Evaluation of levocetirizine in beagle dog and cynomolgus monkey telemetry assays: Defining the no QTc effect profile by timepoint and concentration‐QTc analysis

Abstract

In prior clinical studies, levocetirizine (LEVO) has demonstrated no effect on ventricular repolarization (QTc intervals), therefore it is a relevant negative control to assess in nonclinical assays to define low proarrhythmic risk. LEVO was tested in beagle dog and cynomolgus monkey (nonhuman primate [NHP]) telemetry models to understand the nonclinical‐clinical translation of this negative control. One oral dose of vehicle, LEVO (10 mg/kg/species) or moxifloxacin (MOXI; 30 mg/kg/dog; 80 mg/kg/NHP) was administered to instrumented animals (N = 8/species) using a cross‐over dosing design; MOXI was the in‐study positive control. Corrected QT interval values (QTcI) were calculated using an individual animal correction factor. Blood samples were taken for drug exposure during telemetry and for pharmacokinetic (PK) analysis (same animals; different day) for exposure‐response (C‐QTc) modeling. Statistical analysis of QTc‐by‐timepoint data showed that LEVO treatment was consistent with vehicle, thus no effect on ventricular repolarization was observed over 24 h in both species. PK analysis indicated that LEVO‐maximum concentration levels in dogs (range: 12,300–20,100 ng/ml) and NHPs (range: 4090–12,700 ng/ml) were ≥4‐fold higher than supratherapeutic drug levels in clinical QTc studies. Slope analysis values in dogs (0.00019 ms/ng/ml) and NHPs (0.00016 ms/ng/ml) were similar to the human C‐QTc relationship and indicated no relationship between QTc intervals and plasma levels of LEVO. MOXI treatment caused QTc interval prolongation (dog: 18 ms; NHP: 29 ms). The characterization of LEVO in these non‐rodent telemetry studies further demonstrates the value and impact of the in vivo QTc assay to define a “no QTc effect” profile and support clinical safety assessment.

Study Highlights.

• WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?

Recent International Conference on Harmonization E14/S7B Q&A guidance indicates that high quality in vivo QTc assay data can be used as part of an integrated nonclinical‐clinical pro‐arrhythmia risk assessment to support a TQT waiver.

• WHAT QUESTION DID THIS STUDY ADDRESS?

Levocetirizine (LEVO), an approved drug with a proven clinical QTc safety profile, was used in dog and monkey telemetry studies to characterize a nonclinical “no QTc effect” profile using timepoint and concentration‐QTc analyses.

• WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?

In sensitive dog and monkey telemetry studies, LEVO was devoid of any effect on QTc intervals at exposure levels that exceeded the high clinical exposure observed at the supratherapeutic dose in TQT studies.

• HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE?

Nonclinical in vivo QTc findings with levocetirizine translate well with clinical QTc safety data. The new data demonstrates the high value of well‐designed dog and monkey QTc telemetry studies to support integrated E14/S7B‐based risk assessment and regulatory decision making in drug development.

INTRODUCTION

The nonclinical in vivo QTc assay is a core cardiovascular safety tool used by the pharmaceutical industry to characterize the potential of a new drug to cause delayed ventricular repolarization. The assay is described in the International Council for Harmonization (ICH) S7B guideline ¹ and has demonstrated high translational value for predicting QTc interval prolongation in humans. ² , ³ , ⁴ , ⁵ In conjunction with the in vitro human ether‐a‐go‐go K+ current (hERG) assay, a negative signal in the in vivo QTc assay lowers the probability of clinical ventricular arrhythmia risk for a new drug product. ⁶ Various animal species have been used to conduct the QTc assay, notably the dog and nonhuman primate (NHP), based on similar electrophysiological and pharmacological responses to the pharmacological blockade of cardiac delayed rectifier K+ current (I_Kr). ⁷ , ⁸ , ⁹ In the future, the integration of findings from well‐designed nonclinical QTc assays (e.g., in vivo time‐response, exposure‐response, and in vitro analysis), in conjunction with early clinical QTc assessment, can help waive the need for dedicated TQT assessment ⁶ , ¹⁰ and bring forward beneficial therapeutics in a more efficient manner.

