Metabolism and Pharmacokinetic Drug–Drug Interaction Profile of Vericiguat, A Soluble Guanylate Cyclase Stimulator: Results From Preclinical and Phase I Healthy Volunteer Studies

Identification of the UGT isoforms involved in the glucuronidation of vericiguat, as well as transport proteins for vericiguat, was performed by several in vitro studies (see Methods section and Table 1 of the Electronic Supplementary Material [ESM]). In addition, the potential of vericiguat to act as a perpetrator by inhibiting metabolizing enzymes or transporters as well as inducing CYPs was investigated.

The conduct of all clinical studies met their local legal and regulatory requirements. All studies were conducted in Germany except for the human mass balance study, which was conducted in the Netherlands. The studies were conducted in accordance with the ethical principles that have their origin in the Declaration of Helsinki and the International Conference on Harmonisation guideline E6: Good Clinical Practice. Informed consent was obtained from all individual participants included in the studies. Quantitative measurements of vericiguat and all other drugs were performed using validated methods and in accordance with relevant guidelines (see Table 2 of the ESM for further details) [19, 20].

Healthy white male individuals were eligible in line with accepted criteria for healthy volunteer studies [21]. Across the studies, the general inclusion criteria for healthy volunteers were (with some variability depending on the study): aged 18–65 years, body mass index 18–30 kg/m², resting heart rate 45–100 beats per minute, systolic blood pressure 90–145 mmHg, diastolic blood pressure 45–95 mmHg, and an absence of clinically relevant changes in electrocardiogram findings. Participants were excluded if they were taking regular medications, consuming foods or beverages that could influence the study objectives, unable to abide by study-specific smoking restrictions, or taking supplements with known interaction potential.

One phase I study investigated the clearance mechanisms and elimination of vericiguat and ten phase I studies investigated the potential for DDIs with vericiguat in healthy male volunteers (Table 1). As vericiguat has been demonstrated to exhibit linear pharmacokinetics [16], the use of different doses of vericiguat across the phase I studies is not considered a limitation. In addition, the investigated doses include the target therapeutic dose range of vericiguat in HF [1, 2]. While the studies were conducted in relatively small population sizes and in healthy volunteers, all studies were performed to guideline recommendations and designed to investigate specific DDIs. Further details can be viewed in the Methods section of the ESM.

Pharmacokinetic parameters were estimated using non-compartmental methods by WinNonlin software (Pharsight Corporation, Mountain View, CA, USA). Area under the plasma concentration–time curve (AUC) from 0 to infinity was calculated from the AUC from time 0 to the last quantifiable concentration using the linear-logarithmic trapezoidal method and from the last predicted plasma concentration and apparent terminal rate constant (calculated from the slope of a log-linear regression). The linear-logarithmic trapezoidal method was also used to calculate partial AUCs (e.g., AUC_0–24 [AUC from time 0 to 24 h after administration]) and the AUC within the dosing interval after multiple dosing (AUC_τ,md). The maximum concentration (C_max) and the trough concentration within a dosing interval were directly determined from the concentration–time data for each participant. Summary descriptive statistics were provided for the cumulative excretion of total radioactivity in urine, feces, and urine plus feces in the human mass balance study.

For the DDI studies, PK parameters with and without co-medications were compared. Depending on the design of the study, selected PK parameters were analyzed if applicable and assumed to be log-normally distributed. These included AUC, AUC_0–24, AUC_0–24,md (AUC_0–24 after multiple-dose administration), AUC_0–22 (AUC from time 0 to 22 h), AUC_0–12,md (AUC from time 0 to 12 h after multiple-dose administration), AUC_τ,md, C_max, C_max,md (C_max within a dosing interval after multiple-dose administration), and trough concentration within a dosing interval. The logarithms of these parameters were analyzed using analysis of variance, including sequence, healthy volunteer (sequence), period, and treatment effects, except for the rifampicin, sacubitril/valsartan, and sildenafil DDI studies, which only included healthy volunteer and treatment effects. Based on these analyses, point estimates (least-squares mean) of the geometric mean ratios and 90% confidence intervals for the treatment ratios were calculated by retransformation of the logarithmic data.

