Selective Hcn1 Channels Inhibition by Ivabradine in Mouse Rod Photoreceptors

Abstract

purpose. To evaluate in mammalian rod photoreceptors the selectivity for hyperpolarization-activated cyclic nucleotide-gated (Hcn1, coded by Hcn1) over potassium-selective (Kir 2.4, coded by Kcnj14) channels of ivabradine, a selective inhibitor of the cardiac “funny” current (I _f).

methods. Rods were isolated from the mouse retina and voltage clamped by the perforated-patch technique. The hyperpolarization-activated current (I _h) was blocked by ivabradine during repetitive stimulation with activating/deactivating voltage steps from −80 to −30 mV, from a holding of −35 mV.

results. Full inhibition was observed at a high concentration of ivabradine (30 μM), with intermediate effects at 3 and 0.3 μM. Steady state activation and activation kinetics of the ivabradine- and CsCl-blocked currents were similar, consistent with the block by ivabradine of ion permeation through Hcn1 channels. Hcn1 blockade was also consistent with the lack of current reactivation during long steps at −110 mV. At doses that fully block I _h, ivabradine does not affect the inward rectifier current through potassium-selective Kir 2.4 channels or the outward currents evoked by stepping up from −80 to 50 mV.

conclusions. In mammalian rods, ivabradine is a selective inhibitor of Hcn1 channels. Phosphenes perception in response to abrupt changes in luminance, which has been transiently reported in a dose-dependent way by few patients treated with ivabradine, was consistent with Hcn1 inhibition in rods.

The cardiac pacemaker current (I _f) in the sinus node plays a key role in controlling heart rate, ¹ providing the rationale for the development of selective I _f inhibitors for the treatment of stable angina (for a review see Ref. ² ). Organic compounds such as zatebradine (ULFS 49), ivabradine (S16257), and cilobradine (DK-AH 269) inhibit cardiac I _f by a use-dependent mechanism that requires the drug to access the open pore from the intracellular side. ^{3 4 5 6} . 

The most frequent side effects reported during clinical trials by healthy volunteers and patients treated with I _f inhibitors are luminous phenomena (primarily phosphenes). ^{7 8} Molecular ⁹ and electrophysiological evidence indicates the expression of different hyperpolarization-activated cyclic nucleotide (HCN)-gated channel isoforms by pacing cells ^{10 11} and retinal neurons. ^{12 13 14} However, all HCN isoforms have high structural homology in the pore region, which includes the putative drug-binding pocket, ¹⁵ suggesting that visual symptoms may result from the suppression by I _f inhibitors of the hyperpolarization-activated currents (I _h) in retinal neurons. An additional action of these inhibitors on potassium currents may also play a role in the generation of visual phenomena, as suggested by the observation that zatebradine, an I _f inhibitor structurally related to ivabradine, blocks potassium currents and I _h in amphibian rods. ¹⁶  

The aim of the present work was to investigate in mammalian rods the inhibition of I _h by ivabradine and its selectivity over potassium currents. To this end we compared the effects of ivabradine and CsCl on inward and outward currents of mouse rod photoreceptors. CsCl is an inorganic blocker of inward rectifier currents that has already been used to investigate properties and functional roles of voltage-dependent currents of rod photoreceptors. ^{17 18 19}  

Our results indicate that in mammalian rods, ivabradine is a selective inhibitor of Hcn1 channels that does not affect the inward rectifier current through potassium-selective Kir 2.4 channels coded by Kcnj14 ²⁰ or the outward currents. Phosphenes perception in response to abrupt changes in luminance, which has been transiently reported in a dose-dependent way by a few patients treated with ivabradine, is consistent with the inhibition by I _f inhibitors of Hcn1 channels in rods. 

Materials and Methods

Animals and Cell Dissociation

Adult male C57BL/6 mice, 30 to 60 days old, were reared on a 12-hour dark/12-hour light cycle with ad libitum access to water and food. Rearing and handling complied with institutional guidelines and with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. From a mouse deeply anesthetized by intraperitoneal injection of pentothal sodium (35 mg/kg; Gellini, Aprilia, Italy), the retina was quickly isolated through a corneal slit by gentle squeezing of the eye with curved forceps and then was placed in cold Locke solution, pH 7.5. ²¹ The animal was then killed by an intraperitoneal lethal dose of pentothal. Mouse rod photoreceptors dissociated by gentle mechanical trituration after enzymatic treatment (5′ at 30° in 14–18 U/mL papain, 0.3 mg/mL hyaluronidase, and 100 U/mL DNase) were identified by their characteristic morphology (Figs. 1A 1B)and by rhodopsin immunocytochemistry (Fig. 1C) . 

