TY - JOUR
AU - Wu, Wendy W.
AB - Background and purpose CaV1.2 channels contribute to action potential upstroke in pacemaker cells, plateau potential in working myocytes, and initiate excitation-contraction coupling. Understanding drug action on CaV1.2 channels may inform potential impact on cardiac function. However, literature shows large degrees of variability between CaV1.2 pharmacology generated by different laboratories, casting doubt regarding the utility of these data to predict or interpret clinical outcomes. This study examined experimental factors that may impact CaV1.2 pharmacology. Experimental approach Whole cell recordings were made on CaV1.2 overexpression cells. Current was evoked using a “step-step-ramp” waveform that elicited a step and a ramp current. Experimental factors examined were: 1) near physiological vs. room temperature for recording, 2) drug inhibition of the step vs. the ramp current, and 3) Ca2+ vs. Ba2+ as the charge carrier. Eight drugs were studied. Key results CaV1.2 current exhibited prominent rundown, exquisite temperature sensitivity, and required a high degree of series resistance compensation to optimize voltage control. Temperature-dependent effects were examined for verapamil and methadone. Verapamil’s block potency shifted by up to 4X between room to near physiological temperature. Methadone exhibited facilitatory and inhibitory effects at near physiological temperature, and only inhibitory effect at room temperature. Most drugs inhibited the ramp current more potently than the step current—a preference enhanced when Ba2+ was the charge carrier. The slopes of the concentration-inhibition relationships for many drugs were shallow, temperature-dependent, and differed between the step and the ramp current. Conclusions and implications All experimental factors examined affected CaV1.2 pharmacology. In addition, whole cell CaV1.2 current characteristics—rundown, temperature sensitivity, and impact of series resistance—are also factors that can impact pharmacology. Drug effects on CaV1.2 channels appear more complex than simple pore block mechanism. Normalizing laboratory-specific approaches is key to improve inter-laboratory data reproducibility. Releasing original electrophysiology records is essential to promote transparency and enable the independent evaluation of data quality. Introduction CaV1.2 channels in the heart mediate L-type Ca2+ current that contributes to Ca2+-dependent action potentials (APs) in the pacemaker cells of the sinoatrial and atrioventricular nodes, the plateau phase of the AP in the working myocytes, and triggers cardiac excitation-contraction coupling [1]. Drugs that reduce CaV1.2 channel activity slow heart rate, shorten AP duration in atrial and ventricular myocytes, and decrease contractile force. CaV1.2 channel agonists have not been used in humans. Nonetheless, studies have shown that pharmacologically enhancing L-type Ca2+ current can produce delayed repolarization and ventricular arrhythmias [2, 3]. Therefore, understanding drug effects on CaV1.2 channels may provide insights regarding a drug’s impact on cardiac function. Indeed, a survey of nonclinical safety assessment of proarrhythmia risk used by the pharmaceutical industry found that during early drug discovery, patch clamp characterization of a drug candidate’s effect on CaV1.2 channels is routinely performed, with frequency second only to the hERG assay [4]. Despite of the potential of CaV1.2 data to inform drug effect in the clinical setting, leveraging in vitro patch clamp results has been challenging for drug regulators. Because of the high inter-laboratory data variability reported in the literature, a major concern is that conclusions based on these data are laboratory-dependent. Inter-laboratory data variability is often ascribed to the use of different voltage protocols and stimulation frequencies. However, results from two recent publications suggest that additional factors may be involved [5, 6]. In these studies, CaV1.2 current was recorded using overexpression cell lines, and evoked from the same holding potential, at the same frequency (0.1 Hz), using either a ventricular AP waveform [5] or a ventricular AP-like “step-step-ramp” waveform [6]. Nonetheless, surprisingly different inhibitory potencies were obtained for the same drugs, with difference up to 1743X reported (Table 1). These results underscore the need to understand the conduct and design of CaV1.2 experiments—how each experimental factor may translate into differences in drug effect, especially if these in vitro results are to be used in decision-making regarding risk prediction or mechanistic interpretation of clinical outcomes. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 1. IC50 differences for CaV1.2 channel block reported by Crumb et al. 2016 [5] and Li et al. 2018 [6]. https://doi.org/10.1371/journal.pone.0276995.t001 Comparison of the abovementioned publications revealed several differences in the CaV1.2 experimental conduct and design. One study used a manual patch clamp system, recorded cells at near physiological temperature (PT) using Ba2+ as the charge carrier, and quantified drug effect on the inward current associated with the repolarizing phase of the AP [5]. The other used an automated patch clamp system, recorded cells at ambient temperature using Ca2+ as the charge carrier, and quantified drug effect on the inward current triggered by the initial voltage step [6]. Recording temperature [7, 8] and charge carrier [9–12] are known to affect block potencies for some drugs on CaV1.2 channels in overexpression cells and L-type channels in native myocytes. In addition, measuring drug effects on CaV1.2 current evoked at different time points following the initial depolarization and associated with different membrane voltages (i.e., current resulting from different channel state, due to activation [6] or reactivation following recovery from inactivation [5]), as was done in these two studies is also a likely source of data variability, given that many drugs are known to block CaV1.2 and L-type Ca2+ channels in a state-dependent manner [13–17]. Using manual whole cell patch clamp method to record cells stably expressing hCaV1.2α, β2, and α2δ1 subunits, the present study was conducted to examine CaV1.2 current characteristics and the impact of the recording temperature, charge carrier, and current region where drug effects were quantified on pharmacology. Results show that CaV1.2 current exhibits prominent rundown following whole cell formation, was exquisitely temperature sensitive, and required a high degree of series resistance compensation to optimize voltage control. These characteristics meant that drug-independent changes in the current amplitude can be anticipated during long lasting pharmacology experiments, and laboratory-specific practices to deal with these changes can be sources of data variability. In addition, all experimental factors examined affected drug potency estimations, with drug-specific magnitude and direction of change. For CaV1.2 data intended to support risk prediction or clinical interpretation, normalizing laboratory-specific practices is essential to promote data reproducibility across laboratories—a pivotal step toward engendering confidence amongst regulators for applying these in vitro data in the decision-making process. To support data transparency, the original electrophysiology records, detailed cell culture procedure, and supplemental materials for the present study are available for download at: https://osf.io/g3msb/. Methods Cells CHO cells stably transfected with hCav1.2α, β2, and α2δ1 subunits (Charles River Laboratory; CT3004) were cultured at 5% CO2 and 37°C, following passage in Ham’s F12 media with L-glutamine nutrient mixture (Gibco #11765054) supplemented with 10% tetracycline-screened fetal bovine serum (FBS) (Cytiva Hyclone SH30071.03T) and the following cell selection reagents: Blasticidin (0.01 mg/mL; Gibco #A1113903), Geneticin (G418; 0.25 mg/mL; Sigma G8168), Hygromycin (0.25 mg/mL; Sigma H0654), and Zeocin (0.40 mg/mL; Invitrogen #46–0509). Cells were seeded at low density and kept in culture for 4–7 days before seeding on glass coverslips for electrophysiology use. By the time cells were detached for seeding, they were fully confluent. Twenty-four to 48 hours prior to recording, cultures were washed with DPBS without Ca2+ or Mg2+ (Gibco #14190144), and then detached by applying Accutase (Sigma A6964) for 2 minutes. Cell suspensions of 30,000–40,000 cells/mL were added to 35 mm petri dishes containing 12 mm glass coverslips, in Ham’s F12 media containing only 10% FBS. Cells were kept at 5% CO2 and 37°C until recording. For this cell line, the expressions of β2 and α2δ1 subunits were constitutive, while that of the pore-forming α subunit required tetracycline induction. Three protocols were used for tetracycline induction to accommodate staff schedule. For the first protocol, cells were seeded late in the afternoon the day prior to recording. On the next day, 16–20 hours after seeding, 2.5 μg/mL tetracycline (Sigma T7660) was added to the petri dishes for 4 hours prior to recording. For the second protocol, cells seeded the day before recording were allowed to attach to glass coverslips for ~6 hours, and 0.5 μg/mL tetracycline was added for overnight induction (typically 16–20 hours prior to recording on the following day). For the third protocol, cells were seeded 4 days prior to recording. The day before recordings, cells were fully detached and seeded in media containing 1 μg/mL tetracycline. On the day of the recording, cells were detached again and seeded on glass coverslips. Regarding the first two protocols, after induction of the α subunit cells adopted a very flat morphology, rendering patching and maintaining long lasting recordings challenging. Cells generated using the third protocol were easier to patch due to the more rounded morphology. The use of different cell culture procedures did not impact pharmacology in this study. The amplitude of CaV1.2 current was dependent on both the amount of tetracycline used and the duration of induction. Electrophysiology Voltage clamp recordings were made with Multiclamp 700B amplifier (Molecular Devices, CA) and digitized using a Digidata 1550B (Molecular Devices, CA) interface and the pClamp 10 software (Molecular Devices, CA). Glass coverslips with cells were placed in a recording chamber mounted on an inverted (Zeiss Axiovert 135TV or A1) or an upright microscope (Zeiss AxioExaminer D1), and the recording chamber was continuously perfused using a gravity-fed perfusion system, with an external solution flowing at a rate of 1.5–3 mL/min. Temperature of the recording solution was elevated using a dual channel temperature controller. Two controller models were used: 1) TC2BIP from Cell MicroControls, which elevated the solution temperature with an inline solution heater, and maintained the bath temperature by heating the ITO-coated glass coverslip which formed the bottom of the recording chamber; and 2) TC-344C from Warner Instruments, which elevated the solution temperature with an inline solution heater, and maintained the bath temperature by heating up the anodized aluminum platform (PH-1) that supported the edge of the plastic recording chamber. Two differences were noted regarding these two controller models: 1) at near 37°C, ~2°C difference between the inflow and center and the recording chamber was observed for TC-344C; and 2) TC2BIP provided more stable temperature control near 37ºC than TC-344C, even when the flow rate changed. Bath temperature in the recording chamber was recorded using a thermistor placed in the bath throughout the experiment. Whole-cell current was recorded at near physiological temperature (PT; 37 ± 2°C) or room temperature (RT; 23 ± 1°C). While most of the recordings in this study aimed at 37°C, the term “near PT” was used to acknowledge the few degrees of temperature fluctuations that occurred during the experiments. The internal solution contained (in mM): 120 aspartic acid, 120 CsOH, 10 CsCl, 10 EGTA, 5 MgATP, 0.4 TrisGTP, 10 HEPES; pH adjusted to 7.4 with 5M CsOH; ~290 mOsM. When Ca2+ was used as the charge carrier, the external solution contained (in mM): 137 NaCl, 4 KCl, 1.8 CaCl2, 1 MgCl2, 10 HEPES, 10 glucose; pH adjusted to 7.4 with 5M NaOH; ~290 mOsM. When Ba2+ was used as the charge carrier, the external solution contained (in mM): 137 NaCl, 4 KCl, 4 BaCl2, 1 MgCl2, 10 HEPES, 10 glucose; pH adjusted to 7.4 with 5M NaOH; ~290 mOsM. Recording electrodes were made by pulling borosilicate glass pipettes (BF150-86-10; Sutter Instrument, CA) with a P97 micropipette puller (Sutter Instruments, CA), and had tip resistances in the range of 1.5–2.5 MΩ when filled with the internal solution. The voltage command values were corrected for the 17 mV liquid junction potential (LJP) that resulted from using the above internal solution and Ca2+-containing external solution at 37°C, estimated using the PClamp 10 software. Given that voltage sensed by the membrane equals to voltage at the pipette minus the LJP (or Vm = Vpipette − VLJP), to hold the cell at -80 mV, the input voltage was set at -63 mV. The 17 mV LJP correction was also applied to RT recordings using Ca2+ as the charge carrier (LJP at 23°C was estimated to be 16 mV), and to near PT recordings using Ba2+ as the charge carrier (LJP at 37°C was estimated to be 17 mV). The voltage waveform used was as follows: from a holding potential of -80 mV, the cell was hyperpolarized to -90 mV for 100 ms, repolarized to -80 mV for 100 ms, depolarized to 0 mV for 40 ms, further depolarized to +30 mV for 200 ms, and finally ramped down to -80 mV in 100 ms (-1.1 V/s). This waveform, modified from that used by Li et al., 2018 [6], was chosen as it evoked an inward current peak at the 0 mV step (where Li et al., 2018 characterized drug effects at) and another one at the repolarizing ramp that reflects the channel state of what Crumb et al., 2016 studied [5]. The use of this voltage protocol therefore permitted a direct comparison of drug effects at the step and the ramp current that were separated in time and associated with different membrane voltages/channel states. The voltage waveform was delivered at 5 s intervals or 0.2 Hz. Signals were filtered at 3 or 10 kHz and sampled at 10 kHz. Whole cell capacitance was neutralized. Series resistance (Rs) was measured approximately 2 minutes following whole cell formation, after signs of membrane resealing were no longer evident, using the membrane test function of the pClamp 10 software. Rs was electronically compensated at 80%. The MultiClamp 700B Rs compensation bandwidth control replaces the “lag” control on earlier Axon amplifier series: Bandwidth = 1 / (2 * π * Lag). This study used the default Rs correction bandwidth of 1.02 kHz, which is equivalent to a lag value of 156 μs. For the pharmacology dataset in this manuscript, Rs was 4.5 ± 0.1 MΩ (± SEM; n = 295; RT and near PT data combined); whole cell capacitance was 35.2 ± 0.9 pF (n = 294; value from one cell was not captured). CaV1.2 current showed pronounced rundown in whole cell configuration (see “Results and discussion”). To study drug effects, after CaV1.2 current reached a quasi-steady state level in the control solution, drug solution was perfused as the recording continued. Depending on the cell quality and stability achieved in the control solution, 1 to 2 drug concentrations were tested per cell. Verapamil at 100 μM was used as a full blocker and was applied at the end of the recordings whenever possible. Drugs Naloxone hydrochloride (0599), tolterodine L-tartrate (3761), and diltiazem hydrochloride (0685) were purchased from Tocris Bioscience. Buprenorphine hydrochloride (B9275, USDEA C-III), (±)-methadone hydrochloride (M0267, USDEA C-II), naltrexone hydrochloride (N3136), (±)-verapamil hydrochloride (V4629), and DMSO (D8418) were purchased from Sigma-Aldrich. Norbuprenorphine hydrochloride (USDEA C-II) was purchased from Noramco. Stock solutions of naloxone, naltrexone, and verapamil were dissolved in milliQ water. Stock solutions of methadone, buprenorphine, norbuprenorphine, diltiazem, and tolterodine were dissolved in DMSO. When DMSO was used as a solvent, the % of DMSO exposed to cells was ≤0.3%. Aliquoted stock solutions were stored at -20°C until the day of experiments and were diluted to specific test concentrations in the external solution. Data analysis and reporting Data analysis and curve fitting were done in Clampfit 10.6 (Molecular Devices, CA) and Igor Pro 8.0 (WaveMetrics). Two offline methods were used to isolate CaV1.2 current from the total inward current. The first method was the passive current (Ipassive) subtraction. Ohm’s law was used to calculate the resting input resistance (Rinput) for each current trace: Here I-80 mV refers to the current measured at the holding potential of -80 mV (V-80 mV), and I-90 mV the current measured during hyperpolarizing step to -90 mV (V-90 mV). Ipassive was defined to exhibit linear current-voltage (I-V) relationship. Therefore, assuming that Rinput was constant across all voltages, Ipassive was calculated using the following equation: Here V refers to any voltage within the “step-step-ramp” protocol. Using the custom macros written for Igor Pro, Ipassive was calculated for each current trace and then subtracted from that trace to yield CaV1.2 current. The second current subtraction method was verapamil subtraction. At positive membrane potentials, a population of cells exhibited a non-linear outward current that was not removed by Ipassive subtraction. This outward current was most notable at the +30 mV step and the adjoining repolarizing ramp section. For these cells, the residual current trace in the presence of 100 μM verapamil was subtracted from all current traces to isolate CaV1.2 current. Verapamil subtraction was performed by averaging several traces recorded in verapamil that exhibited full inhibition of the ramp current, and then subtracting this averaged trace from all recorded current traces. For the methadone experiments, some cells did not receive 100 μM verapamil. Nonetheless, complete elimination of the ramp current was achieved by 100 and 300 μM methadone. In these cases, several traces following full ramp current inhibition by methadone were averaged and then subtracted from all recorded traces to isolate CaV1.2 current. S1 Fig illustrates these two subtraction methods and the phenotype of cells to which each was applied. Of note, even when the ramp current was completely eliminated by verapamil or methadone, a sizable portion of the 0 mV step current remained (see “Results and discussions”). Therefore, the step current was always quantified from Ipassive-subtracted traces. The ramp current was quantified using Ipassive-subtracted traces when there was no overt outward current at the positive membrane potentials or using verapamil-subtracted (or methadone-subtracted) traces when there was overt outward current, and the recording showed no time-dependent changes in the passive membrane properties, inferred from stability of Rinput and I-80 mV. From the Ipassive- or verapamil-subtracted current traces, the