Preclinical Assessment of the Absorption and Disposition of the Phosphatidylinositol 3-Kinase/Mammalian Target of Rapamycin Inhibitor GDC-0980 and Prediction of Its Pharmacokinetics and
Efficacy in Human

Laurent Salphati, Jodie Pang, Emile G. Plise, Leslie B. Lee, Alan G. Olivero, Wei Wei Prior,
Deepak Sampath, Susan Wong, and Xiaolin Zhang
Departments of Drug Metabolism and Pharmacokinetics (L.S., J.P., E.G.P., S.W., X.Z.), Chemistry (A.G.O.), and Translational
Oncology (L.B.L., W.W.P., D.S.) Genentech, Inc., South San Francisco, California
Received April 4, 2012; accepted June 13, 2012

ABSTRACT:

(S)-1-{4-[2-(2-Amino-pyrimidin-5-yl)-7-methyl-4-morpholin-4-yl- thieno[3,2-d]pyrimidin-6-ylmethyl]-piperazin-1-yl}-2-hydroxy-propan- 1-one (GDC-0980) is a potent and selective inhibitor of phosphatidyl- inositol 3-kinase (PI3K) and mammalian target of rapamycin, two key components of the PI3K pathway, the deregulation of which is asso- ciated with the development of many cancers. The objectives of these studies were to characterize the absorption and disposition of GDC- 0980 and assess its efficacy in an MCF7-neo/HER2 human breast cancer xenograft model in immunocompromised mice. Studies in parental Madin-Darby canine kidney cells indicated that GDC-0980 had high permeability (Papp ti 18 ti 10ti6 cm/s), suggesting good absorption potential. However, it was found to be a P-glycoprotein and breast cancer resistance protein substrate in transfected cells and in knockout mice studies. Plasma protein binding was low, with the fraction unbound ranging from 29 to 52% across species. GDC-
0980 hepatic clearance (CL) was predicted to be low in all of the species tested from hepatocyte incubations. The plasma CL of GDC- 0980 was low in mouse (6.30 ml ti minti1 ti kgti1), rat (15.4 ml ti minti1 ti kgti1), and dog (6.37 ml ti minti1 ti kgti1) and moderate in cynomolgus monkey (18.9 ml ti minti1 ti kgti1). Oral bioavailability ranged from 14.4% in monkey to 125% in dog. Predicted human plasma CL and volume of distribution using allometry were 5.1 ml ti minti1 ti kgti1 and 1.8 l/kg, respectively. Parameters estimated from the pharmacokinetic/
pharmacodynamic modeling of the MCF7-neo/HER2 xenograft data indicated that the GDC-0980 plasma concentration required for tumor stasis was approximately 0.5 tiM. These parameters, combined with the predicted human pharmacokinetic profile, suggested that 55 mg once daily may be a clinically efficacious dose. GDC-0980 preclinical characterization and the predictions of its human properties sup- ported its clinical development; it is currently in Phase II clinical trials.

Introduction
The phosphatidylinositol 3-kinase (PI3K)/Akt pathway plays a key role in cellular growth, survival, and differentiation (Engelman et al., 2006). Class I PI3Ks, with catalytic subunits ti, ti, ti, or ti , catalyze the phosphorylation of phosphatidylinositol 4,5-bisphosphate to phospha- tidylinositol 3,4,5-trisphosphate. After this initial step, downstream kinases, among which are the serine/threonine protein kinases Akt and mammalian target of rapamycin (mTOR), in turn are activated. Ab- errant regulation of this pathway has been implicated in several types of cancer, including breast, colon, and prostate (Engelman, 2009).

Up-regulation of PI3K can occur through transforming mutations in the p110ti subunit, loss of function of the phosphatase and tensin homolog, which counteracts the function of PI3K, or in response to receptor tyrosine kinase signaling (Chalhoub and Baker, 2009; Wong et al., 2010). Thus, this pathway has been identified as a promising target for the treatment of cancer, and numerous PI3K inhibitors have entered clinical trials (Ciraolo et al., 2011). In addition, inhibitors of the mTOR kinase, such as the rapamycin analogs (rapalogs) everoli- mus and temsirolimus, have shown activity in renal cell carcinoma and pancreatic tumors (Sabbah et al., 2011). However, mTOR func- tions as two protein complexes, mTORC1 and mTORC2, and the

Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.

http://dx.doi.org/10.1124/dmd.112.046052.

selectivity of rapalogs for mTORC1 does not prevent the phosphor- ylation and activation of Akt, possibly by mTORC2, through a feed- back loop (O’Reilly et al., 2006; Guertin and Sabatini, 2009). Hence,

