Transcript Young Innovators 2009

YOUNG INNOVATORS 2009
Exploratory Analysis of Possible Clinical and
Pharmacogenetic Patient Covariates in the
Exposure and Safety of Sorafenib Treatments in
Various Types of Solid Tumors
Lokesh Jain, Sukyung Woo, William L Dahut, Elise C Kohn, Shivaani Kummar,
Robert Yarchoan, Giuseppe Giaccone, Jürgen Venitz, William D. Figg
National Cancer Institute, Bethesda, MD
Virginia Commonwealth University, Richmond, VA
ABSTRACT
•
•
Purpose: Sorafenib, a multikinase inhibitor, acts by inhibiting Ras/Raf and VEGFR2
kinases. High inter-patient variability is observed in its systemic exposure, clinical
efficacy and toxicity. We investigated the factors contributing to variability in sorafenib
exposure and treatment-associated toxicities by population PK analysis, exposuretoxicity and genotype-toxicity correlative studies.
Methods: A total of 112 patients, enrolled in 5 phase I/II clinical trials, received oral
doses of sorafenib 400/200 mg BID in 28-day cycles. 24-hour plasma concentration
profiles were measured on Day 1 and (for selected patients) at steady-state. Genetic
variation in metabolic enzymes (CYP3A4*1B, CYP3A5*3C, UGT1A9*3, UGT1A9*5) and
drug target (VEGFR2 H472Q, VEGFR2 V297I) were assessed. Demographic, liver and
kidney function variables were collected at baseline. A population PK model was
developed using log-transformed concentrations with NONMEM VI using FOCEI. Model
validation was performed by visual predictive check (VPC). Possible associations
between sorafenib exposure and toxicity or genetic variation in target (VEGFR2) and
toxicity were assessed, using the highest grades of treatment-related toxicities,
including the hypertension, hand-foot skin reaction (HFSR), rash/desquamation,
diarrhea and fatigue, by Chi-square tests and Kaplan Meier survival analysis.
Young Innovators 2009
ABSTRACT
•
•
Results: Sorafenib PK was adequately described by a one-compartmental model with
enterohepatic circulation (EHC) and first-order elimination. Square wave function was
used for modeling of gall-bladder emptying. Gastrointestinal absorption was described
by a transit-compartment model (n=4). PK parameters were estimated as CL/F=8.05
L/h (between-subject variability (BSV) 19%; inter-occasion variability (IOV) 48%),
V/F=217 L (BSV 68%) and mean absorption transit time=1.98 h. VPC showed that this
structural model appropriately described the data. Frequency of rash for sorafenib
single agent therapy and HFSR in patients receiving sorafenib in combination with
bevacizumab appeared to be related with sorafenib exposures. Patients carrying the
variant allele for VEGFR2 H472Q had higher incidences of grade ≥2 hypertension and
HFSR compared to carriers of wild type allele (p<0.02). VEGFR2 V297I polymorphism
was not associated with toxicity incidence.
Conclusions: Implementation of EHC and transit-compartmental absorption model
appropriately described the observed full PK profiles. Sorafenib exposures appear to be
associated with treatment-related dermatological toxicities. H472Q polymorphism in
VEGFR2 appears to be associated with increased incidence of hypertension and HFSR,
independent of sorafenib exposures.
Young Innovators 2009
OBJECTIVES
• To characterize the pharmacokinetics of sorafenib by
population pharmacokinetic modeling and to evaluate the
effect of demographic, clinical and pharmacogenetic
covariates on sorafenib exposure.
• To evaluate the exposure-toxicity relationship.
• To study the impact of VEGFR2 SNPs on frequency of
treatment-associated toxicities in patients with solid tumors
receiving sorafenib.
Young Innovators 2009
INTRODUCTION
• Sorafenib is an orally administered, cytostatic multi-kinase
inhibitor.
Mechanism of Action
• prevents tumor cell proliferation by targeting the Raf kinase in
Raf/MEK/ERK pathway and inhibits angiogenesis by blocking
the receptor tyrosine kinases such as vascular endothelial
growth factor receptor-2 (VEGFR 2)1.
• indicated for the treatment of advanced renal cell carcinoma
and unresectable hepatocellular carcinoma1.
