-sec pulses. Data were fitted to a Boltzmann function to determine the V1/2 and slope factor (k) for the relationship. The V1/2 was 19.2 ± 1.6 mV in control and 3.0 ± 0.8 mV and 4.9 ± 3.4 mV at 0.1 and 1.0 µM R-L3, respectively (five cells). The k value was 11.0 ± 1.2 mV in control and was not significantly changed by R-L3.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   R-L3 slows the rate of IKs deactivation in guinea pig isolated ventricular myocytes. Time constants for deactivation were determined for tail currents on return to a variable potential after a 3-sec activating pulse from 50 to +30 mV (four or more cells). *, Significantly different from control (p < 0.05) by paired t test.

R-L3 activates I_Ks independent of -adrenergic receptor activation. I_Ks was activated by 0.5-sec pulses to a V_t ranging from 40 to +50 mV. Iso (10 nM) alone increased I_Ks by 1.75-fold. The addition of 1 µM R-L3 produced a further increase in I_Ks, slowed the rate of deactivation, and shifted the threshold for current activation to more negative potentials (Fig. 6). The effects of R-L3 persisted after washout of the Iso. Similar effects were observed when exposure of cells to R-L3 preceded the addition of Iso. Thus, similar to the decrease in APD caused by R-L3, the increase in I_Ks by R-L3 was additive to that caused by -adrenergic receptor stimulation.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Effects of Iso and R-L3, alone and in combination, on the I-V relationship of IKs in guinea pig isolated ventricular myocytes. A, Traces were recorded during control, after exposure to 10 nM Iso alone, and after the addition of 1 µM R-L3 and then after washout of Iso but in the continued presence of R-L3. Currents were elicited by 0.5-sec pulses from a Vh of 50 mV. B, I-V relationship for time-dependent IKs for each condition (five cells).

R-L3 activates cloned human KvLQT1 channels expressed in X. laevis oocytes. At a concentration of 1 µM, R-L3 increased KvLQT1 elicited with 2-sec pulses to potentials ranging from 70 to +60 mV (Fig. 7, A and B). This increase can partially be accounted for by a 10-mV shift in the voltage dependence of activation caused by the drug (Fig. 7C). R-L3 also slowed the kinetics of KvLQT1, an effect that is easily observed when the time-dependent currents recorded before and after exposure to 1 µM R-L3 are superimposed and scaled to match peak current (Fig. 8A). Activation of KvLQT1 current is best described by a two-exponential function. The effect of R-L3 on these two components varied with V_t. R-L3 slowed the rate of the fast component at V_t 0 mV (Fig. 8B), but increased the rate of the slow component of activation at V_t 30 mV(Fig. 8C). The net effect of R-L3 was to slow the rate of KvLQT1 activation because of a reduction in the relative amplitude of the fast component of activation over the entire voltage range that was examined (Fig. 8D). R-L3 also slowed the rate of KvLQT1 deactivation when assessed at voltages negative to 40 mV (Fig. 8E). Thus, R-L3 increased the magnitude of KvLQT1, shifted the voltage dependence of its activation, and slowed the rates of activation and deactivation. These effects of R-L3 on KvLQT1 current are similar to those observed for I_Ks recorded in guinea pig ventricular myocytes.

View larger version (13K): [in this window] [in a new window]   Fig. 7.   R-L3 activates cloned KvLQT1 channels expressed in X. laevis oocytes. A, Currents were measured in response to 2-sec pulses from a Vh of 80 mV to a Vt of 60 mV to +40 mV, applied in 20-mV increments. Tail currents were measured at 70 mV. B, I-V relationships for peak KvLQT1 current during 2-sec pulses to the indicated test potential before and after 1 µM R-L3. C, Voltage dependence of KvLQT1 activation. Tail current amplitudes were determined from extrapolating a single exponential fit of deactivating currents to the onset of membrane repolarization. Isochronal activation curves were determined by fitting normalized tail current amplitudes to a Boltzmann function. In control, the V1/2 was 28 mV and the slope factor (k) was 11 mV for this relation. In the presence of 1 µM R-L3, the V1/2 was -40 mV and k was 13 mV (eight cells).

View larger version (21K): [in this window] [in a new window]   Fig. 8.   R-L3 slows the rates of activation and deactivation of KvLQT1 expressed in Xenopus oocytes. A, To illustrate the change in KvLQT1 kinetics induced by R-L3, the peak current activated by a 2-sec pulse to +40 mV was scaled to match the peak current recorded after the addition of 1 µM R-L3. Tail current was measured at 70 mV. B, Time constants for fast component of KvLQT1 activation. C, Time constants for slow component of KvLQT1 activation. D, Relative amplitude of the fast component of KvLQT1 activation. E, Time constants of KvLQT1 deactivation (eight cells for all graphs).

