Nonclinical Cardiovascular Assessment of the Soluble Guanylate Cyclase Stimulator Vericiguat S

.


Introduction
Vericiguat is a soluble guanylate cyclase (sGC) stimulator that was recently approved for the treatment of symptomatic chronic heart failure (HF) following a worsening event in adult patients with reduced ejection fraction (European Medicines Agency, 2021;Food and Drug Administration, 2021;McDonagh et al., 2021).Approval was based on results of the pivotal Phase 3 VICTORIA study, showing a reduction in the composite endpoint of cardiovascular death or HF hospitalization in the vericiguat group relative to the placebo group (Armstrong et al., 2020).Chronic HF is associated with endothelial cell dysfunction and an impaired nitric oxide (NO)-sGC-cyclic guanosine monophosphate (cGMP) signaling pathway (Stasch et al., 2011).The reduced availability of cGMP affects physiologic mechanisms including vasorelaxation, platelet aggregation, and myocardial remodeling, which, in combination with other pathophysiological pathways, may ultimately result in HF (Stasch et al., 2011).
In this article, we report the results of standard International Council for Harmonization (ICH) S7B nonclinical studies conducted prior to dosing vericiguat in humans: ventricular repolarization in vitro [human Ether a-go-go Related Gene (hERG) K 1 channel assay] and in vivo [electrocardiogram (ECG) in conscious dogs] (European Medicines Agency, 2006;Center for Drug Evaluation and Research, 2021).Following the completion of the VICTORIA trial we conducted a more comprehensive set of in vitro voltage-clamp recordings on cardiac ion channel recombinant cell lines with vericiguat and its major metabolite M-1, an N-glucuronide.Although no signal for proarrhythmia was observed throughout the vericiguat clinical development program, there was nevertheless recognition of the increased vulnerability to arrhythmia in patients with worsening HF.In light of this and in response to feedback from regulators, we performed these additional electrophysiological studies under conditions consistent with the recently introduced Comprehensive In Vitro Proarrhythmia Assay (CiPA) paradigm, as well as those simulating pathophysiological states, including ischemia (at more depolarized potentials) (Shaw and Rudy, 1997;Carmeliet, 1999;Akar and Akar, 2007) and extremes of heart rate (HR) to mimic tachy-or bradycardia (using high-or low-pacing frequencies).

