Drug Interactions

Evaluation of the Effect of Food and Gastric pH on the Single-Dose Pharmacokinetics of Cabozantinib in Healthy Adult Subjects

The Journal of Clinical Pharmacology 2015, 55(11) 1293–1302 © 2015, The American College of Clinical Pharmacology DOI: 10.1002/jcph.526




Linh Nguyen, PhD1 , Jaymes Holland, RPh1, Richard Mamelok, MD2, Marie-Kristine Laberge, MSc3, Julie Grenier, PhD3, Dennis Swearingen, MD4, Danielle Armas, MD4, and Steven Lacy, PhD1

Abstract Cabozantinib is a small molecule tyrosine kinase inhibitor that has been approved for the treatment of patients with progressive, metastatic medullary thyroid cancer. Cabozantinib exhibits a pH-dependent solubility profile in vitro. Two phase 1 clinical pharmacology studies were conducted in healthy subjects to evaluate whether factors that may affect cabozantinib solubility and gastric pH could alter cabozantinib bioavailability: a food effect study (study 1) and a drug–drug interaction (DDI) study with the proton pump inhibitor (PPI) esomeprazole (study 2). Following a high-fat meal (study 1), cabozantinib Cmax and AUC were increased (40.5% and 57%, respectively), and the median tmax was delayed by 2 hours. Cabozantinib should thus not be taken with food (patients should not eat for at least 2 hours before and at least 1 hour after administration). In the DDI study (study 2), the 90% confidence intervals (CIs) around the ratio of least-squares means of cabozantinib with esomeprazole versus cabozantinib alone for AUC0–inf were within the 80%–125% limits; the upper 90%CI for Cmax was 125.1%. Because of the low apparent risk of a DDI, concomitant use of PPIs or weaker gastric pH-altering agents with cabozantinib is not contraindicated.

Keywords cabozantinib, proton pump inhibitor, drug interactions, food effect, pharmacokinetics

Cabozantinib is an orally administered inhibitor of receptor tyrosine kinases including hepatocyte growth factor receptor, vascular endothelial growth factor receptor 2 (VEGFR2), and the rearranged during transfection receptor.1,2 Cabozantinib has been approved for the treatment of patients with progressive, metastatic medullary thyroid cancer (MTC) and is currently being evaluated in late-stage clinical trials in patients with hepatocellular or kidney cancers.1,3–5 Based on a population pharmacokinetic (PK) analysis of cabozantinib in patients with MTC, the predicted effective half-life is approximately 55 hours, the oral volume of distribution (Vz/F) is approximately 349 L, and the apparent oral clearance (CL/F) at steady state is estimated to be 4.4 L/h.8 The terminal half-life after a single oral dose in healthy volunteers is approximately 120 hours. Following oral administration of cabozantinib in cancer patients, median time to peak cabozantinib plasma concentrations (tmax) ranged from 2 to 5 hours postdose. The intersubject variability (%CV) in exposure for cabozantinib following repeat dose administration was high (maximum concentration [Cmax] %CV range, 37%–43%; area under the curve [AUC] %CV range, 38%–43%). One of the largest sources of intra- and intersubject PK variability is the process of drug absorption, and food and level of acidity in the stomach are 2 of the multiple factors that influence this process.6,7

Cabozantinib is a weak base that belongs to Biopharmaceutics Classification System (BCS) Class II.8 BCS Class II compounds are characterized by low solubility and high permeability.9 Cabozantinib is weakly soluble in aqueous media and demonstrates a pHdependent solubility profile (0.11 mg/mL in 0.01 N HCl and practically insoluble at pH > 4).8 Cabozantinib also exhibits high cell permeability in vitro in MDCK cell assays (Papp values at 70–110 nm/s). Food generally has a positive effect on oral absorption of weak base compounds, where the bioavailability is enhanced through increased dissolution by stimulation of bile and pancreatic enzyme release. A food effect was reported for other small-molecule anticancer drugs 1

Exelixis, Inc., San Francisco, CA, USA Mamelok Consulting, Palo Alto, CA, USA 3 Celerion, Montreal, QC, Canada 4 Celerion, Tempe, AZ, USA 2

Submitted for publication 10 February 2015; accepted 16 April 2015. Corresponding Author: Steven Lacy, PhD, Exelixis, Inc., 210 East Grand Avenue, So. San Francisco, CA 94080 Email: [email protected]