A recent review of publicly available TQT study findings indicated that the majority (i.e., 80%) of therapeutics tested do not cause QTc interval prolongation. ⁶ One well‐known example is levocetirizine (Xyzal; LEVO), a second‐generation antihistamine approved for adult and pediatric patients to relieve seasonal allergy with minimal QTc risk. Since approval in 2001, the drug has not been associated with any cardiac safety liabilities, in contrast with first‐generation antihistamines which also blocked I_Kr. ¹¹ , ¹² More specifically, two dedicated clinical QTc investigations demonstrated that therapeutic (5 mg) and super‐therapeutic (30 mg) doses of LEVO were devoid of any effect on QTc intervals in normal healthy volunteers using time‐course and exposure‐response (C‐QTc) analyses. ¹³ , ¹⁴

To build more nonclinical translational knowledge based on clinical safety data, the goal of this study was to characterize the effect of LEVO on QTc intervals in conscious beagle dogs and cynomolgus monkeys (NHPs), the two animal species used frequently for the cardiovascular safety assessment of new drug candidates. A study design with high statistical power was used to evaluate QTc effects following LEVO treatment using time‐response analysis; in addition, a pharmacokinetic (PK)/pharmacodynamic analysis approach was used to enable C‐QTc analysis. The QTc assays included a moxifloxacin (positive control) arm to confirm detection sensitivity.

METHODS

Animals

All procedures in this protocol complied with the Animal Welfare Act, the Guide for the Care and Use of Laboratory Animals, and the Office of Laboratory Animal Welfare. All studies were reviewed and approved by the Institutional Animal Care and Use Committee. These exploratory (non‐good laboratory practice) telemetry studies were conducted (2021) by LabCorp (Burlington, NC) in non‐naïve beagle dogs (Envigo Global Services, Cumberland, VA) and NHPs (cynomolgus monkeys; Envigo Global Services, Alice, TX). Male beagle dogs (0.8–1.6 years old; 8.7–11.7 kg) and cynomolgus monkeys (NHPs; 2.6–6 years old; 4.8–7.4 kg) were housed individually in stainless steel cages and fed a certified canine diet (#5007C; PMI Nutrition International) or primate diet (#5048; PMI Nutrition International) and water (ad libitum). Water samples were routinely analyzed for specified microorganisms and environmental contaminants. Environmental controls for the animal room were set to maintain 20–26°C, a relative humidity of 30–70%, a minimum of 10 room air changes/h, and a 12 h light/12 h dark cycle. Animals were returned to the colony for re‐use following the completion of the study. Animals were cared for in accordance with the Guide for the Care and Use of Laboratory Animals, 8th Edition (2011) in an AAALAC‐accredited facility in species‐specific housing.

Telemetry preparation

Animals were surgically implanted with telemetry devices at least 2 weeks (dog) or 3 weeks (NHP) prior to study initiation. A L11 transmitter (Data Sciences International, St. Paul, MN) was implanted into the abdomen and sutured within an extra‐peritoneal pocket for collection of electrocardiograms (ECG), blood pressure, and temperature. The ECG leads of the implanted transmitter were arranged in an approximate lead II configuration. The pressure catheter was advanced toward the abdominal aorta via the femoral artery. Detailed surgical procedures were similar to previous reports, with the following changes: the negative ECG lead was placed into the jugular vein and advanced to the right atrium and the positive ECG lead was placed on the diaphragm close to the cardiac apex. ¹⁵ , ¹⁶

Telemetry and pharmacokinetic study designs

The animal studies were conducted in two phases: an initial cardiovascular telemetry study followed by a PK evaluation in the same animals. In the telemetry phase, a double Latin square dosing design (8 × 4 CO; N = 8 animals/dose group) was used to evaluate a single oral dose of four different test articles, as follows: vehicle, LEVO (10 mg/kg/species); moxifloxacin (MOXI; 30 mg/kg/dog; 80 mg/kg/NHP), and quinine (100 mg/kg/species; Table S1). The LEVO dose was evaluated in both non‐rodent species based on a prior report which showed this dosage was tolerated and achieved high exposures in monkeys. ⁴ The typical 8 × 4 CO study is designed to assess three doses of single test article and a vehicle treatment, ¹⁷ , ¹⁸ but an alternate dosing strategy was employed to include an in‐study positive control alongside LEVO treatment. A species‐relevant dose of MOXI was used to confirm QTc sensitivity and the reproducibility of drug‐induced prolongation with prior findings. ¹⁹ As an ancillary component of this study, quinine was also evaluated in both species to explore exposure, tolerability, and overall cardiovascular profile based on a prior report. ⁴ However, due to complex behavioral, autonomic, and high drug exposure variability (due to extensive emesis) in both species, the data were excluded from this report. All animals were acclimated to the study room and dosing procedure for 2 weeks prior to dosing. Levocetirizine hydrochloride (Sigma; Lot No.: ZOSAB; 99.4% purity) was suspended in 0.5% methylcellulose in reverse osmosis water and given to non‐fasted animals. The other test articles were prepared for dosing as previously described. ⁴ , ¹⁹

In the PK phase, the same animals (N = 8) were dosed with each test article (on separate days; Table S1) and blood samples (1 ml; K₂EDTA tubes) collected at 0 (predose), 0.5, 1, 2, 4, 8, 12, and 24 h after treatment. Blood was taken from various sites in each species (dog: cephalic vein; NHP: femoral, saphenous, or cephalic veins) and stored frozen (−80°C) prior to analysis using liquid‐chromatography tandem mass spectrometry. During the telemetry phase, a single blood sample was taken (~7 h postdose) to compare with PK exposures (note: One NHP was dropped from the study on day 17 [unscheduled euthanasia] due to veterinary health concerns that were unrelated to treatment). Plasma protein binding values of LEVO were determined using an ultracentrifugation method for all species (humans: 92.9% bound; dogs 97.2% bound; and NHPs 89.0% bound) and used to calculate free unbound (F _u) drug levels. ²⁰ The LEVO molecular weight value of 388.9 was used to convert between ng/ml and uM.

Cardiovascular telemetry data acquisition and analysis

Telemetry end points (arterial pressure, heart rate [HR], and ECG intervals) were collected for at least 90 min prior to dosing (baseline) and continuously for 24 h following treatment. Telemetry timepoints were analyzed based on a nominal dose time (e.g., 10 a.m. = 0 h). ECG signals were collected at a sampling rate of 500 Hz, and all data were recorded as consecutive 1‐min averages. All signals were acquired and analyzed with a Data Sciences International (DSI) Ponemah system (St. Paul, MN). Cardiovascular parameters (HR; QT, and QTcIs) were analyzed using SAS System software. The QT interval was corrected for HR changes for each animal using a linear model incorporating an individual animal correction factory (IACF), for dog QTcI = QT−IACF × (HR‐75) or NHP QTcI = QT + IACF × (500‐RR). The linear regression slope was established on an individual animal basis using QT/HR or QT/RR interval (the time interval between success R‐waves in the ECG) plots from 1‐min means derived from the vehicle collection. ¹⁹ All hemodynamic and ECG data captured over 24 h are presented in time bins of various sizes (e.g., ≥1 h; see next section) to facilitate data reduction and statistical analysis. Some captured data segments were omitted from analysis as these coincided with study activities (e.g., feeding; blood sampling, etc.) that introduced artifacts.