Material and Guidelines
Vericiguat was synthesized and provided by Bayer AG.For in vitro experiments, a vericiguat stock solution was prepared [10 mM in dimethyl sulfoxide (DMSO)] and diluted appropriately into various extracellular salt solutions to reach the desired final concentrations.For the in vivo study in telemetered dogs, vericiguat was formulated in a vehicle [ethanol/polyethylene glycol 400 (PEG400), 10:90] and administered using gelatin capsules.For the repeat-dose dog toxicity studies, vericiguat was formulated in PEG400 and administered at 1-2 ml/kg.
All in vivo experiments were conducted with local approval and conformed to the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes or the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Study Design
Four separate sets of nonclinical studies were conducted to investigate the effects of vericiguat on blood pressure, HR, ECG, and ion channel currents: (i) a dog telemetry and pharmacokinetic study, (ii) "snapshot" ECG assessments during the 4-and 39-week GLP repeat-dose toxicity studies in dogs, (iii) a GLP-compliant hERG K 1 assay, and (iv) a series of non-GLP-compliant cardiac ion channel studies.
In Vivo Assessments of Systemic Exposure, Blood Pressure, HR, and ECG Intervals in Conscious Beagle Dogs Systemic exposure was investigated in blood samples collected from nontelemetered male and female beagle dogs (n 5 3/dose) at 1, 3, 7, and 24 hours following single oral administration of 0.6, 2, and 6 mg/kg vericiguat.Blood samples were drawn via the jugular or cephalic vein and centrifuged at 4 C at 3600 rpm for 10 minutes.The resulting plasma samples were stored at #À15 C for the duration of the study.C max was defined as the maximum drug concentration in plasma, and T max as the time at which C max was reached.
Arterial blood pressure (abdominal aorta), ECG (subcutaneous electrodes, standard lead II), and body temperature were continuously monitored over a period of 18 hours (2 hours 20 minutes before and 16 hours after dosing) in beagle dogs (two male, two female) that had been surgically implanted with a telemetry system (model TL11M2-D70-PCT; Data Science, Inc., St. Paul, Minnesota).Vericiguat was administered to conscious telemetered beagle dogs at single oral doses of 0 (vehicle), 0.6, 2, and 6 mg/kg body weight following a Latin square study design.Signals were acquired and analyzed with Ponemah P3 Plus, V.4.9 (Data Science, Inc.).Data were processed and averaged over a predefined period (logging rate) of 5 minutes.These data were then averaged over intervals of 15 minutes for each parameter.For calculation of mean maximal changes in cardiovascular parameters, the 15-minute bins were collapsed into a superinterval (2-6 hours post-treatment) that considers magnitude and duration of the response, and pharmacokinetic properties of the compound encompassing T max and several hours thereafter.Appropriately selected superintervals have been shown to improve statistical sensitivity to detect minor changes (Sivarajah et al., 2010).Systolic, diastolic, and mean arterial blood pressure and HR were measured using telemetric pressure signals (n 5 4/dose).PQ and QT intervals and the QRS duration were measured from telemetric ECG signals.QT intervals were corrected for HR by using the formulae of Fridericia (QTcF) (Fridericia, 1920), van de Water (QTcV) (Van de Water et al., 1989), and Matsunaga (QTcM) (Matsunaga et al., 1997).
Dogs did not receive anesthetic agents during the study period as no invasive procedures requiring anesthetics were performed.No dogs were terminated during the conduct of these studies.Beagle dogs employed in the telemetric study had participated in previous experiments (their last treatment was 3-13 weeks before this study).Before the start of the study, dogs received a veterinary health assessment.After surgery, all telemetered dogs were given a recovery period of $10 days.At the end of the study and following an appropriate drug washout period, telemetered dogs were assigned to a pool of dogs to be employed in future experiments.Beagle dogs (n 5 6 per group; three female, three male) initially received a once-daily oral dose of vericiguat of either 0 mg/kg/day, 2.5 mg/kg/day, 7.5 mg/kg/day, or 25 mg/kg/day by gavage.On day 15, the 25 mg/kg dose was reduced to 15 mg/kg for the remainder of the study duration, owing to severe gastrointestinal findings (particularly rectum prolapse).Hereafter, this group is described as the 25/15 mg/kg/ day group.The vehicle was PEG400 at 2.0 ml/kg.Acute short-term (<60 seconds) ECG recordings were conducted prior to dosing (baseline) and 2 hours post-dose (T max ) during week 1 and week 4. Blood pressure was measured using invasive techniques through the femoral artery.

ECG Assessments Following Repeat Dosing in a 39-Week Toxicity Study
Beagle dogs (n 5 8 per group; four female, four male) received a once-daily oral dose of vericiguat of either 0 mg/kg/day, 0.5 mg/kg/day, 1.5 mg/kg/day, or 5 mg/kg/day by gavage.The vehicle used was PEG400 at 1.0 ml/kg.Acute short-term (<60 seconds) ECG recordings were conducted prior to dosing (baseline) and 2 hours post-dose (T max ) during week 13 and week 39.Blood pressure was measured using highdefinition oscillometry.

In Vitro Assessments: Effects of Vericiguat and M-1 on Cardiac Ion Channel Currents
Cell Lines.Stably transfected human embryonic kidney (HEK) cell lines were used for the hERG K 1 , hNav1.5 Na 1 , and hKvLQT/minK channel voltage-clamp studies.Stably transfected Chinese hamster ovary cell lines were used for the hCav1.2Ca 21 and hKv4.3K 1 channel voltage-clamp studies.All cell lines were cultured in a humidified incubator at 37 C and 5% CO 2 .
Manual Voltage-Clamp Technique at Room Temperature (GLP hERG K + Assay).The whole-cell voltage-clamp technique under GLP was used to measure hERG K 1 -mediated inward tail currents elicited by hyperpolarizing voltage steps (repeated every 12 seconds, holding potential À80 mV) from 120 mV to À120 mV (duration 500 milliseconds) at 22 C using standard procedures as previously described (Zhou et al., 1998).Vericiguat delivery to the organ bath was initiated after the initial stabilization period of approximately 6 minutes.When possible, vericiguat concentrations (0.1, 1, and 10 mM) were applied in a cumulative manner using exposure times of approximately 6 minutes per concentration, followed by a washout period and subsequent exposure to a high concentration (1 lM) of the selective hERG K 1 channel blocker E-4031.
Voltage-Clamp Data Analysis.When applicable, the concentration dependence of effects was modeled with a standard four-parameter logistic equation: effect 5 min 1 ðmax=ð1 1 10 Ù ððlogIC 50 À logXÞ Â nHÞÞÞ (1) with minimal and maximal effects (min, max), half-maximal inhibitory drug concentration (IC 50 ), drug concentration (X), and Hill slope (nH).Minimal and maximal effects were usually treated as constants (max 5 100 and min 5 0), and IC 50 and nH as variables.If only one concentration was measured, nH was set to 1.