step and the ramp current were measured as the most negative inward current at the 0 mV step and the entire repolarizing ramp, respectively. Fractional inhibition by the tested drug for each cell was calculated with the following equation: Here Idrug is the averaged current amplitude from the last 10 traces recorded in the drug concentration, and Icontrol is the averaged current amplitude from the last 10 traces recorded in the control solution. Fractional inhibition values for individual cells were plotted against concentrations tested to yield concentration-inhibition graphs, and individual data points were fit with the Hill equation to estimate drug potency: Here IC50 is the concentration that inhibited 50% of the current, [drug] is the drug concentration, and nH is the Hill coefficient. The upper and lower 95% confidence interval (CI) bands were also plotted in the concentration-inhibition graphs to demonstrate uncertainty of the fit parameters. Except for the IC50 and nH values, all data are presented as mean ± SEM. IC50s of buprenorphine, norbuprenorpine, methadone, naltrexone, and naloxone on the ramp current studied in external Ca2+ and at 37ºC were published previously [18]. The IC50 values there differed slightly from those presented in this manuscript because they were estimated by fitting the averaged fractional inhibition values (instead of individual cells’ values) at different concentrations. A t-test with equal variance was performed to compare the extent of Ca2+ and Ba2+ current rundown, using Prism version 8.4 (GraphPad, CA). Cells CHO cells stably transfected with hCav1.2α, β2, and α2δ1 subunits (Charles River Laboratory; CT3004) were cultured at 5% CO2 and 37°C, following passage in Ham’s F12 media with L-glutamine nutrient mixture (Gibco #11765054) supplemented with 10% tetracycline-screened fetal bovine serum (FBS) (Cytiva Hyclone SH30071.03T) and the following cell selection reagents: Blasticidin (0.01 mg/mL; Gibco #A1113903), Geneticin (G418; 0.25 mg/mL; Sigma G8168), Hygromycin (0.25 mg/mL; Sigma H0654), and Zeocin (0.40 mg/mL; Invitrogen #46–0509). Cells were seeded at low density and kept in culture for 4–7 days before seeding on glass coverslips for electrophysiology use. By the time cells were detached for seeding, they were fully confluent. Twenty-four to 48 hours prior to recording, cultures were washed with DPBS without Ca2+ or Mg2+ (Gibco #14190144), and then detached by applying Accutase (Sigma A6964) for 2 minutes. Cell suspensions of 30,000–40,000 cells/mL were added to 35 mm petri dishes containing 12 mm glass coverslips, in Ham’s F12 media containing only 10% FBS. Cells were kept at 5% CO2 and 37°C until recording. For this cell line, the expressions of β2 and α2δ1 subunits were constitutive, while that of the pore-forming α subunit required tetracycline induction. Three protocols were used for tetracycline induction to accommodate staff schedule. For the first protocol, cells were seeded late in the afternoon the day prior to recording. On the next day, 16–20 hours after seeding, 2.5 μg/mL tetracycline (Sigma T7660) was added to the petri dishes for 4 hours prior to recording. For the second protocol, cells seeded the day before recording were allowed to attach to glass coverslips for ~6 hours, and 0.5 μg/mL tetracycline was added for overnight induction (typically 16–20 hours prior to recording on the following day). For the third protocol, cells were seeded 4 days prior to recording. The day before recordings, cells were fully detached and seeded in media containing 1 μg/mL tetracycline. On the day of the recording, cells were detached again and seeded on glass coverslips. Regarding the first two protocols, after induction of the α subunit cells adopted a very flat morphology, rendering patching and maintaining long lasting recordings challenging. Cells generated using the third protocol were easier to patch due to the more rounded morphology. The use of different cell culture procedures did not impact pharmacology in this study. The amplitude of CaV1.2 current was dependent on both the amount of tetracycline used and the duration of induction. Electrophysiology Voltage clamp recordings were made with Multiclamp 700B amplifier (Molecular Devices, CA) and digitized using a Digidata 1550B (Molecular Devices, CA) interface and the pClamp 10 software (Molecular Devices, CA). Glass coverslips with cells were placed in a recording chamber mounted on an inverted (Zeiss Axiovert 135TV or A1) or an upright microscope (Zeiss AxioExaminer D1), and the recording chamber was continuously perfused using a gravity-fed perfusion system, with an external solution flowing at a rate of 1.5–3 mL/min. Temperature of the recording solution was elevated using a dual channel temperature controller. Two controller models were used: 1) TC2BIP from Cell MicroControls, which elevated the solution temperature with an inline solution heater, and maintained the bath temperature by heating the ITO-coated glass coverslip which formed the bottom of the recording chamber; and 2) TC-344C from Warner Instruments, which elevated the solution temperature with an inline solution heater, and maintained the bath temperature by heating up the anodized aluminum platform (PH-1) that supported the edge of the plastic recording chamber. Two differences were noted regarding these two controller models: 1) at near 37°C, ~2°C difference between the inflow and center and the recording chamber was observed for TC-344C; and 2) TC2BIP provided more stable temperature control near 37ºC than TC-344C, even when the flow rate changed. Bath temperature in the recording chamber was recorded using a thermistor placed in the bath throughout the experiment. Whole-cell current was recorded at near physiological temperature (PT; 37 ± 2°C) or room temperature (RT; 23 ± 1°C). While most of the recordings in this study aimed at 37°C, the term “near PT” was used to acknowledge the few degrees of temperature fluctuations that occurred during the experiments. The internal solution contained (in mM): 120 aspartic acid, 120 CsOH, 10 CsCl, 10 EGTA, 5 MgATP, 0.4 TrisGTP, 10 HEPES; pH adjusted to 7.4 with 5M CsOH; ~290 mOsM. When Ca2+ was used as the charge carrier, the external solution contained (in mM): 137 NaCl, 4 KCl, 1.8 CaCl2, 1 MgCl2, 10 HEPES, 10 glucose; pH adjusted to 7.4 with 5M NaOH; ~290 mOsM. When Ba2+ was used as the charge carrier, the external solution contained (in mM): 137 NaCl, 4 KCl, 4 BaCl2, 1 MgCl2, 10 HEPES, 10 glucose; pH adjusted to 7.4 with 5M NaOH; ~290 mOsM. Recording electrodes were made by pulling borosilicate glass pipettes (BF150-86-10; Sutter Instrument, CA) with a P97 micropipette puller (Sutter Instruments, CA), and had tip resistances in the range of 1.5–2.5 MΩ when filled with the internal solution. The voltage command values were corrected for the 17 mV liquid junction potential (LJP) that resulted from using the above internal solution and Ca2+-containing external solution at 37°C, estimated using the PClamp 10 software. Given that voltage sensed by the membrane equals to voltage at the pipette minus the LJP (or Vm = Vpipette − VLJP), to hold the cell at -80 mV, the input voltage was set at -63 mV. The 17 mV LJP correction was also applied to RT recordings using Ca2+ as the charge carrier (LJP at 23°C was estimated to be 16 mV), and to near PT recordings using Ba2+ as the charge carrier (LJP at 37°C was estimated to be 17 mV). The voltage waveform used was as follows: from a holding potential of -80 mV, the cell was hyperpolarized to -90 mV for 100 ms, repolarized to -80 mV for 100 ms, depolarized to 0 mV for 40 ms, further depolarized to +30 mV for 200 ms, and finally ramped down to -80 mV in 100 ms (-1.1 V/s). This waveform, modified from that used by Li et al., 2018 [6], was chosen as it evoked an inward current peak at the 0 mV step (where Li et al., 2018 characterized drug effects at) and another one at the repolarizing ramp that reflects the channel state of what Crumb et al., 2016 studied [5]. The use of this voltage protocol therefore permitted a direct comparison of drug effects at the step and the ramp current that were separated in time and associated with different membrane voltages/channel states. The voltage waveform was delivered at 5 s intervals or 0.2 Hz. Signals were filtered at 3 or 10 kHz and sampled at 10 kHz. Whole cell capacitance was neutralized. Series resistance (Rs) was measured approximately 2 minutes following whole cell formation, after signs of membrane resealing were no longer evident, using the membrane test function of the pClamp 10 software. Rs was electronically compensated at 80%. The MultiClamp 700B Rs compensation bandwidth control replaces the “lag” control on earlier Axon amplifier series: Bandwidth = 1 / (2 * π * Lag). This study used the default Rs correction bandwidth of 1.02 kHz, which is equivalent to a lag value of 156 μs. For the pharmacology dataset in this manuscript, Rs was 4.5 ± 0.1 MΩ (± SEM; n = 295; RT and near PT data combined); whole cell capacitance was 35.2 ± 0.9 pF (n = 294; value from one cell was not captured). CaV1.2 current showed pronounced rundown in whole cell configuration (see “Results and discussion”). To study drug effects, after CaV1.2 current reached a quasi-steady state level in the control solution, drug solution was perfused as the recording continued. Depending on the cell quality and stability achieved in the control solution, 1 to 2 drug concentrations were tested per cell. Verapamil at 100 μM was used as a full blocker and was applied at the end of the recordings whenever possible. Drugs Naloxone hydrochloride (0599), tolterodine L-tartrate (3761), and diltiazem hydrochloride (0685) were purchased from Tocris Bioscience. Buprenorphine hydrochloride (B9275, USDEA C-III), (±)-methadone hydrochloride (M0267, USDEA C-II), naltrexone hydrochloride (N3136), (±)-verapamil hydrochloride (V4629), and DMSO (D8418) were purchased from Sigma-Aldrich. Norbuprenorphine hydrochloride (USDEA C-II) was purchased from Noramco. Stock solutions of naloxone, naltrexone, and verapamil were dissolved in milliQ water. Stock solutions of methadone, buprenorphine, norbuprenorphine, diltiazem, and tolterodine were dissolved in DMSO. When DMSO was used as a solvent, the % of DMSO exposed to cells was ≤0.3%. Aliquoted stock solutions were stored at -20°C until the day of experiments and were diluted to specific test concentrations in the external solution. Data analysis and reporting Data analysis and curve fitting were done in Clampfit 10.6 (Molecular Devices, CA) and Igor Pro 8.0 (WaveMetrics). Two offline methods were used to isolate CaV1.2 current from the total inward current. The first method was the passive current (Ipassive) subtraction. Ohm’s law was used to calculate the resting input resistance (Rinput) for each current trace: Here I-80 mV refers to the current measured at the holding potential of -80 mV (V-80 mV), and I-90 mV the current measured during hyperpolarizing step to -90 mV (V-90 mV). Ipassive was defined to exhibit linear current-voltage (I-V) relationship. Therefore, assuming that Rinput was constant across all voltages, Ipassive was calculated using the following equation: Here V refers to any voltage within the “step-step-ramp” protocol. Using the custom macros written for Igor Pro, Ipassive was calculated for each current trace and then subtracted from that trace to yield CaV1.2 current. The second current subtraction method was verapamil subtraction. At positive membrane potentials, a population of cells exhibited a non-linear outward current that was not removed by Ipassive subtraction. This outward current was most notable at the +30 mV step and the adjoining repolarizing ramp section. For these cells, the residual current trace in the presence of 100 μM verapamil was subtracted from all current traces to isolate CaV1.2 current. Verapamil subtraction was performed by averaging several traces recorded in verapamil that exhibited full inhibition of the ramp current, and then subtracting this averaged trace from all recorded current traces. For the methadone experiments, some cells did not receive 100 μM verapamil. Nonetheless, complete elimination of the ramp current was achieved by 100 and 300 μM methadone. In these cases, several traces following full ramp current inhibition by methadone were averaged and then subtracted from all recorded traces to isolate CaV1.2 current. S1 Fig illustrates these two subtraction methods and the phenotype of cells to which each was applied. Of note, even when the ramp current was completely eliminated by verapamil or methadone, a sizable portion of the 0 mV step current remained (see “Results and discussions”). Therefore, the step current was always quantified from Ipassive-subtracted traces. The ramp current was quantified using Ipassive-subtracted traces when there was no overt outward current at the positive membrane potentials or using verapamil-subtracted (or methadone-subtracted) traces when there was overt outward current, and the recording showed no time-dependent changes in the passive membrane properties, inferred from stability of Rinput and I-80 mV. From the Ipassive- or verapamil-subtracted current traces, the step and the ramp current were measured as the most negative inward current at the 0 mV step and the entire repolarizing ramp, respectively. Fractional inhibition by the tested drug for each cell was calculated with the following equation: Here Idrug is the averaged current amplitude from the last 10 traces recorded in the drug concentration, and Icontrol is the averaged current amplitude from the last 10 traces recorded in the control solution. Fractional inhibition values for individual cells were plotted against concentrations tested to yield concentration-inhibition graphs, and individual data points were fit with the Hill equation to estimate drug potency: Here IC50 is the concentration that inhibited 50% of the current, [drug] is the drug concentration, and nH is the Hill coefficient. The upper and lower 95% confidence interval (CI) bands were also plotted in the concentration-inhibition graphs to demonstrate uncertainty of the fit parameters. Except for the IC50 and nH values, all data are presented as mean ± SEM. IC50s of buprenorphine, norbuprenorpine, methadone, naltrexone, and naloxone on the ramp current studied in external Ca2+ and at 37ºC were published previously [18]. The IC50 values there differed slightly from those presented in this manuscript because they were estimated by fitting the averaged fractional inhibition values (instead of individual cells’ values) at different concentrations. A t-test with equal variance was performed to compare the extent of Ca2+ and Ba2+ current rundown, using Prism version 8.4 (GraphPad, CA). Results and discussion Rundown of CaV1.2 current in external Ca2+ or Ba2+ at near PT Fig 1A and 1B show CaV1.2 current recorded in external Ca2+ and Ba2+, respectively. The maximal peak current evoked by the 0 mV step is referred to as ICa-step or IBa-step depending on the charge carrier used; the maximal ramp current evoked by the repolarizing ramp, ICa-ramp or IBa-ramp. Fig 1C–1F show the time course plots of normalized current amplitudes recorded in the control solution and following verapamil (100 μM) application. After whole cell formation, CaV1.2 current showed prominent rundown regardless of which charge carrier was used, and in most cells eventually reached a quasi-steady state level. Current rundown was seen in every cell and was characterized by a fast phase followed by a slower phase. The normalized ICa-step at the 150th trace was 0.42 ± 0.04 relative to the 1st trace (Fig 1C); the normalized ICa-ramp, 0.44 ± 0.06 (Fig 1D). The normalized IBa-step was 0.34 ± 0.04 (Fig 1E); the normalized IBa-ramp, 0.32 ± 0.04 (Fig 1F). To assess whether the extent of rundown was different between the Ca2+ and the Ba2+ current, a basic statistical analysis of the differences in the amplitude of the 150th trace relative to the first trace was performed. A t-test with equal variance revealed no significant difference between ICa-step and IBa-step and between ICa-ramp and IBa-ramp (p > 0.05). Fig 1G and 1H show the time course of the ratios of ramp-to-step current for traces acquired in the control solution. For Ca2+ current, the averaged ratio was 0.33 ± 0.03; for Ba2+ current, 0.60 ± 0.05. The larger ratio obtained in Ba2+ is consistent with the removal of Ca2+-dependent inactivation [19, 20]. These ratios remained constant throughout the control recording despite current rundown, suggesting that rundown reflects a progressive loss in the available channels. Rundown assessed using the present voltage protocol at 0.2 Hz was not attributed to intracellular Ca2+ accumulation, since it occurred to a similar degree in external Ba2+. Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 1. Rundown of CaV1.2 channel activity at near PT. A, B. Top, representative current traces from 2 cells recorded in either external Ca2+ (A) or Ba2+ (B). Cell ID for (A) was 19327000; for (B), 19404005. Black traces reflect recordings obtained in the control solution, and the 1st, 10th, 90th, and 100th traces recorded after ~2 minutes of whole cell dialysis are shown. Red traces reflect recordings obtained following application of 100 μM verapamil, after steady state inhibition was achieved. Ipassive traces, shown in gray, were calculated using Rinput derived from the verapamil traces. Bottom, the voltage protocol used. C, D) Summary time course plots of normalized ICa-step (C) and ICa-ramp (D) in the control solution and following verapamil application (n = 16). Data points are shown as mean ± sem. Verapamil application did not start at the same time for every cell. Therefore, the X-axes show a break after trace 150 to synchronize the data points obtained following verapamil application. E, F) Summary time course plots of normalized IBa-step (E) and IBa-ramp (F) in the control solution and following verapamil perfusion (n = 17). G, H) Ratio of ICa-ramp-to-ICa-step (G) or IBa-ramp-to-IBa-step (H) for traces acquired in the control solution. Data for individual cells are shown as light gray lines. Mean ± sem are shown as black symbols plus error bars. https://doi.org/10.1371/journal.pone.0276995.g001 Application of verapamil nearly eliminated the ramp current (Fig 1D and 1F) but not the step current (Fig 1C and 1E). Relative to the 150th trace recorded in the control solution, the residual ICa-step in verapamil was 20 ± 2%, while that of ICa-ramp was 4 ± 1%. Likewise, the residual IBa-step in verapamil was 32 ± 4%, while that of IBa-ramp was 2 ± 0%. Therefore, amplitude of the step current was always quantified using Ipassive-subtracted traces, not verapamil-subtracted traces (see “Methods”). Rundown of CaV1.2 current in whole cell configuration is a widely observed phenomenon. Since rundown leads to overestimation of fractional inhibition, laboratory-specific tolerance regarding the rate of rundown and practices to correct rundown can introduce variable degrees of imprecision into drug potency estimation. The diverse practices are exemplified by the two publications that motivated the present study. Crumb et al., 2016 did not correct for rundown, stated that cells with >20% rundown were discarded, but did not define how many current traces the 20% calculation was based [5]. Li et al., 2018 used perforated population recording that presumably reduced rundown, yet still corrected for rundown in the calculation of drug inhibition [6]. Even if the rates of CaV1.2 current rundown were similar between the two publications, these different practices alone would lead to different drug potency estimations. Experience from this laboratory indicates that the rate of rundown, and whether current can reach a quasi-steady state in the control solution for pharmacology experiments are functions of the cell line used and cell culture conditions, and therefore unlikely to be the same across studies. Using the same Ca2+-external solution, internal solution, and voltage protocol, two additional CaV1.2 cell lines that expressed the same channel subunits were tested, and both cell lines showed near complete loss of CaV1.2 current at near PT with time, without reaching a quasi-steady state level (S2 Fig). Can CaV1.2 current rundown under whole cell mode be prevented? Previous studies suggest that rundown of cardiac L-type Ca2+ channel activity reflects channel dephosphorylation. This conclusion is based on evidence from native myocytes that manipulating protein kinase A (PKA)-mediated phosphorylation [21], protein phosphatase activity [21], and increasing the intracellular level of ATP or cyclic AMP—molecules that enhance PKA-mediated phosphorylation [22] all led to expected changes in Ca2+ channel activity. In the present study, inclusion of 5 mM MgATP and 0.4 mM TrisGTP in the internal solution did not prevent current rundown in the three cell lines tested. Therefore, rundown in whole cell configuration could not be prevented by simply supplying ATP. Temperature sensitivity of CaV1.2 current During near PT recordings, the step current sometimes showed amplitude fluctuations that occurred without changes in the passive membrane properties (i.e., Rinput or I-80 mV; S3 Fig). As these amplitude fluctuations were not observed with RT recordings, one hypothesis is that they are associated with temperature fluctuations during the near PT recordings. Therefore, temperature sensitivity of the CaV1.2 current was examined. These experiments were conducted using setups with the TC-344C controller, as amplitude fluctuations were more common when recorded using these setups, and the thermistor measuring the bath temperature was positioned as close to the recorded cell as possible. After the current reached a quasi-steady state level at near PT, temperature control of the aluminum platform was turned off to allow graded bath temperature drop at the recording chamber, and then back on to elevate the bath temperature. Fig 2A shows the time course plots of ICa-step and ICa-ramp from a representative cell. The boxed regions are expanded and shown in Fig 2B. ICa-step decreased and increased as the bath temperature lowered and elevated, respectively (top; note that Ca2+ current amplitudes are expressed as negative values), while ICa-ramp appeared to show the opposite pattern (bottom). No change in Rinput or I-80 mV was observed (Fig 2C). The left panel of Fig 2D shows the plot of ICa-step vs. temperature for this cell, demonstrating a strong linear relationship (r = -0.94). In contrast, no relationship was observed between ICa-ramp and temperature (r = 0.38; Fig 2D, right). Temperature sensitivity of the Ca2+ current was confirmed in 5 more cells (S4A–S4E Fig). In all cells recorded, ICa-step showed strong linear relationship with temperature (r = -0.77 to -0.94). The ICa-ramp-temperature relationship was inconsistent: 4 cells showed no relationship (Fig 2D, right; S4A, S4B and S4D Fig., right), and 2 showed linear relationship (r = -0.73 to -0.88). Fig 2E shows the calculated and normalized ICa-step at 37°C and 34°C. Depending on the cell, 3°C of temperature drop reduced ICa-step amplitude by 12% to 34%. This magnitude of temperature fluctuation was routinely observed during near PT experiments for setups with TC-344 controller. Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 2. Effects of recording temperature on CaV1.2 channel activity. Cell ID for (A) through (D) was 19318007. Recordings were obtained in external Ca2+. A) Time course plots of ICa-step (open circle) and ICa-ramp (open square) for the entire experiment. The boxed regions were expanded and shown in (B). B) Top, time course plots of ICa-step from traces 96 to 216 and bath temperature (gray dotted line). Current amplitudes were corrected for rundown. This could be done since rundown correction was performed on the traces used for the fit. To estimate rundown, data points in (B) were fit with a linear function to yield a slope of 0.00495 nA/trace. This amount of current loss was then added back to the original ICa-step amplitudes. Bottom, time course plots of ICa-ramp and bath temperature. Current amplitudes were also corrected for run rundown (slope = 0.00146 nA/trace). C) Rinput (top) and I-80 mV (bottom). D) Left, rundown-corrected ICa-step vs. bath temperature. These data points were fit with a linear function, yielding a slope of -0.295 nA/°C and a Y-intercept of 7.299 nA (r = -0.94). Right, rundown-corrected ICa-ramp vs. bath temperature (r = 0.38). E) Calculated and normalized ICa-step at 37°C (black circle) and 34°C (gray circle). Cell ID for (F) to (J) was 07_07_0008. Recordings were obtained in external Ba2+. F) Time course plots of IBa-step (open circle) and IBa-ramp (open square) for the entire experiment. The boxed region was expanded and shown in (G). G) Time course plots of IBa-step (top) and IBa-ramp (bottom) from traces 60 to 200 and bath temperature (gray dotted line). Both IBa-step and IBa-ramp were corrected for rundown, using 0.00741 nA/trace and 0.00538 nA/trace, respectively. H) Rinput (top) and I-80 mV (bottom). I) Rundown-corrected IBa-step (left) or IBa-ramp (right) vs. bath temperature. Fitting IBa-step vs. temperature plot with a linear function yielded a slope of -0.147 nA/°C and a Y-intercept of 2.947 nA (r = -0.88). Rundown-corrected IBa-ramp vs. temperature plot (r = -0.36). J) Calculated and normalized IBa-step at 37°C (black circle) and 34°C (gray circle). https://doi.org/10.1371/journal.pone.0276995.g002 Temperature sensitivity of the Ba2+ current was examined as well. Fig 2F shows the time course plots of IBa-step and IBa-ramp from a representative cell, and the boxed region was expanded and shown in Fig 2G. IBa-step changed in the same direction as the bath temperature (top), and IBa-ramp seemed to not respond to temperature changes (bottom). Rinput and I-80 mV remained stable during temperature manipulations (Fig 2H). The left panel of Fig 2I shows the plot of IBa-step vs. temperature, demonstrating a clear linear relationship (r = -0.88). The plot of IBa-ramp vs. temperature, on the other hand, did not reveal any relationship (r = -0.36; Fig 2I, right). Temperature-dependent effect on the Ba2+ was confirmed in 4 more cells (S4F–S4I Fig). In all cells recorded, IBa-step showed strong linear relationship with temperature (r = -0.87 to -0.93). The data for IBa-ramp and temperature were inconsistent: 3 cell showed no-to-weak relationship (Fig 2I, right; S4F and S4G Fig, right), and 2 showed strong relationship (r = -0.80 and -0.88). Fig 2J shows the calculated and normalized IBa-step at 37°C and 34°C. Depending on the cell, 3°C temperature drop reduced IBa-step by 18% to 28%. The step current in the present study is very temperature-sensitive regardless of which charge carrier was used. These results are consistent with prior studies conducted on cloned CaV1.2 channels comprised of α1c, α2/δa, and β1b (or β2c) expressed in xenopus oocytes [23] and L-type Ca2+ channels in native ventricular myocytes [24, 25] that also showed high temperature sensitivity. Increasing PKA-mediated phosphorylation reduced temperature sensitivity of some Ca2+ channel gating parameters, suggesting that the high temperature sensitivity may be due to channels in the unphosphorylated state [25]. Temperature sensitivity of the step current adds an additional challenge to conducting pharmacology experiments at near PT, since several degrees of temperature fluctuations can easily be produced by a slowing of the flow rate due to the presence of bubbles in the perfusion line and/or reduced fluid level in the reservoirs of the gravity-fed perfusion system. Temperature sensitivity of the CaV1.2 current is thus a source of variability for pharmacology data generated within and across laboratories. Consequence of Rs compensation on CaV1.2 current at near PT The cells used to generate Fig 1 were used to estimate the activation kinetics and amplitude of CaV1.2 current at near PT. ICa-step reached peak 1.84 ± 0.17 ms following the voltage jump to 0 mV (range: 1.34 to 3.39 ms; n = 12, using cells with clear separation between capacitive transient and the step current), and was -2210.1 ± 178.3 pA in amplitude (range: -1376.7 to -3362.9 pA, based the average of 91st to 100th traces recorded in the control solution). ICa-ramp reached peak at 0.6 ± 0.6 mV during the repolarizing ramp (range: -4.0 to 3.6 mV; n = 16), and was -814.0 ± 108.4 pA in amplitude (range: -320.2 to -1546.3 pA). The fast kinetics and relatively large amplitudes of the step current prompted for a set of experiments to assess the impact of Rs compensation in these recordings. During these experiments, Rs was measured and compensated at 80% before the start of recording. After CaV1.2 current reached a quasi-steady state level in the control solution, Rs compensation was turned “off” and “on” as the recording continued. Fig 3A and 3B show representative current traces and time course plots of ICa-step and ICa-ramp from one cell obtained with and without Rs compensation. The total initial Rs of this cell was 5.6 MΩ. With Rs compensation, ICa-step was ~600 pA larger than without Rs compensation (Fig 3B, top). In contrast, no difference was observed in ICa-ramp amplitude (bottom) or the ramp voltage at which ICa-ramp reached peak (Fig 3C, top). Rinput and I-80 mV remained stable (Fig 3C, middle and bottom). The impact of Rs compensation was confirmed in 5 additional cells. The left panel of Fig 3D shows the fractional ICa-step loss without Rs compensation for all 6 cells, plotted against either ICa-step with Rs compensation (solid symbols) or the amount of Rs that was compensated (or 80% total Rs; open symbols). No relationship was evident amongst these parameters, as expected since voltage loss through Rs (hence fractional current loss) is a product of both the current amplitude and Rs. For these cells, voltage loss through Rs at ICa-step ranged from 4.7 to 21.3 mV, calculated by using Ohm’s law. The right panel of Fig 3D shows the fractional ICa-ramp loss due to not compensating for Rs, plotted as a function of ICa-ramp with Rs compensation. Collectively, no loss was observed (average: -0.02 ± 0.02). Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 3. Effects of Rs compensation on Ca2+ and Ba2+ currents. Data shown in panels (A) through (D) were obtained in external Ca2+. Cell ID was 21318008 for panels (A) through (D). A) Top, representative Ca2+ current traces obtained from a cell, with (black traces) and without Rs compensation (gray traces). Bottom, the voltage protocol used. B) Time course plots of ICa-step (top) and ICa-ramp (bottom), focusing on traces 30 to 120 for which Rs compensation was turned on and off. No rundown correction was made for these plots. C) Time course plots of ramp voltage at which ICa-ramp reached peak amplitude (top), Rinput (middle), and I-80 mV (bottom) for traces 30 to 120. D) Left, fractional ICa-step loss due to not compensating for Rs, calculated as the ratio of ICa-step without Rs compensation vs. ICa-step with 80% Rs compensation, and plotted against ICa-step obtained with Rs compensation (lower x-axis) or the amount of Rs compensated (upper x-axis). Right, fractional ICa-ramp loss due to not compensating for Rs, plotted against ICa-ramp with Rs compensation. Data shown in panels (E) through (H) were obtained in external Ba2+. Cell ID was 2021_06_25_0004 for panels (E) through (I). E) Representative Ba2+ current traces obtained from a cell, with (black traces) and without Rs compensation (gray traces). F) Time course plots of IBa-step (top) and IBa-ramp (bottom), focusing on traces 40 to 100 during which Rs compensation was turned on and off. No rundown correction was made for these plots. G) Time course plots of ramp voltage (top), Rinput (middle), and I-80 mV (bottom) for traces 40 to 100. H) Left, fractional IBa-step loss due to not compensating for Rs, plotted against IBa-step obtained with Rs compensation (lower x-axis) or the amount of Rs compensated (upper x-axis). Right, delta (Δ) ramp voltage shift for IBa-ramp due to not compensating for Rs, plotted against IBa-ramp with Rs compensation. https://doi.org/10.1371/journal.pone.0276995.g003 The kinetics and amplitude of Ba2+ current at near PT was also quantified. IBa-step reached peak 2.91 ± 0.25 ms following the voltage jump to 0 mV (range: 1.64 to 4.23 ms; n = 11), and was -4147.6 ± 813.1 pA in amplitude (range: -1420.0 to -8395.0 pA). IBa-ramp reached peak at -7.7 ± 0.7 mV (range: -2.5 to -10.5 mV; n = 11) and was -2077.7 ± 309.5 pA in amplitude (range: -555.2 to -3506.4 pA). Fig 3E showed representative Ba2+ current traces obtained from a cell, with and without Rs compensation. The total Rs for this cell was 4.1 MΩ. Similar to Ca2+ current recordings, Rs compensation affected the amplitude of IBa-step (by ~1500 pA; Fig 3F, top) but not the amplitude of IBa-ramp (Fig 3F, bottom). However, the ramp voltage at which IBa-ramp peaked was consistently shifted rightward when Rs was not compensated, toward the more hyperpolarized potential (Fig 3G, top). No change in Rinput or I-80 mV was observed (Fig 3G, middle and bottom). The impact of Rs compensation on the Ba2+ current was verified in 5 more cells. The left panel of Fig 3H shows the fractional IBa-step loss due to not compensating for Rs, plotted against IBa-step amplitude with Rs compensation (solid symbols) or the amount of Rs compensated (open symbols). No relationship was evident amongst these parameters. Voltage loss through Rs at IBa-step ranged from 4.8 to 28.2 mV. The right panel of Fig 3H summarizes the hyperpolarizing shift of the ramp voltage for IBa-ramp without Rs compensation. On average, the shift was -7.7 ± 1.3 mV (range: -4.6 to -12.1 mV) and was not accompanied by IBa-ramp amplitude change (fractional IBa-ramp loss: 0.03 ± 0.00). Using normalized I-V relation generated from 15 cells, the reversal potential of Ba2+ current was estimated to be +35 mV (S5A Fig). The increase in Ba2+ driving force due to the hyperpolarizing shift in the ramp voltage offers an explanation as to why IBa-ramp amplitude was unaltered by not compensating for Rs, since this could compensate for the fewer number of CaV1.2 channels that would be open due to not clamping the membrane potential at the expected levels. These results demonstrate the impact of Rs compensation when recording fast and large Ca2+ and Ba2+ currents at near PT. While Rs compensation is a good practice for voltage clamp experiments, can Rs affect pharmacology results? A recent study that modeled drug block of Na+ channels indicates so, with the degree of rightward shift of the concentration-inhibition graphs dependent on the magnitudes of Rs and Na+ current [26]. Rs compensation is a practice that also differed between Crumb et al., 2016 (compensated) [5] and Li et al., 2018 (did not compensate) [6] (Table 1). Therefore, the decision to apply Rs compensation or not during voltage clamp experiments to characterize drug effects may also be a contributing factor inter-laboratory data variability. Effects of recording temperature on verapamil and methadone inhibition of CaV1.2 current The preceding sections focused on the Ca2+ and Ba2+ current characteristics at near PT. The next two sections focus on the impact of specific experimental factors on CaV1.2 channel pharmacology. The effect of recording temperature on CaV1.2 channel block was examined for verapamil and methadone. Fig 4A and 4C show representative recordings of Ca2+ current obtained in the control solution and following verapamil application at near PT and RT, respectively. Fig 4B and 4D show concentration-inhibition plots of verapamil for ICa-step and ICa-ramp at these temperatures. At near PT, the IC50 of verapamil inhibition of ICa-step was 0.9 μM. This is 2.5X the IC50 of ICa-ramp, which was 0.4 μM. The more potent inhibition at ICa-ramp suggests continued block development throughout the 0 and the +30 mV steps that led to fewer channels available to reactivate during the repolarizing ramp. Since Cav1.2 channels enter inactivation following activation by the 0 mV step, a more potent inhibition of the ramp current relative to the step current using the present voltage protocol suggests inactivated state block in addition to the open state block (inferred by inhibition of the step current). At RT, the IC50s of verapamil for ICa-step and ICa-ramp were 1.5 and 1.6 μM, respectively. The equally potent inhibition of the step and the ramp current at RT suggests a preference for open channel block. A surprisingly finding was that the slope of the concentration-inhibition relationship was temperature-dependent. At near PT, the nH values for ICa-step and ICa-ramp were 0.4 and 0.5, respectively. At RT, these values were 1 (Fig 4B and 4D). Assuming a single binding site, a nH value of 1 as observed at RT indicates a tight coupling between verapamil binding and block of Ca2+ permeation, consistent with a direct pore blocking mechanism. On the other hand, the nH values of 0.4 and 0.5 as observed at near PT suggests a weaker coupling between verapamil binding and effect at the channel pore. This may occur if verapamil inhibition at near PT involves allosteric mechanisms (i.e., drug binding does not directly block Ca2+ permeation, but rather induces channel closure or locking in the inactivated state). Verapamil inhibition of the Ba2+ current was also studied at near PT (Fig 4E and 4F). Relative to the drug effect on the Ca2+ current at near PT, the difference between the IC50s of the step and the ramp current was more pronounced in external Ba2+. The IC50 at IBa-step was 1.5 μM, 5X the IC50 of IBa-ramp which was 0.3 μM. The faster block development at the 0 and +30 mV step in external Ba2+ could be due to verapamil interactions with either open and/or inactivated channel state, since more channels remained open at these voltages due to removal of Ca2+-dependent inactivation. The nH values were low also, 0.3 and 0.5, respectively, similar to the drug effect on the Ca2+ current at near PT and inconsistent with a direct pore block mechanism. Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 4. CaV1.2 channel block by verapamil and methadone at near PT and RT. A) Example current traces recorded from one cell (cell ID: 19820002) in external Ca2+ at near PT. The illustrated traces were obtained in the control solution (black; trace 99), and following application of 1 μM (light red, trace 160), 3 μM (medium red, trace 209), and 100 μM verapamil (dark red; trace 250). This cell exhibited an outward current that was unmasked by 100 μM verapamil. Thus, CaV1.2 current was isolated using the verapamil-subtraction method. B) Concentration-inhibition plots for ICa-step (left) and ICa-ramp (right) at near PT. Data points from individual cells were shown in open symbols (ICa-step, circle; ICa-ramp, square). Group averages (± sem) for different concentrations were shown in solid symbols plus error bars. The solid sigmoidal curve indicates the fit using the Hill equation; the dashed curves, upper and lower limit of the 95% CI of the fit. C) Example current traces recorded from one cell (cell ID: 19d31003) in external Ca2+ at RT. The illustrated traces were obtained in the control solution (black; trace 54), and following application of 1 μM (light red, trace 110) then 100 μM verapamil (dark red; trace 155). This cell also exhibited an outward current that was unmasked by 100 μM verapamil. Thus, CaV1.2 current was isolated by using verapamil-subtraction method. D) Concentration-inhibition plots for ICa-step (left) and ICa-ramp (right) at RT. E) Example current traces recorded from one cell (cell ID: 19d04004) in external Ba2+ at near PT. The illustrated traces were obtained in the control solution (black; trace 82), and following application of 0.3 μM (light red, trace 171), 3 μM (medium red, trace 228), and 100 μM verapamil (dark red; trace 286). Ipassive (gray) was calculated based on Rinput derived from trace 286. This cell had little to no outward current in the presence of verapamil, and CaV1.2 current was isolated using Ipassive-subtraction method. F) Concentration-inhibition plots for IBa-step (left) and IBa-ramp (right) at RT. G) Example current traces recorded from one cell (cell ID: 19508007) in external Ca2+ at near PT. The illustrated traces were obtained in the control solution (black; trace 115), and following application of 30 μM methadone (red, trace 196). Ipassive (gray) was calculated based on Rinput derived from trace 196. Note that in the presence of methadone, ICa-step amplitude was slightly increased while its decay was accelerated. H) Concentration-inhibition plots for ICa-step (left) and ICa-ramp (right) at near PT. For visual presentation only, individual data points at 10, 30, and 60 μM reflecting methadone’s facilitatory effect were fit with the Hill equation (light gray curve). To illustrate the inhibitory effect on ICa-step, individual data points excluding 30 and 60 μM methadone were fit with the Hill equation (dark gray curve). I) Example traces of current recorded from one cell (cell ID: NTCell_2019_08_29_0013) in external Ca2+ at RT. The illustrated traces were obtained in the control solution (black; trace 100), and following application of 30 (medium red, trace 159) and 100 μM methadone (red, trace 250). Ipassive (gray) was calculated based on Rinput derived from trace 250. J) Concentration-inhibition plots of methadone’s effects at RT. https://doi.org/10.1371/journal.pone.0276995.g004 The consequence of recording temperature was also assessed for methadone on the Ca2+ current. Fig 4G through Fig 4J show representative cells and corresponding concentration-inhibition plots at near PT and RT, respectively. At near PT, methadone exhibited complex effects on ICa-step. Collectively, facilitation was seen at 30 and 60 μM, and inhibition was seen at higher concentrations. The facilitatory effect was characterized by an increase in the peak amplitude, accelerated current decay during the 0 mV step (Fig 4G), and was specific for ICa-step (Fig 4H). The latter point is clearly illustrated in Fig 4G: as ICa-step showed a slight increase in 30 μM methadone, ICa-ramp was nearly completely inhibited (Fig 4G). Given the complex effects at ICa-step, methadone’s IC50 was estimated for ICa-ramp only, which was 5.4 μM with nH of 1.4. At RT, only inhibitory effect was observed. The IC50 for ICa-step at RT was 10.7 μM with nH of 0.4, and that for ICa-ramp was 4.6 μM with nH of 0.5. These results suggest that methadone blocks CaV1.2 channels in both open and inactivated states at RT. This conclusion is consistent with a previous study that reported IC50s of 26.6 μM for tonic or open channel block, and 7.7 μM for phasic or inactivated channel block, respectively, for methadone at ambient temperature [27]. In summary, methadone has dual effects at PT but not RT. Like verapamil, the nH value for methadone’s inhibitory effect is temperature-dependent. Dual effects of drugs on L-type Ca2+ channels in native myocytes that depend on the membrane potential have also been reported for dihydropyridine (+)-202-791 [28] and nitrendipine [29]. For nitrendipine, the concentration-facilitation plot clearly illustrated an inflection point, suggesting the existence of two binding sites with different affinities on the Ca2+ channels [29]. Methadone used in the present study is a racemic mixture. Another study has reported that the R- and S-enantiomers of the cyclin-dependent kinase inhibitor roscovitine bind to different sites on CaV1.2 channels to affect activation and inactivation separately [30]. It is tempting to reconcile the present results by proposing distinct binding sites for methadone on CaV1.2 channels that are accessible at near PT but not RT. Follow-up studies are required to test this possibility. A few drugs have been shown to exhibit temperature-dependent block on CaV1.2 channels in overexpression cells [20] or L-type Ca2+ channels in native myocytes [21]. One study showed that nitrendipine and diltiazem inhibited Ca2+ current mediated by CaV1.2, β1, α2/δ, and γ subunits more potently at RT than at 33°C [7]. Another study reported that increasing the recording temperature from 22°C to 37°C increased the block potencies of flavoxate and nifedipine on L-type Ca2+ channels by 2.2X and 7X, respectively [8]. The direction and magnitude of potency shift due to temperature is thus drug-specific. Results of verapamil and methadone from the present study further extend those in the literature, demonstrating that recording temperature is an experimental factor that impacts CaV1.2 pharmacology. Comparisons of drug effects on the Ca2+ and Ba2+ currents at near PT The effects of buprenorphine, norbuprenorphine, naloxone, and diltiazem were studied at near PT on the Ca2+ and Ba2+ currents; of naltrexone and tolterodine, on the Ca2+ current alone. Fig 5 shows the concentration-inhibition plots for these drugs. Fig 6 summarizes the IC50s and nH values for all drugs studied. For near PT recordings, verapamil, buprenorphine, naloxone, diltiazem, and tolterodine inhibited ICa-ramp more potently than ICa-step, suggesting that these drugs all have affinity for open and inactivated channels. Tolterodine showed the largest ICa-ramp vs. ICa-step IC50 difference, by a factor of 8.7, suggesting a stronger preference for the inactivated state comparing with other drugs. For norbuprenorphine, there was no difference in the IC50s for ICa-step and ICa-ramp, suggesting a preference for open channel block when Ca2+ was used as the charge carrier. Naltrexone was the only drug tested that showed higher IC50 for ICa-ramp than for ICa-step. This drug produced a dramatic concentration-dependent hyperpolarizing shift in the ramp voltage (Fig 7A), by -22 mV at 10 mM (Fig 7B), demonstrating an effect on voltage-dependence of channel gating. Using I-V generated from 14 cells, the reversal potential for Ca2+ current under the current experimental condition was estimated to be +46 mV (S5B Fig). The lower fractional inhibition of ICa-ramp relative to ICa-step thus may not indicate drug unbinding during the +30 mV step. Instead, the increased driving force through channels that are available at more hyperpolarized membrane voltages in the presence of naltrexone may also be an explanation of the higher IC50 for ICa-ramp than for ICa-step. S6 and S7 Figs provide time course plots of individual cells tested with select drugs in Ca2+ and Ba2+, respectively. Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 5. Concentration-inhibition plots for CaV1.2 channel block by buprenorphine, norbuprenorphine, naloxone, naltrexone, diltiazem, and tolterodine at near PT. Data for ICa-step are shown in circles; ICa-ramp, squares; IBa-step, upright triangles; IBa-ramp, inverted triangles. Open symbols reflect individual data points; filled symbols plus error bars, mean ± sem. The solid sigmoidal curve indicates the fit with the Hill equation; the dashed curves, upper and lower limit of the 95% CI of the fit. https://doi.org/10.1371/journal.pone.0276995.g005 Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 6. Summary of IC50 and nH values. Top. The voltage protocol used and example traces recorded at near PT from two cells, one recorded in external Ca2+ (left; cell ID: 21515003c) and the other in external Ba2+ (right; cell ID: 21617012). For the cell on the left, the illustrated traces were obtained in the control solution (black; trace 206), 0.3 μM diltiazem (light red; trace 278), 10 μM diltiazem (medium red; trace 339), and 100 μM verapamil solutions (dark red; trace 397). For the cell on the right, the illustrated traces were obtained in the control solution (black; trace 50), 0.3 μM diltiazem (light red; trace 130), 10 μM diltiazem (medium red; trace 240), and 100 μM verapamil solutions (dark red; trace 286). Ipassive traces, shown in gray, were calculated based on Rinput derived from the verapamil traces shown. The voltage protocol was overlaid on top of the current traces. Note that for the cell on the right, Ba2+ current isolation was done using Ipassive subtraction, as the outward current seen in control solution was no longer apparent in diltiazem and verapamil solutions. Bottom, summary IC50 and nH values, (mean ± 95% CI) obtained at the current regions indicated by the dotted arrows. https://doi.org/10.1371/journal.pone.0276995.g006 Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 7. Concentration-dependent shift of voltage at which ICa-ramp peaked. A) Changes in the ramp voltage vs. naltrexone concentrations. Gray open symbols indicated data points from individual cells; black filled symbols plus error bars indicate mean ± sem. B) Representative current traces obtained from 1 cell (cell ID: 20109006) in external Ca2+ at near PT. The 90th trace (black) was the last trace obtained in the control solution; the 140th trace (red) was obtained following 10 mM naltrexone application. The voltage protocol was shown below the current traces. In this cell, the ramp voltage that ICa-ramp reached peak shifted from –2 mV in the control solution to -18 mV following naltrexone application. For the naltrexone data set, the ramp voltage at which peak ICa-ramp occurred in the control solution was 0.15 ± 0.35 mV (n = 21), consistent with the cells used to generate Fig 1C, 1D and 1G. https://doi.org/10.1371/journal.pone.0276995.g007 These results demonstrate that even when drug effects were analyzed within the same cell using the same traces, different drug potencies could be obtained depending on which current region was analyzed. Similar analyses have been performed for (-)-menthol and nimodipine [15]. In rabbit ventricular myocytes, these drugs inhibited the peak Ca2+ current evoked by a step depolarization less potently than the late Ca2+ current that remained at the end of the step depolarization, with nimodipine showing 13.1X difference in the IC50s. The difference in IC50s obtained for the step and the ramp current are compatible with literature findings of drugs exhibiting state- and/or use-dependent block of CaV1.2 channels in overexpression cell lines and L-type Ca2+ channels in native myocytes. Nitrendipine [13, 17], nisoldipine, nicardipine [17], nifedipine, verapamil [16], and mibefradil [14] all showed more block when the cells were held at depolarized membrane potential than when cells were held at hyperpolarized membrane potential, suggesting preferential block of channels in the inactivated state. Verapamil and diltiazem showed block increasing at higher stimulation frequencies and higher depolarizations, suggesting a preference for the open and inactivated state over closed state [9–11, 31]. Figs 5 and 6 also show data summarizing inhibition of the Ba2+ current by verapamil, buprenorphine, norbuprenorphine, naloxone, and diltiazem. All drugs showed greater inhibition of IBa-ramp than IBa-step (Figs 5 and 6) The largest difference was seen for verapamil, for which IC50s between the step and the ramp current differed by a factor of 5. In external Ba2+, the differences between the IC50s of the step and the ramp current were more pronounced than in external Ca2+ for verapamil, buprenorphine, norbuprenorpine, and naloxone. Even norbuprenorphine, which showed no difference between the IC50s of ICa-step and ICa-ramp, showed a difference of 2.4X between IBa-step and IBa-ramp. These results thus demonstrate that state-dependent interactions between drugs and CaV1.2 channels, as well as relative preference for different channel states, are dependent on the charge carrier used. Generalizations regarding drug-CaV1.2 channel interactions based on different studies should thus take the charge carrier used into consideration. Comparisons of the IC50s on the same current region obtained in external Ba2+ and Ca2+ showed a small impact on the step current. Based on point estimate comparison, diltiazem showed the biggest difference, with an IC50 at ICa-step 2.1X higher than that at IBa-step. Likewise, the impact of charge carrier on the ramp current and the direction of shift were also drug-specific. Verapamil showed no difference, while naloxone showed a 5X difference, with IBa-ramp being more sensitive to block than ICa-ramp. Results of the present study thus add to those in the literature demonstrating an impact of charge carrier on drug inhibition of Ca2+ and Ba2+ currents. In ventricular myocytes, diltiazem [11], D600 (a methoxy derivative of verapamil [11]), and verapamil [12] were less effective in inhibiting Ba2+ current than Ca2+ current through Ca2+ channels [11]. Similar findings were reported for verapamil [10] and diltiazem [9] studied using CaV1.2 channels with β1b and α2δ subunits using overexpression cells. Data in Figs 5 and 6 also show that the nH values were quite variable amongst the drugs, ranging from very low (nH = 0.2 for naloxone on IBa-step) to quite steep (nH = 1.4 for methadone on ICa-ramp). Shallow slopes of the concentration-inhibition relationship may be reflective of technical challenges in the experiments. These challenges include current rundown that can inflate estimation of fractional inhibition at low drug concentration, and insolubility at high drug concentration that leads to fewer than expected free drug molecules to block ion channels. While these possibilities cannot be ruled out for tolterodine, they cannot explain data of other drugs. For verapamil and methadone, changing the recording temperature greatly altered the steepness of the concentration-inhibition curves, with nH values at one temperature being 0.4 to 0.6 and at the other temperature being 1 and above. For naltrexone and diltiazem, nH values differ between the step and the ramp current, with that for one current approaching 1 while the other remaining at or below 0.6. For naloxone and naltrexone, this laboratory has previous tested similar concentrations on NaV1.5 channels and obtained nH values equal to or greater than 0.8 on the peak and the late current [18]. Therefore, the simplest explanation is that for these drugs, the shallowness of the concentration-inhibition relationship reflects more complex drug-CaV1.2 channel interactions that cannot be readily explained by 1:1 drug-receptor binding scheme that leads to immediate current inhibition. It is difficult to relate these nH results to existing literature, since nH values for concentration-inhibition plots are often not reported. In a study that assessed the structural basis of diltiazem block of voltage-gated Ca2+ channels, the resting state block was found to have an IC50 of 41 μM and a much steeper slope for the concentration-inhibition relationship than use-dependent block, which had a shallower slope but more potent block with an IC50 of 10.4 μM [32]. These results therefore demonstrate that nH values for diltiazem are state-dependent, consistent with the present findings. Based on X-ray crystallographic analysis, the study showed that diltiazem has two distinct binding poses with the Ca2+ channels: upon entering the channel pore, this drug forms a loose channel-blocking complex that appears to be a low affinity binding mode, and then rearranges within the channel to a tighter binding, more stably blocking complex with diltiazem projecting into the selectivity filter from the central cavity upon voltage-dependent inactivation [32]. Importantly, diltiazem binding also allosterically modulates Ca2+ binding in the selectivity filter, suggesting this mechanism may also contribute to reduction of current. Therefore, for a “pore blocker” like diltiazem, the different binding poses associated with different channel states may explain different nH values. Table 2 summarizes the IC50 values reported in the literature for methadone, diltiazem, verapamil, and tolterodine as well as experimental protocol used. A wide range of IC50s were reported, with the max-to-min ratios for diltiazem being 397X (0.63 μM vs. 250 μM) and for verapamil 294X (0.16 μM vs. 47 μM). The only study that reported CaV1.2 channel block by buprenorphine, norbuprenorphine, naltrexone, and naloxone was by this laboratory [18]. In Tran et al., 2020, the IC50 values of drug block on the ramp current measured using Ca2+ as the charge carrier was derived from the same cells used in the present study. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 2. Comparisons of IC50 values for CaV1.2 channel block for methadone, diltiazem, and verapamil generated by different experimental protocols. https://doi.org/10.1371/journal.pone.0276995.t002 Limitation, lessons learned, protocol standardization, and conclusion There are several limitations in the present study that can impact drug potency estimation. The first is that the drug concentrations exposed to the recorded cell were not measured using an analytical method. Drug concentrations can deviate from the target concentration due to compound-specific factors and human errors. The former include nonspecific binding to the plastic and glass substrates within the patch clamp perfusion apparatus, potential insolubility at higher concentrations tested, and instability in the perfusion solution under the experimental condition. Notably, verapamil and diltiazem are compounds stated to be light sensitive (https://www.sigmaaldrich.com/US/en/sds/sigma/v4629) and advised to keep away from direct sunlight (https://documents.tocris.com/pdfs/tocris_msds/0685_sds.pdf?1646647276&_ga=2.29634103.924187291.1646647248-600272910.1596464753), respectively, on the Safety Data Sheet from their distributors. While verapamil solutions and stocks were protected from light in this laboratory, diltiazem was not, and time-dependent degradation for these as well as other compounds tested throughout the recording day cannot be ruled out. Human errors can also occur during drug stock and drug solution preparation. Concentration verification, if possible, should be included as a part of the study design to rule out the possibility that deviations of drug concentrations translate into variability in drug potency estimation. The second limitation is that CaV1.2 current rundown under whole cell configuration could not be prevented, and current rundown was not corrected for drug potency estimation in this study. Although drugs were applied after the initial fast phase of rundown had ended, the IC50s may have still been underestimated. Given cell-specific rundown profile, if rundown correction were to be implemented, then accounting for individual cell’s rundown time course by fitting as many data points obtained in the control solution as possible, rather than using the time course plots derived from separate cells is recommended. Of note, although recordings in perforated patch configuration decrease cell dialysis-dependent rundown, this technique leads to higher Rs compared to whole cell recordings. Current rundown was a trade-off for voltage control in the present study, and the method to optimize voltage control was to use whole cell recording to obtain as small of Rs as possible followed by high degree of compensation. The third limitation is that Rs was not measured throughout the recordings. Not having this measure across time raises a logical concern that rundown of Ba2+ and Ca2+ current is due to large Rs changes. In this study, Ca2+ and Ba2+ currents recorded at near PT activate extremely fast, and large changes in Rs manifest as slowing of time-to-peak for the step current (for Ba2+ and Ca2+ currents) and shift in the voltage at which ramp peak occurred (for the Ba2+ current; Fig 3). Therefore, whether Rs changed dramatically or not during the control recording period was based on offline assessment of these current profiles. Progressive decreases in the current amplitude not accompanied by kinetic changes seem incompatible with the conjecture that rundown is secondary to large changes in Rs. As the original electrophysiology records are available at https://osf.io/g3msb/, interested readers are encouraged to assess these files to draw independent conclusions regarding the mechanisms subserving rundown. The fourth limitation is that the reversal potentials of Ca2+ and Ba2+ currents were extrapolated from currents obtained between 0 and +20 mV steps (S5 Fig). These I-V relations were generated to assess adequacy of voltage control, inferred from graded increases in the current amplitudes to increasing voltages between -60 to -20 mV. For the purpose of measuring the reversal potential, extending the voltage steps to beyond the reversal potential would provide direct measurement for each cell. The fifth limitation is the uncertainty that the recorded CaV1.2 current indeed reflects activity of channels with β2 and α2δ1 auxiliary subunits. The gray traces in Fig 1G and 1H show that the ratios of ramp-to-step current are quite variable across cells. Likewise, temperature sensitivity of the whole cell current was also quite different across cells (Fig 2E and 2J). Since auxiliary subunits modulate CaV1.2 channel gating, it is possible that recordings in this study were from heterogenous CaV1.2 channels with either one or both auxiliary subunits or channels in different states of phosphorylation. A few lessons were learned by conducting the present study. First, data variability of CaV1.2 channel block was collectively larger than those observed for cardiac hERG and NaV1.5 channel block based on this laboratory’s concurrent work. The concentration-inhibition plots presented in Figs 4 and 5 show variable degrees of current inhibition for individual cells to a given drug concentration, with no clear outliers observed. This level of data spread was not seen for hERG and NaV1.5 current inhibition by the same drugs [18]. Within-the-study data variability for CaV1.2 current inhibition may be due to variable degrees of current rundown and heterogenous coupling between the channel α subunit with β2 and α2δ1 subunits, as the latter can also be targets of drugs. Second, state-dependent block of drugs on CaV1.2 channels are common, inferred from the different IC50s obtained for the step and the ramp current. Understanding state-dependent block using voltage protocols that recapitulate cardiac AP may be important when trying to predict drug impact on cardiac electrophysiology. In the sinoatrial nodal cells, CaV1.2 channels are activated rapidly upon membrane depolarization and contribute to the upstroke of the AP. Drug block of open CaV1.2 channels (i.e., inhibition of the step current) may thus inform the potential of drug in affecting the heart rate and the PR interval. On the other hand, during a ventricular AP, CaV1.2 channels are activated by the abrupt depolarization from rest, enter inactivated state during the plateau potential, and then become reactivated during the repolarizing phase of the AP before entering closed state. Information regarding how CaV1.2 channel block develops during a ventricular AP, as well as drug effect on voltage-dependence of channel gating may inform potential change in AP shape (i.e., triangulation or simply shortening), thereby allowing better prediction of proarrhythmia risk. Some statistical [33] and in silico myocyte models have incorporated drug block of CaV1.2 channels to assess the risk of Torsade de Pointes imposed by hERG channel block [34]. Incorporating IC50s measured at distinct current regions or accounting for state-dependent block of CaV1.2 channels may lead to better model performance. The third lessons learned is that many drugs have nH values much smaller than 1, and these values are dependent on the recording temperature and channel state (Fig 6). Once the contributions of current rundown and insolubility at higher tested concentrations to shallow concentration-inhibition graphs are ruled out, the most straightforward interpretation of these results is that drug-CaV1.2 channel interactions leading to current inhibition are complex processes that may involve multiple binding poses (i.e., diltiazem), multiple binding sites (i.e., nitrendipine, roscovitine), or through allosteric mechanisms (i.e., diltiazem; naltrexone, Fig 7). The present results showed that even with the same voltage protocol presented at the same stimulation rate, Cav1.2 pharmacology can still be sensitive to a variety of factors encountered during the experiments and during data analysis. When numerous experimental factors are different between two studies, as seen Crumb et al., 2016 [5] and Li et al., 2018 [6], drug potency estimates can be very different even for the same drugs. For CaV1.2 data intended to support risk prediction or clinical interpretation, normalizing laboratory-specific practices is essential toward promoting data reproducibility across laboratories—a pivotal step toward engendering confidence amongst regulators for applying these in vitro data in the decision-making process. Toward this end, the FDA Cardiac Safety Studies Interdisciplinary Review Team (CSS-IRT) has posted a document regarding the recommended voltage protocols for cardiac ion channels, including CaV1.2 channels, on its website (https://www.fda.gov/media/151418/download). The voltage waveform, stimulation frequency, compositions of the internal and Ca2+-based external solutions, and data analysis method are consistent with those used in the present study. For cardiac safety assessment, drug developers and regulators are following the guidelines released by the International Council for Harmonisation for Technical Requirements for Pharmaceuticals for Human Use: ICH S7B for nonclinical [35] and ICH E14 for clinical studies [36]. The newly released ICH E14/S7B Questions and Answers guideline offers best practice recommendations for patch clamp ion channel studies intended to support cardiac safety assessment [37] (ICH S7B Q&A 2.1), and the protocol used in this manuscript is consistent with these recommendations. ICH S7B Q&A 2.1 also recommends recording at near PT. This study tested two temperature controller models. Given the gravity-fed perfusion method and shallow bath chambers used here, the temperature controller model that heats the chamber bottom uniformly provided more stable temperature control. However, if a perfusion pump were used to maintain flow rate, then conceivably the temperature controller model that maintained bath temperature by heating the anodized aluminum platform would have also achieved stable temperature control. Experimenters interested in measuring drug block of CaV1.2 channels at near PT are recommended to consider how bath temperature may fluctuate given the rig design, and importantly measure bath temperature near the recorded cell throughout the recordings to enable subsequent analysis of temperature fluctuations on within-the-study data variability. In conclusion, results from the present study offer rationale for the best practice recommendations regarding experimental design, conduct, and data quality consideration, and may benefit stakeholders considering utilizing CaV1.2 channel data to support regulatory decision-making. This table summarizes the IC50 and nH values for CaV1.2 channel block determined using the manual patch clamp method [5] and automated patch clamp system [6]. While these studies examined more overlapping drugs, only those for which Crumb et al., 2016 provided IC50 and nH values are shown for comparison. The ratios are calculated as maximum vs. minimum IC50s. Crumb et al., 2016 used a CaV1.2-CHO cell line from Cytocentrics Bioscience GmbH (Rostock, Germany), and did not provide information regarding subunits expressed. Experiments were conducted using whole cell patch clamp method at 36 ± 1ºC, and current was evoked using a rabbit ventricular AP waveform repeated at 10 s interval. Ba2+ (4 mM) was used as the charge carrier. The automated patch clamp data from Li et al., 2018 were generated using a CHO cell line that expressed hCaV1.2α, β2, α2δ1 subunits from Charles River Laboratories (Wilmington, MA). Recordings were performed using IonWorks Barracuda system operating in population perforated patch clamp mode. Recording temperature was not controlled and was expected to be higher than RT due to the heat produced during system operation. Inward current was evoked using the same “step-step-ramp” voltage waveform as used in the present study but repeated at 10 s interval. The current elicited by the 0 mV step was used to quantify drug effects. Ca2+ (6.8 mM) was used as the charge carrier. Compositions of the external recording solution were the same for both studies except for the charge carrier. For internal solution, Crumb et al., 2016 used 130 mM CsCl as the main salt, while Li et al., 2018 used 90 mM CsF + 50 mM CsCl. Current mediated by CaV1.2 channels exhibit prominent rundown when recorded under the whole cell configuration. Percent current inhibition by drug reported by Li et al. 2018 was adjusted for current run down, using data from vehicle and positive control wells, even though recordings were obtained using perforated patch mode. Crumb et al., 2016 did not correct for current rundown nor specified the rate of rundown for the cells used. The predicted logP values for these drugs based on ChemAxon are provided as estimates of lipophilicity. The sources are as follows: bepridil (https://go.drugbank.com/drugs/DB01244), chlorpromazine (https://go.drugbank.com/drugs/DB00477), diltiazem (https://go.drugbank.com/drugs/DB00343), ondansetron (https://go.drugbank.com/drugs/DB00904), terfenadine (https://go.drugbank.com/drugs/DB00342), and verapamil (https://go.drugbank.com/drugs/DB00661). There is no relationship between the ratio of max-to-min IC50s and logP values for these drugs. Rundown of CaV1.2 current in external Ca2+ or Ba2+ at near PT Fig 1A and 1B show CaV1.2 current recorded in external Ca2+ and Ba2+, respectively. The maximal peak current evoked by the 0 mV step is referred to as ICa-step or IBa-step depending on the charge carrier used; the maximal ramp current evoked by the repolarizing ramp, ICa-ramp or IBa-ramp. Fig 1C–1F show the time course plots of normalized current amplitudes recorded in the control solution and following verapamil (100 μM) application. After whole cell formation, CaV1.2 current showed prominent rundown regardless of which charge carrier was used, and in most cells eventually reached a quasi-steady state level. Current rundown was seen in every cell and was characterized by a fast phase followed by a slower phase. The normalized ICa-step at the 150th trace was 0.42 ± 0.04 relative to the 1st trace (Fig 1C); the normalized ICa-ramp, 0.44 ± 0.06 (Fig 1D). The normalized IBa-step was 0.34 ± 0.04 (Fig 1E); the normalized IBa-ramp, 0.32 ± 0.04 (Fig 1F). To assess whether the extent of rundown was different between the Ca2+ and the Ba2+ current, a basic statistical analysis of the differences in the amplitude of the 150th trace relative to the first trace was performed. A t-test with equal variance revealed no significant difference between ICa-step and IBa-step and between ICa-ramp and IBa-ramp (p > 0.05). Fig 1G and 1H show the time course of the ratios of ramp-to-step current for traces acquired in the control solution. For Ca2+ current, the averaged ratio was 0.33 ± 0.03; for Ba2+ current, 0.60 ± 0.05. The larger ratio obtained in Ba2+ is consistent with the removal of Ca2+-dependent inactivation [19, 20]. These ratios remained constant throughout the control recording despite current rundown, suggesting that rundown reflects a progressive loss in the available channels. Rundown assessed using the present voltage protocol at 0.2 Hz was not attributed to intracellular Ca2+ accumulation, since it occurred to a similar degree in external Ba2+. Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 1. Rundown of CaV1.2 channel activity at near PT. A, B. Top, representative current traces from 2 cells recorded in either external Ca2+ (A) or Ba2+ (B). Cell ID for (A) was 19327000; for (B), 19404005. Black traces reflect recordings obtained in the control solution, and the 1st, 10th, 90th, and 100th traces recorded after ~2 minutes of whole cell dialysis are shown. Red traces reflect recordings obtained following application of 100 μM verapamil, after steady state inhibition was achieved. Ipassive traces, shown in gray, were calculated using Rinput derived from the verapamil traces. Bottom, the voltage protocol used. C, D) Summary time course plots of normalized ICa-step (C) and ICa-ramp (D) in the control solution and following verapamil application (n = 16). Data points are shown as mean ± sem. Verapamil application did not start at the same time for every cell. Therefore, the X-axes show a break after trace 150 to synchronize the data points obtained following verapamil application. E, F) Summary time course plots of normalized IBa-step (E) and IBa-ramp (F) in the control solution and following verapamil perfusion (n = 17). G, H) Ratio of ICa-ramp-to-ICa-step (G) or IBa-ramp-to-IBa-step (H) for traces acquired in the control solution. Data for individual cells are shown as light gray lines. Mean ± sem are shown as black symbols plus error bars. https://doi.org/10.1371/journal.pone.0276995.g001 Application of verapamil nearly eliminated the ramp current (Fig 1D and 1F) but not the step current (Fig 1C and 1E). Relative to the 150th trace recorded in the control solution, the residual ICa-step in verapamil was 20 ± 2%, while that of ICa-ramp was 4 ± 1%. Likewise, the residual IBa-step in verapamil was 32 ± 4%, while that of IBa-ramp was 2 ± 0%. Therefore, amplitude of the step current was always quantified using Ipassive-subtracted traces, not verapamil-subtracted traces (see “Methods”). Rundown of CaV1.2 current in whole cell configuration is a widely observed phenomenon. Since rundown leads to overestimation of fractional inhibition, laboratory-specific tolerance regarding the rate of rundown and practices to correct rundown can introduce variable degrees of imprecision into drug potency estimation. The diverse practices are exemplified by the two publications that motivated the present study. Crumb et al., 2016 did not correct for rundown, stated that cells with >20% rundown were discarded, but did not define how many current traces the 20% calculation was based [5]. Li et al., 2018 used perforated population recording that presumably reduced rundown, yet still corrected for rundown in the calculation of drug inhibition [6]. Even if the rates of CaV1.2 current rundown were similar between the two publications, these different practices alone would lead to different drug potency estimations. Experience from this laboratory indicates that the rate of rundown, and whether current can reach a quasi-steady state in the control solution for pharmacology experiments are functions of the cell line used and cell culture conditions, and therefore unlikely to be the same across studies. Using the same Ca2+-external solution, internal solution, and voltage protocol, two additional CaV1.2 cell lines that expressed the same channel subunits were tested, and both cell lines showed near complete loss of CaV1.2 current at near PT with time, without reaching a quasi-steady state level (S2 Fig). Can CaV1.2 current rundown under whole cell mode be prevented? Previous studies suggest that rundown of cardiac L-type Ca2+ channel activity reflects channel dephosphorylation. This conclusion is based on evidence from native myocytes that manipulating protein kinase A (PKA)-mediated phosphorylation [21], protein phosphatase activity [21], and increasing the intracellular level of ATP or cyclic AMP—molecules that enhance PKA-mediated phosphorylation [22] all led to expected changes in Ca2+ channel activity. In the present study, inclusion of 5 mM MgATP and 0.4 mM TrisGTP in the internal solution did not prevent current rundown in the three cell lines tested. Therefore, rundown in whole cell configuration could not be prevented by simply supplying ATP. Temperature sensitivity of CaV1.2 current During near PT recordings, the step current sometimes showed amplitude fluctuations that occurred without changes in the passive membrane properties (i.e., Rinput or I-80 mV; S3 Fig). As these amplitude fluctuations were not observed with RT recordings, one hypothesis is that they are associated with temperature fluctuations during the near PT recordings. Therefore, temperature sensitivity of the CaV1.2 current was examined. These experiments were conducted using setups with the TC-344C controller, as amplitude fluctuations were more common when recorded using these setups, and the thermistor measuring the bath temperature was positioned as close to the recorded cell as possible. After the current reached a quasi-steady state level at near PT, temperature control of the aluminum platform was turned off to allow graded bath temperature drop at the recording chamber, and then back on to elevate the bath temperature. Fig 2A shows the time course plots of ICa-step and ICa-ramp from a representative cell. The boxed regions are expanded and shown in Fig 2B. ICa-step decreased and increased as the bath temperature lowered and elevated, respectively (top; note that Ca2+ current amplitudes are expressed as negative values), while ICa-ramp appeared to show the opposite pattern (bottom). No change in Rinput or I-80 mV was observed (Fig 2C). The left panel of Fig 2D shows the plot of ICa-step vs. temperature for this cell, demonstrating a strong linear relationship (r = -0.94). In contrast, no relationship was observed between ICa-ramp and temperature (r = 0.38; Fig 2D, right). Temperature sensitivity of the Ca2+ current was confirmed in 5 more cells (S4A–S4E Fig). In all cells recorded, ICa-step showed strong linear relationship with temperature (r = -0.77 to -0.94). The ICa-ramp-temperature relationship was inconsistent: 4 cells showed no relationship (Fig 2D, right; S4A, S4B and S4D Fig., right), and 2 showed linear relationship (r = -0.73 to -0.88). Fig 2E shows the calculated and normalized ICa-step at 37°C and 34°C. Depending on the cell, 3°C of temperature drop reduced ICa-step amplitude by 12% to 34%. This magnitude of temperature fluctuation was routinely observed during near PT experiments for setups with TC-344 controller. Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 2. Effects of recording temperature on CaV1.2 channel activity. Cell ID for (A) through (D) was 19318007. Recordings were obtained in external Ca2+. A) Time course plots of ICa-step (open circle) and ICa-ramp (open square) for the entire experiment. The boxed regions were expanded and shown in (B). B) Top, time course plots of ICa-step from traces 96 to 216 and bath temperature (gray dotted line). Current amplitudes were corrected for rundown. This could be done since rundown correction was performed on the traces used for the fit. To estimate rundown, data points in (B) were fit with a linear function to yield a slope of 0.00495 nA/trace. This amount of current loss was then added back to the original ICa-step amplitudes. Bottom, time course plots of ICa-ramp and bath temperature. Current amplitudes were also corrected for run rundown (slope = 0.00146 nA/trace). C) Rinput (top) and I-80 mV (bottom). D) Left, rundown-corrected ICa-step vs. bath temperature. These data points were fit with a linear function, yielding a slope of -0.295 nA/°C and a Y-intercept of 7.299 nA (r = -0.94). Right, rundown-corrected ICa-ramp vs. bath temperature (r = 0.38). E) Calculated and normalized ICa-step at 37°C (black circle) and 34°C (gray circle). Cell ID for (F) to (J) was 07_07_0008. Recordings were obtained in external Ba2+. F) Time course plots of IBa-step (open circle) and IBa-ramp (open square) for the entire experiment. The boxed region was expanded and shown in (G). G) Time course plots of IBa-step (top) and IBa-ramp (bottom) from traces 60 to 200 and bath temperature (gray dotted line). Both IBa-step and IBa-ramp were corrected for rundown, using 0.00741 nA/trace and 0.00538 nA/trace, respectively. H) Rinput (top) and I-80 mV (bottom). I) Rundown-corrected IBa-step (left) or IBa-ramp (right) vs. bath temperature. Fitting IBa-step vs. temperature plot with a linear function yielded a slope of -0.147 nA/°C and a Y-intercept of 2.947 nA (r = -0.88). Rundown-corrected IBa-ramp vs. temperature plot (r = -0.36). J) Calculated and normalized IBa-step at 37°C (black circle) and 34°C (gray circle). https://doi.org/10.1371/journal.pone.0276995.g002 Temperature sensitivity of the Ba2+ current was examined as well. Fig 2F shows the time course plots of IBa-step and IBa-ramp from a representative cell, and the boxed region was expanded and shown in Fig 2G. IBa-step changed in the same direction as the bath temperature (top), and IBa-ramp seemed to not respond to temperature changes (bottom). Rinput and I-80 mV remained stable during temperature manipulations (Fig 2H). The left panel of Fig 2I shows the plot of IBa-step vs. temperature, demonstrating a clear linear relationship (r = -0.88). The plot of IBa-ramp vs. temperature, on the other hand, did not reveal any relationship (r = -0.36; Fig 2I, right). Temperature-dependent effect on the Ba2+ was confirmed in 4 more cells (S4F–S4I Fig). In all cells recorded, IBa-step showed strong linear relationship with temperature (r = -0.87 to -0.93). The data for IBa-ramp and temperature were inconsistent: 3 cell showed no-to-weak relationship (Fig 2I, right; S4F and S4G Fig, right), and 2 showed strong relationship (r = -0.80 and -0.88). Fig 2J shows the calculated and normalized IBa-step at 37°C and 34°C. Depending on the cell, 3°C temperature drop reduced IBa-step by 18% to 28%. The step current in the present study is very temperature-sensitive regardless of which charge carrier was used. These results are consistent with prior studies conducted on cloned CaV1.2 channels comprised of α1c, α2/δa, and β1b (or β2c) expressed in xenopus oocytes [23] and L-type Ca2+ channels in native ventricular myocytes [24, 25] that also showed high temperature sensitivity. Increasing PKA-mediated phosphorylation reduced temperature sensitivity of some Ca2+ channel gating parameters, suggesting that the high temperature sensitivity may be due to channels in the unphosphorylated state [25]. Temperature sensitivity of the step current adds an additional challenge to conducting pharmacology experiments at near PT, since several degrees of temperature fluctuations can easily be produced by a slowing of the flow rate due to the presence of bubbles in the perfusion line and/or reduced fluid level in the reservoirs of the gravity-fed perfusion system. Temperature sensitivity of the CaV1.2 current is thus a source of variability for pharmacology data generated within and across laboratories. Consequence of Rs compensation on CaV1.2 current at near PT The cells used to generate Fig 1 were used to estimate the activation kinetics and amplitude of CaV1.2 current at near PT. ICa-step reached peak 1.84 ± 0.17 ms following the voltage jump to 0 mV (range: 1.34 to 3.39 ms; n = 12, using cells with clear separation between capacitive transient and the step current), and was -2210.1 ± 178.3 pA in amplitude (range: -1376.7 to -3362.9 pA, based the average of 91st to 100th traces recorded in the control solution). ICa-ramp reached peak at 0.6 ± 0.6 mV during the repolarizing ramp (range: -4.0 to 3.6 mV; n = 16), and was -814.0 ± 108.4 pA in amplitude (range: -320.2 to -1546.3 pA). The fast kinetics and relatively large amplitudes of the step current prompted for a set of experiments to assess the impact of Rs compensation in these recordings. During these experiments, Rs was measured and compensated at 80% before the start of recording. After CaV1.2 current reached a quasi-steady state level in the control solution, Rs compensation was turned “off” and “on” as the recording continued. Fig 3A and 3B show representative current traces and time course plots of ICa-step and ICa-ramp from one cell obtained with and without Rs compensation. The total initial Rs of this cell was 5.6 MΩ. With Rs compensation, ICa-step was ~600 pA larger than without Rs compensation (Fig 3B, top). In contrast, no difference was observed in ICa-ramp amplitude (bottom) or the ramp voltage at which ICa-ramp reached peak (Fig 3C, top). Rinput and I-80 mV remained stable (Fig 3C, middle and bottom). The impact of Rs compensation was confirmed in 5 additional cells. The left panel of Fig 3D shows the fractional ICa-step loss without Rs compensation for all 6 cells, plotted against either ICa-step with Rs compensation (solid symbols) or the amount of Rs that was compensated (or 80% total Rs; open symbols). No relationship was evident amongst these parameters, as expected since voltage loss through Rs (hence fractional current loss) is a product of both the current amplitude and Rs. For these cells, voltage loss through Rs at ICa-step ranged from 4.7 to 21.3 mV, calculated by using Ohm’s law. The right panel of Fig 3D shows the fractional ICa-ramp loss due to not compensating for Rs, plotted as a function of ICa-ramp with Rs compensation. Collectively, no loss was observed (average: -0.02 ± 0.02). Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 3. Effects of Rs compensation on Ca2+ and Ba2+ currents. Data shown in panels (A) through (D) were obtained in external Ca2+. Cell ID was 21318008 for panels (A) through (D). A) Top, representative Ca2+ current traces obtained from a cell, with (black traces) and without Rs compensation (gray traces). Bottom, the voltage protocol used. B) Time course plots of ICa-step (top) and ICa-ramp (bottom), focusing on traces 30 to 120 for which Rs compensation was turned on and off. No rundown correction was made for these plots. C) Time course plots of ramp voltage at which ICa-ramp reached peak amplitude (top), Rinput (middle), and I-80 mV (bottom) for traces 30 to 120. D) Left, fractional ICa-step loss due to not compensating for Rs, calculated as the ratio of ICa-step without Rs compensation vs. ICa-step with 80% Rs compensation, and plotted against ICa-step obtained with Rs compensation (lower x-axis) or the amount of Rs compensated (upper x-axis). Right, fractional ICa-ramp loss due to not compensating for Rs, plotted against ICa-ramp with Rs compensation. Data shown in panels (E) through (H) were obtained in external Ba2+. Cell ID was 2021_06_25_0004 for panels (E) through (I). E) Representative Ba2+ current traces obtained from a cell, with (black traces) and without Rs compensation (gray traces). F) Time course plots of IBa-step (top) and IBa-ramp (bottom), focusing on traces 40 to 100 during which Rs compensation was turned on and off. No rundown correction was made for these plots. G) Time course plots of ramp voltage (top), Rinput (middle), and I-80 mV (bottom) for traces 40 to 100. H) Left, fractional IBa-step loss due to not compensating for Rs, plotted against IBa-step obtained with Rs compensation (lower x-axis) or the amount of Rs compensated (upper x-axis). Right, delta (Δ) ramp voltage shift for IBa-ramp due to not compensating for Rs, plotted against IBa-ramp with Rs compensation. https://doi.org/10.1371/journal.pone.0276995.g003 The kinetics and amplitude of Ba2+ current at near PT was also quantified. IBa-step reached peak 2.91 ± 0.25 ms following the voltage jump to 0 mV (range: 1.64 to 4.23 ms; n = 11), and was -4147.6 ± 813.1 pA in amplitude (range: -1420.0 to -8395.0 pA). IBa-ramp reached peak at -7.7 ± 0.7 mV (range: -2.5 to -10.5 mV; n = 11) and was -2077.7 ± 309.5 pA in amplitude (range: -555.2 to -3506.4 pA). Fig 3E showed representative Ba2+ current traces obtained from a cell, with and without Rs compensation. The total Rs for this cell was 4.1 MΩ. Similar to Ca2+ current recordings, Rs compensation affected the amplitude of IBa-step (by ~1500 pA; Fig 3F, top) but not the amplitude of IBa-ramp (Fig 3F, bottom). However, the ramp voltage at which IBa-ramp peaked was consistently shifted rightward when Rs was not compensated, toward the more hyperpolarized potential (Fig 3G, top). No change in Rinput or I-80 mV was observed (Fig 3G, middle and bottom). The impact of Rs compensation on the Ba2+ current was verified in 5 more cells. The left panel of Fig 3H shows the fractional IBa-step loss due to not compensating for Rs, plotted against IBa-step amplitude with Rs compensation (solid symbols) or the amount of Rs compensated (open symbols). No relationship was evident amongst these parameters. Voltage loss through Rs at IBa-step ranged from 4.8 to 28.2 mV. The right panel of Fig 3H summarizes the hyperpolarizing shift of the ramp voltage for IBa-ramp without Rs compensation. On average, the shift was -7.7 ± 1.3 mV (range: -4.6 to -12.1 mV) and was not accompanied by IBa-ramp amplitude change (fractional IBa-ramp loss: 0.03 ± 0.00). Using normalized I-V relation generated from 15 cells, the reversal potential of Ba2+ current was estimated to be +35 mV (S5A Fig). The increase in Ba2+ driving force due to the hyperpolarizing shift in the ramp voltage offers an explanation as to why IBa-ramp amplitude was unaltered by not compensating for Rs, since this could compensate for the fewer number of CaV1.2 channels that would be open due to not clamping the membrane potential at the expected levels. These results demonstrate the impact of Rs compensation when recording fast and large Ca2+ and Ba2+ currents at near PT. While Rs compensation is a good practice for voltage clamp experiments, can Rs affect pharmacology results? A recent study that modeled drug block of Na+ channels indicates so, with the degree of rightward shift of the concentration-inhibition graphs dependent on the magnitudes of Rs and Na+ current [26]. Rs compensation is a practice that also differed between Crumb et al., 2016 (compensated) [5] and Li et al., 2018 (did not compensate) [6] (Table 1). Therefore, the decision to apply Rs compensation or not during voltage clamp experiments to characterize drug effects may also be a contributing factor inter-laboratory data variability. Effects of recording temperature on verapamil and methadone inhibition of CaV1.2 current The preceding sections focused on the Ca2+ and Ba2+ current characteristics at near PT. The next two sections focus on the impact of specific experimental factors on CaV1.2 channel pharmacology. The effect of recording temperature on CaV1.2 channel block was examined for verapamil and methadone. Fig 4A and 4C show representative recordings of Ca2+ current obtained in the control solution and following verapamil application at near PT and RT, respectively. Fig 4B and 4D show concentration-inhibition plots of verapamil for ICa-step and ICa-ramp at these temperatures. At near PT, the IC50 of verapamil inhibition of ICa-step was 0.9 μM. This is 2.5X the IC50 of ICa-ramp, which was 0.4 μM. The more potent inhibition at ICa-ramp suggests continued block development throughout the 0 and the +30 mV steps that led to fewer channels available to reactivate during the repolarizing ramp. Since Cav1.2 channels enter inactivation following activation by the 0 mV step, a more potent inhibition of the ramp current relative to the step current using the present voltage protocol suggests inactivated state block in addition to the open state block (inferred by inhibition of the step current). At RT, the IC50s of verapamil for ICa-step and ICa-ramp were 1.5 and 1.6 μM, respectively. The equally potent inhibition of the step and the ramp current at RT suggests a preference for open channel block. A surprisingly finding was that the slope of the concentration-inhibition relationship was temperature-dependent. At near PT, the nH values for ICa-step and ICa-ramp were 0.4 and 0.5, respectively. At RT, these values were 1 (Fig 4B and 4D). Assuming a single binding site, a nH value of 1 as observed at RT indicates a tight coupling between verapamil binding and block of Ca2+ permeation, consistent with a direct pore blocking mechanism. On the other hand, the nH values of 0.4 and 0.5 as observed at near PT suggests a weaker coupling between verapamil binding and effect at the channel pore. This may occur if verapamil inhibition at near PT involves allosteric mechanisms (i.e., drug binding does not directly block Ca2+ permeation, but rather induces channel closure or locking in the inactivated state). Verapamil inhibition of the Ba2+ current was also studied at near PT (Fig 4E and 4F). Relative to the drug effect on the Ca2+ current at near PT, the difference between the IC50s of the step and the ramp current was more pronounced in external Ba2+. The IC50 at IBa-step was 1.5 μM, 5X the IC50 of IBa-ramp which was 0.3 μM. The faster block development at the 0 and +30 mV step in external Ba2+ could be due to verapamil interactions with either open and/or inactivated channel state, since more channels remained open at these voltages due to removal of Ca2+-dependent inactivation. The nH values were low also, 0.3 and 0.5, respectively, similar to the drug effect on the Ca2+ current at near PT and inconsistent with a direct pore block mechanism. Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 4. CaV1.2 channel block by verapamil and methadone at near PT and RT. A) Example current traces recorded from one cell (cell ID: 19820002) in external Ca2+ at near PT. The illustrated traces were obtained in the control solution (black; trace 99), and following application of 1 μM (light red, trace 160), 3 μM (medium red, trace 209), and 100 μM verapamil (dark red; trace 250). This cell exhibited an outward current that was unmasked by 100 μM verapamil. Thus, CaV1.2 current was isolated using the verapamil-subtraction method. B) Concentration-inhibition plots for ICa-step (left) and ICa-ramp (right) at near PT. Data points from individual cells were shown in open symbols (ICa-step, circle; ICa-ramp, square). Group averages (± sem) for different concentrations were shown in solid symbols plus error bars. The solid sigmoidal curve indicates the fit using the Hill equation; the dashed curves, upper and lower limit of the 95% CI of the fit. C) Example current traces recorded from one cell (cell ID: 19d31003) in external Ca2+ at RT. The illustrated traces were obtained in the control solution (black; trace 54), and following application of 1 μM (light red, trace 110) then 100 μM verapamil (dark red; trace 155). This cell also exhibited an outward current that was unmasked by 100 μM verapamil. Thus, CaV1.2 current was isolated by using verapamil-subtraction method. D) Concentration-inhibition plots for ICa-step (left) and ICa-ramp (right) at RT. E) Example current traces recorded from one cell (cell ID: 19d04004) in external Ba2+ at near PT. The illustrated traces were obtained in the control solution (black; trace 82), and following application of 0.3 μM (light red, trace 171), 3 μM (medium red, trace 228), and 100 μM verapamil (dark red; trace 286). Ipassive (gray) was calculated based on Rinput derived from trace 286. This cell had little to no outward current in the presence of verapamil, and CaV1.2 current was isolated using Ipassive-subtraction method. F) Concentration-inhibition plots for IBa-step (left) and IBa-ramp (right) at RT. G) Example current traces recorded from one cell (cell ID: 19508007) in external Ca2+ at near PT. The illustrated traces were obtained in the control solution (black; trace 115), and following application of 30 μM methadone (red, trace 196). Ipassive (gray) was calculated based on Rinput derived from trace 196. Note that in the presence of methadone, ICa-step amplitude was slightly increased while its decay was accelerated. H) Concentration-inhibition plots for ICa-step (left) and ICa-ramp (right) at near PT. For visual presentation only, individual data points at 10, 30, and 60 μM reflecting methadone’s facilitatory effect were fit with the Hill equation (light gray curve). To illustrate the inhibitory effect on ICa-step, individual data points excluding 30 and 60 μM methadone were fit with the Hill equation (dark gray curve). I) Example traces of current recorded from one cell (cell ID: NTCell_2019_08_29_0013) in external Ca2+ at RT. The illustrated traces were obtained in the control solution (black; trace 100), and following application of 30 (medium red, trace 