Chemicals

CsCl, proteolytic enzymes, amphotericin, and salts were from Sigma-Aldrich Italia. Ivabradine (S-16257), chemical name (S)-3-(3-(((3,4-dimethoxybicyclo(4.2.0)octa-1,3,5-trien-7-yl)methyl)methylamino)propyl)-1,3,4,5-tetrahydro-7,8-dimethoxy-2H-3-benzazepin-2-one hydrochloride, was provided by Institut de Recherches Internationales Servier (France). The drug was dissolved in ultrapure-grade water (MilliQ; Millipore, Billerica, MA) as a 30-mM stock solution and stored up to 5 days at 4°C and diluted in Locke solution at the desired final concentration for cell superfusion. 

Patch-Clamp Recordings

Unless specified, perforated-patch recordings were carried out in Locke solution as previously reported for guinea pig rods, ²¹ and voltage-dependent currents were measured using an amplifier (Axopatch 1D; Axon Instruments, Sunnyvale, CA), low-pass filtered at 200 Hz by a four-pole Bessel filter, and digitized online at 1 kHz (Digidata 1320 board driven by pClamp 8.0 software; Axon Instruments). 

The temperature of the 1-mL recording chamber was set at 32°C by a 51-W Peltier device (RS Components, Milan, Italy). The superfusing Locke solution (flow rate, 3 mL/min), was heated by a 1.2-W Peltier device (RS components). 

Analysis of Records

The fractional inhibition of I _h was estimated by the ratio of current amplitudes at −80 in the presence and in the absence of the drug. To estimate I _h activation, net currents were computed by subtracting voltage-clamp records in the presence of the blocker from those in control saline (difference records). Conductance values were computed from net currents amplitude during the last 100 ms of the 2-second step using the I _h reversal potential of −30 mV measured in preliminary experiments, as previously reported for guinea pig rods in similar recording conditions. ²¹ Conductance values for each step were normalized to those at −120 mV to generate fractional activation values (F(V)). F(V) values were interpolated by  

\[F(V)\ {=}\ 1/{\{}1\ {+}\ \mathrm{exp}{[}(V\ {-}\ V_{1/2})/s{]}{\}}\]

where V _1/2 is the half-activation voltage and s is the inverse slope factor of I _h. 

The time constants of activation were estimated by fitting net currents by  

\[I(t)\ {=}\ I_{\mathrm{MAX}}\ {+}\ A_{0}\mathrm{exp}{[}{-}(t\ {-}\ t_{0})/{\tau}{]}\]

where t ₀ is the time of application of the activating step, I _MAX is the maximum current amplitude, A ₀ is the initial current amplitude, and τ is the time constant of activation. 

Statistical Analysis

After assessing that current amplitudes were normally distributed (data not shown), parametric statistical analysis was applied to compare current amplitudes, before and after drug administration, obtained by averaging 10 individual sweeps. The effects of ivabradine concentration on I _h inhibition were analyzed by one-way ANOVA using the function implemented in data analysis software (Origin 6.0; Microcal, GE Healthcare, Chalfont, St. Giles, UK). 

Comparison between control and ivabradine concentrations was carried out by multiple t-tests, adopting the correction by Bonferroni. Student’s t-test was used to compare current amplitudes at −120 after block by either ivabradine 30 μM or CsCl 3 mM (see Figure 5C ). 

Results

In Figure 2A , an inward current of −9.6 pA (CNTR) activates in response to a hyperpolarizing voltage step to −80 mV. The average value from 21 rods was −14.9 ± 2.2 pA (range, −4.7/−50.9 pA). Note that after block by CsCl, a known I _h inhibitor, the membrane current approaches the zero-current level, indicating that I _h is the main inward current at −80 mV. 