Currently at Medivation, San Francisco, CA USA

1294 belonging to BCS Class II, such as nilotinib10 and erlotinib,11 but not for dasatinib.12 Weakly basic drugs may also show decreased absorption under conditions of elevated gastric pH (ie, with coadministration of acidreducing agents). Coadministration with acid-reducing agents (eg, proton pump inhibitors [PPIs], H2 receptor blockers) resulted in marked reductions in plasma drug exposures (AUC) for a variety of other small-molecule kinase inhibitors including dasatinib, imatinib, erlotinib, gefitinib, lapatinib, and nilotinib.13 As cabozantinib belongs to BCS Class II and exhibits a pH-dependent solubility profile in vitro, 2 clinical studies were conducted to evaluate whether factors that may affect gastric pH and cabozantinib solubility would influence cabozantinib bioavailability: a food effect study and a drug– drug interaction (DDI) study with a PPI. The S-enantiomer of omeprazole, esomeprazole, is a PPI that inhibits gastric acid secretion by suppressing Hþ ion formation by specific inhibition of the Hþ/Kþ ATPase enzyme system at the secretory surface of gastric parietal cells14 and was selected as the gastric pH-modifying agent for evaluation in the DDI study. Esomeprazole at 40 mg once daily has been shown to maintain an intragastric pH > 4 for a significantly longer duration as well as to have higher median intragastric pH over 24 hours, compared with standard doses of omeprazole, lansoprazole, pantoprazole, and rabeprazole.15,16 Yin et al has shown that after esomeprazole administration at 40 mg once daily for 5 days, mean gastric pH was shown to be elevated to 4.0–5.4 from the baseline of 1.17 Here, the results are presented from these 2 phase 1 studies that evaluated the effect of a high-fat meal (study 1, food effect) and of an increase in gastric pH

Figure 1. Schematic diagrams of study designs.

The Journal of Clinical Pharmacology / Vol 55 No 11 (2015)

(study 2, DDI with esomeprazole) on the single-dose PK of cabozantinib.

Methods Ethics The design and monitoring of these clinical studies complied with the ethical principles of Good Clinical Practice, in accordance with the Declaration of Helsinki. Study protocols and informed consent documents were reviewed and approved by the institutional review board of participating institutions, and informed consent was obtained from all subjects before participation. Each study was conducted at 1 study site. Prior to study initiation, all pertinent study documents for both studies were reviewed by the institutional review board of Chesapeake Research Review, Inc. Study Design Both studies were phase 1, single-center, open-label, 2treatment in nonsmoking healthy male and female subjects, aged 18–55 years. The design of both studies is shown in Figure 1. In the 2-way crossover food effect study (study 1), 56 subjects were randomized to receive a single dose of 175 mg of cabozantinib malate salt (140 mg free base equivalent [FBE]) in capsule form on day 1 under fasting or fed (30 minutes after administration of a high-fat meal)18 conditions, separated by a washout of 28 days. The high-fat meal used contained approximately 138 calories from protein, 217 calories from carbohydrate, and 720 calories from fat (approximately 50% of total meal caloric content).

Nguyen et al

The clinical cabozantinib dose approved for the treatment of progressive, metastatic MTC is 140 mg FBE and was thus chosen for use in this food effect study. The safety profile for cabozantinib (nonclinical and clinical characterizations) and the single 140-mg FBE dose design of this proposed study (yielding exposures predicted to be lower than steady-state exposure at maximum tolerated dose [140-mg FBE once daily] and lower than the Food and Drug Administration acceptable genotoxicity impurity exposure limit) supported conducting the proposed study in normal healthy subjects. In the fixed-sequence DDI study (study 2), 22 subjects received a single cabozantinib dose of 100 mg FBE administered as a single tablet on day 1 of period 1. This dose was selected for use in this study as it was anticipated to provide measurable plasma concentrations even under the circumstances of potentially lower exposures following esomeprazole administration. In period 2, following once-daily oral doses of 40 mg esomeprazole delayedrelease capsule for 5 days, subjects received a single dose of 40 mg esomeprazole delayed-release capsule followed 1 hour later by a single 100-mg FBE dose of cabozantinib tablet. There was a washout of 31 days between cabozantinib dosing in period 1 and period 2. In both studies, blood samples (3 mL each) were collected up to 15 minutes before cabozantinib dosing and 0.5, 1, 2, 3, 4, 5, 6, 8, 10, 14, 24, 48, 72, 120, 168, 240, 336, 408, and 504 hours after dosing cabozantinib for each treatment period for analysis of cabozantinib plasma concentrations. PK sampling up to 504 hours postdose was planned to characterize > 80% of the cabozantinib AUC0–inf. All subjects underwent pre- and poststudy assessments including physical examinations, routine laboratory tests, vital signs, and 12-lead electrocardiograms. A pregnancy test (women of childbearing potential) and alcohol and drug screen were conducted before each cabozantinib dosing period. All subjects (male and female) from both studies were between 18 and 55 years old with a body mass index (BMI)  18 and  33.0 kg/m2. Subjects had to refrain from the use of any prescription medications, including drugs known to be a strong or moderate inhibitor or inducer of cytochrome P450 (CYP) enzymes within 28 days of dosing. Subjects were to also refrain from the use of nonprescription or over-the-counter preparations (including H2 blockers, PPIs, vitamins, minerals, and herbal supplements) within 7 days of dosing or any tobacco- or nicotine-containing products within 3 months of dosing. Subjects were excluded from participation in the studies if any abnormalities were found during the medical examination. Women of childbearing potential were included in the studies if they agreed to use consistently during the course of the study (starting at or before screening) and for 120 days following the last dose of study drug one of the