Exposure Multiple Calculation
Exposure multiples were calculated based on human clinical C max unbound plasma concentrations of 18 nM and 43 nM for vericiguat and M-1, respectively, at the maximum recommended human dose of 10 mg and in comparison with the IC 50 or highest concentration tested for each channel (for the in vitro studies) or with the maximum unbound drug concentration in plasma (C max.u ) in the dogs (for the in vivo studies).

Statistical Methods
In the blood pressure and HR study in telemetered dogs, the calculations included determination of arithmetical mean and standard deviation (SD) or standard error of mean (SEM).Effects were assessed regarding changes after administration versus baseline values compared with changes in the vehicle group; this was done for 15-minute bins (results not shown) as well as for the 2-to 6-hour superinterval that was chosen to enhance sensitivity (Sivarajah et al., 2010).The data reported are group mean values and corresponding SD or SEM of n experiments.Subsequent data analysis with a one-way ANOVA followed by Dunnett's multiple comparisons test versus vehicle control [differences significant if P < 0.05 (multiplicity-adjusted)] was done for the 15-minute bins (results not shown) as well as for the 2-to 6hour superinterval.The results of the repeat-dose toxicity studies were not subject to statistical analysis because of the short (<60 seconds) duration of recording.Calculations and graphical presentation of data were conducted with GraphPad Prism v8.

In Vivo Studies in Beagle Dogs
Systemic Exposure Following a Single Oral Dose.Following single oral administration, C max levels for vericiguat were 270 lg/l, 848 lg/L, and 1949 lg/L for the 0.6, 2.0, and 6.0 mg/kg doses, respectively, with calculated T max of 2.1 hours (0.6 mg/kg) and 3 hours (2 and 6 mg/kg) (Table 1).Trough plasma levels were < 10% of C max after 24 hours.In the corresponding protein-unbound plasma, C max.u levels were 60 nM, 189 nM, and 434 nM for the 0.6, 2.0, and 6.0 mg/kg doses, respectively (Table 1).Plasma concentrations of M-1 were neither determined in this single-dose study nor in the two repeat-dose toxicity studies mentioned later.
Blood Pressure, HR, and ECG Intervals Following a Single Oral Dose.Vericiguat was associated with expected pharmacology-mediated dose-dependent changes in cardiovascular function (Table 2).The effects of vericiguat on mean arterial blood pressure in conscious telemetered dogs are shown in Fig. 1.A change was observed in arterial blood pressure, particularly in systolic arterial blood pressure (up to À20%), that was not fully reversible within 16 hours.Vericiguat was associated with a dose-dependent increase in HR, with changes from baseline of 17%, 28%, and 53% for 0.6, 2.0, and 6.0 mg/kg doses, respectively.Along with an increased HR (Table 2; Fig. 1), the PQ and QT intervals were shortened (Table 2) with administration of vericiguat.PQ interval changed by À7%, À15%, and À17% for 0.6, 2.0, and 6.0 mg/kg doses, respectively.QT interval changed by À5%, À7%, and À12% for 0.6, 2.0, and 6.0 mg/kg doses, respectively (Table 2; Fig. 2).
In the 39-week toxicity study, C max.u at week 13 was 59 nM, 192 nM, and 481 nM for the 0.5 mg/kg/day, 1.5 mg/kg/day, and 5 mg/kg/day dosing regimens, respectively.When measured at week 38, C max.u changed from week 13 by 114 nM, 116 nM, and À149 nM for the 0.5 mg/kg/day, 1.5 mg/kg/day, and 5 mg/kg/day dosing regimens, respectively (Supplemental Table 4).