I _h inhibition by the use-dependent I _f inhibitor ivabradine ⁵ was assessed by repeated application of the activation/deactivation protocol, shown in Figure 2B . In preliminary experiments, a slower and less complete block was also obtained with the use of shorter and more frequent stimuli (not shown) and in the absence of repetitive stimulation (Fig. 2D) . Data in Figures 2C 2D 2E 2F 2G 2Hillustrate the time and dose dependence of I _h inhibition by ivabradine. Sweeps in Figures 2C 2E and 2Gwere acquired at selected times before and during application of ivabradine 30 (Fig. 2C) , 3 (Fig. 2E) , and 0.3 μM (Fig. 2G) . Note that in Figure 2C , at −30 mV, all sweeps superimpose, as expected for the selective block of a current reversing close to −30 mV. As shown in Figure 2D , full I _h inhibition by 30 μM ivabradine takes place in approximately 6 minutes, and the current approaches the horizontal dashed line that marks the zero-current level expected for a cell with fully blocked I _h (Fig. 2A) . Longer times are required for partial inhibition, as shown in (Figs. 2E 2F)and (Figs. 2G 2H)for 3 and 0.3 μM ivabradine, respectively. The block was occasionally observed to reverse on prolonged washout of the drug, as shown in Figure 2F , although the time course of current recovery often exceeded the limited recording time allowed by the small size of mammalian rods. 

The inhibition illustrated in Figures 2C 2D 2E 2F 2G 2Hwas observed during repetitive application of the stimulation protocol illustrated in Figure 2B . The inhibition was not strictly use dependent, however, because a current reduction was also observed when the repetitive stimulation was briefly stopped (see horizontal dotted lines in Fig. 2D ). Data in Figure 2Hplot the average I _h inhibition, with the SEM, measured in response to the application of three ivabradine concentrations and 3 mM CsCl. 

Use-independent block by ivabradine has been reported ²² for Hcn1 homomeric channels. To evaluate the hypothesis that ivabradine inhibits I _h carried through Hcn1 homomeric channels, we compared the voltage dependence and time constants of activation of ivabradine-blocked and of Cs-blocked currents; the data are illustrated in Figure 3 . Steady state I _h activation was investigated by the stimulation protocol shown in Figure 3A . Difference records (see Materials and Methods) in Figures 3B and 3Cplot currents blocked by ivabradine 30 μM and CsCl 3 mM, respectively (when the current inhibition was complete). Data plotted in Figure 3Bindicate that ivabradine inhibits the slowly activating inward currents induced in response to voltage steps from −60 to −120 mV. On the other hand, Cs is not very effective at blocking I _h activated by voltage steps at −60 and −70 mV, and difference records for these voltages in Figure 3Cdo not show clear evidence for slowly activating inward currents. Average conductance values (see Materials and Methods), with their SEM, are plotted in Figure 3Dand fitted with equation 1 . Ivabradine- and Cs-blocked currents have half-activation voltages close to −75 mV, but the inverse slope factor s is larger for ivabradine-blocked (7.6 mV) than for Cs-blocked (5.1 mV) currents. 

Dotted traces in Figures 3E and 3Fplot on an expanded time scale the ivabradine (Fig. 3E)and Cs-blocked (Fig. 3F)currents at −120 mV and are fitted by equation 2(continuous smooth lines). Differences between sweeps and fit curves are plotted in the upper part of the panels and stay close to the zero-current level (dashed horizontal lines), indicating that a single exponential decay provides a reasonably good fit to I _h activation. Average activation time constants are plotted in Figure 3Gfor ivabradine-blocked (filled circles) and Cs-blocked (open circles) currents, indicating that both drugs block currents with similar fast-activation kinetics. It has been reported ²² that after ivabradine blockade, long hyperpolarizing pulses fail to reactivate the current through Hcn1 homomeric channels, in contrast to the behavior of the f-channels of pacing cells. ²³ As a further test for the selective inhibition of rod Hcn1 channels by ivabradine, after attaining full I _h inhibition by 30 μM ivabradine (not shown), we applied an 18.75-second long step from −35 to −110 mV. As shown in Figure 3I , no current reactivation was observed in response to the long step at −110 mV. 