1295 approved birth control methods such as a non-hormonereleasing intrauterine device with spermicide, a male or female condom with spermicide, a contraceptive sponge with spermicide, a diaphragm with spermicide, or a cervical cap with spermicide. In the 2-way crossover food effect study (study 1), an intravaginal system, an oral, implantable, transdermal, or injectable hormonal contraceptive for at least 3 months prior to dosing was also accepted as a birth control method. Subjects were excluded from the study if they had a history of any medical or surgical conditions (eg, stomach or intestinal surgery or resection) that would potentially interfere with or alter the gastrointestinal (GI) absorption, distribution, metabolism, or excretion of the study drug. Subjects were also excluded if they had any intercurrent illness and/or a history or clinical manifestations of urological, GI, renal, hepatic, neurological, hematological, metabolic, oncologic, pulmonary, immunologic, psychiatric, or cardiovascular disease. Analytical Methods Plasma cabozantinib concentrations were measured using a validated liquid chromatographic–tandem mass spectrometry method, with the lower limit of quantitation for cabozantinib of 0.5 ng/mL.8 Cabozantinib concentrations were measured and reported as FBE. In study 1, concentration range was 0.5 to 500 ng/mL. Assay performance was monitored using quality control (QC) samples at concentrations of 0.5, 1.5, 20, 250, and 400 ng/ mL for study 1; the %CV for the QCs ranged from 2.6% to 5.9%, and accuracy was 95.9% to 102.2%. In study 2, the concentration range was 0.5 to 1000 ng/mL. QC samples at concentrations of 1.5, 80, 400, and 800 ng/mL were used; the %CV for the QCs ranged from 3.7% to 6.2%, and accuracy was 92.4% to 101.3%. PK Parameter Estimates As data allowed, noncompartmental PK parameters AUC extrapolated to infinity (AUC0–inf), Cmax, time to Cmax (tmax), half-life (t1/2), CL/F, and apparent volume of distribution (Vz/F) were calculated from cabozantinib concentration–time data in plasma using WinNonlin Professional (versions 5.3 and 6.3; Pharsight, Mountain View, California). AUC0–inf was calculated as the sum of AUC from time 0 to the time (t) of the last observed/ measured nonzero concentration (AUC0–t) plus the ratio of the last measurable plasma concentration to the elimination rate constant (kel); AUC0–t was calculated by the linear trapezoidal summation method; Cmax and tmax were reported as observed values; t1/2 was calculated as ln(2)/kel; kel was calculated by linear least-squares regression analysis using the maximum number of points in the terminal log-linear phase (using 3 or more nonzero plasma concentrations); CL/F was calculated as dose/ AUC0–inf;Vz/F was calculated as dose/(AUC0–inf  kel).

The Journal of Clinical Pharmacology / Vol 55 No 11 (2015)

1296 The primary PK end point measures were Cmax and AUC0–inf. Statistical Analyses Statistical analyses were conducted to compare cabozantinib primary PK end points in the presence and absence of food (study 1) and prior to and following esomeprazole administration (study 2). For both studies, a SAS mixedmodel procedure was performed on the natural log (ln)– transformed PK parameters Cmax and AUC0–inf. The mixed model included treatment as a fixed effect and subject as a random effect. Each mixed model included calculation of the geometric mean, the difference between geometric mean, and the standard error associated with this difference. Consistent with the two 1-sided tests approach for bioavailability, 90% confidence intervals (CIs) for the differences in between-treatment geometric means were derived from the analyses on the lntransformed PK parameters Cmax and AUC0–inf. The CIs were expressed as a percentage relative to the geometric mean of the reference treatment (cabozantinib alone or under fasting conditions). Ratios of geometric means (GMRs) and the associated 90%CIs were reported by exponentiating the difference between geometric means and its 90%CIs for ln-transformed PK parameters Cmax and AUC0–inf. No food effect or drug interaction would be claimed if the 90%CI for the GMR of PK parameters fell within 80%–125%. The statistical power calculation19 indicated that 44 evaluable subjects would be required to provide 80% power to show no food effect in study 1. In study 2, 18 evaluable subjects were estimated to provide 80% power to detect a 30% difference between geometric means of the ln-transformed PK parameter AUC0–inf of test to reference treatment.

Results Demographics Baseline demographics of the participants in both studies are shown in Table 1. Effect of Food on the PK of Cabozantinib (Study 1) A total of 56 subjects were enrolled, and 50 completed the study. One subject discontinued during period 1 following Table 1. Demographic Characteristics of Subjects at Baseline

Parameter Age (y)a Weight (kg)a BMI (kg/m2)a Sex (male/female) a

Values are mean (range).

