Differential Response to Na⁺ Channel Blockade, β-Adrenergic Stimulation, and Rapid Pacing in a Cellular Model Mimicking the SCN5A and HERG Defects Present in the Long-QT Syndrome

Myocyte Isolation

Guinea pig ventricular myocytes were obtained by enzymatic digestion via aortic retrograde perfusion of the coronary bed, by following a modified protocol previously published by Powell.¹⁶ Guinea pigs weighing 250 to 300 g were lightly anesthetized with sodium thiopental and then stunned (decerebration via cervical dislocation). After thoracotomy, the heart, including the ascending aorta, was quickly removed from the chest and placed in a cold (4°C) physiological (saline) solution. In order to reduce the oxygen consumption and lower the metabolic activity, the heart was kept in the cold solution during the subsequent surgical procedure, consisting of the removal of the connective tissue and the isolation of the ascending aorta. A silk suture was then used to bind the heart to the perfusion apparatus. A 5-minute retrograde aortic perfusion on a Langendorff apparatus, with an oxygenated (100% O₂) HEPES-buffered solution (containing the following concentrations [mmol/L]: NaCl 133.5, KCl 3.4, MgCl₂ 1.2, NaH₂PO₄ 1.2, HEPES 10, glucose 10, and CaCl₂ 1.2), at 37°C and pH 7.38, was performed to wash out the blood from the coronary bed. Only hearts in which the onset of this procedure restored spontaneous, rhythmic, and vigorous contractions were used for cell dissociation. Subsequently, on a separate apparatus the heart was perfused with Ca²⁺-free oxygenated HEPES-buffered solution (37°C, pH 7.35) for 10 minutes after cessation of the spontaneous contractions. In this step, as well as in the subsequent collagenase perfusion, a constant perfusion flow was maintained by means of a peristaltic pump at 6 to 7 mL/min. After 10 minutes of nominally 0 Ca²⁺ perfusion, collagenase (Worthington type II) was added at a concentration of 0.28 to 0.30 mg/mL for another 15 minutes.

Experimental Protocol

Only rod-shaped myocytes (with clear cross striations) showing no spontaneous contractions were considered for the study protocol. Transmembrane action potentials were recorded with glass microelectrodes (with an input resistance of 30 to 50 MΩ) filled with 3 mol/L KCl. Action potentials were recorded through an electrometer (Axoclamp A2, Axon Instruments), analog-to-digital conversion was performed using a Labmaster TL-1 125 (Axon Instruments), and data were subsequently stored on an optical disk (Laser Bank 1000, Micro Design International) at an acquisition frequency of 1.5 kHz.

After cell impalement, a 10-minute equilibration period was allowed in order to reach a stable action potential morphology, duration, and resting membrane potential. Only cells with a normal resting membrane potential (−90 mV), with no spontaneous activity or afterdepolarizations during pacing, were used for the subsequent study protocol. Cells were studied in an experimental bath (volume, 5 mL), with temperature and pH kept in a physiological range (37.5°C, pH 7.38), and continuously superfused (constant flow, 5 mL/min) with oxygenated HEPES-buffered solution containing 1.3 mmol/L Ca²⁺. Cells were paced at a cycle length of 1000 ms (stimulator model S8800, Grass Instruments); in 20 cells exposed to anthopleurin (n=10) and dofetilide (n=10) and in 10 cells in control conditions, pacing protocols were performed at 2.5 Hz for 40 consecutive beats in order to assess APD adaptation to changes in cycle length. In 16 cells previously studied with pacing at 1 and 2.5 Hz, pacing was performed at 0.5, 1, 1.5, 2, and 2.5 Hz in order to evaluate APD90 adaptation at intermediate rates (control, n=5; anthopleurin, n=5; and dofetilide, n=6).