1. Nexavar® (Sorafenib) tablet prescribing information (2007)
Young Innovators 2009
INTRODUCTION
•
•
•
•
ADME
Metabolized primarily by hepatic CYP3A4 and UGT1A9; both
parent drug and metabolites undergo biliary excretion1.
Subject to GI solubility-limited absorption as evidenced by
less than proportional increase in exposure (AUC) with
escalating doses; a plateau is reached at 600 mg BID2.
Associated with high between-subject variability in
pharmacokinetics as observed in various phase I and phase II
clinical studies2.
Undergoes enterohepatic circulation (EHC), resulting in typical
double peaks in the plasma concentration – time profiles from
patients treated with sorafenib.
1.
2.
Nexavar® (Sorafenib) tablet prescribing information (2007)
Strumberg et al., Oncologist, 2007 Apr;12(4):426-37.
Young Innovators 2009
INTRODUCTION
• The common sorafenib treatment-associated toxicities are1:
Toxicity
Incidences
Hypertension
30-40%
Hand-foot skin reaction
20-30%
Rash: desquamation
25-40%
Diarrhea
35-45%
Fatigue
30-40%
1. Nexavar® (Sorafenib) tablet prescribing information (2007)
Young Innovators 2009
MATERIALS AND METHODS
• Study Design
Cancer
Type
Phase
mCRPC
Phase II
NSCLC
Course
No. of patients
Sample collection time (hr)
ID
SS
C1D1
46
-
0, 0.25, 0.5, 1, 2, 4, 6, 8, 12 & 24
Phase II
C1D1 &
C1D15
18
17
0, 0.25, 0.5, 1, 2, 4, 6, 8, 12 & 24
ST
Phase I
C1D1 &
C2D1
28
12
0, 0.25, 0.5, 1, 2, 4, 6, 8, 12 & 24
CR
Phase II
C1D1
18
--
0, 1, 2, 4, 8, 12, 16 & 24
KS
Phase I
C1D7
-
2
0, 1, 2, 4, 8, 12, 16 & 24
ID: Initial doses, SS: Steady-state, mCRPC: metastatic castrate-resistant prostate cancer, C: Cycle, D: Day, NSCLC:
Non-small cell lung cancer, ST: refractory solid tumors, CR: Colorectal cancer, KS: Kaposi’s sarcoma
Young Innovators 2009
MATERIALS AND METHODS
• Pharmacokinetic Analysis
• Sample analysis
: LC-MS/MS method with an LLOQ and LLOD of 5
and 0.2 ng/mL1
• Software
: NONMEM v6.0 (FOCE INTER)
• Statistical methods
: IIV and IOV – exponential model
Residual variability – proportional and additive model
• Covariate analysis
: Mixed step-wise forward addition (p<0.05) and
step-wise backward elimination (p<0.001)
• Model evaluation
: Visual predictive check & Posterior predictive check
1Jain
L et al, J Pharm Biomedical Analysis 2008 Jan; 46(2): 362-367.
Young Innovators 2009
MATERIALS AND METHODS
• Patient Characteristics (N=111)
Variable
Value
Demographics
Median (Range) or N (%)
Age, years
63.9 (30-85)
BSA, m2
1.9 (1.2-2.5)
Weight, kg
81.4 (35-133)
Gender (F/M)
34 (31%) / 77 (69%)
Race (Caucasian/African-American/Others)
Clinical
90(81%) / 12(11%) / 9(8%)
Median (Range)
Albumin, g/dL
3.6 (2.2-4.4)
Total protein, g/dL
6.6 (4.6-8.0)
Alakaline phosphatase, U/L
82 (34-414)
Bilirubin total, mg/dL
0.6 (0.1-1.7)
SGOT, U/L
26 (13-90)
SGPT, U/L
21 (8-75)
Creatinine clearance, mL/min
Young Innovators 2009
95.3 (26-226)
MATERIALS AND METHODS
• Genetic Variation in Metabolic Enzymes and Drug Target
Genetic Variants
Genotype Frequencies
N (%)
N
Allele Frequencies
(proportion)
Wt
Het
Var
p
q
Metabolic enzymes
CYP3A4*1B
108
89 (82.4)
10 (9.3)
9 (8.3)
0.87
0.13
CYP3A5*3C
108
8 (7.4)
17 (15.7)
83 (76.9)
0.15
0.85
UGT1A9*3
107
103 (96.3)
3 (2.8)
1 (0.9)
0.98
0.02
UGT1A9*5
107
107 (100)
0 (0)
0 (0)
1
0
VEGFR2 H472Q
106
66 (62.2)
35 (33.1)
5 (4.7)
0.79
0.21
VEGFR2 V297I
106
78 (73.6)
25 (23.6)
3 (2.8)
0.85
0.15
Drug Target
Wt: wild-type, Het: heterozygous, Var: variant genotype
p, q are allele frequencies as per Hardy Weinberg Equilibrium nomenclature
Young Innovators 2009
MATERIALS AND METHODS
• Exposure-Toxicity Relationship Analysis
• Patients treated with only sorafenib, sorafenib + bevacizumab and sorafenib +
cetuximab combination, were divided into four groups based on distribution
quartiles of sorafenib systemic exposure (AUC).