Study 1 Food Effect (N ¼ 56)

Study 2 DDI (N ¼ 22)

38 (18–55) 75.8 (48.5–95.9) 28.1 (20.1–33.1) 26/30

38.0 (25–50) 72.2 (55.7–99.9) 26.5 (20.9–31.9) 9/13

treatment under fed conditions because of personal reasons (classified as “subject request other than adverse event”). Five subjects were withdrawn during the washout period because of protocol-directed reasons (classified as “other”) including hemoglobin level below lower limit of normal (2 subjects in each treatment sequence) and out of range BMI (1 subject following treatment under fasted conditions). Plasma concentrations and PK parameters for cabozantinib are provided in Figure 2 and Table 2, respectively. In the presence of a high-fat meal, cabozantinib PK parameter values for Cmax and AUC0-inf were increased by 40.5% and 57.0%, respectively, and the median time to maximum cabozantinib plasma concentration (tmax) was longer (increased from 4 [fasted] to 6 [fed] hours). The PK values for t1/2, CL/F, and Vz/F appear to be similar under fed and fasted conditions. One outlier subject in the fed group had CL/F and Vz/F values approximately 21- to 22-fold higher than the mean values (CL/F, 2.50 L/h [mean] vs 52.6 L/h [outlier subject]; Vz/F, 450 L [mean] vs 9838 L [outlier subject]). If the CL/F and Vz/F values for this subject were excluded from the analysis, the mean values were 1.39 L/h and 242 L, respectively, which are approximately 42% lower than the mean values observed in the fasted group; these lower CL/F and Vz/F values reflect the magnitude increase in the AUC0–inf values after a high-fat meal. Statistical comparisons of cabozantinib PK parameters under fed and fasted conditions including the outlier subject are provided in Table 3. The 90%CIs for the GMRs of Cmax (117.9–167.4) and AUC0–inf (135.1– 182.3) between test (fed) and reference (fasted) were not within the 80% to 125% boundary, demonstrating that bioequivalent standards were not met. Effect of Esomeprazole on the PK of Cabozantinib (Study 2) A total of 22 subjects were enrolled, and 21 completed the study. One subject was withdrawn on day 2 of period 2 after receiving all doses of study medication, because of an adverse event (AE) of vomiting. Plasma concentrations and PK parameters for cabozantinib are provided in Figure 3 and Table 4, respectively. Mean plasma peak concentration (Cmax) and overall exposure (AUC0–inf) of cabozantinib were similar following administration alone or in combination with esomeprazole. The median tmax occurred 1 hour later when coadministered with esomeprazole (median, 4.0 hours) compared with cabozantinib alone (median, 3.0 hours), whereas the terminal t1/2, CL/F, and Vz/F of cabozantinib were comparable when cabozantinib was administered alone or in combination with esomeprazole. Statistical comparisons of cabozantinib PK parameters with and without esomeprazole are provided in Table 5.

Nguyen et al

1297

Figure 2. Mean (SD) cabozantinib plasma concentration–time profiles following administration of a 140-mg FBE capsule of cabozantinib under fed or fasting conditions in healthy subjects (linear axes).

For the 21 subjects who completed both periods of the study, the geometric means for Cmax and AUC0–inf were similar between treatments, with GMR values 110.5% and 109.1%, respectively, for cabozantinib coadministered with esomeprazole relative to cabozantinib alone. The 90%CIs around GMRs were within the limits of 80%–125% for the AUC0–inf parameter; however, the upper 90%CI for Cmax was determined to be 125.1%. Intrasubject variability (%CV) was estimated for AUC0–inf (20%) and Cmax (24%). Safety Summary In both studies, cabozantinib was well tolerated under each treatment condition. Overall, the most common AEs were headache, somnolence, abdominal pain, constipation, and diarrhea. All AEs were mild or moderate in severity, and no serious AEs were reported. In study 1, the

incidence of treatment-emergent adverse events (TEAE; mostly grade 1) was similar following treatment under fed and fasted conditions; 33 subjects (62.3%) reported TEAEs under each treatment condition. In study 2, overall TEAEs were reported by 13 subjects (59%); the incidence of TEAEs was higher for cabozantinib þ esomeprazole than for cabozantinib alone. The majority of TEAEs were of mild severity. In study 2 during period 2, TEAEs occurred while cabozantinib was being administered.

Discussion Bioavailability of medications is affected by many factors such as food, pH, presystemic metabolism, and/or uptake or efflux transporters. Clinical studies that investigate these factors are important to appropriately characterize possible drug–drug and food–drug interactions that could

Table 2. Summary of Pharmacokinetic Parameters for Cabozantinib Following a Single Dose Administration of 140 mg FBE Cabozantinib Under Fed and Fasted Conditions Fed (Test) Parameters Cmax (ng/mL) (n ¼ 47) tmax (h)a (n ¼ 47) AUC0–inf (ng  hr/mL) (n ¼ 46) t1/2 (h) (n ¼ 46) CL/F (L/h) (n ¼ 46) Vz/F (L) (n ¼ 46)

Mean 794 105000 124 2.50 450

SD 283 6.0 [4.0–24.0] 30100 29.7 7.56 1420

Fasted (Reference) CV%

Mean

SD

CV%

36

536

38

29 24 302 315

63200 124 2.38 421

201 4.0 [2.0–24.0] 17500 29.1 0.65 145

28 24 27 35

Cmax, maximum observed concentration; tmax, time to reach Cmax; AUC0–inf, area under the concentration-time curve from time zero to infinity; t1/2, apparent terminal elimination half-life; CL/F, apparent total body clearance (calculated as 140 mg FBE dose/AUC0-inf); Vz/F, apparent total volume of distribution (calculated as 140 mg FBE dose/(AUC0-inf
kel); SD, standard deviation; CV%, coefficient of variation. Fed (Test): 140 mg FBE cabozantinib dose administered 30 minutes after consuming a high-fat meal. Fasted (Reference): 140 mg FBE cabozantinib dose administered under fasting conditions. a Expressed as median [range] for tmax.