ABBREVIATIONS: PI3K, phophatidylinositol 3-kinase; A, apical; AUCinf, area under the plasma concentration-time curve extrapolated to infinity; B, basolateral; BDC, bile duct-cannulated; BQL, below quantitation limit; CL, clearance; ER, efflux ratio; GDC-0941, 2-(1H-indazol-4-yl)-6-(4- methanesulfonyl-piperazin-1-ylmethyl)-4-morpholin-4-yl-thieno[3,2-d]pyrimidine; GDC-0980, (S)-1-{4-[2-(2-amino-pyrimidin-5-yl)-7-methyl-4- morpholin-4-yl-thieno[3,2-d]pyrimidin-6-ylmethyl]-piperazin-1-yl}-2-hydroxy-propan-1-one; GF120918, N-(4-[2-(1,2,3,4-tetrahydro-6,7-dimethoxy-2- isoquinolinyl)ethyl]-phenyl)-9,10-dihydro-5-methoxy-9-oxo-4-acridine carboxamide; LSC, liquid scintillation counting; MCT, 0.5% methylcellulose with 0.2% Tween 80; MDCK, Madin-Darby canine kidney; MLP, maximum lifespan potential; mTOR, mammalian target of rapamycin; N.D., not determined; PBS, phosphate-buffered saline; PD, pharmacodynamics; PK, pharmacokinetics; QWBA, quantitative whole-body autoradiography; TV, tumor volume; MDR1, multidrug resistance 1; Bcrp1, breast cancer resistance protein 1.
1785

HO

O

N

O

N

N S
N
FIG. 1. Chemical structure of GDC-0980.

N N

N NH2

it is expected that the inhibition of both mTOR complexes would prevent PI3K pathway activation by this mechanism.
The implication of both PI3K and mTOR in cancer and the results obtained through their separate inhibition have led to the development of compounds able to inhibit both kinases, which are being tested currently (Liu et al., 2009a).
(S)-1-{4-[2-(2-Amino-pyrimidin-5-yl)-7-methyl-4-morpholin-4-yl- thieno[3,2-d]pyrimidin-6-ylmethyl]-piperazin-1-yl}-2-hydroxy- propan-1-one (GDC-0980) (Fig. 1) is a small molecule inhibitor of class I PI3K and mTOR (mTORC1 and mTORC2) that is being developed by Genentech, Inc. for the treatment of various malig- nancies. GDC-0980 was shown to be selective against a large panel of related kinases, including DNA-dependent protein kinase, VPS34, and c2alpha and c2beta (Castanedo et al., 2008), and can be considered equipotent against the four class I PI3K isoforms, with IC values of 0.005, 0.027, 0.007, and 0.014 tiM against
50
p110ti, ti, ti , and ti, respectively. In addition, it is a potent inhibitor of mTOR, with a K of 0.017 tiM (Sutherlin et al., 2011). GDC-
i
0980 also is able to inhibit the proliferation of MCF7-neo/HER2 breast and PC3-NCI prostate cancer cells with IC50 values of 0.24 and 0.12 tiM, respectively, and was efficacious against these tumor cell lines grown as xenografts in immunocompromised mice (Sutherlin et al., 2011; Wallin et al., 2011).
The purposes of the present studies were to assess the absorption and disposition properties of GDC-0980 and to model its efficacy in the MCF7-neo/HER2 (breast, PI3K mutant) xenograft model. The results obtained in this preclinical evaluation were used to predict human GDC-0980 pharmacokinetic (PK) parameters and profile as well as its potential efficacy. When possible, these predictions were compared with the data from the Phase I study.
Materials and Methods
Chemicals and Reagents. GDC-0980 was synthesized at Genentech (South San Francisco, CA) and [14C]GDC-0980 was synthesized at Selcia Limited (Essex, UK). All of the other reagents or materials used in these studies were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise stated.
In Vitro Studies. Madin-Darby canine kidney cell permeability and trans- port studies. The Madin-Darby canine kidney I (MDCKI) cell line, used in the permeability assay, was acquired from the American Type Culture Collection (Manassas, VA). For transport studies, MDR1-MDCKI cells were licensed from the National Cancer Institute (Bethesda, MD), whereas Bcrp1-MDCKII cells were obtained from the Netherlands Cancer Institute (Amsterdam, The Netherlands). The cells were seeded at a density of 1.3 ti 105 cells/ml in Transwell plates (12-well, polyester membrane, 0.4-tim pore size, 1.0 cm2 growth area; Corning Life Sciences, Lowell, MA) and cultured for 5 days at 37°C with 5% CO2 and 95% humidity. GDC-0980 was tested at a concentra-
tion of 10 tiM in the apical-to-basolateral (A-B) and basolateral-to-apical (B-A) directions. In the transfected cells, the studies were conducted in the absence and presence of N-(4-[2-(1,2,3,4-tetrahydro-6,7-dimethoxy-2-iso- quinolinyl)ethyl]-phenyl)-9,10-dihydro-5-methoxy-9-oxo-4-acridine carbox- amide (GF120918) (2 tiM; inhibitor of P-gp) or fumitremorgin C (10 tiM; inhibitor of Bcrp1). The compound was dissolved in transport buffer consisting of Hanks’ balanced salt solution and 10 mM HEPES (Invitrogen, Carlsbad, CA). Transepithelial electrical resistance and lucifer yellow permeability were used to monitor monolayer integrity at the beginning and the end of the experiments, respectively. GDC-0980 was analyzed by liquid chromatogra- phy-tandem mass spectrometry (LC-MS/MS). The apparent permeability (Papp), in the A-B and B-A directions, was calculated as follows: Papp ti (dQ/dt)(1/AC0), where dQ/dt is the rate of compound appearance in the receiver compartment, A is the surface area of the insert, and C0 is the initial substrate concentration at T0. The efflux ratio (ER) was calculated as Papp,B-A/Papp,A-B.
Metabolic stability study in cryopreserved hepatocytes. The metabolic sta- bility of GDC-0980 was evaluated in pooled cryopreserved hepatocytes from CD-1 mice (n ti 10 animals), Sprague-Dawley rats (n ti 3 animals), cyno- molgus monkeys (n ti 3 animals), beagle dogs (n ti 3 animals) (Invitrogen), and humans (n ti 10; Celsis, Baltimore, MD). The cells were resuspended at a density of 0.5 ti 106 cells/ml, and the reactions were initiated with the addition of GDC-0980 at a final concentration of 1 tiM. Samples were incubated at 37°C in 5% CO2 with saturating humidity, and aliquots were sampled at 0, 1, 2, and 3 h. Reactions were quenched with acetonitrile at each time point. The samples were centrifuged at 2000g for 10 min, the supernatant was diluted with water (1:2 ratio), and the percentage of GDC-0980 remaining was determined by LC-MS/MS. With the t ti 0 peak area ratio values as 100%, the in vitro intrinsic clearance (CL) and scaled hepatic CL were determined as described by Obach et al. (1997).
Blood-to-plasma partitioning. The blood-to-plasma partitioning of GDC- 0980 was assessed in pooled whole blood with K2EDTA anticoagulant from CD-1 mice, Sprague-Dawley rats, beagle dogs, cynomolgus monkeys, and humans (Bioreclamation, Inc., Hicksville, NY). Blood from all of the species was obtained from at least three individual donors. GDC-0980 and [14C]GDC- 0980 were added to whole blood at total concentrations of 1, 10, and 40 tiM. Blood samples were incubated at 37°C for 60 min in a shaking water bath. After the incubation, an aliquot of blood was sampled, and the remaining blood was centrifuged to obtain plasma. Radioactivity in plasma and blood was determined using a Packard Tri-Carb 2900TR liquid scintillation counter (PerkinElmer Life and Analytical Sciences, Waltham, MA). The blood-to- plasma ratio was calculated by dividing the measured radioactivity in blood by that measured in plasma. Incubations were performed in triplicate. Parameters are presented as mean ti S.D.
Plasma protein binding. The extent of plasma protein binding of GDC-0980 was determined in vitro, in CD-1 mouse, Sprague-Dawley rat, cynomolgus monkey, beagle dog, and human plasma (Bioreclamation, Inc.), by equilibrium dialysis using a 96-well block (HTDialysis LLC, Gales Ferry, CT). GDC-0980 and [14C]GDC-0980 were added to pooled plasma (n ti 3) at total concentra-

tions of 1, 10, and 40 tiM. These concentrations were selected to cover the range of potential clinical levels and preclinical concentrations measured. Plasma samples were equilibrated with phosphate-buffered saline (PBS) (pH 7.4) at 37°C in 95% humidity and 5% CO2 for 6 h. After the dialysis, radioactivity in plasma and buffer was measured using a Packard Tri-Carb 2900TR liquid scintillation counter. The percentage GDC-0980 unbound in plasma was determined by dividing the radioactivity measured in the postdi- alysis buffer by that measured in the postdialysis plasma and multiplying by 100. Incubations were performed in quadruplicate. Parameters are presented as mean ti S.D.
Brain tissue binding. The extent of brain tissue binding of GDC-0980 was determined in vitro, in CD-1 mouse brain (Bioreclamation, Inc.), using a Rapid Equilibrium Dialysis device (Thermo Fisher Scientific, Waltham, MA). Brain homogenate was prepared by homogenizing 1 g of brain with 3 ml of PBS (pH 7.4) using a BeadBeater (BioSpec Products, Bartlesville, OK). GDC-0980 was added to brain homogenate (n ti 3) to a final concentration of 10 tiM. Brain homogenate samples then were equilibrated with PBS at 37°C in 95% humid- ity and 5% CO2 for 4 h at a shaking speed of 150 rpm. After the incubation, a quenching mixture containing an internal standard, blank PBS, or blank brain homogenate was added to the postdialysis PBS sample and postdialysis brain sample so that the matrices of both sides were equivalent. Samples were centrifuged, and the supernatants were analyzed using LC-MS/MS for GDC- 0980 concentrations. Incubations were performed in triplicate. Calculations of the free fraction were performed as described by Kalvass et al. (2007). Parameters are presented as mean ti S.D.
In Vivo Studies. All of the studies performed were approved by the Institutional Animal Care and Use Committees at Genentech, Inc., Harlan Bioproducts for Science Inc. (Indianapolis, IN), Covance Laboratories Inc. (Madison, WI), Covance Research Products (Princeton, NJ), MPI Research, Inc. (Mattawan, MI), or QPS, LLC (Newark, DE). When GDC-0980 was administered orally, animals were fasted overnight until 4 h postdose.
Pharmacokinetic study in mouse. Twenty-seven female NCr nude mice (Taconic Farms, Germantown, NY) were given a 1 mg/kg intravenous bolus dose of GDC-0980 in 5% dimethyl sulfoxide with 5% cremophor. Three additional groups of 27 mice each received a 1, 5, or 10 mg/kg p.o. dose of GDC-0980 as a 0.5% methylcellulose with 0.2% Tween 80 (MCT) suspension. At the initiation of the study, the mice weighed from 16.5 to 26.9 g. One blood sample of approximately 0.2 ml was collected from each mouse (n ti 3 mice per time point) by terminal cardiac puncture, while the animals were anesthe- tized with isoflurane. Blood samples were collected in tubes containing K2EDTA as the anticoagulant, predose, and at 0.033, 0.16, 0.5, 1, 3, 6, 9, and 24 h after the intravenous administration, and predose and at 0.083, 0.25, 0.5, 1, 3, 6, 9, and 24 h after the oral dose. Samples were centrifuged within 30 min of collection, and plasma was collected and stored at ti80°C until analysis. The concentration of GDC-0980 in each plasma sample was determined by LC- MS/MS analysis.
Pharmacokinetic study in rat. Three days before the study, jugular and femoral vein cannulae were implanted in male Sprague-Dawley rats (Charles River Laboratories, Inc., Wilmington, MA) assigned to the intravenous group, and only jugular vein cannulae were implanted in rats assigned to the oral group. At the initiation of the study, the rats weighed from 262 to 294 g. Three rats were given a single intravenous dose of 1 mg/kg GDC-0980 in 5% dimethyl sulfoxide with 5% cremophor via the femoral vein cannulae. Three additional rats were given a single oral dose of 5 mg/kg GDC-0980 in 80% (w/v) polyethylene glycol 400 in water as a solution. Blood samples (approx- imately 0.2 ml per sample) were drawn from each animal via the jugular vein cannulae predose and at 0.033, 0.083, 0.25, 0.5, 1, 2, 4, 8, and 24 h postdose. Plasma samples were collected and analyzed as described above.
Pharmacokinetic study in monkey. Three male cynomolgus monkeys (Har- lan Bioproducts for Science Inc.) were given GDC-0980 in a crossover study. The two phases were separated by a 7-day washout period. In the first phase of the study, monkeys were given a single intravenous dose of 1 mg/kg of GDC-0980 in 30% hydroxypropyl-ti-cyclodextrin solution via a saphenous vein. In the second phase, the same monkeys each were given a single oral dose of 2 mg/kg GDC-0980 as an MCT suspension via nasogastric intubation. At the initiation of the study, the monkeys weighed from 2.29 to 3.32 kg. Blood samples (approximately 1 ml per sample) were collected predose and at 0.033, 0.083, 0.25, 0.5, 1, 2, 4, 8, 12, and 24 h after the intravenous dose and

predose and at 0.083, 0.25, 0.5, 1, 2, 4, 8, 12, and 24 h after the oral dose. Plasma was isolated within 1 h of blood collection and stored at ti80°C until analysis. Urine was collected overnight predose and 0 to 24 h postdose from the animals receiving the intravenous dose. Urine volume was measured and recorded. A 10-ml aliquot of urine for each sample (when available) was taken from the bulk and was stored at ti80°C until analysis. The concentration of GDC-0980 in each plasma and urine sample was determined using an LC- MS/MS assay.
Pharmacokinetic study in dog. Six male beagle dogs (9.1 to 13.2 kg; Harlan Bioproducts for Science Inc.) were given GDC-0980 in a parallel study. The first group of three dogs received a single intravenous dose of 1 mg/kg GDC-0980 in 30% hydroxypropyl-ti-cyclodextrin solution via a cephalic vein. The second group of three dogs each received a 2 mg/kg p.o. dose of GDC-0980 as an MCT suspension. Blood samples (approximately 3 ml per sample) were collected predose and at 0.033, 0.083, 0.25, 0.5, 1, 2, 4, 8, 12, and 24 h after the intravenous dose and predose and at 0.083, 0.25, 0.5, 1, 2, 4, 8, 12, and 24 h after the oral dose. Blood and plasma were processed as described above.
LC-MS/MS Analysis. GDC-0980 plasma and urine concentrations were determined by LC-MS/MS using a nonvalidated method. After plasma protein precipitation with acetonitrile, the supernatant was injected onto the column, a Gemini C18 column (30 ti 2 mm, 5 tim particle size; Phenomenex, Torrance, CA). A CTC HTS PAL autosampler (LEAP Technologies, Carrboro, NC) linked to a SCL-10A controller with LC-10AD pumps (Kyoto, Japan), coupled with an API 4000 triple quadrupole mass spectrometer (Applied Biosystems/
MDS Sciex, Foster City, CA) were used for the LC-MS/MS assay. The aqueous mobile phase was water with 0.1% formic acid, and the organic mobile phase was acetonitrile with 0.1% formic acid (B). The gradient was as follows: 10% B for the first 0.5 min, increased to 90% B from 0.5 to 2 min, maintained at 90% B for 1.5 min, and decreased to 10% B within 0.1 min. The total run time was 5 min with a flow rate of 0.5 ml/min, and the ionization was conducted in the positive ion mode using the transition m/z 499.3 3 341.1 in atmospheric pressure chemical ionization mode. GDC-0980 retention time was
1.2min. The injection volume was 20 til. The lower and upper limits of quantitation of the assay were 0.005 and 10 tiM, respectively. The internal standard for the plasma and urine assays was the deuterated (d8) analog (dog and monkey studies) or a closely related analog (mouse and rat studies) of GDC-0980.
Pharmacokinetic Analysis. Pharmacokinetic parameters were calculated by noncompartmental methods as described by Gibaldi and Perrier (1982) using WinNonlin (version 5.2; Pharsight, Mountain View, CA). Parameters are presented as mean ti S.D. Bioavailability (F) in monkeys was determined for each animal by dividing the dose-normalized area under the plasma concen- tration-time curve extrapolated to infinity (AUCinf) obtained after each oral dose by the dose-normalized AUCinf obtained after the intravenous dose. In dogs, rats, and mice, F was determined by dividing the dose-normalized AUCinf for each animal dosed orally by the dose-normalized mean AUCinf determined from the animals dosed intravenously (a pooled profile was used in mice). Renal clearance was estimated in monkeys after the intravenous dose by dividing the cumulative amount of GDC-0980 excreted over the 24-h urine collection period by plasma AUC0–24 obtained after the intravenous dose.
Mass balance and routes of elimination in rat. A single oral dose of [14C]GDC-0980 (2 mg/kg; 100 tiCi/kg) was administered to bile duct-intact and bile duct-cannulated (BDC) male and female Sprague-Dawley rats (n ti 3 per sex; Covance Laboratories Inc.). The dose was prepared in MCT. Urine and feces were collected from the bile duct-intact rats in plastic containers surrounded by dry ice predose (overnight for at least 12 h) and at 0 to 8 and 8 to 24 h postdose and at 24-h intervals through 168 h postdose. Urine and bile were collected from BDC animals in plastic containers surrounded by dry ice predose (overnight for at least 12 h) and at 0 to 8 and 8 to 24 h postdose and at 24-h intervals through 120 h postdose. All of the samples were stored at approximately ti70°C before and after analysis.
Urine, bile, cage rinse, and cage wash samples were mixed by shaking, and aliquots were analyzed directly by liquid scintillation counting (LSC). Feces samples were mixed with a sufficient amount of solvent (ethanol/water, 1:1, v/v) to facilitate homogenization, and aliquots were combusted and analyzed by LSC. All of the samples were analyzed in duplicate if sample size allowed.
All of the sample combustions were done in a model 307 sample oxidizer (PerkinElmer Life and Analytical Sciences), and the resulting 14CO2 was

trapped in a mixture of Permafluor and Carbo-Sorb (PerkinElmer Life and Analytical Sciences). Ultima Gold XR scintillation cocktail was used for samples analyzed directly. All of the samples were analyzed for radioactivity in Packard Tri-Carb 2900TR liquid scintillation counters for at least 5 min or 100,000 counts.

TABLE 2
Hepatic clearance of GDC-0980 in multiple species predicted from
hepatocyte incubations
Results reported as mean of duplicate.

Mass balance and routes of elimination in dog. A single dose of [14C]GDC- 0980 at a target dose of 2 mg/kg (20 tiCi/kg) was administered to two male and two female bile duct-intact and two male BDC beagle dogs (MPI Research, Inc., Mattawan, MI). The dose prepared with both radiolabeled and nonradio- labeled GDC-0980 in MCT was administered via oral gavage at a dose volume of 5 ml/kg. Urine and feces were collected in plastic containers (surrounded by dry ice for urine only) predose and at 0 to 8 and 8 to 24 h postdose and at 24-h intervals through 240 h postdose. Bile was collected from BDC animals in
Species

Mouse
Rat Monkey Dog Human
Predicted Hepatic CL
ml ti minti1 ti kgti1
5.1
3.3
9.7
3.8
3.3

plastic containers surrounded by dry ice predose and at 0 to 8 and 8 to 24 h postdose and at 24-h intervals through 168 h postdose. For urine and bile, triplicate aliquots (approximately 0.5 and 0.2 g, respectively) were weighed into glass scintillation vials, and an appropriate amount of Ultima Gold LSC cocktail (PerkinElmer Life and Analytical Sciences) was added before direct analysis by LSC. For feces, the appropriate amount of deionized water was added to the samples to create a 4:1 homogenate. Triplicate aliquots (approx- imately 0.5 g) were weighed into combustion cones and oxidized by combus- tion. The required amounts of Carbo-Sorb E and Permafluor Eti were added to the oxidized aliquots before the analysis by LSC. For cage rinses, triplicate aliquots (approximately 1 g each) were weighed into glass scintillation vials, and appropriate amounts of Ultima Gold LSC cocktail were added before the analyses by LSC. All of the radioactive samples collected were counted by LSC for at least 5 min or 100,000 counts in triplicate, sample size allowing.
Quantitative whole-body autoradiography in rats. The tissue distribution of [14C]GDC-0980-related radioactivity in Sprague-Dawley rats (Hilltop Labo- ratory Animals, Inc., Scottdale, PA) was investigated after a single oral dose of 2 mg/kg (100 tiCi/kg) in MCT using whole-body autoradiography. Animals were euthanized at 1, 8, 24, and 120 h postdose. The carcasses were frozen in hexane dry ice and stored at ti20°C before processing for quantitative whole- body autoradiography (QWBA). Each carcass was embedded, cut into sagittal sections, and mounted for QWBA. Selected sections were exposed to phosphor image screens, and tissue radioactivity concentrations were quantified from the whole-body autoradiograms using an image analysis system [Typhoon 9410 image acquisition system (GE Healthcare, Little Chalfont, Buckinghamshire, UK); MCID image analysis software, version 7.0; GE Healthcare]. Concen- trations of radioactivity were expressed as tiCi/g and converted to microgram equivalents of GDC-0980 per gram of matrix (tig equiv/g) using the specific activity (47.1 tiCi/mg) of the administered formulated [14C]GDC-0980.
Studies in FVBn and Mdr1a/b(ti/ti)/Bcrp1(ti/ti) mice. Male FVBn (wild- type) and Mdr1a/b(ti/ti)/Bcrp1(ti/ti) mice were obtained from Taconic Farms (Germantown, NY). The mice weighed between 19 and 28 g at the beginning of the study. GDC-0980, prepared in MCT, was administered by oral gavage at a dose volume of 10 ml/kg. The animals received 20 mg/kg p.o. GDC-0980. Blood samples (0.2 ml) were collected 1 and 6 h postdose. Samples were taken from three mice at each time point. Each mouse was sampled twice (retro-orbital collection and cardiac puncture). After being
of GDC-0980 were determined by LC-MS/MS. Brains also were collected at
1.2.1and 6 h postdose from three mice at each time point. Each brain was rinsed with ice-cold saline, blotted dry, and cut into two halves along the sagittal plane for GDC-0980 concentration measurement and pharmacodynamic (PD) analysis. Each half of the brain was weighed, then snap-frozen in liquid nitrogen, and stored at ti80°C until analysis. The right halves of the brains were homogenized in three volumes of water. The homogenates were extracted by protein precipitation with acetonitrile containing the internal standard and then analyzed by LC-MS/MS. Brain homogenate concentrations were con- verted to brain concentrations for the calculations of brain-to-plasma ratios. The internal standard for the plasma and brain assays was the deuterated (d8) analog of GDC-0980.
For PD analysis, cell extraction buffer (Invitrogen) containing 10 mM Tris (pH 7.4), 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM NaF, 20 mM Na4P2O7, 2 mM Na3VO4, 1% Triton X-100, 10% glycerol, 0.1% SDS, and 0.5% deoxycholate was supplemented with phosphatase, protease inhibitors (Sigma-Aldrich), and 1 mM phenylmethylsulfonyl fluoride and added to frozen brain biopsies. Brains were homogenized with a small pestle (Konte Glass Company, Vineland, NJ), sonicated briefly on ice, and centrifuged at 20,000g for 20 min at 4°C. Protein concentration was determined using a bicinchoninic acid protein assay (Thermo Fisher Scientific). Proteins were separated by electrophoresis and transferred to NuPAGE nitrocellulose mem- branes (Invitrogen). An Odyssey infrared detection system (LI-COR Biosci- ences, Lincoln, NE) was used to assess and quantify protein expression. PI3K and mTOR pathway markers were evaluated by immunoblotting using anti- bodies against pAktSer473, total Akt, pS6Ser235/236, and total S6 (Cell Signaling Technology, Danvers, MA). Inhibition (%) of pAkt and pS6 in wild-type and triple-knockout mice treated with GDC-0980 was calculated by comparing pAkt and pS6 signal with that measured in untreated mice.
MCF7-neo/HER2 Breast Cancer Xenograft Studies. The MCF7-neo/
HER2 cell line was developed at Genentech, Inc. by expressing human HER2

TABLE 3
Blood-to-plasma ratio of [14C] GDC-0980 in mouse, rat, dog, monkey, and
human whole blood
Results reported as mean ti S.D. (n ti 3).

mixed with K2EDTA (anticoagulant), the blood samples were stored on ice and, within 1 h of collection, were centrifuged at 2000g for 5 min at 2 to 8°C. Plasma was collected and stored at ti80°C until analysis. Total concentrations

Species
Total GDC-0980 Concentrationa
Blood-Plasma Ratio of ti 14CtiGDC-0980

tiM
Mouse 1 0.75 ti 0.016

TABLE 1
Apparent permeability (P ) of GDC-0980 in MDCK, MDR1-MDCK, and
app
Bcrp1-MDCK cells
Results reported as mean of duplicate.

Rat
10
40
1
10
40
0.78 ti 0.049 0.85 ti 0.041 0.97 ti 0.017 0.92 ti 0.045 0.90 ti 0.027

Cell Line
Papp
Papp Ratio
Cynomolgus monkey
1
10
1.1 ti 0.0043
1.2.1.1ti 0.035

A to B B to A B-A/A-B 40 1.0 ti 0.0087
Dog 1 0.94 ti 0.039

MDCKI
MDR1-MDCKI
MDR1-MDCKI ti GF120918

18
1.0
17
10ti 6 cm/s
30
19
19

1.7
19
1.1

Human
10
40
1
10
0.85 ti 0.013 0.83 ti 0.0042
1.3ti 0.015 1.1 ti 0.081

Bcrp1-MDCKII 0.93 6.2 6.7 40 1.1 ti 0.055

Bcrp1-MDCKII ti Fumitremorgin C 8.1 6.0 0.74
a Unlabeled and 14C-labeled GDC-0980.

TABLE 4
Percentage of [14C] GDC-0980 free in mouse, rat, dog, monkey, and human plasma
Results reported as mean ti S.D. (n ti 4).

3 to 7 days, 20 million human breast cancer MCF7-neo/HER2 cells, resus- pended in a 1:1 mixture of Hanks’ balanced salt solution and Matrigel base- ment membrane matrix (BD Biosciences, San Jose, CA), were implanted subcutaneously into the right flank of each mouse. After the implantation, tumors were monitored until they reached a mean tumor volume (TV) of 200

Species
Total GDC-0980 Concentrationa
tiM
Free ti14CtiGDC-0980

%
to 300 mm3. Tumor size and body weight were recorded twice per week during the study. Animal body weights were measured using an Adventurer Pro AV812 scale (Ohaus Corporation, Pine Brook, NJ). Tumor lengths and widths

Mouse

Rat

Dog

Cynomolgus monkey

Human

a Unlabeled and 14C-labeled GDC-0980.
1
10
40
1
10
40
1
10
40
1
10
40
1
10
40
29 ti 3.1 29 ti 2.6 32 ti 1.2 42 ti 3.5
41ti 2.9
42ti 0.94 49 ti 1.8 52 ti 2.4 52 ti 1.1 46 ti 3.8 48 ti 1.5 51 ti 1.9 39 ti 2.8 41 ti 0.29 45 ti 2.6
were measured using Ultra-Cal IV calipers (model 54-10-111, Fred V. Fowler Company, Inc., Newton, MA). Mice were euthanized if body weight loss was ti20% from their initial body weight or if the tumors exceeded 2000 mm3. The mean TV across all of the groups was 226 mm3 at the initiation of dosing. Ten mice were assigned to each dose group.
TV was calculated using the following equation with Excel (version 11.2; Microsoft, Redmond, WA): TV (mm3) ti length ti width2 ti 0.5.
GDC-0980 was administered orally in MCT for 21 days. Animals in the control groups received the vehicle, MCT (100 til). The animals treated with GDC-0980 received the following doses in 100 til of MCT: 0.25, 0.5, 1, 2.5, 5, and 7.5 mg/kg once daily. The TVs at the end of the study for each group were compared with the TVs of the vehicle group.
Pharmacokinetic-Pharmacodynamic Modeling. PK and PK-PD model- ing was performed using SAAM II (SAAM Institute, University of Washing- ton, Seattle, WA).

in the MCF7 human breast cancer cell line (American Type Tissue Collection, Manassas, VA).
Because the tumorigenicity of the MCF7-neo/HER2 cells in mice is estro- gen dependent, estrogen pellets (17ti-estradiol, 0.36 mg/pellet, 60-day release) obtained from Innovative Research of America, Inc. (Sarasota, FL) were implanted into the dorsal shoulder blade area of female NCr nude mice. After
Pharmacokinetic analysis. A one-compartment model with first-order oral absorption was used to fit the mean plasma concentration-time data from female NCr nude mice. The pharmacokinetics of GDC-0980 appeared linear over the range of doses tested, and the three single doses (1, 5, and 10 mg/kg) were fitted simultaneously. The mean estimates of the absorption rate constant (ka), the elimination rate constant (ke), and the apparent volume of distribution

10

1

0.1

0.01

A

10

1

0.1

B

IV, 1 mg/kg
PO, 5 mg/kg (80% PEG400)

0.001

0.0001

IV, 1 mg/kg
PO, 5 mg/kg (MCT)

0.01

0.001

0
4
8
12
Time (hr)
16
20
24
0
4
8
12
Time (hr)
16
20
24

C D

10

1
IV, 1 mg/kg
PO, 2 mg/kg (MCT)
10

1
IV, 1 mg/kg
PO, 2 mg/kg (MCT)

0.1

0.1
0.01

0.001 0.01
0 4 8 12 0 4 8 12 16 20 24
Time (hr) Time (hr)
FIG. 2. Mean plasma concentration-time profiles after intravenous (1 mg/kg) and oral (5 mg/kg in rodents; 2 mg/kg in monkey and dog) administrations of GDC-0980. A, mouse; B, rat; C, cynomolgus monkey; D, dog. Error bars represent S.D.

TABLE 5
Pharmacokinetic parameters (mean ti S.D.) of GDC-0980 in mouse, rat, dog, and monkey after intravenous and oral administrations
Parameters NCr Nude Mice Sprague-Dawley Rats Beagle Dogs Cynomolgus Monkeys
No. of animals 27* (3/time point) 3* 3† 3†
Sex Female Male Male Male
Dose, mg/kg 1 1 1 1
CL, ml ti minti1 ti kgti1 6.30 15.4 ti 2.85 6.38 ti 0.657 19.0 ti 1.31

t
, h
1/2
2.05
3.52 ti 2.39
6.41 ti 0.885
0.558 ti 0.101

Mean residence time, h 3.84 4.38 ti 2.50 7.57 ti 0.780 0.654 ti 0.106

V
, l/kg
ss
1.45
3.78 ti 1.43
2.90 ti 0.397
0.739 ti 0.0678

Renal clearance, ml ti minti1 ti kgti1 Oral administration
N.D. N.D. N.D. 0.405 ti 0.268

No. of animals 27§ (3/time point) 3‡ 3§ 3§
Sex Female Male Male Male
Dose, mg/kg 5 5 2 2

AUC
, tiMtih
inf
26.9
11.8 ti 3.31
13.2 ti 0.545
0.813 ti 0.302

Cmax, tiM 4.65 1.31 ti 0.533 2.71 ti 0.524 0.438 ti 0.184
tmax, h 0.25 1.17 ti 0.764 0.333 ti 0.144 1.67 ti 0.577
F, % 101 106 ti 29.7 125 ti 5.16 22.9 ti 8.56
N.D., not determined.
* