Using Accelerator Mass Spectrometry to Explain the Pharmacokinetics of Vismodegib Cornelis E.C.A. Hop

Using Accelerator Mass Spectrometry to Explain the Pharmacokinetics of Vismodegib Cornelis E.C.A. Hop

Topics to be Addressed Why AMS? AMS for mass balance studies with vismodegib AMS for absolute bioavailability studies with vismodegib

Vismodegib - Approved by the FDA for the Treatment of BCC in January 2012 MW = 421 pka = 2.8 clogp = 3.9

Human PK Predictions Minimal turnover in microsomes and hepatocytes for all species except cyno; Human Cl hep = 0.5 ml/min/kg Low clearance in rat and dog, but moderate in cyno Predicted human PK: Low clearance: < 1 ml/min/kg Lower volume of distribution: < 1 ml/min/kg Long half life Incomplete absorption: F = F a * F g * F h Rat F = 53%; Dog F = 33 % Based on BDC rat and dog data, F a is 42% and 31%, respectively

Clinical PK of Vismodegib Phase I Cancer Patient Study 5 Graham et al., Clin Cancer Res 2011, 17, 2512 A B C (A) Increase in exposure after single 150 or 270 mg dose with sustained [plasma] (B) (C) (D) Steady state reached quickly; unexpected with 10-14 day half-life Steady state concentration independent of dose Particle size effect on day 1 that was not present at steady state (data not shown)

EOP1 FDA Comments According to 21 CFR 320.25, the bioavailability (fraction absorbed by IV/PO routes) of GDC-0449 should be assessed. Absolute bioavailability study Address the absorption, distribution, metabolism and excretion (ADME study) of GDC-0449 in humans and if needed characterize the effect of renal and or hepatic impairment on the PK of GDC-0449. Mass balance study

But How? Traditional mass balance study: How ethical is it to have healthy volunteers exposed to 50 100 mci for several months? Can you keep them in the clinic that long? The amount excreted each day is likely very small and may not be detectable with liquid scintillation counting Poor mass balance? Can metabolites in plasma, feces and urine be detected?

Mass Balance Studies for Compounds with Long T 1/2 vismodegib Poor mass balance is likely for vismodegib Roffey et al., Drug Metab Rev 2007, 39, 17

But How? Traditional mass balance study: How ethical is it to have healthy volunteers exposed to 50 100 mci for several months? Can you keep them in the clinic that long? The amount excreted each day is likely very small and may not be detectable with liquid scintillation counting Poor mass balance Can metabolites in plasma, feces and urine be detected? Study is likely to fail 14 C tracer study with accelerator mass spectrometry detection!

Accelerator Mass Spectrometer AMS allows quantitative separation of 12 C, 13 C and 14 C from a sample containing a very low level 14 C-tagged drug or biopharmaceutical

Flow Diagram of AMS Procedures for Biomedical Samples Administer nci 14 C-drug Collect blood etc Analyse by AMS Graphitise Combust

Types of AMS Studies in Drug Development Phase 0 Human micro-dosing study Drug candidate selection using human micro-dosing data Investment decisions Phase I/II IV PK, Human Absolute Bioavailability study Intravenous tracer + concomitant oral dose Mass balance and nano-tracing human metabolism study Low dose of 14 C incorporated into the therapeutic oral dose of drug Mass balance and metabolite profiling As an alternative to conventional Human Radiolabel Study? Addresses MIST guidelines To answer specific ADME questions To investigate/mitigate metabolism liabilities & retention issues Mass balance and metabolite profiling

Microtracer Absolute Bioavailability Studies with NCEs Intended for Regulatory Submission 13 Courtesy of Xceleron & Graham Lappin

Cp/Dose AMS Mass Balance Study Design Amount of radioactivity greatly reduced: 1,000 nci It is feasible to send volunteers home after two weeks Hot Cold 150 mg PO on day 1 (unlabeled) + 1,000 nci/10 µg 14 C PO Time (days) How to ensure acceptable mass balance? Continuous sampling until day 14 and interim sampling on days 21, 28, 35, 42, 49, and 56 and interpolation Dose as homogenous suspension

Vismodegib Mass Balance Results Confined D1-D14 Total Recovery = 86.6% Feces Recovery = 82.2% Urine Recovery = 4.4% Renal impairment study not needed - widens exclusion criteria for ongoing trials Need for hepatic impairment study dependent on metabolite profiling

Vismodegib Predominant in Plasma Plasma Concentration (ng or ng equiv per ml) 10000 1000 100 10 1 0.1 Plasma GDC-0449 "cold" (ng/ml) Plasma total radioactivity "hot" (ng equiv/ml) 0 200 400 600 800 1000 1200 1400 Time (hr) GDC-0449 is major drug-species in plasma - no significant circulating metabolites Graham et al., Drug Metab Dispos 2011, 39, 1460

Metabolic Profiling in Feces feces 0-72 72-312 hrs hrs GDC-0449 M3 dpm/g feces homogenate M18 GDC-0449 M1 M13 M3 Mainly vismodegib in circulation Mainly excreted in feces as parent compound (including un-absorbed material) and oxidative metabolites Graham et al., Drug Metab Dispos 2011, 39, 1460

Metabolism of Vismodegib Oxidative metabolism and unique pyridine ring opening Graham et al., Drug Metab Dispos 2011, 39, 1460