Block of the inactivation of I_Na was obtained by exposure of cells to 10 nmol/L anthopleurin (Sigma Chemical Co) for 15 minutes, and the effect of I_Na blockade (n=5) was assessed by adding 10 μmol/L mexiletine to the solution containing anthopleurin. Washout was then performed by washing the solution with the active substances and replacing it with HEPES-buffered solution. K⁺ channel blockade was performed (n=5) by exposure to 10 μmol/L dofetilide, a selective I_Kr blocker, for 15 minutes; 10 μmol/L mexiletine was then added to the dofetilide solution, and combined exposure was maintained for an additional 15 minutes before washout. Thirteen cells pretreated with dofetilide were exposed to mexiletine at higher dosages (50 μmol/L [n=6] and 100 μmol/L [n=7]) in order to evaluate whether the response to Na⁺ channel blockade was shifted to the right compared with anthopleurin-treated cells.

Six cells treated with dofetilide and six cells treated with anthopleurin were exposed to 10 nmol/L isoproterenol. Combined exposure to dofetilide and isoproterenol or to anthopleurin and isoproterenol was maintained for 15 minutes until steady state APD values were obtained.

Throughout the study, cells were kept at a temperature of 37.5°C, and APD was measured at APD90 in at least five consecutive beats at 5-minute intervals by Axotape software (Axon Instruments).

Differences between the means in the groups of cells studied were analyzed by paired and unpaired t tests, and one-way ANOVA was used for multiple comparisons. Post hoc analysis was performed using Scheffé’s method. A value of P<.05 was accepted as statistically significant.

Effect of Anthopleurin and of Dofetilide

The effect of anthopleurin or dofetilide was studied in 46 cells; 35 of them were studied also with either mexiletine or isoproterenol (Table 1⇓), and 11 cells were used for the pacing protocol from 0.5 to 2.5 Hz. In the 46 cells, the mean duration of APD90 in control conditions at a basic drive of 1 Hz was 213±33 ms. Sixteen cells were exposed to anthopleurin (10 nmol/L), and 30 cells were exposed to dofetilide (10 μmol/L). Both substances prolonged APD. The mean steady state prolongation was 54±13 ms in cells treated with anthopleurin (n=16, from 220±29 to 274±25 ms; ANOVA, P<.001) and 62±16 ms in cells treated with dofetilide (n=30, from 209±35 to 272±36 ms; ANOVA, P<.001).

Effect of Mexiletine

The effect of the Na⁺ channel blocker mexiletine (10 μmol/L)¹⁷ was studied in 23 cells. Mexiletine was added to the perfusion solution, and it reduced the APD prolongation induced by anthopleurin (n=5, from 264±38 to 226±32 ms; ANOVA, P<.001), whereas it did not affect the prolongation induced by dofetilide (n=5, from 314±61 to 314±64 ms; P=NS) (Figs 1⇓ and 2⇓). The effect of mexiletine in the two groups of cells was significantly different, as the mean APD difference after mexiletine in the anthopleurin group was −37±8 versus +1±12 ms in the dofetilide group (P<.001). Thirteen additional cells pretreated with dofetilide were also exposed to higher dosages of mexiletine. In the six cells exposed to 50 μmol/L mexiletine and in the seven cells exposed to 100 μmol/L mexiletine, APD90 remained unchanged (Table 1⇑).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Change in APD90 during cell superfusion with anthopleurin (○, n=5) and dofetilide (•, n=5) and during combined exposure to both drugs plus mexiletine (mex) (n=5 for both groups). There was a significant difference (P<.01) between the two groups during combined exposure with mexiletine.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Left, Example of APD prolongation induced in a control cell (A) by anthopleurin (B) and subsequent reduction of APD after combined exposure to anthopleurin and mexiletine (C). Right, Example of APD prolongation induced in a control cell (A) by dofetilide (B′) and subsequent reduction of APD after combined exposure to dofetilide and mexiletine (C′).