Downloaded from
In Vitro Whole-Cell Voltage-Clamp Investigations Effects of Vericiguat and M-1 on the hERG K + Current Channel.In the GLP manual voltage-clamp study at approximately 22 C, vericiguat blocked hERG K 1 -mediated tail currents of stably transfected HEK293 cells in a dosedependent manner.Based on mean values ± SD, curve fitting with a standard four-parameter logistic equation yielded a 20% threshold inhibitory concentration (IC 20 ) of approximately 1.9 lM and IC 50 of 9.9 lM (Table 3).
In automated voltage-clamp recordings conducted at room temperature, vericiguat inhibited hERG K 1 outward and inward tail currents by 65% at 10 lM at 22 C, with an IC 50 of 4.4 lM and 4.3 lM, respectively (Table 3; Fig. 3, A and B).Exposure multiples were calculated as 244 and 239 for hERG K 1 outward and inward tail currents, respectively.M-1 inhibited hERG K 1 outward and inward tail currents by 20% and 17%, respectively, at 10 lM at 22 C. Exposure multiples were calculated as >233 (based on an IC 50 >10 mM) for both outward and inward tail currents.
In manual patch clamp experiments performed to characterize the rate dependence of hERG K 1 current reduction at physiologic temperature, vericiguat inhibited hERG K 1 with an IC 50 of 2.9 lM at 0.03 Hz (73% at 10 lM), compared with 20% at 10 lM at 1 Hz.At 10 lM, inhibition by M-1 was À3.2% at 1 Hz and À0.0% at 0.03 Hz (Table 3).
Effects of Vericiguat and M-1 on Other Current Channels.Vericiguat and M-1 did not cause biologically relevant inhibition of hNav1.5, hCav1.2,hKvLQT1, and hKv4.3 at 10 lM, neither at 22 C nor at physiologic temperature or in experiments simulating extreme depolarization or pacing rates (Table 3; Fig. 3, C-F).