Adverse events, vital signs, electrocardiogram, and laboratory parameters (i.e., hematology, coagulation, clinical chemistry, and urinalysis) in relation to the study drug were captured for all phase I studies.

In vitro studies showed that glucuronidation of vericiguat was catalyzed primarily by UGT1A9 and UGT1A1. Human liver, kidney, and intestinal microsomes fortified with uridine 5ʹ-diphospho-glucuronic acid mediate glucuronidation of vericiguat. In human liver microsomes of single donors (n = 20), the formation of M-1 correlated slightly better with propofol glucuronidation (r² = 0.73), a highly selective UGT1A9-mediated pathway, than with estradiol 3-glucuronidation (r² = 0.61), a highly selective UGT1A1-mediated pathway. Niflumic acid, a potent UGT1A9 inhibitor, showed stronger inhibition on vericiguat glucuronidation than a potent UGT1A1 inhibitor (atazanavir), suggesting that glucuronidation of vericiguat could be preferentially catalyzed via UGT1A9 than via UGT1A1. Further results are presented in Tables 3 and 4 of the ESM.

Additional in vitro studies have determined that vericiguat is a substrate of the transporter proteins P-glycoprotein and breast cancer resistance protein. Vericiguat is not a substrate of organic anion transporting polypeptide (OATP) 1B1, OATP1B3, or organic cation transporter 1. Metabolite M-1 is not a substrate of efflux (P-glycoprotein, breast cancer resistance protein) or uptake transporters (OATP1B1, OATP1B3). The design and interpretation of the in vitro studies were in line with official guidance [22, 23].

Furthermore, no inhibition was observed for vericiguat and its metabolite (M-1) at the highest tested concentration for efflux transporters P-glycoprotein and bile salt export pump (half maximal inhibitory concentration [IC₅₀] > 100 µM), multidrug and toxin extrusion protein (MATE) 1 and MATE2K (IC₅₀ > 10 µM), as well as for uptake of organic cation transporter 1 (IC₅₀ > 50 µM), organic anion transporter (OAT) 1, OAT3, and organic cation transporter 2 (all IC₅₀ > 5 µM). Both vericiguat and M-1 did not inhibit major UGT isoforms (UGT1A1, 1A4, 1A6, 2B4, and 2B7; all IC₅₀ > 50 µM), except for UGT1A9 inhibition by vericiguat (IC₅₀ = 10.6 µM, inhibitory constant [K_i] = 6.6 µM). In addition, vericiguat is an inhibitor of breast cancer resistance protein (IC₅₀ = 20 µM) as well as OATP1B1 and 1B3 (approximate IC₅₀ of 16 and 30 µM, respectively). Metabolite M-1 is an inhibitor of OATP1B1 and OATP1B3 (IC₅₀ = 25.6 and 16.6 µM, respectively). However, considering the clinical exposure of vericiguat (total C_max approximately 0.8 µM, unbound C_max approximately 0.02 µM, and projected unbound inlet C_max 0.03 µM; data on file) and M-1 (total C_max 2.7 µM and unbound C_max 0.04 µM; data on file) following a 10-mg dose once daily, the DDI risk assessment according to the US Food and Drug Administration and European Medicines Agency guidelines indicated a low potential for vericiguat and M-1 to cause DDIs when co-administered with substrates of major CYP and UGT isoforms as well as transporters.

The CYP induction potential of vericiguat and M-1 was evaluated in cultured human hepatocytes from three different donors. Vericiguat showed no significant induction of CYP1A2 and CYP2B6 up to the highest tested concentration of 10,000 μg/L (23.5 µM). For CYP3A4, a 2.7-fold increase in CYP3A4 messenger RNA was observed in one of the three donors at the highest tested concentration (4% of the rifampicin [positive control] response). The potential risk for CYP3A4 induction in the clinic was further assessed using the relative induction score method according to the US Food and Drug Administration and European Medicines Agency DDI guidance [22, 23]. The in vitro induction data in human hepatocytes (CYP3A4 messenger RNA data) and the clinical exposure data from multiple-dose oral administration of 10 mg of vericiguat once daily in patients with HFrEF (total C_max approximately 1 µM and unbound C_max approximately 0.02 µM; data on file) were used as input data. The relative induction score model predicted a 15% decrease in the AUC of midazolam, indicating that vericiguat is likely not a clinically meaningful inducer of CYP3A4 in vivo.