QTc by timepoint statistical analysis

QTc values following drug treatment were compared to time‐matched vehicle treatment in the same conscious animals using an analysis of variance (ANOVA) model comprised of factors including dosing day, animal, and treatment group. ²¹ The 24 h dataset for each treatment was divided into five time bins or super‐interval periods for analysis. ²² The super‐interval periods were based on the PK profile of LEVO and relative to dosing at 0 h, as follows: 0.5–3 h (period 1); 3–5 h (period 2); 8–12 h (period 3); 12–18 h (period 4), and 20–24 h (period 5). The 5–8 h time block was excluded to avoid data noise associated with animal husbandry activity. Fitted period means for the LEVO and MOXI treatments were calculated using the parameter estimates from the ANOVA model and compared to vehicle fitted mean values in each period. Confidence intervals (95 % CIs) were also calculated to compare treatments as CI values reflect the range of true effect sizes. ¹⁷ The exclusion of zero from the 95% CI indicated statistical significance at the p ≤ 0.05 level. A statistical analysis was conducted to determine the minimal detectable difference (MDD) in the QTc interval and calculated as the half‐width of the 95% CI. The MDD value represents the smallest difference between group means that would be detectable in the t‐test at a probability level <0.05 (e.g., the border between significance and nonsignificance). ²³ The inclusion of an in‐study statistical metric is a best practice consideration for new cardiovascular telemetry studies to quantify QTc assay sensitivity and for potential comparison to prior studies. ¹⁸ , ²⁴

Concentration‐QTc statistical analysis

A pharmacometric analysis of LEVO was conducted in conscious beagles and cynomolgus monkeys and guided by established methods. ⁴ , ¹⁹ For the C‐QTc analysis, control‐corrected QTc (ΔQTc) was calculated by subtracting vehicle QTc values from each animal at each timepoint and dose level. The ΔQTc values were further corrected by subtracting predose QTc values (an average of the 90‐min period prior to dosing) from each animal to obtain ΔΔQTc values used for C‐QTc analysis. Regression analysis used total plasma concentrations matched with the corresponding vehicle‐ and baseline‐adjusted QTcIs (ΔΔQTcI), with a total of 64 matched pairs in dogs and 56 matched pairs in NHPs. QTcI data are reported as the median value of 1‐min data bins for the 30‐min period prior to the plasma collection timepoint. ⁴ , ¹⁹ The critical plasma concentration necessary to produce a 10 ms increase in ΔΔQTcI was determined based on the derived regression line and 90% CIs around the regression parameters. The 10‐ms effect size was used as the reference threshold to assess drug‐induced effects and species differences in QTc sensitivity ⁴ , ¹⁹ and perform clinical translation. ¹⁴ Linear regression analysis of the C‐QTc relationship (i.e., slope and intercept), projected critical concentration values (i.e., plasma concentration that produces a 10‐ms increase in QTc) and CIs (90%) were calculated as described previously. ¹⁹ , ²⁵

RESULTS

Time‐response relationships

Baseline (predose) values for QT/QTcI and HR parameters collected from conscious (N = 8) beagle dogs and NHPs (N = 7) are shown in Table S2. LEVO was well‐tolerated in both dogs and NHPs at the 10 mg/kg dose, although small amounts of vomitus were noted postdose in four of eight dogs during telemetry. Figure 1 illustrates the pharmacological effect of LEVO on QTcIs and ΔΔQTcI in dogs (Figure 1a,b, respectively). The QTcI did not change significantly over the course of 24 h (Figure 1a; Table 1). Vehicle‐ and baseline‐adjusted QTcI (ΔΔQTcI) values confirmed that LEVO mean values did not exceed 4 ms and that CIs overlapped with zero (Figure 1b; Table 1). The in‐study sensitivity (MDD) for QTc signal detection in the dog assay was 4.4 ms. The MOXI positive control produced significant QTc prolongation (Figure 1a,b; Table 1, and Table S3). Plasma levels of LEVO showed peak exposures during the 0.5–3 h time period (Figure 1c; Table 2) and samples collected at 7 h postdose in the telemetry phase were similar to those observed in the PK phase. Figure 2 illustrates QTcIs and ΔΔQTcIs following LEVO treatment in conscious NHPs. In this species, QTcIs were similar to vehicle treatment and remained unchanged over the 24 h observation period and the mean ΔΔQTcI values did not exceed 3 ms (Figure 2a,b; Table 2). The in‐study sensitivity (MDD) for QTc signal detection in the NHP assay was 4.2 ms. LEVO plasma levels were largest between 0.5 and 4 h (Figure 2c, Table 2). The free plasma levels of LEVO (corrected for protein binding) in both dogs and NHPs achieved exposures that were greater than four‐fold larger than the high clinical exposure of LEVO in humans (F _u: 92.4 ng/ml; Figure 3). As expected, the MOXI positive control produced significant QTc prolongation in the cynomolgus monkey (Figure 2; Table 2, and Table S3).