The Journal of Clinical Pharmacology / Vol 55 No 11 (2015)

1298

Table 3. Statistical Comparisons of Cabozantinib Pharmacokinetic Parameters Following a Single Dose Administration of 140 mg FBE Capsule Cabozantinib Under Fed and Fasted Conditions Pharmacokinetic Parameter Cmax (ng/mL) (n ¼ 47) AUC0–inf (ng  hr/mL) (n ¼ 46)

Geometric Mean Fed (Test)

Geometric Mean Tasted (Reference)

GMR (%) (Fed/Fasted)

90%CI of Ratio

709 95200

505 60700

140.5 157.0

117.9–167.4 135.1–182.3

GMR, geometric mean ratio; CI, confidence interval; Cmax, maximum observed concentration; AUC0–inf, area under the concentration-time curve from time zero to infinity. Test: 140 mg FBE cabozantinib dose administered 30 minutes after consuming a high-fat meal. Reference: 140 mg FBE cabozantinib dose administered under fasting conditions.

give rise to complications in targeted patient populations. Because cabozantinib exhibits a pH-dependent solubility profile in vitro, a food effect study and a DDI study with esomeprazole were conducted in healthy subjects to evaluate whether factors that may affect gastric pH and cabozantinib solubility may also affect the bioavailability of cabozantinib.

The first study was conducted to evaluate the effect of a high-fat meal on the PK of cabozantinib. Each subject received cabozantinib under fed (high-fat meal) and fasting conditions according to a randomization schedule. Food can affect the PK of a drug via several mechanisms, such as delay in gastric emptying, stimulation of bile flow, physical or chemical interaction with dosage form, changes in

Figure 3. Mean (SD) cabozantinib plasma concentration–time profiles following administration of a 100-mg FBE tablet of cabozantinib with or without esomeprazole in healthy subjects: (a) linear axes; (b) semi logarithmic axes.

Nguyen et al

1299

Table 4. Summary of Pharmacokinetic Parameters for Cabozantinib Following a Single Dose Administration of 100 mg FBE Tablet Cabozantinib With and Without Esomeprazole Period 1 Cabozantinib Alone (Reference) n ¼ 21 Parameters Cmax (ng/mL) tmax (h)a AUC0–inf (ng  hr/mL) t1/2 (h) CL/F (L/h) Vz/F (L)

Period 2 Cabozantinib þ Esomeprazole (Test) n ¼ 21

Mean

SD

CV%

Mean

SD

CV%

647

197 3.0 [2.0–5.0] 14800 29.6 0.56 109

30

706

27

25 25 31 36

64200 123 1.67 283

189 4.0 [2.0–5.0] 17000 39.9 0.48 81.1

58900 117 1.83 305

27 32 29 29

Cmax, maximum observed concentration; tmax, time to reach Cmax; AUC0–inf, area under the concentration-time curve from time zero to infinity; t1/2, apparent terminal elimination half-life; CL/F, apparent total body clearance (calculated as 100 mg FBE dose/AUC0–inf); Vz/F, apparent total volume of distribution (calculated as 100 mg FBE dose/(AUC0–inf
kel); SD, standard deviation; CV%, coefficient of variation. Period 1 (Reference): 100 mg cabozantinib tablet dose administered under fasting conditions. Period 2 (Test): Multiple oral dose of 40 mg esomeprazole delayed-release capsule qd in the morning of Days -5 to -1, a single dose of 40 mg esomeprazole delayed-release capsule on Day 1, and a single dose of 100 mg cabozantinib tablet one hour after esomeprazole dosing on Day 1 under fasting conditions. a Expressed as median [range] for tmax.

gastrointestinal pH, alterations in luminal metabolism, change in blood flow, or interactions of the drug with the food itself. The type of food, including calorie content, nutrient composition (protein, carbohydrate-rich, or highfat meals), volume, temperature of the meal, and fluid ingestion are factors that may modulate the effect. In general, meals that are high in total calories and fat content are more likely to affect the gastrointestinal physiology and thereby result in a larger effect on the bioavailability of a drug.18 Therefore, a high-fat meal was used in this study to maximize the potential for an observable food effect. BCS Class II generally comprises weak acids, weak bases, and lipophilic compounds that typically respond differently to food. For weak acids and bases with high pKa, which represent the majority of the compounds from this class, a meal stimulates bile flow that, in turn, increases secretion of intestinal and pancreatic enzymes (eg, lipase) to enhance dissolution20 and thus drug absorption. Cabozantinib is a BCS Class II weak base and was therefore expected to demonstrate a food effect.