The characteristics of the healthy volunteers enrolled in each of the studies who were valid for PK analyses are shown in Tables 2, 3. Studies were conducted between 2012 and 2017.

A total of six healthy volunteers received a single oral solution containing ¹⁴C-labeled vericiguat 5 mg. Mean total radioactivity (range) recovered in urine and feces within 336 h was 98.3% (95.6–99.8%) of the dose administered. Of the administered dose, 53.1% (49.9–56.1%) and 45.2% (40.1–49.9%; mean values [range]) were excreted via urine and feces, respectively.

N-glucuronide metabolite M-1, a pharmacologically inactive major plasma metabolite with no known off-target activities, represented the majority of the radioactivity in urine (Fig. 1, Fig. 1 of the ESM, and Table 4; 40.8% of the dose) together with vericiguat (9.0% of the dose). Metabolite M-15, resulting from oxidative metabolism (presumably by demethylation/decarboxylation/acetylation), was detected as a minor component in urine and feces (1.9% and 1.6%, respectively). Vericiguat represented most of the radioactivity in feces (42.6% of the dose). Metabolite M-1 was the predominant component in human plasma, accounting for 72% of total radioactivity AUC, compared with 27% for the parent drug.

The clinical interaction studies confirmed and extended findings from the in vitro studies (Fig. 2, Tables 5 and 6, and Tables 5 and 6 of the ESM).

Vericiguat was well tolerated in all studies, and most treatment-emergent adverse events were of mild-to-moderate severity and as expected from the mechanism of action of vericiguat. A total of seven treatment-emergent adverse event-related discontinuations were reported across the DDI studies; however, only one (moderate headache) was related to treatment with vericiguat in the sildenafil DDI study. The treatment-emergent adverse event of moderate headache was considered to be related to both vericiguat and sildenafil.

The glucuronidation of vericiguat is mediated by UGT1A1 and UGT1A9. The preclinical biotransformation experiments demonstrated the dominant role of UGT1A9. Considering the different abundances of these UGT isoforms in tissues, UGT1A9 is predominantly responsible for M-1 formation in the kidneys, whereas both UGT1A9 and UGT1A1 are involved in the liver [24, 25]. In the human mass balance study, M-1 represented the majority of the radioactivity in urine (41% of the dose), while vericiguat represented the majority of the radioactivity in feces (43% of the dose). Given the high oral bioavailability (93%; with food) [14], the unchanged drug recovered in feces is not due to non-absorption. Metabolite M-1 was not detected in human feces; ex vivo incubation with human feces under anaerobic conditions showed that M-1 could undergo hydrolytic cleavage to form vericiguat by microbial flora (Fig. 2 of the ESM). Based on a preclinical mass balance study in bile-duct cannulated rats, metabolite M-1 can be excreted via bile (data on file). Therefore, the parent drug observed in feces could result from biliary/intestinal secretion of the parent, as well as possible hydrolysis of M-1 in the gut lumen. Thus, although 41% of the total vericiguat dose is found in the urine as M-1, glucuronidation may contribute to a greater percentage of the overall clearance.

Ten phase I DDI studies in healthy volunteers investigated the PK interactions between vericiguat and drugs that are known to affect: intestinal pH and NO signaling; inhibit/induce metabolic pathways; and drugs that are commonly administered to patients with cardiovascular disease. Polypharmacy is common in patients with HF and therefore drugs intended to be used in HF are required to possess a low potential for DDIs as these may result in uncontrolled changes in drug plasma concentrations [11‐13].