R-L3 activates cloned human I_Ks currents depending on hminK/KvLQT1 ratio. I_Ks currents are formed by coassembly of KvLQT1 and minK subunits (Barhanin et al., 1996; Sanguinetti et al., 1996b). Therefore, in addition to its effects on KvLQT1 channels, we determined the effects of R-L3 on KvLQT1 plus hminK (I_Ks) currents expressed in X. laevis oocytes. When oocytes were injected with 11.5 ng of KvLQT1 and 1 ng of hminK cRNAs, amounts similar to previous studies, we expected to observe an increase in the magnitude of cloned I_Ks similar to that described above for I_Ks recorded from guinea pig myocytes. Surprisingly, we found that R-L3 had no obvious effect on channels formed by coassembly of KvLQT1 plus hminK (Fig. 9). The only statistically significant effect of the drug was a slowing of the rate of deactivation. At 50 mV, deactivation was 470 ± 18 msec in control and 507 ± 12 msec after 1 µM R-L3 (p < 0.05). We observed a similar lack of effect of R-L3 on expressed I_Ks when either human or guinea pig minK cRNA alone was injected into X. laevis oocytes (data not shown), Presumably, these currents represent channels formed from coassembly of exogenous minK and endogenous KvLQT1. These data indicate that association of KvLQT1 with minK subunits prevents the activation of channel activity by R-L3 that occurs when only KvLQT1 channels are overexpressed in oocytes.

View larger version (15K): [in this window] [in a new window]   Fig. 9.   R-L3 does not activate cloned IKs currents expressed in X. laevis oocytes. A, Currents recorded in an oocyte during 7.5-sec pulses applied to Vt between 20 mV and +40 mV. Oocyte was injected with 11.5 ng of KvLQT1 and 1 ng of hminK cRNAs, and currents were recorded in control and 7 min after the addition of 1 µM R-L3. B, I-V relationships for peak currents (eight cells).

View larger version (18K): [in this window] [in a new window]   Fig. 10.   R-L3 activates current induced by injection of KvLQT1 and hminK cRNAs when amount of minK cRNA is limiting. A, Currents recorded in an oocyte during 7.5-sec pulses applied in 20-mV increments to a Vt between 40 and +40 mV. Oocyte was injected with 11.5 ng of KvLQT1 and 0.1 ng of hminK cRNAs, and currents were recorded in control and 7 min after the addition of 1 µM R-L3. B, I-V relationships for peak currents (seven cells).

R-L3 blocks I_Kr and I_Ca but not I_K1. To determine the selectivity of R-L3, we measured its effects on three other currents, I_Kr, I_K1, and I_Ca, that modulate cardiac APD.

View larger version (19K): [in this window] [in a new window]   Fig. 11.   R-L3 blocks ICa in guinea pig isolated ventricular myocytes at 10 µM. ICa was measured at room temperature during 100-msec pulses from a Vh of 50 mV to Vt between 40 and +60 mV during control and after the addition of 1 and 10 µM R-L3.

View larger version (25K): [in this window] [in a new window]   Fig. 12.   R-L3 blocks IKr in AT-1 cells. Currents were measured at room temperature during 1-sec pulses from a Vh of 40 mV to Vt between 30 and +50 mV during control and after the addition of 1 and 10 µM R-L3.

	Discussion

Top Summary Introduction Procedures Results Discussion References

At concentrations of 1 µM, R-L3 shortened APD of cardiac myocytes by selective activation of I_Ks. At membrane potentials and pulse durations typical for a cardiac AP, the most important mechanisms of action of R-L3 were a negative shift in the voltage dependence of activation and a slowing of I_Ks deactivation. R-L3 also caused a modest increase in I_Ks beyond what could be explained by these two mechanisms. The molecular mechanism of these effects on I_Ks current-gating is not known, but it is not mediated through the -adrenergic receptor activation pathway.

At concentrations of >1 µM, the effect of R-L3 on APD would reflect multiple mechanisms, including activation of I_Ks (especially at potentials of <0 mV), block of I_Kr, and block of L-type I_Ca. Block of I_Kr would lengthen APD, whereas block of I_Ca would contribute to a shortening of APD. Because APD₉₀ was unchanged, whereas APD₅₀ was further shortened and the plateau height (APA_50ms) was reduced when R-L3 was increased from 1 to 10 µM, it is likely that the block of I_Ca was more important than the block of I_Kr at 10 µM in guinea pig ventricular myocytes. A high concentration (10 µM) of R-L3 also caused a block of I_Ks after long pulses to very positive potentials. However, even at 10 µM, R-L3 would be expected to cause only an increase in I_Ks during the limited time of depolarization of a cardiac AP. We could not test the effects of R-L3 at concentrations of >10 µM because of its limited aqueous solubility.

We compared and contrasted the effects of R-L3 and Iso on configuration. These studies revealed a similar shortening of APD₉₀ but other important differences, especially on plateau height. Effects of -adrenergic receptor stimulation on APs are difficult to interpret because of the multitude of effects on ion channels and pumps. Nevertheless, the persistent effects of R-L3 in the presence of timolol, a -adrenergic receptor blocker, demonstrate that its effects are not mediated via this signaling pathway.

The findings that R-L3 activated I_Ks at low concentrations but blocked I_Ks at high concentrations, as did the S-enantiomer, suggest that there may be multiple binding sites for L-3 on the I_Ks channel. Multiple binding sites have been proposed to explain the dual action of dihydropyridines such as Bay K 8644 on the L-type Ca²⁺ channel (Brown et al., 1986; Kokubun et al., 1986).