Effect of Isoproterenol

The effect of isoproterenol was studied in 15 cells. Three cells received isoproterenol (10 nmol/L) in control conditions, and in none of them was APD modified (from 208±34 to 205±47 ms, P=NS),¹⁸ nor did EADs develop.¹⁹ In the cells (n=6) that were exposed first to anthopleurin and subsequently to anthopleurin plus isoproterenol, a prompt shortening of APD was obtained after 1 minute (from 277±19 to 240±22 ms; ANOVA, P<.001). By contrast, in the cells (n=6) exposed to dofetilide, the addition of isoproterenol initially further prolonged APD from 280±25 to 313±20 ms (ANOVA, P<.001) but after 15 minutes reduced it to control values (229±22 ms; ANOVA, P<.001) (Figs 3⇓ and 4⇓). During the initial prolongation induced by isoproterenol added to dofetilide, EADs always developed; in contrast, EADs did not develop in any of the cells exposed to anthopleurin and isoproterenol.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Change in APD90 during cell superfusion with anthopleurin (•, n=6) and dofetilide (○, n=6). Isoproterenol (Iso, 10 nmol/L) was then added to the superfusion solution. In the early phase (3±1.8 minutes), cells previously exposed to dofetilide showed a further prolongation of APD90 (P<.01 between the two treatment groups at early phase), and cells developed EADs; at steady state, Iso shortened APD, and EADs disappeared. In contrast, the cells exposed to anthopleurin (n=6) showed a prompt shortening of APD90, and EADs were never observed (P=NS at steady state). *P<.01.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Left, Exposure of a control cell (A) to anthopleurin (B) results in APD prolongation; subsequent exposure to isoproterenol (C) is accompanied by shortening of APD. Right, Exposure of a control cell (A) to dofetilide (B′) induces APD prolongation; addition of isoproterenol to the perfusate further prolongs APD (C′) in the first minutes (early phase) of exposure, whereas at steady state (late phase), APD is markedly reduced (D).

Effect of Pacing

The effect of pacing was studied in 30 cells. Pacing at 1 Hz and at 2.5 Hz was performed in all of them. Sixteen cells were subsequently studied with incremental pacing from 0.5 to 2.5 Hz by steps of 0.5 Hz, whereas the remaining 14 cells after the pacing protocol were entered in either the mexiletine or the isoproterenol protocols (Table 2⇓, Fig 5⇓).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Individual and mean±SD values of the response to fast pacing (from 1 to 2.5 Hz) in control cells (n=10) and in anthopleurin-treated (n=10) and dofetilide-treated (n=10) cells. Data are expressed as APD90 shortening (in milliseconds). P<.001 (ANOVA); *P<.05 vs control cells (Scheffé’s post hoc test); and **P<.05 vs dofetilide-treated and control cells.

The 30 cells (10 in control conditions, 10 exposed to anthopleurin, and 10 exposed to dofetilide) were studied with pacing protocols to evaluate APD shortening at faster rates. At a 1-Hz driving frequency, the mean steady state APD was 202±15 ms in control cells, and it was longer (P<.005) in cells exposed to anthopleurin (277±19 ms) and in cells exposed to dofetilide (278±31 ms). When the pacing rate was increased to 2.5 Hz, mean APD decreased by 10% to 182±11 ms in control cells, by 13% to 240±26 ms in the dofetilide group, and by 22% to 217±18 ms in the anthopleurin group (ANOVA, P<.001) (Fig 5⇑).

Pacing from 0.5 to 2.5 Hz by steps of 0.5 Hz was performed in 16 cells: five control cells, six cells treated with dofetilide, and five cells treated with anthopleurin. APD90 shortening was higher in dofetilide- and anthopleurin-treated cells than in control cells, and anthopleurin-treated cells had a more pronounced adaptation than did dofetilide-treated cells. Data are shown in Fig 6⇓.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Shortening of APD90 in control cells (•, n=5) and in dofetilide-treated (▴, n=6) and anthopleurin-treated (▪, n=5) cells during pacing at 0.5, 1, 1.5, 2, and 2.5 Hz. *P<.05 vs control cells (Scheffé’s post hoc test); **P<.05 vs control and dofetilide-treated cells (Scheffé’s post hoc test).