Discussion
The four sets of nonclinical studies demonstrated no evidence of an increased proarrhythmic risk from the in vitro and in vivo assessment of vericiguat or its major N-glucuronide metabolite M-1.
In dogs, administration of vericiguat as single oral doses at a maximum dose of 6.0 mg/kg was associated with dose-dependent decreases in arterial blood pressure and compensatory increases in HR.This is consistent with long-lasting vasodilation attributable to the mode of action of sGC stimulators, which mediate relaxation of smooth muscle cells by increasing cGMP levels (Stasch et al., 2011).QTc intervals using QTcF, QTcV, and QTcM formulae were not prolonged to a meaningful extent in beagle dogs.Recorded exposure multiples of 3.3, 10.5, and 23.6 at 0.6, 2.0, and 6.0 mg/kg, respectively, indicate proarrhythmic risk did not occur at clinically relevant doses.
In several in vitro electrophysiological studies, potential effects on cardiac ion channels of vericiguat and M-1 were assessed in a comprehensive manner, covering the most important cardiac ion channels (hERG K 1 , hNav1.5, hCav1.2,hKvLQT1/minK, and hKv4.3) at various temperatures (ambient and physiologic) and a range of stimulation rates (0.03-3 Hz), simulating extreme pathophysiological conditions that might be observed in patients with HF.Neither vericiguat nor M-1 inhibited cardiac ion channels (hERG K 1 , hNav1.5, hCav1.2,hKvLQT1/minK, and hKv4.3) at substantial exposure multiples of therapeutically relevant concentrations.The results for M-1 are consistent with the preponderance of scientific literature that  Smith et al., 2018).Therefore, these data contribute to the current scientific understanding that there is no cause for concern regarding exposure to N-glucuronide metabolites.Vericiguat inhibited the hERG K 1 ion channel with an IC 20 approximately 105-fold higher than the human clinical C max.u of 18 nM at 10 mg at room temperature.Some technical challenges were encountered during the study of pharmacological effects by vericiguat or M-1 under conditions mimicking pathophysiological conditions.Excessive current rundown was particularly noted in studies measuring KvLQT1/minK currents at physiologic temperature at extreme pacing rates, confounding pharmacological assessments.For this reason, a separate assessment was conducted at room temperature, in which little or no effects were noted at the extreme rates that could be tested.Similarly, no assessments of stable hNav1.5 currents were possible from holding potentials depolarized lower than À80 mV, owing to excessive current rundown attributed to the steady state inactivation properties of hNav1.5 currents.
At supratherapeutic concentrations, reverse frequency dependence was observed during the study of hERG K 1 inhibition by vericiguat at extreme pacing rates.The inhibitory potency at the two extremes (1 Hz and 0.033 Hz) was in line with previous assessments at room temperature, and sufficient margins were calculated for the lowest inhibitory concentration.This phenomenon of a so-called "reverse-frequencydependent" hERG K 1 inhibition is well known (Weirich and Antoni, 1998) and has been described for many hERG K 1 blockers in vitro (Baskin and Lynch, 1994) and in vivo, including in humans (D emolis et al., 1996).Although reverse frequency dependence on action potentials has been correlated with drugs potently blocking the hERG K 1 current, when frequency dependence of blockade on hERG K 1 channels has been systematically studied and modeled, it is clear that the major difference between drugs with high versus low torsadogenic risk is the impact of inward currents mitigating the effect of hERG K 1 and not a systematic difference in the intrinsic frequency dependence of hERG K 1 inhibition (Li et al., 2017).Other important considerations in the overall reverse frequency dependence are the effect of a drug on the KvLQT1/ minK current, which contributes to the repolarization reserve (Weirich and Antoni, 1998), and safety margins.Drugs such as loratadine, sold over the counter, have been shown to display rate frequency dependence on the hERG K 1 current (Crumb, 2000), but these potential concerns are obviated by the large safety margin to the effective clinical C max (Redfern et al., 2003).Similarly, our study of vericiguat indicates intrinsic blocking properties on the hERG K 1 channel do not translate into adverse effects when sufficient safety margins are established.
Vericiguat had no meaningful effect on QTc interval in dogs when administered at doses #7.5 mg/kg/day.This absence of an effect was observed in 4-and 39-week repeat-dose toxicity studies in beagle dogs.It is important to note a positive control was not used in these studies.As in vitro hERG K 1 channel studies and in vivo telemetered beagle dog assays were deemed negative in terms of meaningful QT prolongation and altered ventricular repolarization at the therapeutic dose, a positive control was waived and is supported by recent recommendations in a cardiac safety regulation protocol (Lester, 2021).Moxifloxacin is often used as a positive control in QT prolongation investigations to determine the sensitivity of an assigned assay owing to its expected prolongation of the QT interval, regardless of therapy area (Carlson et al., 2011;Langenickel et al., 2016;Demmel et al., 2018;Sun et al., 2020).Although a positive control was not per protocol in these experiments, a clinical QTc interval study in patients with chronic coronary heart disease did use moxifloxacin as a positive control and similarly concluded that vericiguat 10 mg was not associated with a proarrhythmic risk (Boettcher et al., 2021).In summary, there was no nonclinical evidence of an increased proarrhythmic risk based on comprehensive in vitro and in vivo assessment of vericiguat or its major N-glucuronide metabolite M-1-findings consistent with the lack of any evidence of proarrhythmia seen during the vericiguat clinical development program.The integrated risk assessment of these nonclinical data combined with existing clinical data (Boettcher et al., 2020) demonstrate that administration of vericiguat 10 mg once daily in patients with HF with reduced ejection fraction is not associated with a proarrhythmic risk.
Exposure multiples calculated based on human clinical C max (unbound) of 18 nM. b

TABLE 3
In vitro electrophysiological assessments of vericiguat and M-1 Exposure multiples calculated based on human clinical C max (unbound) of 18 nM, compared with the IC 50 or highest concentration tested for each channel at 10 mg dosage.Exposure multiples calculated based on human clinical C max (unbound) of 43 nM, compared with IC 50 or highest concentration tested for each channel at 10 mg dosage.Interpreted as small, reverse frequency dependence of hERG potency, comparing effects at the different stimulation rates tested in the hERG studies listed.
a Manual voltage clamp.b Automated voltage clamp.c d e