FIGURE 1.

Time‐response and concentration‐QTc (C‐QTc) relationship evaluation of oral levocetirizine and moxifloxacin‐induced QTc prolongation in conscious beagle dogs. Vehicle (open blue circles), LEVO (black circles), and moxifloxacin (red circles) were administered at 0 h. The plots represent timepoint analysis of absolute QTcI (a) and baseline‐ and vehicle‐corrected QTcI effects (ΔΔQTcI) (b) following treatment with MOXI (30 mg/kg) or LEVO (10 mg/kg). The LEVO PK curve (c) and C‐QTc relationship for LEVO (d) are also shown. The PK point at 7 h was taken during the telemetry phase. Group sizes were eight for all figures and values are mean ± SD. For panel d, data were fitted by linear regression (solid line) and dotted lines represent 90% confidence interval of the model‐predicted mean ΔΔQTcI. LEVO, levocetirizine; MOXI, moxifloxacin; PK, pharmacokinetic; QTcI, Qtc interval

TABLE 1.

Effect of LEVO and MOXI on QTcIs based on timepoint analysis in the beagle dogs and cynomolgus monkeys

Dog (N = 8)	Absolute QTcIs (ms ± SEM; fitted means) per period
Period, h	Control	LEVO (10 mg/kg)	MOXI (30 mg/kg)
0.5–3	243 ± 2	247 ± 2	261 ± 2 a
3–5	243 ± 2	243 ± 2	261 ± 2 a
8–12	239 ± 2	240 ± 2	256 ± 2 a
12–18	238 ± 1	239 ± 1	252 ± 1 a
20–24	239 ± 2	237 ± 2	247 ± 2 a
	QTc differences from control (ms and 95% CI) per period
0.5–3	–	4 (−2 to 10)	18 (12–24) a
3–5	–	0 (−6 to7)	18 (12–25) a
8–12	–	1 (−4 to 6)	17 (12–22) a
12–18	–	1 (−3 to 5)	14 (9–18) a
20–24	–	−1 (−6 to 3)	8 (4–13) a

Monkey (N = 7)	Absolute QTcIs (ms ± SEM; fitted means) per period
Period, h	Control	LEVO (10 mg/kg)	MOXI (80 mg/kg)
0.5–3	244 ± 3	246 ± 3	266 ± 3 a
3–5	244 ± 2	245 ± 3	273 ± 2 a
8–12	248 ± 3	250 ± 3	272 ± 3 a
12–18	255 ± 2	258 ± 3	275 ± 2 a
20–24	254 ± 1	252 ± 2	259 ± 1 a
	QTc differences from control (ms and 95% CI) per period
0.5–3	–	2 (−6 to 10)	22 (15–30) a
3–5	–	1 (−6 to 9)	29 (22–36) a
8–12	–	2 (−7 to 11)	24 (16–33) a
12–18	–	3 (−5 to 10)	20 (12–27) a
20–24	–	−2 (−7 to 2)	5 (1–9) a

Note: Drug effects are shown as absolute QTcIs and treatment differences.

Abbreviations: CI, confidence interval; LEVO, levocetirizine; MOXI, moxifloxacin; QTcI, Qtc interval.

^a Statistically significant at 5% level.

TABLE 2.

Evaluation of levocetirizine in QTc‐by‐timepoint and concentration‐QTc (C‐QTc) analyses in conscious beagle dogs and cynomolgus monkeys

Species	Timepoint analysis			C‐QTc analysis
	ΔQTcI (ms) from vehicle a	Cmax‐total, ng/ml	Tmax‐mean, h	Slope, ms/ng/ml	Intercept	Predicted ΔΔQTcI (ms) at Cmax	Predicted concentration (ng/ml) for 10 ms increase
Dog N = 8	4 [−2–10]	16,775 ± 2371	0.5	0.000199 [−0.000002–0.00040]	−0.197 [−1.89–1.50]	3.14 [0.62–5.66]	51,241 [25,493‐ND]
NHP N = 7	2 [−6–10]	6731 ± 3248	2	−0.00016 [−0.00035–0.00003]	−0.0075 [−4.85–4.84]	−1.08 [−4.85–2.69]	ND [333,563‐ND]

Note: Timepoint and C‐QTc analysis values are mean and 95% confidence intervals [in brackets]. C_max values are mean and SEM.

