Influence of mild and moderate hepatic impairment on axitinib pharmacokinetics

In order to assist in the development of the clinical study, predictions of the effects of hepatic impairment on the pharmacokinetics of axitinib were conducted using Simcyp Population-Based ADME Simulator version 8.2 SP2. The model for hepatic impairment was developed using parameters obtained from preclinical assessments as well as built-in software assumptions and standard physiological parameter estimates. Plasma concentrations expected after administration of a single 5-mg oral dose of axitinib in subjects with normal hepatic function or mild or moderate hepatic impairment were simulated in ten clinical trials (with ten subjects per clinical trial) under fed conditions using a one-compartmental distribution and first-order absorption model.

Input parameters specific to axitinib estimated in vitro included the apparent intrinsic clearance for each recombinant CYP isoform: CYP3A4 (maximum velocity [V_max]: 9.6 pmol/min/pmol of isoform; Michaelis-Menten constant [K_m]: 4 μM), CYP3A5 (V_max: 1.41 pmol/min/pmol of isoform; K_m: 1.9 μM), CYP2C19 (V_max: 0.11 pmol/min/pmol of isoform; K_m : 5.9 μM), and CYP1A2 (apparent intrinsic clearance: 0.17 μL/min/pmol of isoform); the V_max and K_m for UGT1A1 (132.5 pmol/min/mg of microsomal protein and 44.2 μM, respectively); fraction unbound in plasma (0.010); blood-to-plasma partition coefficient (0.79); fraction absorbed (1.0);—Log₁₀ dissociation constant (4.2); absorption rate (1.98 L/h); and volume under steady state conditions (1 L/kg). Based on observations from a human mass balance study (unpublished data), the renal clearance for axitinib was assigned a value of zero. The Simcyp hepatic-impairment model assumes that the activity of CYP isoforms decreases with worsening hepatic function, specifically, that the activity of CYP3A4/5 is decreased by 42% and 62% in mild and moderate hepatic impairment, respectively, compared with normal hepatic function.

This phase I, open-label, single-dose, parallel-group study assessed the pharmacokinetics, safety, and tolerability of a 5-mg single oral dose of axitinib in subjects with normal hepatic function or mild or moderate hepatic impairment according to the US Food and Drug Administration (FDA) guidelines on hepatic impairment studies [8]. The study was carried out at two centers (Orlando, FL, and Miami, FL) from May 16, 2008, to October 26, 2008.

Participants were male or female subjects aged ≥ 18 years with normal hepatic function or mild or moderate hepatic impairment, according to the Child-Pugh (CP) classification [9, 10]. Subjects were screened for participation within 21 days prior to the first dose of study treatment. Subjects with normal hepatic function were recruited after subjects with mild (CP class A [CP-A] score 5–6) and moderate (CP class B [CP-B] score 7–9) hepatic impairment had completed the study. In all, eight subjects enrolled into each of the three groups. Control subjects with normal hepatic function were matched to subjects with hepatic impairment based first on weight, and then on age and gender. In order to match subjects based on body weight, four of the subjects with normal hepatic function were matched to four subjects in each of the mild and moderate hepatic-impairment groups. Individual subjects to match were those with the lowest (n = 1), highest (n = 1), and median (n = 2) weights (±10 kg) in each hepatic-impaired group. After matching subjects with normal hepatic function by body weight, they were then matched for age (±10 years) and gender to subjects with mild or moderate hepatic impairment.

Subjects with normal hepatic function were required to be healthy, defined as no clinically relevant abnormalities, which were identified by a detailed medical history and full physical examination, including blood pressure, pulse rate, 12-lead electrocardiogram (ECG), and clinical laboratory tests. Liver function tests, albumin, and prothrombin time were all required to be within normal range. Eligible subjects had a body mass index (BMI) of 18–32 kg/m² and a total body weight >50 kg.

Subjects with hepatic impairment were required to have hepatic dysfunction due to hepatocellular disease (not secondary to other diseases) as documented by medical history, physical examination, liver biopsy, hepatic ultrasound, computed tomography scan, or magnetic resonance imaging, and have adequate renal function, determined by a 24-h creatinine clearance >75 ml/min. Hepatic dysfunction was required to be stable, defined as no clinically significant change in disease status within the last 30 days, which was documented by recent medical history, including no worsening of clinical signs of hepatic impairment or no worsening of total bilirubin or prothrombin time >50%. Subjects with hepatic impairment also had to be on a stable treatment regimen or dose of medication, or a fluctuating treatment regimen if the underlying disease was under control, provided sponsor approval was granted prior to the first dose of study treatment. In addition to the general exclusion criteria for all groups outlined below, exclusion criteria specific to subjects with hepatic impairment included any other clinically significant diseases that contraindicated the use of axitinib or may have affected its pharmacokinetics; the use of food or prescription or non-prescription drugs that are potent inducers or inhibitors of CYP3A4/5 within 7 days or 5 half-lives (whichever was longer) prior to dosing; any clinically significant laboratory abnormality, except for those parameters influenced by hepatic impairment; or significant hepatic encephalopathy (grade > 2 portal systemic encephalopathy score and severe ascites and/or pleural effusion).