Previous studies have shown an increase in AUC and C_max by approximately 44% and 41%, respectively, and lower variability (geometric coefficient of variation in AUC was 19% vs 41%) of vericiguat (alone) after administration with food compared to without [14, 16]. Here, a reduction by about 30% in the bioavailability of vericiguat was observed following co-administration in fasting conditions of omeprazole (40 mg once daily) and magnesium/aluminum hydroxide (600/900 mg) relative to vericiguat alone. This is primarily driven by the increase in gastric pH, as vericiguat is a basic compound with a pK_a of 4.7 and its solubility is lower in neutral conditions compared with acidic conditions. The increased exposure following administration with food is believed to be due to enhanced wettability and solubility of vericiguat by stimulation of bile salt secretion and the presence of food components enabling better solubilization. Therefore, vericiguat is recommended to be taken with food, irrespective of food type and the concomitant administration of proton pump inhibitors, as well as antacids, were allowed in phase II and III studies.

No relevant PK interaction was observed with the strong CYP3A and broad-spectrum transporter inhibitor ketoconazole (200 mg twice daily; vericiguat AUC increased by 13%) or the inducer rifampicin (600 mg once daily; vericiguat AUC decreased by 29%). Therefore, similar concomitant medications were not excluded in the clinical studies. In addition, inhibition of the major metabolizing enzyme of vericiguat, UGT1A9, by mefenamic acid (500 mg four times daily) led to a non-relevant increase in the AUC of vericiguat (by 20%), thereby allowing UGT inhibitors as concomitant medications in further clinical studies.

The lack of a DDI with midazolam (7.5 mg, single dose) in the clinic is consistent with the in vitro data indicating that vericiguat is not a CYP3A4 inhibitor [18]. Moreover, results from the relative induction score model indicate that vericiguat is likely not a clinically meaningful inducer of CYP3A4 in vivo.

Consistent with the preclinical data, vericiguat had no effect on the pharmacokinetics of narrow therapeutic index drugs commonly prescribed in cardiovascular disease (warfarin 25 mg and digoxin 0.375 mg). Similarly, warfarin, digoxin, aspirin (500 mg; single dose), and sacubitril/valsartan (co-administered at 97/103 mg twice daily) had no effect on the pharmacokinetics of vericiguat.

To rule out that a potential pharmacodynamic interaction is driven by an underlying PK interaction in the sacubitril/valsartan interaction study with vericiguat, the plasma concentrations of both compounds were measured. Exposure and peak concentration of the active sacubitril metabolite LBQ657 were unaffected by vericiguat. For valsartan, increases in exposure and peak concentration were similar following co-administration with vericiguat and co-administration with placebo. Therefore, a PK interaction between vericiguat and sacubitril/valsartan was ruled out. Furthermore, multiple oral doses of vericiguat 2.5 mg administered 2 h prior to sacubitril/valsartan had no additional effect on seated blood pressure [15].

The sildenafil DDI study indicated an increase in exposure for sildenafil co-administered with vericiguat of 22.4% or less, compared with sildenafil co-administered with placebo; however, all 90% confidence intervals included one. No meaningful change in the pharmacokinetics of vericiguat in response to sildenafil treatment was observed. No PK dose-dependent trend was observed with sildenafil (25, 50, or 100 mg).

Vericiguat was generally well tolerated and had a favorable safety profile across all phase I studies when given for a short duration or as a single dose in healthy volunteers. Most treatment-emergent adverse events were mild or moderate in severity and in line with those previously reported for vericiguat [26, 27]. However, as the studies were performed in healthy volunteers, the findings may not reflect how the combination of vericiguat and other co-medications will be tolerated in patients. The concomitant use of vericiguat and phospodiesterase-5 inhibitors, such as sildenafil, is not recommended because of the lack of clinical experience in patients with HF and the potential increased risk for symptomatic hypotension.

Overall, this series of DDI studies demonstrates that vericiguat has a low potential for PK DDIs across the intended therapeutic dose range. Vericiguat has predictable pharmacokinetics, which can be described by a one-compartment model parameterized by apparent clearance and apparent volume of distribution [17]. Based on this population PK model from patients with HFrEF [17], variability in steady-state AUC is estimated to be about 30%. Therefore, the observed maximum changes in exposure in the reported studies were within the range of reported PK vericiguat variability in patients with HFrEF and are considered not clinically relevant. The concurrent or anticipated use of drugs affecting gastric pH and known metabolic pathways, as well as drugs with a narrow therapeutic index or commonly prescribed in patients with cardiovascular disease were permitted in the phase III VICTORIA study.