We next addressed the selectivity toward I_h by testing ivabradine effects on other voltage-dependent currents of rod photoreceptors. Sweeps in Figures 4A and 4Bplot currents activated by stepping from −80 mV to voltages ranging from −10 to +50 mV in 20-mV steps. Data in Figure 4Ashow the current recorded before (thin traces) and after (thick traces) full I_h inhibition by 30 μM ivabradine (note the difference in the holding currents at −80 mV between control and I_h-blocked sweeps). For comparison, the effect of 3 mM CsCl is shown in Figure 4B . Data points in Figures 4C and 4Dplot the average amplitudes of difference records from three cells, with each SEM, indicating that neither ivabradine nor CsCl affects currents generated in response to voltages more positive than −50 mV. 

An additional outward current of rods is carried by anions flowing through a voltage- and calcium-dependent conductance. ¹⁷ Activation of this conductance by membrane depolarization above −30 mV generates a slowly activating, noninactivating, outward currents, as shown in Figure 4E . Note that on stepping to +50, close to the calcium reversal potential, the chloride current is suppressed. 

Data in Figure 4F , recorded from a cell with a small outward chloride current, suggest that the drug does not block the current through voltage and calcium-dependent chloride channels. The expression of Kcnj14 in mouse rods has recently been reported. ²⁰ Considering that Kcnj14 codes for Kir 2.4 channels, selective inhibition of Hcn1 channels is expected to reveal an inward rectifier current. The known sensitivity of Kir channels to Cs predicts that the inward rectification should not be present in records acquired after CsCl application. Figures 5A and 5Bplot currents measured in response to voltage steps ranging from −90 to −120 mV, in 10-mV steps, after complete inhibition of the slowly activating I _h by ivabradine 30 μM (Fig. 5A)or 3 mM CsCl (Fig. 5B) . 

Note that although these rods had similar membrane resistance (compare current amplitudes at −35 and −70 mV in Figs. 5A and 5B ) and resting potential in bright light, for voltages negative to the potassium reversal potential (approximately −90 mV), current amplitudes increased in the ivabradine-treated cell compared with the Cs-blocked rod. Average data from three experiments are plotted in the Figure 5C , showing that for voltages negative to −80 mV, an inward rectification is apparent after full I _h inhibition by ivabradine, whereas in the presence of CsCl, the rectification was reduced. Activation at −70 mV of a slowly inactivating calcium current may contribute to the inward rectification after ivabradine. To evaluate this possibility, we investigated the properties of inward currents through calcium channels, isolated in the presence of CsCl and other potassium channel blockers, as shown in Figure 5D(see legend). The block by 200 μM CdCl₂ (arrow) confirmed that in these recording conditions, the current was indeed flowing through calcium channels. As shown in Figure 5D , the current turns inward at approximately −50 mV, suggesting that calcium current activation may not explain the residual inward rectification after complete I _h inhibition by ivabradine, which turns inward negatively to −80 mV. 

In a fraction of rods (approximately 20%), we observed an inward peak of 2 to 5 pA (arrow) in the current activated by a step to −120 mV, as shown in Figure 5E . For comparison, Figure 5Fplots the inward current at −80 and −120 mV for a second rod that did not display this transient peak at −120 mV. Note that rods in Figures 5E and 5Fhad similar access resistance (approximately 100 MΩ) and outward currents (Figure 5G) . 

The current decays in Figures 5A and 5Ehave a similarly slow time course, with a decrease of several pA within 2 seconds, as shown in Figure 5H , which plots on expanded scales data from the −120 mV step in Figures 5A and 5E . 

Discussion

The present results show that ivabradine, a use-dependent inhibitor of cardiac pacemaker I _f, ^{5 23} is a selective inhibitor of Hcn1 channels in mouse rod photoreceptors. 

Ivabradine at a high dose that fully blocks I _h (see data in Figs. 2 and 3 ) does not affect membrane currents that gate positively to −50 mV, as shown in Figures 4A 4B 4C 4D . The absence of ivabradine effects on outward currents is in sharp contrast to the inhibition observed in amphibian rods with zatebradine, ¹⁶ a molecule with structural analogies to ivabradine. Although species differences (amphibia vs. mammals) may contribute to the different selectivity of ivabradine and zatebradine in rod photoreceptors, zatebradine was less selective for I _f than ivabradine in mammalian pacing cells. ²⁴  