Exclusion criteria for subjects in all groups included a history of febrile illness ≤ 5 days prior to dosing; any condition, such as gastrectomy, that could affect drug absorption; 12-lead ECG demonstrating corrected QT interval >450 msec in healthy subjects (i.e., normal hepatic function) and >470 msec in subjects with mild or moderate hepatic impairment at screening; a history of regular alcohol consumption (>7 drinks/week for females or >14 drinks/week for males) within 6 months of screening; and the use of prescription or non-prescription drugs (except acetaminophen ≤ 1 g/day), vitamins, or dietary supplements within 7 days or 5 half-lives (whichever was longer) prior to dosing in the normal hepatic-function group. Herbal supplements and hormone replacement therapy were to have been discontinued 28 days prior to dosing. Female subjects of childbearing potential, including those with tubal ligation, were not included in the study.

All subjects were required to abstain from grapefruit-containing products for 7 days prior to and during the study until collection of the final blood sample for pharmacokinetic analysis, strenuous exercise 48 h prior to and during the study, and alcohol and caffeine consumption 24 h prior to and during the study. The use of tobacco- or nicotine-containing products equivalent to ≤ 5 cigarettes/day was permitted throughout the study period.

This study was performed in accordance with ethical principles originating in or derived from the Declaration of Helsinki and in compliance with the International Conference on Harmonization Good Clinical Practice guidelines and applicable local regulatory requirements and laws. An Institutional Review Board approved the protocol, and all participants gave written informed consent. This trial was registered on ClinicalTrials.gov (NCT00692341).

Subjects with hepatic impairment were admitted to the clinical research unit 2 days before dosing (day-1) in order to obtain a baseline 24-h urine collection for accurate assessment of renal function. Subjects with normal hepatic function were admitted the day before dosing (day 0).

Subjects fasted for ≥ 10 h prior to consumption of a standardized meal on day 1. The standardized meal consisted of 30% fat, 15% protein, and 55% carbohydrate, equivalent to 500–700 calories, and represented a typical meal consumed by an oncology patient. Subjects received a 5-mg single oral dose of axitinib administered as a film-coated immediate-release, dry-granulated tablet within 30 min of the start of the meal with 240 ml of water. All subjects were required to remain within the clinical research unit until day 3 and return for follow-up visits on days 5 and 7.

Blood samples (5 ml) were collected in K₃-EDTA anticoagulant tubes at 0, 0.5, 1, 1.5, 2, 2.5, 3, 4, 6, 12, 16, 24, 36, 48, 96, and 144 h after the dose of axitinib was administered on day 1. Separate blood samples (10 ml) in K₃-EDTA anticoagulant were also collected on day 1 (0 and 4 h after dosing) for measurement of protein binding of axitinib in plasma. Blood samples for axitinib protein binding were initially scheduled to be collected 0 and 1–2 h after axitinib dosing (pre- and post-dosing, respectively). However, the observed time to maximum observed concentration of axitinib was delayed in patients with hepatic impairment (occurring 4 h after dosing). Thus, blood samples collected 0 and 4 h after dosing were used to determine protein binding of axitinib in plasma. To protect from photodegradation, axitinib in all blood and plasma samples was protected from visible and ultraviolet light exposure during collection, storage, processing, and analysis. Plasma samples were analyzed for axitinib using a high-performance liquid chromatography with tandem mass spectrometric (LC/MS/MS) detection method (Charles River Laboratories, Shrewsbury, MA) validated for accuracy, precision, sensitivity, selectivity, matrix effects, recovery, and stability in compliance with Pfizer standard operating procedures, which conform to the requirements of the FDA bioanalytical method validation guidance [11], as well as a subsequent bioanalytical white paper generated by the FDA and industry representatives [12]. The standard curve was linear with weighted (1/concentration²) regression and included nine concentrations in duplicate (range 0.5–100 ng/ml). Each individual standard was within 12% of nominal concentration, and inter-assay precision at each standard level was within 4% coefficient of variation (CV). Duplicate quality control (QC) samples at four concentration levels spanning the assay range (1.5, 7.0, 40, and 80 ng/ml) were included in each analytical run to measure assay acceptance. At each QC level, inter-assay accuracy was within 6% of nominal concentration and inter-assay precision was within 6% CV.

Standard axitinib pharmacokinetic parameters were estimated using non-compartmental methods with Pfizer proprietary software (eNCA). Pharmacokinetic parameters assessed were area under the plasma concentration-time curve (AUC) from time zero to infinity \( \left( {{\hbox{AU}}{{\hbox{C}}_{0 - \infty }}} \right) \), AUC from time zero to time of the last measurable concentration (AUC_0-last), maximum observed plasma concentration (C_max), time to C_max, t_1/2, apparent oral clearance (CL/F), and apparent volume of distribution (Vz/F).

Plasma samples were processed to determine axitinib protein binding using an equilibrium dialysis method, and concentrations of resulting unbound and bound axitinib were measured by high-performance LC/MS/MS detection methods (Charles River Laboratories, Shrewsbury, MA) validated in compliance with Pfizer standard operating procedures. Dialysis was performed in a 96-well Teflon dialysis plate (Equilibrium Dialysis Device HTD 96; HTDialysis LLC, Gales Ferry, CT) with wells divided by Spectra/Por®-regenerated cellulose membranes with a molecular weight cut-off of 12,000–14,000 Da (Spectrum® Laboratories, Rancho Dominguez, CA). Plasma samples were dialyzed in duplicate against equal volumes of phosphate-buffered saline (70 mM sodium chloride, 50 mM sodium phosphate, pH 7.4) for approximately 20 h at 37°C with orbital shaking at 80 rpm. Axitinib concentrations in plasma dialysate (representing total bound and unbound axitinib) were measured with the LC/MS/MS assay used for plasma pharmacokinetic samples. Each individual standard was within 10% of nominal concentration and intra-assay precision at each standard level was within 8% CV. At each QC level, intra-assay accuracy was within 9% of nominal concentration and intra-assay precision was within 6% CV.

Axitinib concentrations in buffer dialysate (representing unbound axitinib) were measured using a validated LC/MS/MS assay in K₃-EDTA plasma “water”, an aqueous layer of the supernatant generated by high-speed ultracentrifugation of plasma [13] (440,000 × g for 2.5 h at 37°C) using a Beckman TL100 Micro UltraCentrifuge (Beckman Coulter Inc, Brea, CA). The plasma “water” assay had been developed previously and was considered fit-for-purpose for buffer dialysate study sample analysis due to matrix similarity. The standard curve was linear with weighted (1/concentration²) regression and included nine concentrations in duplicate (range 20–1,000 pg/ml). Each acceptable individual standard was within 13% of nominal concentration, and intra-assay precision at each standard level was within 8% CV (17% CV at the lower limit of quantitation). Duplicate QC samples at three concentration levels spanning the assay range (60, 150, and 800 pg/ml) were included in each analytical run to measure assay acceptance. At each QC level, intra-assay accuracy was within 6% of nominal concentration and intra-assay precision was within 1% CV. In addition, QCs consisting of dialyzed plasma:dialyzed buffer (1:1, v/v) at concentrations corresponding to low and high areas of the standard curve (n = 6 at 60 and 800 pg/ml) were included to determine whether the matrix difference between standards and study samples affected quantitation accuracy. For the mixed plasma:buffer QCs, intra-assay accuracy was within 2% of nominal concentration and intra-assay precision was within 3% CV at each QC level.

Fraction unbound (fu) was calculated as axitinib in buffer dialysate (ng/ml) / (axitinib in plasma dialysate [ng/ml] + axitinib in buffer dialysate [ng/ml]). Mean fu for each sample was used to calculate pharmacokinetic parameters. Unbound axitinib pharmacokinetic parameters assessed included unbound \( {\hbox{AU}}{{\hbox{C}}_{0 - \infty }} \) and AUC_0-last (calculated as \( AU{C_{0 - \infty }} \) * fu and AUC _0-last * fu, respectively), C_max unbound (calculated as C _max * fu), unbound CL/F (calculated as CL/F / fu), and unbound Vz/F (calculated as Vz/F / fu).

Safety assessments included recording of adverse events, ECGs, blood pressure, pulse rate, clinical laboratory tests (including liver and renal function tests, prothrombin time, and partial thromboplastin time), physical examinations, and vital signs. Blood pressure and pulse rate were measured at screening, pre-dose on day 1, and at 1, 2, 16, and 144 h post dose. A single standard 12-lead ECG was obtained at screening, post dose on day 1, and at 1.5, 48, and 144 h post dose.

Sample size was based on recommendations from the FDA Guidance for Industry on hepatic impairment studies [8]. Pharmacokinetic parameters were estimated by using non-compartmental methods as described above, and summarized by descriptive statistics. Using one-way analysis of variance (ANOVA), natural log-transformed pharmacokinetic parameters (\( {\hbox{AU}}{{\hbox{C}}_{0 - \infty }} \), AUC_0-last, and C_max) were compared between the group of subjects with normal hepatic function and each group of subjects with hepatic impairment. Point estimates and 90% confidence intervals (CIs) for the differences between these groups on the log scale were constructed then back-transformed to provide estimates of the ratio of adjusted geometric means (i.e., moderate hepatic impairment/normal function and mild hepatic impairment/normal function) and 90% CIs for the ratios. The differences in pharmacokinetic parameters between subjects with hepatic impairment and with normal hepatic function were evaluated by examination of the ratios and 90% CIs. Relationships between axitinib exposure and laboratory components of CP score, including bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), albumin, and prothrombin time, were assessed graphically and by simple linear regression.