In the present study, the peak and extent of bioavailability (Cmax and AUC0–inf) were increased by 40.5% and 57%, respectively. Food effect studies with other tyrosine kinase inhibitors (TKIs) demonstrated similar impacts of food intake on absorption and exposure. When icotinib was administered after a high-fat and high-calorie meal in healthy subjects, the mean Cmax and AUC values were increased by 59% and 79%, respectively.21 Lapatinib (BCS Class IV) demonstrated a dramatic increase in peak and overall plasma exposures;13 Cmax and AUC values were increased by 3- and 4.25-fold, respectively, when given following a high-fat meal.22 Other TKIs in the BCS Class II family that demonstrated a food effect are erlotinib (AUC values increased by 2-fold under fed vs fasted conditions), nilotinib (AUC values increased by 50% in chronic myeloid leukemia patients and by 82% in healthy subjects), and pazopanib (AUC and Cmax values increased by 2-fold under fed vs fasted conditions).11,23,24 Conversely, no food effect was observed with the TKIs vandetanib or axitinib.25,26

Table 5. Statistical Comparisons of Cabozantinib Pharmacokinetic Parameters Following a Single Dose Administration of 100 mg FBE Tablet Cabozantinib With and Without Esomeprazole Pharmacokinetic Parameter Cmax (ng/mL) AUC0–inf (ng  hr/mL)

Geometric Mean Cabozantinib þ Esomeprazole (Test) n ¼ 21

Geometric Mean Cabozantinib Alone (Reference) n ¼ 21

GMR (%) (Test/reference)

90%CI of Ratio

679 62000

614 56900

110.5 109.1

97.6–125.1 98.0–121.4

GMR, geometric mean ratio; CI, confidence interval; Cmax, maximum observed concentration; AUC0–inf, area under the concentration-time curve from time zero to infinity. Period 1 (Reference): A single oral dose of 100 mg cabozantinib tablet administered under fasting conditions. Period 2 (Test): Multiple oral dose of 40 mg esomeprazole delayed-release capsule qd in the morning of Days -5 to -1, a single dose of 40 mg esomeprazole delayed-release capsule on Day 1, and a single dose of 100 mg cabozantinib tablet one hour after esomeprazole dosing on Day 1 under fasting conditions.

1300 There has been some controversy in recent years regarding product label requirements for dose administration under fasting conditions for a high proportion of oral oncology drugs for which food can substantially affect bioavailability. The arguments against such label requirements for fasting generally center around 2 concepts: (1) that variability in exposure could increase under fasting conditions because of noncompliance with the fasting requirement, and (2) that lower bioavailability with fasting will directly result in higher levels of unabsorbed drug, which has been characterized as poor pharmacoeconomic efficiency and could also lead to higher levels of gastrointestinal toxicity.27–29 Arguments in support of fasting conditions for oral oncology drugs generally focus on the need to control variability in drug absorption from the standpoint of safety and efficacy, especially for drugs with potentially life-threatening toxicities, and/or drugs for which the therapeutic index is an issue.30,31 Such variability in absorption is typically not apparent in the fed versus fasted states in the context of a clinical trial, in which the fed state is tightly controlled with a standardized diet. Instead, the primary concern is with the administration of a drug in the fed state in the context of a community setting, where diets can be highly variable across ethnic groups and geographic regions.31 In addition, it has been posited that the pricing of oral oncology drugs is generally not affected by the cost of manufacture or dosage size, which is inconsistent with the pharmacoeconomic argument against label requirements for fasting.32 The product label for cabozantinib carries a boxed warning for perforations and fistulas and hemorrhage, which are potentially life-threatening AEs. Accordingly, controlling bioavailability and minimizing exposure variability are particularly important safety considerations for this agent. Therefore, because dietary content can greatly influence the extent of the food effect on the bioavailability of some drugs and because the effects of normal or low-fat meals on the PK of cabozantinib have not yet been studied, the product label for cabozantinib indicates that it is not to be taken with a meal.1 Oral bioavailability can be affected by factors other than food, such as gastric pH. Acid-reducing agents are the most commonly prescribed medications in North America and Western Europe, and PPIs are the most commonly prescribed acid-reducing agents in cancer patients (65%–79% of all cancer patients).33 More than 50% of approved orally administered cancer therapeutics are weak bases and are characterized as having pHdependent solubility, including erlotinib, dasatinib, and nilotinib.13,34,35 Thus, there is potential for a DDI when weak base anticancer drugs with pH-dependent solubility are coadministered with agents such as PPIs; the lower drug solubility at higher gastric pH levels following PPI administration may lead to decreases in drug absorption, exposure, and therapeutic effect.