R-L3 increased the magnitude of cloned KvLQT1 current, but its activation of human I_Ks channels formed by coassembly of KvLQT1 and hminK subunits depended on their ratio. When minK and KvLQT1 were coexpressed at relatively high ratio, such that the expression of minK was not limiting, R-L3 had no significant effect on I_Ks. However, when coexpressed at a 10-fold lower ratio, where minK was limiting, R-L3 increased I_Ks. The resulting currents at the low minK/KvLQT1 ratio were larger and activated at a rate slower than that of the current induced by KvLQT1 alone but were faster than those induced by the high subunit ratio. The stoichiometry of I_Ks channels is unknown. Because KvLQT1 alone or coexpressed with minK can form functional channels, it may be possible for minK subunits to coassemble with individual KvLQT1 subunits or KvLQT1 tetramers in variable ratios and thereby impart differing biophysical characteristics and pharmacological sensitivity. Alternatively, there may be only one functional I_Ks channel type with a fixed stoichiometry of KvLQT1 and minK subunits, and these may coexist in variable ratios with KvLQT1 homotetramers. A clearer understanding of this stoichiometry may help to explain why the effects of the drug on guinea pig I_Ks are better mimicked by cloned KvLQT1 channels, whereas the biophysical properties of I_Ks are more closely mimicked by channels formed by coassembly of KvLQT1 + hminK. Nevertheless, these findings indicate that the binding site for R-L3 is located on the KvLQT1 subunit and that coassembly with hminK diminishes or abolishes the agonist activity of R-L3. These studies are consistent with, but not proof of, the ideas that minK and R-L3 bind to a common region of KvLQT1 subunits and that association of minK with KvLQT1 precludes the drug-induced alteration of gating kinetics observed when only KvLQT1 channels are overexpressed in oocytes. Regardless of the exact mechanism, these findings suggest the possibility that I_Ks recorded in cardiac myocytes represents the sum of current mediated by KvLQT1 homotetrameric channels and heteromultimeric channels formed by coassembly of KvLQT1 plus minK subunits.

This and all studies of I_Ks face the same possible limitations, including the potential for K⁺ accumulation during long depolarizing pulses (Boyett et al., 1980, but see also Sanguinetti and Jurkiewicz, 1990), a lack of steady state activation of I_Ks, and interference or overlap with other cardiac currents. Our voltage-clamp protocols and conditions were designed to reduce or eliminate these concerns. To measure activation of I_Ks channels, relatively long pulse durations of 7.5 sec were used as in previous studies (Sanguinetti and Jurkiewicz, 1990) to approach steady state activation, whereas longer pulses were avoided to minimize the potential for K⁺ accumulation. I_Kr blockers were used in excess (100 × IC₅₀ values at 4 mM [K⁺]_o) to completely block I_Kr. Although an increase in [K⁺]_o has been reported to decrease the potency of I_Kr blockers (e.g., dofetilide; (Yang and Roden, 1996), the 100-fold excess used to block I_Kr in this study would not be overcome by potential elevations of [K⁺]_o. Most currents other than I_Ks were also eliminated by pharmacological blockade or inactivation (or both) with V_h.

A drug that activates I_Ks could be beneficial for the treatment of certain arrhythmias. The proarrhythmic potential of class III antiarrhythmic drugs that block I_Kr prompted the search for agents that would produce more modest prolongation of APD at high doses. The proarrhythmic liability associated with block of I_Kr currents was recently confirmed by the finding that mutations in HERG, the gene encoding I_Kr channel subunits, can cause inherited LQT (Curran et al., 1995). Recently, I_Ks blockers have been described that lengthen APD and have antiarrhythmic effects in animal models. Chromanol 293B (Busch et al., 1996) and the more potent L-735,821, a 1,4-benzodiazepine (Salata et al., 1996), inhibit I_Ks and cause a self-limiting prolongation of APD that is less than that produced by I_Kr blockers. However, mutations in KvLQT1 can also cause inherited LQT (Wang et al., 1996; Neyroud et al., 1997). This latter finding could dampen enthusiasm for the development of I_Ks blockers for prophylactic treatment of ventricular tachyarrhythmias. If I_Kr or I_Ks blockers are used as antiarrhythmic drugs, it would be useful to have a channel activator that could reverse the effects of overdose. Moreover, pharmacological activation of I_Ks might also be useful for treatment of inherited LQT and abnormally delayed repolarization associated with heart failure (Beuckelmann et al., 1993; Tomaselli et al., 1994).

In summary, we described the properties of R-L3, a novel benzodiazepine that selectively activates I_Ks in myocytes at low concentrations. This compound represents a new pharmacological probe for the study of I_Ks in cardiac myocytes. In addition, R-L3 can be used as a probe to define the physiological role of I_Ks and KvLQT1 currents in other tissues, such as the pancreas (Wang et al., 1996) and stria vascularis of the inner ear (Neyroud et al., 1997).