• Percent incidences of five common sorafenib treatment-associated toxicities,
hypertension, rash/desquamation, hand-foot skin reaction (HFSR), diarrhea and
fatigue, were compared among exposure quartiles using Chi-squares test.
• Genotype-Toxicity Relationship Analysis
• Association of VEGFR2 genotype with sorafenib treatment-associated toxicities
was assessed by Fisher’s exact test.
RESULTS
• Representative Plasma Concentration-Time Profiles for Sorafenib
1st Dose
2nd Dose
1st Dose
2nd Dose
Young Innovators 2009
Important characteristics:
• Delayed absorption
• Enterohepatic circulation
RESULTS
Key features:
• GI transit compartments (N=4)
• Enterohepatic circulation
• Structural PK Model
GI transit compartments
Gut
Absorption
Compartment (A0)
ka
A1
ka
A2
ka
An-1
ka
An
ka
Central
compartment/
Plasma (Ac.c.)
Fent
Ehc* kEhc
Ka
Ke
Kb
kEhc
Fent
Ehc
A0
An
Ac.c.
Ag.b.
N
ke
(1-Fent)
kb
Gall bladder (A
)
g.b.
: first-order absorption rate constant
: first-order elimination rate constant (=CL/V)
: first-order rate constant for excretion of drug from central compartment to gallbladder
: first-order rate constant for recirculation of drug from gallbladder to absorption compartments
: fraction of dose undergoing entero-hepatic recirculation
: square-wave function (on-off switch, controls emptying at regular intervals)
: amount in absorption compartment
: amount in nth G.I. transit compartment
: amount in central compartment
: amount in gall bladder
: Number of G.I. transit compartment
Young Innovators 2009
RESULTS
• PK Model Equations
dA0
 k a  A0
dt
dAn 1
 k a  ( An1  An2 ) where n= 2, 3,….N
dt
dAn
 k a  ( An  An1 )  Ehc  k Ehc  Ag .b.
dt
dAc.c.
CL
 k a  An  Fent  k b  Ac.c.  (1  Fent ) 
 Ac.c.
dt
V
dAg .b.
 Fent  k b  Ac.c.  Ehc  k Ehc  Ag .b.