Analysis of the effects of ivabradine and CsCl (Figs. 2 3)indicates similar activation kinetics and half-activation voltages for ivabradine and Cs-sensitive currents. These results are consistent with ion permeation through homomeric Hcn1 channels, ^{12 25 26 27 28} as expected from in situ hybridization ⁹ and immunolabeling with an antibody anti-HCN1 in mouse retina. ¹⁴ Furthermore, in general agreement with the effects of ivabradine on heterologously expressed homomeric Hcn1 channels, ²² in rods the blocked channels do not reopen during prolonged hyperpolarization (Fig. 3I) . Last, ivabradine-induced block is not strictly use dependent in rods (Fig. 2D) , similar to homomeric Hcn1 channels and in contrast with the use-dependent inhibition of I _f in pacing tissue, when HCN4 is the main isoform. ^{10 11} These observations lend further support to the notion that ivabradine blocks homomeric Hcn1 channels in mouse rods. 

Some differences between CsCl and ivabradine were noted. More negative activation thresholds (Figs. 3B 3C)and steeper inverse slope factors (Fig. 3D)were found for CsCl-subtracted than for ivabradine-subtracted records. Similar negative threshold and steep slope factors for the activation of Cs-subtracted currents have been reported in salamander rods, ¹⁹ suggesting that ivabradine is a more effective I _h inhibitor than Cs at membrane potentials close to the I _h activation threshold. These effects in rods are consistent with a voltage-dependent I _h block by Cs and in agreement with the known voltage dependence of I _f block by Cs. ²⁹  

Selective I _h inhibition by ivabradine revealed an inward rectification that has never previously been reported in rods. The inward rectification observed in the presence of ivabradine (Figs. 5A 5B 5C)is consistent with the expression by rods of the potassium-selective and Cs-sensitive inward rectifier Kir2.4 (coded by Kcnj14). ²⁰ Our data suggest that in mouse rods, ivabradine is a more selective inhibitor of I _h than CsCl, which blocks I _h and Kir2.4-mediated currents. 

Our results indicate considerable variability in amplitude and kinetics of inward rectification. The presence in some rods (Fig. 5E)of a peak in the current activated by a voltage step to −120 mV suggests some variability in the voltage-dependent inactivation of Kir currents. ³⁰ Regarding I _h variability (Figs. 1 5) , cAMP has been reported to increase I _h amplitude in rabbit rods, ¹² and the observed variability may result from changes in cAMP levels. Although no data are available in rods, changes in phosphoinositides may also affect I _h gating. ³¹ The transcription factors coded by Nrl and Nr2e3 are required for Kcnj14 ²⁰ but dispensable for Hcn1 expression in mouse rods (see the GEO database in Medline, http://www.ncbi.nlm.nih.gov/sites/entrez?db=geo; use the search terms: Kcnj14 and Nrl or Hcn1 and Nrl), and the uncoordinated expression of Kcnj14 and Hcn1 may also contribute to the observed differences in the characteristics of inward rectification. 

Phosphene perception has been occasionally reported by patients treated with I _f-inhibitors, ^{7 8} and evidence based on ERG recordings indicate that I _f inhibitors interfere with signal processing across the rod to rod-bipolar synapse. ³² Inhibition of Hcn1 and Hcn2 channels expressed by rods ^{9 12} and rod-bipolar cells, ^{9 13 14} respectively, may increase the amplitude of voltage responses to light and spontaneous voltage fluctuations (for a review see Ref. ³³ ), thus preventing their rejection by the threshold filter implemented in rod bipolar cells. ³⁴ Recent evidence has shown that in rod bipolar cells, Hcn2 channels open in darkness may endow the cell with resonant behavior and shape the voltage response to dim light stimuli. ¹⁴ The observation that in many cases phosphenes perception is triggered by an abrupt change in luminance suggests that the inhibition of Hcn2 channels open in darkness may not be sufficient to induce phosphenes. Data in Figure 3Dindicate that Hcn1 channels are closed at −40 mV, the rod membrane potential in darkness, ³⁵ and a 10- to 15-mV hyperpolarization is required for I _h to reach an intensity of 2 to 3 pA and to have functional relevance. These observations are consistent with the notion that phospenes are triggered by the block of Hcn1 channels in rods. 

Phosphenes may disappear despite continued treatment, suggesting that Hcn1 inhibition is required but not sufficient to trigger phosphenes. In this context, it is tempting to speculate about the role Kir-mediated currents described in Figures 5E 5F 5G 5Hmay play in phosphene perception. In other sensory neurons, a low ratio between I _h and Kir-mediated currents is associated with spontaneous low-frequency oscillations, ³⁶ which are typical of amplified resonance. ³⁷ Considering the variability in I _h amplitude we have reported, in a fraction of rods I _f inhibitors may reduce the I _h/Kir ratio below a critical level, triggering an amplified resonance mode in response to large hyperpolarization. In these rods, the amplification of spontaneous voltage fluctuations may generate phosphenes until the I _h/Kir ratio rises above the critical level. 

In conclusion, the demonstration of selective I _h inhibition in rods by ivabradine paves the way to future work investigating the impact of Hcn1 and Kir 2.4 channel interaction on the processing of visual signals by retinal networks. 

 Supported by the Italian Ministry of University and Education Grants MIUR 2004–2004057720_004 and MIUR 2006–2006053302_002.

 Submitted for publication August 3, 2008; revised October 23, 2008; accepted February 2, 2009.

 Disclosure: G.C. Demontis, Institute de Recherches Servier (F, C); C. Gargini, Institute de Recherches Servier (F, C); T.G. Paoli, Institute de Recherches Servier (F); L. Cervetto, Institute de Recherches Servier (F, C)

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

morphologic and functional properties of isolated rods. (A, B) Living mouse rods, freshly isolated by enzymatic treatment followed by mechanical dissociation, are identified by their characteristic morphology, with slender outer segments (OS) and inner segments (IS) and a cell body (CB). (C) A rod fixed with 1% paraformaldehyde in phosphate-buffered saline for 5 minutes at 4°C was stained by the anti–rhodopsin antibody in the OS. A long tiny axon connects the CB to the rod spherule (RS) that represents the rod axon terminal. In most cases, however, the RS is lost during the dissociation. The small size of mammalian rod IS and CB (diameter, 2–4 μm) makes it difficult to obtain stable recordings from these cells.

Figure 2.

 

Time- and dose-dependent I _h inhibition by ivabradine. (A) Sweeps plot the currents activated by a 2-second long hyperpolarizing step to −80, imposed from a holding voltage of −35 mV, close to the membrane potential of mammalian rods in darkness. ³⁵ CNTR and CsCl indicate sweeps recorded before and after application of 3 mM CsCl, respectively. (B) The use-dependent I _h block by ivabradine was investigated by the application every 5 seconds of a stimulation protocol that alternated a 1.2-second activating step at −80 mV, with a deactivating step at −30 mV, close to the I _h reversal. (C, E, G) Voltage-clamp responses to the application of the stimulation protocol illustrated in (B). Horizontal dashed lines in (A, E, G) indicate the zero-current level. Lowercase letters close to the sweeps indicate records acquired at different times, either before or during the application of ivabradine 30 μM (C), 3 μM (E), and 0.3 μM (G). (C) Transient increase in inward in current (labeled b) was a perfusion artifact because it was not systematically observed with 30 μM ivabradine but was observed in some control records after switching to different perfusion lines. (D, F, H) Data points plot the amplitude of inward current measured at the end of the 1.2-second hyperpolarizing pulse at −80 mV. Dotted lines in (D) indicate current levels at the moment of transient interruption of the stimulation protocol to assess the use dependence of ivabradine blockade. (F) Ten sweeps were averaged to reduce the recording noise; each data point is therefore the mean value from 10 records. Note that on ivabradine washout, slow and partial recovery is present. (I) Columns plot the mean fractional current (with SEM) remaining after the application of ivabradine or CsCl. Three cells were used for each dose, and each cell was tested once with a single dose. Significant differences (P < 0.001) were found between 0 μM ivabradine and both 3 and 30 μM ivabradine.

Figure 3.

 

Ivabradine inhibition of HCN1 channels. (A) Traces plot the stimulation protocol used for the study of I _h steady state activation. From a holding of −35 mV, 2-second long voltage steps, ranging from −60 to −120 mV in 10-mV steps, were followed by a 0.5-second step at −70 mV for tail current analysis. (B, C) Difference records obtained by subtracting records in the presence of 30 μM ivabradine (B) or 3 mM CsCl (C) from their corresponding control records. (D) Current amplitudes during the last 100 ms of each difference record were converted to conductances and normalized to the value at −120 mV to generate fractional activation curves. Average fractional activation plotted from ivabradine-sensitive (filled circles) and Cs-sensitive (open circles) currents. Smooth curves through data points plot best fits by equation 1 . Ivabradine data points were best fit by the continuous curve using the following parameters: V _1/2 = −75.3 ± 2.1 mV; s = 7.6 ± 0.7 mV. CsCl data points were best fit by the dotted curve using the following parameters: V _1/2 = −75.8 ± 2.3 mV; s = 5.1 ± 0.6 mV. (E, F) Dotted traces plot on an expanded scale time difference records at −120 mV from (B) and (C) in (E) and (F), respectively. The smooth continuous lines are best fits of equation 2to difference records, with time constants of 38 and 33 ms for the ivabradine and the Cs-sensitive currents, respectively. Horizontal dotted traces were computed as the difference between fit and difference records. (G) Data points plot average activation time constants as a function of activating voltages from three independent experiments for ivabradine-blocked (filled circles) and Cs-blocked (open circles) currents. (H, I) The response to a 18.75-second long step at −110 (H) applied after full I _h inhibition by 30 μM ivabradine is plotted (I).

Figure 4.

 

Selective inhibition of I _h by ivabradine. (A, B) Outward currents were generated in response to 800-ms long voltage-steps, ranging from −70 to +50 mV, from a holding voltage of −80 mV. The step duration was selected to attain steady state deactivation of the transient outward current. Data in (A) and (B) plot currents recorded in response to voltage stimuli ranging from −10 to +50 mV in 20-mV steps before (thin traces) and after (thick traces) application of 30 μM ivabradine (A) or 3 mM CsCl (B). Horizontal dashed lines indicate zero-current levels. (C, D) Data points plot average difference-record amplitudes from three independent experiments using 30 μM ivabradine (C) or 3 mM CsCl (D). Currents were measured at the end of the 0.8-second long voltage steps stimuli, ranging from −50 to +50 mV in 10-mV steps. (E) Sweeps plot currents generated in response to voltage-clamp stimuli ranging from −30 to +50 mV in 20-mV steps. Large inward tail currents were measured on stepping back to −35 mV from +10 (∗) and +30 mV (∗∗) steps, indicating that the outward current is carried in part by chloride, whose reversal potential is close to zero. No inward tail was measured after stepping back from +50 mV. Note that this chloride current was erratic, was expressed by a fraction of rods, and may in addition spontaneously appear and fade out during recording in the absence of ivabradine. (F) Sweeps plot current generated in response to voltage stimuli ranging from −30 to +30 mV in 20-mV steps before (thin traces) and after (thick traces) I _h inhibition by 30 mM ivabradine (note the differences in the holding currents at −90 mV). *Records in ivabradine.

Figure 5.

 

Inward rectification in the presence of ivabradine. (A, B) Currents were recorded in response to voltage stimuli ranging from −90 to −120 mV in 10-mV steps after the application of 30 μM ivabradine (A) or 3 mM CsCl (B). Horizontal dashed lines indicate zero-current levels. (A) Vertical deflections are perfusion-related artifacts. (C) Data points plot average current amplitudes measured at the end of the 2-second steps, ranging from −60 to −120 mV in 10-mV steps, in the presence of ivabradine (open squares) and CsCl (filled squares), respectively. (D) Sweeps plot current generated in response to a 225-ms long voltage ramp from −75 to +50 mV in the presence of 20 mM TEA, 2.5 mM BaCl₂, and 3 mM CsCl (Ba/Cs/TEA solution) (thin trace) or in the presence of 200 μM CdCl₂ in addition to the Ba/Cs/TEA solution (thick trace). NaCl was reduced to 120 mM to prevent cell shrinking, and records were low-pass filtered at 5 kHz and sampled at 33 kHz. (E, F) Sweeps plot the inward currents activated by 2-second voltage steps at −80 and −120 mV for two different rods. Horizontal dashed lines in (E, F) plot zero-current level. The arrow in (E) points to the peak in the current evoked by the voltage step at −120 mV. Calibration bars in (E) apply to (E) and (F). (G) Thin and thick traces plot the outward currents evoked by a voltage step from −70 to +40 mV for the cells in (E) and (F), respectively. (H) Sweeps plot on expanded scales the responses to the −120 mV voltage step in (A, thick trace) and (E, dotted trace).

The authors thank Dario DiFrancesco for helpful comments and critical reading of an earlier version of this manuscript.