The Journal of Clinical Pharmacology / Vol 55 No 11 (2015)

The second study therefore investigated the effect of change of gastric pH on cabozantinib bioavailability. The selected PPI esomeprazole, the S-enantiomer of omeprazole, is a strong and widely used gastric pH-modifying agent that inhibits gastric acid secretion. In addition, cabozantinib is predominantly metabolized via cytochrome P450 (CYP) 3A with CYP2C9 contributing as a minor pathway.8 Esomeprazole is extensively metabolized in the liver by CYP2C19 and CYP3A, but is not likely to inhibit CYP3A and CYP2C914; therefore, no DDI between cabozantinib and esomeprazole through a CYP-related mechanism is expected. Because esomeprazole may have a prolonged effect on acid secretion due to its irreversible binding to the Hþ/Kþ ATPase pump, a fixed sequence of drug administration was used in this study. If a 2-way crossover design had been employed, subjects might have had different levels of acid secretion during the 2 periods, thereby confounding the results. This risk was removed by dosing all subjects first with cabozantinib alone in period 1, then concomitantly with esomeprazole in period 2. The selected dose of 40 mg once daily for 6 consecutive days for esomeprazole was considered adequate to evaluate the maximal effect of acid suppression on cabozantinib bioavailability15,16,17; the percentage of time over a 24-hour period with gastric pH > 4 for this dosing regimen was shown to be 70%.14 Overall, based on the GMRs, cabozantinib plasma PK parameters (AUC0–inf and Cmax) were 9.1% to 10.5% higher when cabozantinib was coadministered with esomeprazole compared with cabozantinib administered alone. Although the upper 90%CI for Cmax was determined to be 125.1%, the 90%CIs around the GMRs of AUC0–inf parameter were within the limit of 80%–125%, suggesting there was no DDI between cabozantinib and esomeprazole. The upper 90%CI for Cmax of 125.1% of the GMR is not clinically relevant as it represents a minimal risk for resulting in a treatmentemergent toxicity. In addition, a dose adjustment based on a 25.1% increase in Cmax would not be warranted as the intersubject variability of Cmax at steady state is higher (37%–43%). Several published studies demonstrated that bioavailability of some TKIs may be affected by PPIs. For nilotinib, a BCS Class IV TKI weak base, there was a reported decrease in peak and overall exposure by 26% and 34%, respectively, when administered with esomeprazole; however, nilotinib PK parameters were not affected in the presence of famotidine (an H2 blocker) or an antacid.35,36 Omeprazole administration also resulted in approximately 40% reductions in peak and overall plasma exposure of dasatinib, a BCS Class II TKI weak base.35 Concomitant use of a PPI is therefore not recommended for nilotinib and dasatinib.10,12 In contrast

Nguyen et al

plasma exposures for the TKIs imatinib (BCS Class II), crizotinib (BCS Class II or IV), and sorafenib (BCS Class II) showed no clinically meaningful change following esomeprazole administration.35,37 In vitro, cabozantinib demonstrates a pH-dependent solubility profile (0.11 mg/mL in 0.01 N HCl and practically insoluble at pH > 4). Lack of effect of the acid-reducing agent, esomeprazole, on the exposure of cabozantinib in the present study could be due to the low solubility of cabozantinib even in acidified (pH 1–2) gastric fluid; the magnitude of solubility decrease from pH 1–2 to pH > 4 may not result in a change in cabozantinib exposure. This phenomenon was postulated by Smelick et al. and similar observations have also been reported for other weak bases such as darunavir, dronedadone, and sorafenib.33

Conclusion In summary, the present studies provide a more detailed understanding of the factors that may affect cabozantinib oral bioavailability and as a consequence may also affect safety and efficacy in patients receiving cabozantinib. Following a high-fat meal, peak (Cmax) and overall exposure (AUC0–inf) were increased by 40.5% and 57%, respectively. Cabozantinib should be taken on an empty stomach; patients should not eat for at least 2 hours before and at least 1 hour after cabozantinib administration. Coadministration of multiple doses of esomeprazole (40 mg once daily) with a single dose of cabozantinib did not alter cabozantinib plasma exposure. Therefore, concomitant use of cabozantinib and PPIs or weaker gastric pH-altering agents is not contraindicated because of the low risk of a DDI. Acknowledgments

1301

9.

10. 11.

12. 13.

14. 15.

16.

17.

18.

19.

20. 21.

This study was funded by Exelixis, Inc.

References 1. COMETRIQ (cabozantinib) capsules [US prescribing information]; 2012. 2. Yakes FM, Chen J, Tan J, et al. Cabozantinib (XL184), a novel MET and VEGFR2 inhibitor, simultaneously suppresses metastasis, angiogenesis, and tumor growth. Mol Cancer Ther. 2011;10:2298–2308. 3. Exelixis NDA 203756 for COMETRIQ (cabozantinib) 20 mg and 80 mg capsules; 2012. 4. Hart C, De Boer R. Profile of cabozantinib and its potential in the treatment of advanced medullary thyroid cancer. Onco Targets Ther. 2013;6:1–7. 5. Nagilla M, Brown R, Cohen E. Cabozantinib for the treatment of advanced medullary thyroid cancer. Adv Ther. 2012;29:925–934. 6. Jamel M, Turner D, Yang J, et al. Population-based mechanistic prediction of oral drug absorption The AAPS J. 2009;11(2):225–237 7. Burton PS, Goodwin JT, Vidmar TJ, et al. Predicting drug absorption: how nature made it difficult. J Pharmacol Exp Therapeut. 2002;303:889–895. 8. Food and Drug Administration. CDER clinical pharmacology and biophamaceutics reviews[s] for cabozantinib [COMETRIQ];

22.

23. 24. 25. 26.