dt
(t  DT ) 40
Ehc 
(t  DT ) 40  (t )
where DT is the dosing time
Young Innovators 2009
RESULTS
• Goodness-of-fit Plots: model predictions are reasonably consistent with
2
4
6
8
10
8
6
4
2
6
0
2
4
DV
0
0
•
•
•• ••
• ••••••
•
•
• • •• •
• •• •••
• •• •
•
•
• •• ••
•
• ••
•• • ••
••
• • ••
•
•
••
••
• ••• • •
••
• ••••• ••
••
• • •• •••
• ••
•
• • ••
•• • •
•
••
••••• •••••••••
•
• •••• ••••••• •••••••
• • •• ••• •••••
• ••• •••••••••••••••••••••••••••••••••••••••••••••••••••••••
•
•
• • •••• ••••••••••••••••••••••••••••••••••••••••••••••••••••••• ••••••••••••
••• • • •••••••••••••••••••••••••• •••
••••• •• ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• ••
• • ••••••••••••• • ••
•• ••••••••••••••• • •••••••••••••••••••••••••••••••• • •
•
• • ••• • •• ••••••••••••• •••••
•
•• •••
•••••• •• ••••••••••••
••••• • • • •••• ••• •
• • •••• • • •• •
••
• •• •••••• • •
••
•
• ••• •••
•
• •••• •• ••
•
•
•
• •
DV
8
10
measured concentrations
10
••
•••••••••••••••
•
•
•
•
• ••••••••••••••••••
•••••• •••••••
•• ••••••••••••••••••••••••••••••••••••••••••••••
•
•
•
•
•
•
•
•
•
•
•
• •••••••••••••••••••• ••••
• • • •••••••••••••••••••••••••••••••••••••••••••••••••••••••••• •
•
•
•
•
•
•
• • •••••••••••••••••••••••••
• •••••••••••••••••••••••••••••••••••••••••••••••
•• •••••••••••••••••••••••••••••••••••
•••• •
• • •••••••••••••
•••••••••••••••••••••••••
•••••••••••• ••• ••
•
•• •••••••• ••• ••
••••• ••••• •
•
•
• ••••• ••
•• • • • •
••••• •• ••
•
•
•••• ••••• ••
•••
• •
•••••• ••• •
•
•
•
•
•
• •••• ••••
•
•• •• • •• • •
0
2
4
8
10
IPRED
0
-2
IWRES
IWRES
-1
-3
-4
-2
CWRES
CWRES
0
-2
-4
CWRES
CWRES
0
2
2
1
4
4
2
PRED
6
0
5
10
TSLD
15
TSLD
20
0
5
10
15
TSLD
TSLD
20
25
0
2
4
IPR E
6
IPRED
8
10
RESULTS
• Model Evaluation – Visual Predictive Check (VPC)
After initial doses
At steady-state
0.060
Dose-normalized concentrations
(ng/mL/mg)
0.050
0.050
0.040
0.040
0.030
0.030
0.020
0.020
0.010
0.010
0.000
0.000
-0.010
-0.010
0
4
8
12
16
20
24
0
4
8
12
Time (hr)
Time (hr)
Measured concentrations ( x ). Median and 90% prediction interval for dose-normalized model predicted
(
) and measured (
) concentrations.
Young Innovators 2009
16
20
24
RESULTS
• Parameter Estimates from the Final Model
Parameter
CL/F (L/hr)
V/F‡ (L)
Mean absorption transit time* (hr)
NONMEM estimate
8.05
217
1.98
KEhc (hr-1)
Fent
t′
Correlation CL/F–V/F
Proportional residual error (%CV)
Additive residual error (ng/mL)
0.998
0.50
6.66
0.77
51.4%
1
CL / Fi  CL / F  exp(CL / Fi   CL / Fij )
‡Baseline
IIV (%CV)
18.5
68.7
†61.8
IOV (%CV)
47.8
V / Fi  V / F  (weighti / 81.5)1  exp(V / Fi )
body weight accounted for 4% of IIV in V/F
*Mean absorption transit time = (number of transit compartments+1)/ ka = 5/2.53 = 1.98
†IIV estimated for k
a
t′ : time post-dose administration at which EHC starts
RESULTS
• Exposure – Toxicity Relationship Analysis
Sorafenib single agent
Sorafenib and bevacizumab combination
1st Q (AUC <=23.85 mg/L*h)
2nd Q (AUC >23.85 & <=27.72)
3rd Q (AUC >27.72 & <=31.79)
4th Q (AUC >31.79 & <=41.75)
% patients with toxicity grade ≥2
1st Q (AUC <=43.15 mg/L*h)
2nd Q (AUC >43.15 & <=48.54)
3rd Q (AUC >48.54 & <=55.33)
4th Q (AUC >55.33 & <=99.26)
p=0.004*
80
80
P=0.020*
60
60
40
40
20
20
0
0
Fatigue
Rash
HFSR
Diarrhea
HTN
Young Innovators 2009
Fatigue
Rash
HFSR
Diarrhea
HTN
RESULTS
% patients with HTN grade ≥2
p=0.021*
% patients with HFSR grade ≥2
• Genotype – Toxicity Relationship Analysis
wt
het+var
40
30
20
10
0
N: 15/71 19/45
H472Q
27/85
40
30
20
10
0
9/31
V297I
VEGFR2
*Fisher’s
wt
het+var
p=0.006*
14/71 20/45
24/85 10/31
H472Q
V297I
VEGFR2
exact test
Young Innovators 2009
DISCUSSION & CONCLUSIONS
• A mechanism-based population PK model for sorafenib in a
diverse oncology population was developed, accounting for
known disposition characteristics of sorafenib, such as delayed,
solubility-limited absorption and enterohepatic circulation.