159) and 100 μM methadone (red, trace 250). Ipassive (gray) was calculated based on Rinput derived from trace 250. J) Concentration-inhibition plots of methadone’s effects at RT. https://doi.org/10.1371/journal.pone.0276995.g004 The consequence of recording temperature was also assessed for methadone on the Ca2+ current. Fig 4G through Fig 4J show representative cells and corresponding concentration-inhibition plots at near PT and RT, respectively. At near PT, methadone exhibited complex effects on ICa-step. Collectively, facilitation was seen at 30 and 60 μM, and inhibition was seen at higher concentrations. The facilitatory effect was characterized by an increase in the peak amplitude, accelerated current decay during the 0 mV step (Fig 4G), and was specific for ICa-step (Fig 4H). The latter point is clearly illustrated in Fig 4G: as ICa-step showed a slight increase in 30 μM methadone, ICa-ramp was nearly completely inhibited (Fig 4G). Given the complex effects at ICa-step, methadone’s IC50 was estimated for ICa-ramp only, which was 5.4 μM with nH of 1.4. At RT, only inhibitory effect was observed. The IC50 for ICa-step at RT was 10.7 μM with nH of 0.4, and that for ICa-ramp was 4.6 μM with nH of 0.5. These results suggest that methadone blocks CaV1.2 channels in both open and inactivated states at RT. This conclusion is consistent with a previous study that reported IC50s of 26.6 μM for tonic or open channel block, and 7.7 μM for phasic or inactivated channel block, respectively, for methadone at ambient temperature [27]. In summary, methadone has dual effects at PT but not RT. Like verapamil, the nH value for methadone’s inhibitory effect is temperature-dependent. Dual effects of drugs on L-type Ca2+ channels in native myocytes that depend on the membrane potential have also been reported for dihydropyridine (+)-202-791 [28] and nitrendipine [29]. For nitrendipine, the concentration-facilitation plot clearly illustrated an inflection point, suggesting the existence of two binding sites with different affinities on the Ca2+ channels [29]. Methadone used in the present study is a racemic mixture. Another study has reported that the R- and S-enantiomers of the cyclin-dependent kinase inhibitor roscovitine bind to different sites on CaV1.2 channels to affect activation and inactivation separately [30]. It is tempting to reconcile the present results by proposing distinct binding sites for methadone on CaV1.2 channels that are accessible at near PT but not RT. Follow-up studies are required to test this possibility. A few drugs have been shown to exhibit temperature-dependent block on CaV1.2 channels in overexpression cells [20] or L-type Ca2+ channels in native myocytes [21]. One study showed that nitrendipine and diltiazem inhibited Ca2+ current mediated by CaV1.2, β1, α2/δ, and γ subunits more potently at RT than at 33°C [7]. Another study reported that increasing the recording temperature from 22°C to 37°C increased the block potencies of flavoxate and nifedipine on L-type Ca2+ channels by 2.2X and 7X, respectively [8]. The direction and magnitude of potency shift due to temperature is thus drug-specific. Results of verapamil and methadone from the present study further extend those in the literature, demonstrating that recording temperature is an experimental factor that impacts CaV1.2 pharmacology. Comparisons of drug effects on the Ca2+ and Ba2+ currents at near PT The effects of buprenorphine, norbuprenorphine, naloxone, and diltiazem were studied at near PT on the Ca2+ and Ba2+ currents; of naltrexone and tolterodine, on the Ca2+ current alone. Fig 5 shows the concentration-inhibition plots for these drugs. Fig 6 summarizes the IC50s and nH values for all drugs studied. For near PT recordings, verapamil, buprenorphine, naloxone, diltiazem, and tolterodine inhibited ICa-ramp more potently than ICa-step, suggesting that these drugs all have affinity for open and inactivated channels. Tolterodine showed the largest ICa-ramp vs. ICa-step IC50 difference, by a factor of 8.7, suggesting a stronger preference for the inactivated state comparing with other drugs. For norbuprenorphine, there was no difference in the IC50s for ICa-step and ICa-ramp, suggesting a preference for open channel block when Ca2+ was used as the charge carrier. Naltrexone was the only drug tested that showed higher IC50 for ICa-ramp than for ICa-step. This drug produced a dramatic concentration-dependent hyperpolarizing shift in the ramp voltage (Fig 7A), by -22 mV at 10 mM (Fig 7B), demonstrating an effect on voltage-dependence of channel gating. Using I-V generated from 14 cells, the reversal potential for Ca2+ current under the current experimental condition was estimated to be +46 mV (S5B Fig). The lower fractional inhibition of ICa-ramp relative to ICa-step thus may not indicate drug unbinding during the +30 mV step. Instead, the increased driving force through channels that are available at more hyperpolarized membrane voltages in the presence of naltrexone may also be an explanation of the higher IC50 for ICa-ramp than for ICa-step. S6 and S7 Figs provide time course plots of individual cells tested with select drugs in Ca2+ and Ba2+, respectively. Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 5. Concentration-inhibition plots for CaV1.2 channel block by buprenorphine, norbuprenorphine, naloxone, naltrexone, diltiazem, and tolterodine at near PT. Data for ICa-step are shown in circles; ICa-ramp, squares; IBa-step, upright triangles; IBa-ramp, inverted triangles. Open symbols reflect individual data points; filled symbols plus error bars, mean ± sem. The solid sigmoidal curve indicates the fit with the Hill equation; the dashed curves, upper and lower limit of the 95% CI of the fit. https://doi.org/10.1371/journal.pone.0276995.g005 Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 6. Summary of IC50 and nH values. Top. The voltage protocol used and example traces recorded at near PT from two cells, one recorded in external Ca2+ (left; cell ID: 21515003c) and the other in external Ba2+ (right; cell ID: 21617012). For the cell on the left, the illustrated traces were obtained in the control solution (black; trace 206), 0.3 μM diltiazem (light red; trace 278), 10 μM diltiazem (medium red; trace 339), and 100 μM verapamil solutions (dark red; trace 397). For the cell on the right, the illustrated traces were obtained in the control solution (black; trace 50), 0.3 μM diltiazem (light red; trace 130), 10 μM diltiazem (medium red; trace 240), and 100 μM verapamil solutions (dark red; trace 286). Ipassive traces, shown in gray, were calculated based on Rinput derived from the verapamil traces shown. The voltage protocol was overlaid on top of the current traces. Note that for the cell on the right, Ba2+ current isolation was done using Ipassive subtraction, as the outward current seen in control solution was no longer apparent in diltiazem and verapamil solutions. Bottom, summary IC50 and nH values, (mean ± 95% CI) obtained at the current regions indicated by the dotted arrows. https://doi.org/10.1371/journal.pone.0276995.g006 Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 7. Concentration-dependent shift of voltage at which ICa-ramp peaked. A) Changes in the ramp voltage vs. naltrexone concentrations. Gray open symbols indicated data points from individual cells; black filled symbols plus error bars indicate mean ± sem. B) Representative current traces obtained from 1 cell (cell ID: 20109006) in external Ca2+ at near PT. The 90th trace (black) was the last trace obtained in the control solution; the 140th trace (red) was obtained following 10 mM naltrexone application. The voltage protocol was shown below the current traces. In this cell, the ramp voltage that ICa-ramp reached peak shifted from –2 mV in the control solution to -18 mV following naltrexone application. For the naltrexone data set, the ramp voltage at which peak ICa-ramp occurred in the control solution was 0.15 ± 0.35 mV (n = 21), consistent with the cells used to generate Fig 1C, 1D and 1G. https://doi.org/10.1371/journal.pone.0276995.g007 These results demonstrate that even when drug effects were analyzed within the same cell using the same traces, different drug potencies could be obtained depending on which current region was analyzed. Similar analyses have been performed for (-)-menthol and nimodipine [15]. In rabbit ventricular myocytes, these drugs inhibited the peak Ca2+ current evoked by a step depolarization less potently than the late Ca2+ current that remained at the end of the step depolarization, with nimodipine showing 13.1X difference in the IC50s. The difference in IC50s obtained for the step and the ramp current are compatible with literature findings of drugs exhibiting state- and/or use-dependent block of CaV1.2 channels in overexpression cell lines and L-type Ca2+ channels in native myocytes. Nitrendipine [13, 17], nisoldipine, nicardipine [17], nifedipine, verapamil [16], and mibefradil [14] all showed more block when the cells were held at depolarized membrane potential than when cells were held at hyperpolarized membrane potential, suggesting preferential block of channels in the inactivated state. Verapamil and diltiazem showed block increasing at higher stimulation frequencies and higher depolarizations, suggesting a preference for the open and inactivated state over closed state [9–11, 31]. Figs 5 and 6 also show data summarizing inhibition of the Ba2+ current by verapamil, buprenorphine, norbuprenorphine, naloxone, and diltiazem. All drugs showed greater inhibition of IBa-ramp than IBa-step (Figs 5 and 6) The largest difference was seen for verapamil, for which IC50s between the step and the ramp current differed by a factor of 5. In external Ba2+, the differences between the IC50s of the step and the ramp current were more pronounced than in external Ca2+ for verapamil, buprenorphine, norbuprenorpine, and naloxone. Even norbuprenorphine, which showed no difference between the IC50s of ICa-step and ICa-ramp, showed a difference of 2.4X between IBa-step and IBa-ramp. These results thus demonstrate that state-dependent interactions between drugs and CaV1.2 channels, as well as relative preference for different channel states, are dependent on the charge carrier used. Generalizations regarding drug-CaV1.2 channel interactions based on different studies should thus take the charge carrier used into consideration. Comparisons of the IC50s on the same current region obtained in external Ba2+ and Ca2+ showed a small impact on the step current. Based on point estimate comparison, diltiazem showed the biggest difference, with an IC50 at ICa-step 2.1X higher than that at IBa-step. Likewise, the impact of charge carrier on the ramp current and the direction of shift were also drug-specific. Verapamil showed no difference, while naloxone showed a 5X difference, with IBa-ramp being more sensitive to block than ICa-ramp. Results of the present study thus add to those in the literature demonstrating an impact of charge carrier on drug inhibition of Ca2+ and Ba2+ currents. In ventricular myocytes, diltiazem [11], D600 (a methoxy derivative of verapamil [11]), and verapamil [12] were less effective in inhibiting Ba2+ current than Ca2+ current through Ca2+ channels [11]. Similar findings were reported for verapamil [10] and diltiazem [9] studied using CaV1.2 channels with β1b and α2δ subunits using overexpression cells. Data in Figs 5 and 6 also show that the nH values were quite variable amongst the drugs, ranging from very low (nH = 0.2 for naloxone on IBa-step) to quite steep (nH = 1.4 for methadone on ICa-ramp). Shallow slopes of the concentration-inhibition relationship may be reflective of technical challenges in the experiments. These challenges include current rundown that can inflate estimation of fractional inhibition at low drug concentration, and insolubility at high drug concentration that leads to fewer than expected free drug molecules to block ion channels. While these possibilities cannot be ruled out for tolterodine, they cannot explain data of other drugs. For verapamil and methadone, changing the recording temperature greatly altered the steepness of the concentration-inhibition curves, with nH values at one temperature being 0.4 to 0.6 and at the other temperature being 1 and above. For naltrexone and diltiazem, nH values differ between the step and the ramp current, with that for one current approaching 1 while the other remaining at or below 0.6. For naloxone and naltrexone, this laboratory has previous tested similar concentrations on NaV1.5 channels and obtained nH values equal to or greater than 0.8 on the peak and the late current [18]. Therefore, the simplest explanation is that for these drugs, the shallowness of the concentration-inhibition relationship reflects more complex drug-CaV1.2 channel interactions that cannot be readily explained by 1:1 drug-receptor binding scheme that leads to immediate current inhibition. It is difficult to relate these nH results to existing literature, since nH values for concentration-inhibition plots are often not reported. In a study that assessed the structural basis of diltiazem block of voltage-gated Ca2+ channels, the resting state block was found to have an IC50 of 41 μM and a much steeper slope for the concentration-inhibition relationship than use-dependent block, which had a shallower slope but more potent block with an IC50 of 10.4 μM [32]. These results therefore demonstrate that nH values for diltiazem are state-dependent, consistent with the present findings. Based on X-ray crystallographic analysis, the study showed that diltiazem has two distinct binding poses with the Ca2+ channels: upon entering the channel pore, this drug forms a loose channel-blocking complex that appears to be a low affinity binding mode, and then rearranges within the channel to a tighter binding, more stably blocking complex with diltiazem projecting into the selectivity filter from the central cavity upon voltage-dependent inactivation [32]. Importantly, diltiazem binding also allosterically modulates Ca2+ binding in the selectivity filter, suggesting this mechanism may also contribute to reduction of current. Therefore, for a “pore blocker” like diltiazem, the different binding poses associated with different channel states may explain different nH values. Table 2 summarizes the IC50 values reported in the literature for methadone, diltiazem, verapamil, and tolterodine as well as experimental protocol used. A wide range of IC50s were reported, with the max-to-min ratios for diltiazem being 397X (0.63 μM vs. 250 μM) and for verapamil 294X (0.16 μM vs. 47 μM). The only study that reported CaV1.2 channel block by buprenorphine, norbuprenorphine, naltrexone, and naloxone was by this laboratory [18]. In Tran et al., 2020, the IC50 values of drug block on the ramp current measured using Ca2+ as the charge carrier was derived from the same cells used in the present study. Download: PPT PowerPoint slide PNG larger image TIFF original image Table 2. Comparisons of IC50 values for CaV1.2 channel block for methadone, diltiazem, and verapamil generated by different experimental protocols. https://doi.org/10.1371/journal.pone.0276995.t002 Limitation, lessons learned, protocol standardization, and conclusion There are several limitations in the present study that can impact drug potency estimation. The first is that the drug concentrations exposed to the recorded cell were not measured using an analytical method. Drug concentrations can deviate from the target concentration due to compound-specific factors and human errors. The former include nonspecific binding to the plastic and glass substrates within the patch clamp perfusion apparatus, potential insolubility at higher concentrations tested, and instability in the perfusion solution under the experimental condition. Notably, verapamil and diltiazem are compounds stated to be light sensitive (https://www.sigmaaldrich.com/US/en/sds/sigma/v4629) and advised to keep away from direct sunlight (https://documents.tocris.com/pdfs/tocris_msds/0685_sds.pdf?1646647276&_ga=2.29634103.924187291.1646647248-600272910.1596464753), respectively, on the Safety Data Sheet from their distributors. While verapamil solutions and stocks were protected from light in this laboratory, diltiazem was not, and time-dependent degradation for these as well as other compounds tested throughout the recording day cannot be ruled out. Human errors can also occur during drug stock and drug solution preparation. Concentration verification, if possible, should be included as a part of the study design to rule out the possibility that deviations of drug concentrations translate into variability in drug potency estimation. The second limitation is that CaV1.2 current rundown under whole cell configuration could not be prevented, and current rundown was not corrected for drug potency estimation in this study. Although drugs were applied after the initial fast phase of rundown had ended, the IC50s may have still been underestimated. Given cell-specific rundown profile, if rundown correction were to be implemented, then accounting for individual cell’s rundown time course by fitting as many data points obtained in the control solution as possible, rather than using the time course plots derived from separate cells is recommended. Of note, although recordings in perforated patch configuration decrease cell dialysis-dependent rundown, this technique leads to higher Rs compared to whole cell recordings. Current rundown was a trade-off for voltage control in the present study, and the method to optimize voltage control was to use whole cell recording to obtain as small of Rs as possible followed by high degree of compensation. The third limitation is that Rs was not measured throughout the recordings. Not having this measure across time raises a logical concern that rundown of Ba2+ and Ca2+ current is due to large Rs changes. In this study, Ca2+ and Ba2+ currents recorded at near PT activate extremely fast, and large changes in Rs manifest as slowing of time-to-peak for the step current (for Ba2+ and Ca2+ currents) and shift in the voltage at which ramp peak occurred (for the Ba2+ current; Fig 3). Therefore, whether Rs changed dramatically or not during the control recording period was based on offline assessment of these current profiles. Progressive decreases in the current amplitude not accompanied by kinetic changes seem incompatible with the conjecture that rundown is secondary to large changes in Rs. As the original electrophysiology records are available at https://osf.io/g3msb/, interested readers are encouraged to assess these files to draw independent conclusions regarding the mechanisms subserving rundown. The fourth limitation is that the reversal potentials of Ca2+ and Ba2+ currents were extrapolated from currents obtained between 0 and +20 mV steps (S5 Fig). These I-V relations were generated to assess adequacy of voltage control, inferred from graded increases in the current amplitudes to increasing voltages between -60 to -20 mV. For the purpose of measuring the reversal potential, extending the voltage steps to beyond the reversal potential would provide direct measurement for each cell. The fifth limitation is the uncertainty that the recorded CaV1.2 current indeed reflects activity of channels with β2 and α2δ1 auxiliary subunits. The gray traces in Fig 1G and 1H show that the ratios of ramp-to-step current are quite variable across cells. Likewise, temperature sensitivity of the whole cell current was also quite different across cells (Fig 2E and 2J). Since auxiliary subunits modulate CaV1.2 channel gating, it is possible that recordings in this study were from heterogenous CaV1.2 channels with either one or both auxiliary subunits or channels in different states of phosphorylation. A few lessons were learned by conducting the present study. First, data variability of CaV1.2 channel block was collectively larger than those observed for cardiac hERG and NaV1.5 channel block based on this laboratory’s concurrent work. The concentration-inhibition plots presented in Figs 4 and 5 show variable degrees of current inhibition for individual cells to a given drug concentration, with no clear outliers observed. This level of data spread was not seen for hERG and NaV1.5 current inhibition by the same drugs [18]. Within-the-study data variability for CaV1.2 current inhibition may be due to variable degrees of current rundown and heterogenous coupling between the channel α subunit with β2 and α2δ1 subunits, as the latter can also be targets of drugs. Second, state-dependent block of drugs on CaV1.2 channels are common, inferred from the different IC50s obtained for the step and the ramp current. Understanding state-dependent block using voltage protocols that recapitulate cardiac AP may be important when trying to predict drug impact on cardiac electrophysiology. In the sinoatrial nodal cells, CaV1.2 channels are activated rapidly upon membrane depolarization and contribute to the upstroke of the AP. Drug block of open CaV1.2 channels (i.e., inhibition of the step current) may thus inform the potential of drug in affecting the heart rate and the PR interval. On the other hand, during a ventricular AP, CaV1.2 channels are activated by the abrupt depolarization from rest, enter inactivated state during the plateau potential, and then become reactivated during the repolarizing phase of the AP before entering closed state. Information regarding how CaV1.2 channel block develops during a ventricular AP, as well as drug effect on voltage-dependence of channel gating may inform potential change in AP shape (i.e., triangulation or simply shortening), thereby allowing better prediction of proarrhythmia risk. Some statistical [33] and in silico myocyte models have incorporated drug block of CaV1.2 channels to assess the risk of Torsade de Pointes imposed by hERG channel block [34]. Incorporating IC50s measured at distinct current regions or accounting for state-dependent block of CaV1.2 channels may lead to better model performance. The third lessons learned is that many drugs have nH values much smaller than 1, and these values are dependent on the recording temperature and channel state (Fig 6). Once the contributions of current rundown and insolubility at higher tested concentrations to shallow concentration-inhibition graphs are ruled out, the most straightforward interpretation of these results is that drug-CaV1.2 channel interactions leading to current inhibition are complex processes that may involve multiple binding poses (i.e., diltiazem), multiple binding sites (i.e., nitrendipine, roscovitine), or through allosteric mechanisms (i.e., diltiazem; naltrexone, Fig 7). The present results showed that even with the same voltage protocol presented at the same stimulation rate, Cav1.2 pharmacology can still be sensitive to a variety of factors encountered during the experiments and during data analysis. When numerous experimental factors are different between two studies, as seen Crumb et al., 2016 [5] and Li et al., 2018 [6], drug potency estimates can be very different even for the same drugs. For CaV1.2 data intended to support risk prediction or clinical interpretation, normalizing laboratory-specific practices is essential toward promoting data reproducibility across laboratories—a pivotal step toward engendering confidence amongst regulators for applying these in vitro data in the decision-making process. Toward this end, the FDA Cardiac Safety Studies Interdisciplinary Review Team (CSS-IRT) has posted a document regarding the recommended voltage protocols for cardiac ion channels, including CaV1.2 channels, on its website (https://www.fda.gov/media/151418/download). The voltage waveform, stimulation frequency, compositions of the internal and Ca2+-based external solutions, and data analysis method are consistent with those used in the present study. For cardiac safety assessment, drug developers and regulators are following the guidelines released by the International Council for Harmonisation for Technical Requirements for Pharmaceuticals for Human Use: ICH S7B for nonclinical [35] and ICH E14 for clinical studies [36]. The newly released ICH E14/S7B Questions and Answers guideline offers best practice recommendations for patch clamp ion channel studies intended to support cardiac safety assessment [37] (ICH S7B Q&A 2.1), and the protocol used in this manuscript is consistent with these recommendations. ICH S7B Q&A 2.1 also recommends recording at near PT. This study tested two temperature controller models. Given the gravity-fed perfusion method and shallow bath chambers used here, the temperature controller model that heats the chamber bottom uniformly provided more stable temperature control. However, if a perfusion pump were used to maintain flow rate, then conceivably the temperature controller model that maintained bath temperature by heating the anodized aluminum platform would have also achieved stable temperature control. Experimenters interested in measuring drug block of CaV1.2 channels at near PT are recommended to consider how bath temperature may fluctuate given the rig design, and importantly measure bath temperature near the recorded cell throughout the recordings to enable subsequent analysis of temperature fluctuations on within-the-study data variability. In conclusion, results from the present study offer rationale for the best practice recommendations regarding experimental design, conduct, and data quality consideration, and may benefit stakeholders considering utilizing CaV1.2 channel data to support regulatory decision-making. This table summarizes the IC50 and nH values for CaV1.2 channel block determined using the manual patch clamp method [5] and automated patch clamp system [6]. While these studies examined more overlapping drugs, only those for which Crumb et al., 2016 provided IC50 and nH values are shown for comparison. The ratios are calculated as maximum vs. minimum IC50s. Crumb et al., 2016 used a CaV1.2-CHO cell line from Cytocentrics Bioscience GmbH (Rostock, Germany), and did not provide information regarding subunits expressed. Experiments were conducted using whole cell patch clamp method at 36 ± 1ºC, and current was evoked using a rabbit ventricular AP waveform repeated at 10 s interval. Ba2+ (4 mM) was used as the charge carrier. The automated patch clamp data from Li et al., 2018 were generated using a CHO cell line that expressed hCaV1.2α, β2, α2δ1 subunits from Charles River Laboratories (Wilmington, MA). Recordings were performed using IonWorks Barracuda system operating in population perforated patch clamp mode. Recording temperature was not controlled and was expected to be higher than RT due to the heat produced during system operation. Inward current was evoked using the same “step-step-ramp” voltage waveform as used in the present study but repeated at 10 s interval. The current elicited by the 0 mV step was used to quantify drug effects. Ca2+ (6.8 mM) was used as the charge carrier. Compositions of the external recording solution were the same for both studies except for the charge carrier. For internal solution, Crumb et al., 2016 used 130 mM CsCl as the main salt, while Li et al., 2018 used 90 mM CsF + 50 mM CsCl. Current mediated by CaV1.2 channels exhibit prominent rundown when recorded under the whole cell configuration. Percent current inhibition by drug reported by Li et al. 2018 was adjusted for current run down, using data from vehicle and positive control wells, even though recordings were obtained using perforated patch mode. Crumb et al., 2016 did not correct for current rundown nor specified the rate of rundown for the cells used. The predicted logP values for these drugs based on ChemAxon are provided as estimates of lipophilicity. The sources are as follows: bepridil (https://go.drugbank.com/drugs/DB01244), chlorpromazine (https://go.drugbank.com/drugs/DB00477), diltiazem (https://go.drugbank.com/drugs/DB00343), ondansetron (https://go.drugbank.com/drugs/DB00904), terfenadine (https://go.drugbank.com/drugs/DB00342), and verapamil (https://go.drugbank.com/drugs/DB00661). There is no relationship between the ratio of max-to-min IC50s and logP values for these drugs. Supporting information S1 Fig. Ipassive-subtraction vs. verapamil-subtraction. Recordings were obtained in external Ca2+. Cell ID was 18n20001 for panels (A) through (C); 18n14007, panels (D) through (I). Dashed lines in panels (A), (B), (D), (F), and (H) mark the 0 pA level. A) Original unsubtracted traces from a cell for which Ipassive subtraction method worked well to quantify ICa-ramp. Traces 1 and 35 were the 1st and last recorded traces in control solution (black). Trace 85 was the last trace recorded in 30 nM tolterodine (light red). Trace 126 was the last trace recorded in 100 μM verapamil (red). Ipassive, calculated using Rinput derived from trace 126, is shown in gray. Note the good alignment between Ipassive and trace 126 at all voltages, suggestive of little to no endogenous voltage-dependent current in this cell under the specified experimental condition. B) Ipassive-subtracted traces show little to no outward current that is above 0 pA. ICa-ramp for this cell is quantified as the peak inward current during the voltage ramp down phase using Ipassive-subtracted traces. C) Time course plots of ICa-ramp (top panel), Rinput (middle panel), and I-80 mV (lower panel) for the same cell. D) Original unsubtracted traces from a cell for which verapamil subtraction method was used to quantify ICa-ramp. Traces 1 and 62 were the 1st and last recorded traces in control solution (black). Trace 114 was the last trace recorded in 3 μM tolerodine (medium red). Trace 143 was the last trace recorded in 100 μM verapamil (red). Note that this cell had larger outward current at the +30 mV step and the voltage ramp down phase than the cell illustrated above that was unmasked when the inward current was reduced (see medium red and red traces). E) Top, time course plots of absolute current amplitude, measured with a 20 ms window around the inward current peak (corresponding to ramp voltage 10.8 to -39.5 mV). Given the presence of outward current, as inward Ca2+ current was suppressed with verapamil, polarity of the absolute current reversed from being negative to positive. The middle and lower panels are time course plots of Rinput and I-80 mV for this cell. F) Ipassive-subtracted traces for the same cell showed that outward current remained. G) Voltages at which the maximal negative ramp current was identified from Ipassive-subtracted traces and by searching the entire voltage ramp down phase. Similar to panel (E), there was a jump of voltage at which the maximal negative ramp current was identified when Ca2+ current was largely inhibited (i.e., ramp current approaching linear). H, I) Verapamil-subtracted traces for this cell (H) and time course plot of ICa-ramp quantified using verapamil-subtracted traces (I). https://doi.org/10.1371/journal.pone.0276995.s001 (EPS) S2 Fig. Rundown of Ca2+ current for two additional cell lines. A, B) These panels show Ca2+ currents recorded from 2 additional cell lines. Currents were also mediated by hCav1.2α, β2, and α2δ1 subunits expressed in HEK293 cells. Verapamil was not applied for cell in (B) since there was little CaV1.2 current remaining after 150 traces. C, D) Summary time course plots of ICa-step (C) and ICa-ramp (D) recorded in control solution for the cell line represented by (A). E, F) Summary time course plots of ICa-step (E) and ICa-ramp (F) for the cell line represented by (B). Note that some cells did not last for all 200 traces of recording. Verapamil was not applied for this cell line. https://doi.org/10.1371/journal.pone.0276995.s002 (EPS) S3 Fig. Time course plots demonstrating amplitude fluctuations for ICa-0 mV but not ICa-ramp. Recordings occurred in the control solution followed by bath application of 100 μM verapamil. Cell ID was 19322002. A) Top, current traces 100 and 114 (black) were obtained in the control solution; trace 200 (red), following steady state current inhibition by verapamil. Ipassive, shown in gray, was derived from the verapamil trace. Bottom, the voltage protocol used. B) The time course plots of ICa-0 mV and ICa-ramp. C) Rinput (top) and I-80 mV (bottom). https://doi.org/10.1371/journal.pone.0276995.s003 (EPS) S4 Fig. Current-temperature relationship for additional cells shown in Fig 2E and 2J. Cells shown in panels (A) through (E) were recorded in 1.8 mM Ca2+; panels (F) through (I), in 4.0 mM Ba2+. For each panel, the cell ID is shown on the top of the plots. A) This cell had 3 cycles of temperature manipulations and had rundown correction performed for both ICa-0 mV and ICa-ramp. Fitting the rundown-corrected ICa-step-temperature relation with a linear function yielded a slope of -0.0405 nA/°C and a Y-intercept of 0.704 nA. No relation was observed between ICa-ramp and temperature. B) This cell had 5 cycles of temperature manipulations and did not require rundown correction of the current amplitudes. The slope of the ICa-0 mV-temperature relation was -0.121 nA/°C, and the Y-intercept was 3.406 nA. No relation was observed between ICa-ramp and temperature. C) This cell had 5.5 cycles of temperature manipulations and did not require of rundown correction of the current amplitudes. The slope of the ICa-step-temperature relation was -0.0278 nA/°C, and the Y-intercept was 0.342 nA. ICa-ramp was also inversely related to temperature for this cell. The slope of the ICa-ramp-temperature relation was -0.0160 nA/°C, and the Y-intercept was 0.0694 nA. D) This cell had 2 cycles of temperature manipulations and did not require rundown correction of current amplitudes. The slope of the ICa-step-temperature relation was -0.129 nA/°C, and the Y-intercept was 2.66 nA. No relation was observed between ICa-ramp and temperature (r = 0.27). E) This cell had 2 cycles of temperature manipulations and did not require rundown correction of current amplitudes. The slope of ICa-0 mV-temperature relation was -0.288 nA/°C, and the Y-intercept was 8.059 nA. ICa-ramp was also inversely related to the bath temperature for this cell. The slope of the ICa-ramp-temperature relation was -0.0301 nA/°C, and the Y-intercept was 0.808 nA. F) This cell had 4 cycles of temperature manipulations. Rundown correction was performed for both IBa-step and IBa-ramp. The slope of rundown corrected IBa-step-temperature relation was -0.117 nA/°C, and the Y-intercept was 1.414 nA. IBa-ramp appeared to be inversely related to the bath temperature. The slope of rundown corrected IBa-ramp-temperature relation was -0.0292 nA/°C, and the Y-intercept was -0.771 nA. G) This cell had two cycles of temperature manipulations. Rundown correction was performed on both IBa-0 mV and IBa-ramp. The slope of the IBa-0 mV-temperature relation was -0.146 nA/°C, and the Y-intercept was 2.360 nA. IBa-ramp appeared to be inversely related to the bath temperature. The slope of IBa-ramp-temperature relation was -0.0225nA/°C, and the Y-intercept was -0.558 nA. H) This cell had 5 cycles of temperature manipulations and did not require rundown correction of current amplitudes. The slope of the IBa-step-temperature relation was -0.120 nA/°C, and the Y-intercept was 3.165 nA. IBa-ramp was also inversely related to the bath temperature. The slope of the linear fit was -0.0603 nA/°C, and the Y-intercept was 1.426 nA. I) This cell had 5 cycles of temperature manipulations and did not require rundown correction of current amplitudes. The slope of IBa-step-temperature relation was -0.287 nA/°C, and the Y-intercept was 7.098 nA. IBa-ramp was also inversely related to the bath temperature. The slope of the IBa-ramp-temperature relation was -0.0770 nA/°C, and the Y-intercept was 0.774 nA. https://doi.org/10.1371/journal.pone.0276995.s004 (EPS) S5 Fig. Normalized I-V relation of Ba2+ and Ca2+ current. Currents were evoked from a holding potential of -80 mV to +20 mV in external Ba2+ (A) or +15 mV in external Ca2+ (B) with 5 ms voltage steps in 5 mV increments. For each cell, the peak current evoked by each voltage step was normalized to the maximum current recorded. The normalized and averaged current from all cells for a particular voltage step was then plotted against that membrane voltage to generate the normalized I-V plots. Reversal potentials (Erev) of the Ba2+ and Ca2+ currents were estimated by fitting the averaged data points from -5 mV to +20 mV in external Ba2+ (A) or -5 mV to +15 mV in external Ca2+ (B) with linear functions in the form of y = a + bx, and then solving for x when y = 0. For (A), a = 0.794; b = -0.023. For (B) a = 0.799; b = -0.017. https://doi.org/10.1371/journal.pone.0276995.s005 (EPS) S6 Fig. Exemplar time course plots of Ca2+ current recorded at near PT following applications of: A) buprenorphine, B) norbuprenorphine, C) methadone, D) naltrexone, E) naloxone, and F) tolterodine. Open black symbols reflect absolute current amplitudes obtained by analyzing unsubtracted current traces; open red symbols reflect Ipassive-subtracted current amplitudes. All plots show absolute and Ipassive-subtracted amplitudes. Due to current rundown, data points for earlier traces for some cells are off the scale hence not illustrated. https://doi.org/10.1371/journal.pone.0276995.s006 (EPS) S7 Fig. Exemplar time course plots of Ba2+ current recorded at near PT following applications of: A) buprenorphine, B) norbuprenorphine, and C) naloxone. Open black symbols reflect absolute current amplitudes obtained by analyzing unsubtracted current traces; open red symbols reflect Ipassive-subtracted current amplitudes. All plots show absolute and Ipassive-subtracted amplitudes. Due to current rundown, data points for earlier traces for these cells are off the scale hence not illustrated. https://doi.org/10.1371/journal.pone.0276995.s007 (EPS)
TI - Experimental factors that impact CaV1.2 channel pharmacology—Effects of recording temperature, charge carrier, and quantification of drug effects on the step and ramp currents elicited by the “step-step-ramp” voltage protocol
JO - PLoS ONE
DO - 10.1371/journal.pone.0276995
DA - 2022-11-23
UR - https://www.deepdyve.com/lp/public-library-of-science-plos-journal/experimental-factors-that-impact-cav1-2-channel-pharmacology-effects-qQetzdULPV
SP - e0276995
VL - 17
IS - 11
DP - DeepDyve
ER -