Data for all groups are presented in Tables 1⇑ and 2⇑.

Genetic Background

LQTS is the first recognized inherited myocardial ion channel disease.²¹ Recently, three defective ion channel genes have been identified in LQTS patients: the gene for the cardiac Na⁺ channel (SCN5A),^{9 10} the gene for the HERG K⁺ channel,¹¹ and the gene for a yet undefined K⁺ channel.⁸ Prolongation of the QT interval is the common phenotypic manifestation.

Because the LQTS deletion appears to be located on the III-IV linker of the Na⁺ channel putative topology, it was initially predicted that the mutation would affect the inactivation of I_Na.⁹ However, Bennett et al²¹ have expressed the first identified mutation (ΔKPQ) and have found that mutant channels fluctuate between normal and noninactivating gating modes. These channels burst for prolonged periods of time, producing persistent inward current, which explains the prolongation of cardiac action potentials. Moreover, Dumaine et al²² have recently expressed not only the ΔKPQ mutation but also two more recently described¹⁰ single residue substitutions (R1644H and N1325S) and found that the late I_Na produced by all three mutations is blocked by mexiletine.

The mutations identified in LQT2 affect HERG (human Ether a-go-go–related gene), a newly defined human K⁺ channel described in 1994.²³ Sanguinetti et al¹² expressed HERG in Xenopus oocytes and recorded a delayed rectifier K⁺ current that was interpreted as encoding the human protein for the I_Kr channel. They also showed that the different mutations observed in HERG lead to a reduced function of the channel by diminishing K⁺ current.²⁴

The Present Study

On the basis of evidence that the mutant SCN5A Na⁺ channels produce an increased amount of slowly inactivating inward I_Na,^{21 22} anthopleurin was selected as the pharmacological intervention that might mimic LQT3.

Since LQT2 consists of an alteration of the I_Kr channel encoded by the HERG gene,¹² it is expected that this would reduce the expression of functional channels. Accordingly, in an attempt to mimic the condition of LQT2, we used dofetilide, a selective blocker of I_Kr.²⁵

In these two models, we tested the response of APD to (1) exposure to the Na⁺ channel blocker mexiletine, (2) exposure to isoproterenol, and (3) its adaptation to increases in stimulation frequency.

Exposure to a Na⁺ Channel Blocking Agent

Both anthopleurin and dofetilide induced a prolongation of APD, confirming that either a reduced I_Kr or a persistent inward I_Na can cause the classic phenotypic alteration of LQTS, ie, prolongation of the QT interval. Exposure to the Na⁺ channel blocker mexiletine significantly reduced APD in cells treated with anthopleurin, whereas it did not modify the prolongation induced by dofetilide. Furthermore, we exposed dofetilide-treated cells to higher dosages of mexiletine in order to evaluate whether the dose-response curve to the Na⁺ channel blockers was shifted to the right compared with the response in anthopleurin-treated cells. The lack of shortening induced by mexiletine at the higher dosages supports a lack of effect on APD90 prolongation induced by K⁺ channel blockade. The latter result may appear in contrast with the reports of shortening of APD induced by mexiletine^{26 27} and lidocaine²⁸ in Purkinje fibers pretreated with the K⁺ channel blocker sotalol. However, the use of Purkinje fibers whose plateau currents differ from those of myocytes^{29 30} and of sotalol (which, at variance with dofetilide,^{31 32} also exerts an inhibitory effect on I_Na³³ ) makes comparisons hazardous.

Our observations suggested that a Na⁺ channel blocker might reduce APD, and therefore the QT interval, in LQT3 patients. In striking confirmation of this hypothesis, when we administered mexiletine to 13 LQTS patients,²⁰ we found that the QT interval shortened significantly, by an average value of 90 ms, in the LQT3 patients (n=6), whereas it remained almost unchanged in the LQT2 patients (n=7).

Effect of β-Adrenergic Stimulation

β-Adrenergic stimulation performed in control cells did not modify APD, as previously described.³⁴ However, when isoproterenol was added to anthopleurin, a prompt shortening of APD toward control values was observed. Also, the cells treated with dofetilide shortened their APD to control values when exposed to isoproterenol; however, during the first 3 minutes of exposure to isoproterenol, APD was further prolonged and EADs developed. Only after 15 minutes of exposure did APD shorten to control values.

This differential effect of β-adrenergic stimulation according to pretreatment with anthopleurin or dofetilide may suggest that vulnerability to catecholamines may also differ in LQTS patients according to the gene involved. In agreement with this, we found that the majority of cardiac arrests occur during stress among LQT2 patients and during sleep or at rest among LQT3 patients.²⁰

Adaptation of APD to Changes in Rate

As expected,^{35 36} cells exposed to dofetilide shortened APD during rapid pacing more than the control cells did. However, anthopleurin-treated cells had an even greater shortening of APD to rapid pacing than did the control and dofetilide-treated cells. A pacing protocol at frequencies ranging from 0.5 to 2.5 Hz was performed in 16 cells and confirmed the more pronounced adaptation to the fast rate of anthopleurin-treated cells. Consistent with these findings are data showing that tetrodotoxin prolongs the restitution of cardiac APD, thus suggesting that blockade of Na⁺ channels alters adaptation.³⁷ This suggests that increased shortening in response to fast rate is probably not a specific consequence of anthopleurin but rather an effect of increased I_Na. Once more, our observations in myocytes fit with the finding that chromosome 3–linked patients shortened their QT intervals more during physiologically induced increases in heart rate than do chromosome 7–linked patients and healthy control subjects.²⁰

Limitations of the Study

The model used in the present study represents the first attempt to replicate in vitro the electrophysiological consequences of the genetic abnormalities identified in patients affected by two variants (LQT2 and LQT3) of LQTS. As with many experimental preparations, data should be interpreted with caution, especially when attempting clinical extrapolation. Four potential limitations apply to our preparation: (1) The model is based on guinea pig ventricular cells, which have several important differences compared with myocytes of other species,³⁸ including humans.³⁹ The transient outward current is absent in guinea pigs, and the slow component of the delayed rectifier is more prominent than in most species.³⁸ (2) Differences have been described in myocytes originating from epicardium, endocardium, and midmyocardium,⁴⁰ which may influence the results; we have not attempted any characterization of the site of origin of ventricular cells used in the present study. (3) The use of dofetilide to block I_Kr is made on the assumption²⁴ that genetic defects in HERG result in a reduction of the native current. (4) Experiments performed with mexiletine are based on the assumption that the binding affinity to Na⁺ channels is not modified by anthopleurin and dofetilide. We have addressed this issue by performing dose-response curves showing that mexiletine, even at higher dosages (50 and 100 μmol/L), does not affect APD prolongation induced by dofetilide; however, since pharmacological data on the effects of anthopleurin and dofetilide on mexiletine binding are not available, we cannot rule this out as a potential confounding factor in our model.

Conclusions

Pharmacological tools in isolated myocytes have allowed an attempt to develop a cellular model designed to mimic, within the constraints of experimental electrophysiology, the two forms of hereditary LQTS dependent on mutations on the SCN5A (LQT3) and HERG (LQT2) genes. The two preparations have shown differential responses to the interventions tested. When, on this experimental basis, similar interventions have been assessed in LQT3 and LQT2 patients, we have observed responses very similar to those present in the cells pretreated with anthopleurin and with dofetilide, respectively. These findings are encouraging for the exciting possibility of using cellular models, which have originated from the progress in molecular genetics, to guide novel therapeutic approaches for a life-threatening disease such as LQTS.