• Baseline body weight was found to be a statistically significant
covariate for volume of distribution, accounting for 4% of IIV.
• None of the studied clinical (liver and kidney function) and
demographic covariates were found to be clinically important.
Young Innovators 2009
DISCUSSION & CONCLUSIONS
• The genetic variation in selected metabolic enzymes
(CYP3A4*1B, CYP3A5*3, UGT1A9*3, and UGT1A9*5) did not
explain the variability in sorafenib disposition, in this
population.
• Model evaluation by post-hoc visual and posterior predictive
checks confirmed that model-predicted concentrations and
systemic exposures were consistent with measuredconcentrations and systemic exposures.
• Incidences of dermatological toxicities appear to be
associated with sorafenib systemic exposures.
Young Innovators 2009
DISCUSSION & CONCLUSIONS
• Incidences of treatment-related hypertension and HFSR
increased to almost double in patients carrying the VEGFR2
H472Q variant allele than wild-type allele.
Young Innovators 2009
ACKNOWLEDGMENTS
VCU
• Jürgen Venitz, M.D., Ph.D
NCI
• William D. Figg, Pharm.D., M.B.A.
• Douglas K. Price, Ph.D.
• Jeanny Aragon-Ching, M.D.
• PI’s for clinical trials
–
–
–
–
–
–
–
–
William Dahut, M.D.
Elise C Kohn, M.D.
Giuseppe Giaccone, M.D.
Heidi Kong, M.D.
Shivaani Kumaar, M.D.
Robert Yarchoan, M.D.
Martin E. Gutierrez, M.D.
Nilofar Azad, M.D.
University of Pisa
• Romano Danesi, M.D., Ph.D.
Dr. Figg’s lab
• Su Woo, Ph.D.
• Erin R. Gardner, Ph.D.
• Tristan Sissung, Ph.D.
Projections Research, Inc.
• Diane R. Mould, Ph.D.
VCU
• Pravin Jadhav, Ph.D. (FDA)
Young Innovators 2009
BIOS/CONTACT INFO
Lokesh Jain, Pre-Doctoral Visiting Research Fellow, National Cancer
Institute, NIH – [email protected], [email protected]
Lokesh Jain is a pre-doctoral fellow in Clinical Pharmacology Program at National Cancer Institute
(NCI). Lokesh received his bachelor of pharmacy (B. Pharm.) degree in 2002 from L. M. College
of Science and Technology, Jodhpur affiliated with Rajasthan University, India. He received his
master of pharmacy (M. Pharm.) degree in 2005 from Birla Institute of Technology and Science,
Pilani, India. In the same year, he joined the Virginia Commonwealth University/NCI joint track
clinical pharmacology Ph.D. program and has been continuing his research under supervision of
Drs. Jürgen Venitz, M.D., Ph.D. and William D. Figg, Pharm. D., M.B.A.
His graduate research is focused on identifying the clinical, laboratory and pharmacogenetic
covariates for efficacy and toxicity of anti-cancer drugs. He has worked on drug development
aspects related to bioanalysis, pharmacokinetic and pharmacogenetic analysis, exposure-genotyperesponse studies as well as translational studies. His research work has been published in peerreviewed journals. He has received several awards during his graduate career, including the AAPS2009 CPTR graduate symposium award, ACCP-2009 student/trainee abstract award, VCU School
of Pharmacy Dean’s award-2009, VCU Department of Pharmaceutics John Wood award-2009 and
Thacker award-2007.
Young Innovators 2009