Cp/dose Cp/dose AMS Absolute Bioavailability Study Design Dose cold vismodegib orally and 14 C labeled vismodegib intravenously at the T max No concerns about linearity of iv PK because hot and cold present at the same time Single dose and steady state (one week dosing) 500 nci/10 µg 14 C IV dose on day 1 at T max = 2 hr 500 nci/10 µg 14 C IV dose on day 7 150 mg PO on day 1 150 mg PO qd for 7 days Time (days) Time (days)

Graham et al., Br J Clin.Pharmacol 2012, 73, 788 Dose Normalized GDC-0449 Concentration vs. Time Vismodegib Absolute Bioavailability Results Dose normalized Dose normalized GDC-0449 GDC-0449 Plasma Concentration Plasma Concentration (ng/ml/dose) (ng/ml/dose) 1000 1000 100 100 10 10 1 1 0.1 0.1 0.01 Dose Normalized GDC-0449 Concentration vs. Time dose-normalized data 0 200 400 600 800 1000 1200 1400 Time (hours) Time (hours) Absolute bioavailability determined to be 31.8 + 4.6% Oral Mean IV Mean Oral Mean IV Mean 0.01 0 200 400 600 800 1000 1200 1400 Slow IV CL consistent with rate limited systemic elimination (not flip-flop ) Does F and/or CL change with daily dosing? Total clearance increased close to two-fold Absolute bioavailability decreased to 7.4 + 2.5%

Summary of Vismodegib IV PK 0.01 ml/min/kg T 1/2 (d) CL (ml/h) V ss (L) Single dose Multiple dose 13.0 43.4 16.4 +81% 10.3 78.5 26.8 No evidence for CYP induction in human hepatocytes Graham et al., Br J Clin.Pharmacol 2012, 73, 788

Graham et al., Br J Clin.Pharmacol 2012, 73, 788 Plasma Protein Binding Binding to both HSA and AAG, but AAG binding can be saturated The higher concentration at steady state results in a larger free fraction

Summary of Vismodegib IV PK T 1/2 (d) CL (ml/h) CL u (ml/h) V ss (L) V ss,u (L) Single dose 13.0 43.4 13,152 16.4 4,970 Multiple dose 10.3 78.5 9,937 26.8 3,392 The intrinsic clearance and volume of distribution does not change significantly from single to multiple dose Graham et al., Br J Clin.Pharmacol 2012, 73, 788

Summary of Vismodegib IV & Oral PK T 1/2 (d) CL (ml/h) CL u (ml/h) V ss (L) V ss,u (L) Oral AUC (mm.hr) F (%) Single dose Multiple dose 13.0 43.4 13,152 16.4 4,970 2,850 31.8-88% -77% 10.3 78.5 9,937 26.8 3,392 356 7.4 The total oral exposure is reduced by close to eight-fold due to non-sink absorption conditions Graham et al., Br J Clin.Pharmacol 2012, 73, 788

Summary of Vismodegib Human Pharmacokinetics Vismodegib is absorbed slowly after a single oral dose mainly due to poor solubility Terminal elimination half-life is very long after both oral and IV administration due to very slow elimination The CL is incredibly low, 0.01 ml/min/kg Metabolism is via oxidation and pyridine ring opening The V ss of 0.23 L/kg is low, but is still indicative of distribution out of the plasma space. Mean bioavailability is moderate, 31.8%, with little inter-individual variability; micromolar plasma concentrations were achieved with a single 150 mg oral dose At steady state, the absorption and clearance are reduced, but the intrinsic clearance has not changed All of this is driven by very slow clearance, non-sink absorption and non-linear plasma protein binding AMS enabled these studies and absolute bioavailability and mass balance studies using AMS are now a standard part of drug development

Graham et al., Clin Cancer Res 2011, 17, 2512 We Can Rationalize the Clinical PK of Vismodegib! Phase I Cancer Patient Study A B C (A) Increase in exposure after single 150 or 270 mg dose with sustained [plasma] (B) (C) (D) Steady state reached quickly; unexpected with 10-14 day half-life Steady state concentration independent of dose Particle size effect on day 1 that was not present at steady state (data not shown)

Acknowledgements DMPK, Clinical Pharmacology and Clinical Operations colleagues and others on the HH team PRA and Xceleron and all healthy volunteers

History of Accelerator Mass Spectrometry Highly specialised nuclear physics instrument originally developed for carbon dating in the 1970 s It uses very high energies to separate rare isotopes which are then measured with high precision Willard Frank Libby won the Nobel Prize for Chemistry for his method to use carbon-14 for age determination in archaeology, geology, geophysics Approximately 160 AMS instruments worldwide First biological application described in 1989 (LLNL) Xceleron first company in the world to use AMS for biomedical research

Principle of LSC vs AMS Detection LSC -decay of 14 C atom Detected by LSC as photons of light in photomultiplier tube 0.012% of 14 C decays per annum; 1 billion 14 C atoms 1 dpm AMS Atoms separated by differences in mass, charge and energy 12 C, 13 C and 14 C atoms individually counted 1000 14 C atoms required for valid measurement Key Sample containing 12 C 13 C and 14 C atoms

1,000,000 more sensitive than LSC 100,000 more sensitive than MS

Xceleron s 5 MV AMS

Interpolation to Ensure Adequate Radioactivity Recovery for Long Half-Life Drugs Log percent of total excreted per 24 h period 1 0.1 0.01 Up to 168 h all excreta collected AUC 0- = 90.5 % recovery After 168 h, collections over selected 24 h periods 0 500 1000 1500 2000 2500 Time (h) Subjects were administered 3 mg/1 μci of drug X orally