27. 28.

2012. http://www.accessdata.fda.gov/drugsatfda_docs/nda/2012/ 203756Orig1s000ClinPharmR.pdf US FDA: Center for Drug Evaluation and Research (CDER). Guidance for Industry: Waiver of In Vivo Bioavailability and Bioequivalence Studies for Immediate-Release Solid Oral Dosage Forms Based on a Biopharmaceutics Classification System; 2000. http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm070246.pdf Tasigna (nilotinib) [US prescribing information]; 2014. Ling J, Fettner S, Lum B, Riek M, Rakhit A. Effect of food on the pharmacokinetics of erlotinib, an orally active epidermal growth factor receptor tyrosine-kinase inhibitor, in healthy individuals. Anticancer Drugs. 2008;19(2):209–216. Sprycel (dasatinib) [US prescribing information]; 2014. Budha NR, Frymoyer A, Smelick GS, et al. Drug absorption interactions between oral targeted anticancer agents and PPIs: is pHdependent solubility the Achilles heel of targeted therapy? Clin Pharmacol Ther. 2012;92(2):203–213. Nexium (Esomeprazole magnesium) [US prescribing information]; 2014. Miner P, Katz PO, Chen Y, et al. Gastric acid control with esomeprazole, lansoprazole, omeprazole, pantoprazole, and rabeprazole: a five-way crossover study. Am J Gastroenterol. 2003; 98:2616–2620. Roehss K, Lind T, Wilder-Smith C. Esomeprazole 40 mg provides more effective intragastric acid control than lansoprazole 30 omeprazole 20 mg pantoprazole 40 mg and rabeprazole 20mg in patients with gastro-oesophageal reflux symptoms. Eur J Clin Pharmacol. 2004;60:531–539. Yin OQ, Gallagher N, Fischer D, et al. Effect of the proton pump inhibitor esomeprazole on the oral absorption and pharmacokinetics of nilotinib. J Clin Pharmacol. 2010;50(8):960–967. US FDA: Center for Drug Evaluation and Research (CDER). Guidance for Industry: Food-Effect Bioavailability and Fed Bioequivalence Studies. 2002. http://www.fda.gov/downloads/ regulatoryinformation/guidances/ucm126833.pdf Hauschke D, Steinijans VW, Diletti E, Burke M. Sample size determination for bioequivalence assessment using a multiplicative model. J Pharmacokinet Biopharm. 1992;20(5):557–561. Kimberley AL. Current methods for predicting human food effect. AAPS. J. 2008;10(2):282–288. Liu D, Jiang J, Zhang L, et al. Clinical pharmacokinetics of Icotinib, an anti-cancer drug: evaluation of dose proportionality, food effect, and tolerability in healthy subjects. Cancer Chemother Pharmacol. 2014;73(4):721–727. Koch K, Reddy N, Cohen R, et al. Effects of food on the relative bioavailability of lapatinib in cancer patients. J Clin Oncol. 2009; 27(8): Kang SP, Ratain MJ. Inconsistent labeling of food effect for oral agents across therapeutic areas: differences between oncology and non-oncology products. Clin Cancer Res. 2010;16(17): 4446–4451. Ratain MJ. Flushing oral oncology drugs down the toilet. J Clin Oncol. 2011;29(30):3958–3959. Szmulewitz RZ, Ratain MJ. Vemurafenib oral bioavailability: an insoluble problem. J Clin Pharmacol. 2014;54(4):375–377. Jain RK, Brar SS, Lesko L. Food and oral antineoplastics: more than meets the eye. Clin Cancer Res. 2010;16(17):4305–4307. Todd M, Meyer ML, Charnas R, Acharya M, Molina A. Fast and flawed or scientifically sound: the argument for administering oral oncology drugs during fasting. J Clin Oncol. 2012;30(8):888–894. Needle MN. The virtue of low-flow toilets. J Clin Oncol. 2012;30:890. Smelick GS, Heffron TP, Chu L, et al. Prevalence of acid-reducing agents (ARA) in cancer populations and ARA drug-drug interaction

1302 potential for molecular targeted agents in clinical development. Mol. Pharmaceutics. 2013;10(11):4055–4062. 29. Eley T, Luo F, Agrawal S, et al. Phase I study of the effect of gastric acid pH modulators on the bioavailability of oral dasatinib in healthy subjects. J Clin Pharmacol. 2009;49(6):700–709. 30. Zhang L, Wu F, Lee S, Zhao H, Zhang L. pH-dependent drugdrug interactions for weak base drugs: potential implications for new drug development. Clin Pharmacol Ther. 2014;96(2):266–277.

The Journal of Clinical Pharmacology / Vol 55 No 11 (2015) 31. Yin O, Bedoucha V, McCulloch T, et al. Effects of famotidine or an antacid preparation on the pharmacokinetics of nilotinib in healthy volunteers. Cancer Chemother Pharmacol. 2013;71(1): 219–226. 32. Egorin M, Shah D, Christner S, et al. Effect of a proton pump inhibitor on the pharmacokinetics of imatinib. Br J Clin Pharmacol. 2009;68(3):370–374.