Abbreviations: C_max, maximum concentration; ND, not determined; NHP, nonhuman primate; QTcI, QTc interval; T_max, time to maximum concentration.

^aEffect measured during super‐interval period (0.5–3 h) associated with C_max.

FIGURE 2.

Time‐response and concentration‐QTc (C‐QTc) relationship evaluation of oral LEVO and MOXI‐induced QTc prolongation in conscious cynomolgus monkeys (NHPs). Vehicle (open blue circles), LEVO (black circles), and MOXI (red circles) were administered at 0 h. The plots represent timepoint analysis of absolute QTcI (a) and baseline‐and vehicle‐corrected QTcI effects (ΔΔQTcI) (b) following treatment with MOXI (80 mg/kg) or LEVO (10 mg/kg). The LEVO PK curve (c) and C‐QTc relationship for LEVO (d) are also shown. The PK point at 7 h was taken during the telemetry phase. Group sizes were seven for all figures and values are mean ± SD. For panel d, data were fitted by linear regression (solid line) and dotted lines represent 90% confidence interval of the model‐predicted mean ΔΔQTcI. LEVO, levocetirizine; MOXI, moxifloxacin; NHP, nonhuman primate; PK, pharmacokinetic; QTcI, Qtc interval

FIGURE 3.

Plasma exposure of LEVO (F _u) in dog and cynomolgus monkeys (NHPs). Free unbound levels were estimated by multiplying total plasma exposure with fraction unbound values for each species (human: 0.071, dog: 0.028, and NHP: 0.11). Human C_max values at 5 mg (F _u: 16.2 ng/ml) and 30 mg (F _u: 92.4 ng/ml) doses are depicted. Plasma samples collected during the telemetry phase (at 7 h postdose) are shown as open square (NHP) and open circle (dog). C_max, maximum concentration; F _u, free unbound; LEVO, levocetirizine; NHP, nonhuman primate

Concentration‐QTc relationship

Linear regression and slope analysis of the total concentration‐QTc slope and the predicted change in ΔΔQTcI for maximum concentration (C_max) values are shown for each in vivo QTc assay (Figure 1d, dog and Figure 2d, NHP; Table 2). Slopes were flat for both conscious dogs (0.000199 ms/ng/ml) and NHPs (0.00016 ms/ng/ml; Table 2). The slopes were extrapolated to calculate a predicted exposure that might result in a 10 ms prolongation of ΔΔQTcI: 51241 ng/ml (90% CI: 25493‐NA) in dogs and for the NHPs, no intercept could be determined (Table 2). Based on the dog C‐QTc regression, LEVO would cause a 10 ms increase at a plasma level (F _u: 1440 ng/ml) that was 15‐times larger than the highest clinical exposure (F _u: 92.4 ng/ml; Figure 3; Table 3).

TABLE 3.

Comparison of levocetirizine C‐QTc relationships in cynomolgus monkeys (NHPs) and humans

In vivo QTc assay	C‐QTc slope, ms/ng/ml	Intercept	Cmax (ng/ml, total)	Predicted change in ΔΔQTcI (ms, mean)	Predicted concentration (ng/ml) needed to produce a 10 ms increase in ΔΔQTcI
NHP a	0.00031	0.31	14,900	4.9	ND
Human (5 and 30 mg) b	0.0014	ND	160 (5 mg) 1024 (30 mg)	ND	ND
Human (30 mg) b	0.00042	ND	1024	2.1	ND
Human (5 and 30 mg) c	0.002	1.47	228 (5 mg) 1302 (30 mg)	1.9 4.1	ND

Abbreviations: C_max, maximum concentration; ND, Not determined (projected to be >C_max); NHPs, nonhuman primates; QTcI, QTc interval.

^a Komatsu et al. ⁴

^b Darpo et al. ¹⁴

^c Hulhoven et al. ¹³

For comparative purposes, the c‐QTc slope and intercept values based on a single dose of MOXI in both animal species were generally consistent with our prior study that examined QTc changes over a higher exposure range (Table S3).

DISCUSSION

The main goal of this study was to define the quantitative relationship between LEVO exposure and QTcI changes in the in vivo QTc assay to define a “no effect profile” using two statistical methods to quantify ventricular repolarization risk. The primary findings were that at the dose level used in this study, LEVO had no effect on QTcIs as determined by QTc‐by‐timepoint and C‐QTc analyses at clinically relevant plasma drug levels. Thus, LEVO demonstrated a low‐risk profile in both the beagle dog and cynomolgus monkey telemetry models used typically to profile the cardiovascular safety of new drug candidates. The QTc sensitivity of the nonrodent models was confirmed with the inclusion of MOXI, a clinically relevant positive control known to cause QTc prolongation in these two animal species ¹⁹ and the estimation of MDD, ¹⁸ , ²³ which reinforce that the animal studies were effectively powered to detect small effect sizes. The actual magnitude of QTc prolongation following MOXI treatment in the dog (18 ms) and NHP (29 ms) models exceeded the in‐study MDD (i.e., 4–5 ms) and were consistent with our prior experience using the same animal species, study design, dose levels (30 mg/kg‐dog; 80 mg/kg‐NHP), analysis methodologies and laboratory facilities. ¹⁹ This new nonclinical cardiovascular safety pharmacology evidence revealed no statistical effect on QTcIs following peak LEVO exposures that were much higher (≥4‐fold) than supratherapeutic drug levels obtained in clinical safety pharmacology studies. ¹³ , ¹⁴

Nonclinical‐clinical translation

Robust nonclinical‐clinical QTc translation is based in part on the need to utilize high quality nonclinical data sets generated under standardized conditions to minimize QTc variability (i.e., noise) and maximize QTc detection (i.e., signal). The use of experimental designs that have sensitivity to detect clinically relevant QTc changes (e.g., 10 ms prolongation) and demonstrate responsiveness to positive controls are examples of safety pharmacology best practices that promote the quality of the in vivo QTc assay for risk assessment. ²⁶ Quality improvements, in turn, will provide the foundation for enlisting nonclinical data as a supplement to early clinical data to support regulatory decision making in regard to QTc prolongation risk. ⁶ , ¹⁰ , ²⁴ For the LEVO studies in dogs and NHPs, the crossover study design and animal group size attained the power to detect small changes in the QTcI. ¹⁸

For these validation studies, the LEVO dose used (10 mg/kg p.o.) was based on a prior evaluation of exposure and QTc effect of this agent in the NHP. ⁴ It should be mentioned that LEVO was evaluated in conventional general toxicity studies in beagle dogs at oral doses (8–75 mg/kg/day for 13 week) that achieved very high multiples over the clinical therapeutic exposure (e.g., >500×) with no cardiac safety signals determined by episodic ECG recordings, but a dedicated nonclinical in vivo QTc assay was not performed. ¹¹ In the dog and NHP studies, the QTc‐by‐time‐course analyses found LEVO to be equivalent to vehicle treatment (no effect) as determined by super‐interval ANOVA, a sensitive method for optimizing QTc signal detection. ²² In addition, an exposure‐QTc response slope analysis conducted in both species confirmed that LEVO had no effect on QTcIs, which further indicate the lack of a QTc effect in these sensitive PK/QTc experiments (N = 7–8 animals and 56–64 PK/QTc data pairs). Because LEVO has not been evaluated in a dedicated dog telemetry study, these new findings verify that high exposures of LEVO are devoid of QTc effects in this species as well. Last, the dog C_max values (total) were of similar magnitude to the Komatsu et al. study (16,775 vs. 14,900 ng/ml), but our cynomolgus monkey C_max values (total) were two‐fold lower in the same species (Table 3). ⁴

An important consideration in the design of nonclinical cardiovascular safety pharmacology telemetry studies is that dose levels are chosen to achieve plasma exposure ranges that exceed the anticipated exposure range in humans, allowing the establishment of a clinical risk assessment based on safety margin. ² , ⁵ , ²⁷ In these studies, the average LEVO plasma drug level (based on F _u estimates) attained in the in vivo QTc assays were 4.3× (dog) and 6.6× (NHP) over the largest human C_max attained at the supratherapeutic dose (30 mg) in two TQT studies (total – 1302 ng/ml; F _u – 92.4 ng/ml; Table 3).

The clean QTc safety profile of LEVO in beagle dogs, cynomolgus monkeys, and humans is attributed to its poor ability to cause hERG channel blockade. For example, Delanois et al. ¹¹ indicated that LEVO exhibited no ability to inhibit hERG current expressed in Xenopus laevis oocytes at a high concentration of 30 μM. However, a recent in vitro safety pharmacology investigation demonstrated LEVO to have low potency in a sensitive hERG binding assay (Ki: ≈40 μM or 15,556 ng/ml free). ²⁸ Based on the latter hERG potency estimate, LEVO has a very large in vitro margin of safety (168‐fold) relative to the free drug level attained at the human supratherapeutic dose of 30 mg (see red line in Figure 3; 92.4 ng/ml). Overall, LEVO represents the prototypical “double‐negative” profile in the ICH S7B core assays, which illustrates the predictive safety value of both assays for characterizing drugs that lack ventricular repolarization risk. ⁶

In summary, the beagle dog and cynomolgus monkey QTcI data presented here confirm the ability of QTc‐by‐timepoint analysis and pharmacometric (C‐QTc) evaluation to define “no effect” for proarrhythmic risk assessment of new drug candidates using an approved human therapeutic with a proven QTcI safety profile. This case study highlights the value of nonclinical QTc data obtained from these animal species to identify new drug agents with low risk for causing clinical QTc prolongation. Last, the use of powered and sensitive nonrodent cardiovascular telemetry assays will improve the interpretation of drug‐mediated effects, and instill confidence in study findings, which will be needed to support integrated nonclinical‐clinical QTc risk assessment and regulatory decisions related to proarrhythmic risk. ⁶ , ¹⁰

Limitations

In retrospect, these safety pharmacology studies had some limitations for consideration. First, the studies used a single high dose of levocetirizine and an in‐study positive control to evaluate a specific hypothesis, which is a departure from conventional study designs that use dose–response analysis to fully characterize the cardiovascular safety of novel drug candidates prior to first‐in‐human testing. ²⁶ Second, the inclusion of a separate PK arm following a telemetry study is needed for optimal exposure‐response data‐pairing in the same animal, but may introduce variability in drug exposure between sessions, and error into pharmacometrics analysis, especially if the treatment impacts animal behavior or absorption (i.e., emesis). Third, the margin calculations based on unbound (F _u) drug are dependent on protein binding assessment and methodologies that may not be standardized across the industry (i.e., another source of error). In this study, the dog protein binding obtained (F _u: 0.028) was different than previously reported (F _u: 0.11 ²⁹ ) which led to a 3.9‐fold smaller margin in this species. Thus, variation in protein binding estimates will directly impact safety margin estimates based on F _u calculations for commonly used reference drugs and new drug products.

AUTHOR CONTRIBUTIONS

M.J.E., H.M.V., F.A.C., Z.J.J., J.W., and R.W.C. wrote the manuscript. M.J.E., H.M.V., J.W., and J.B. designed and performed the research. M.J.E., H.M.V., F.A.C., Z.J.J., R.W.C., and J.B. analyzed the data and contributed new analytical tools.

FUNDING INFORMATION

Amgen Inc. provided study funding.

CONFLICT OF INTEREST

All authors are employees and shareholders of Amgen Inc. All authors declared no competing interests for this work.

Supporting information

Table S1.

Engwall MJ, Baublits J, Chandra FA, et al. Evaluation of levocetirizine in beagle dog and cynomolgus monkey telemetry assays: Defining the no QTc effect profile by timepoint and concentration‐QTc analysis. Clin Transl Sci. 2023;16:436‐446. doi: 10.1111/cts.13454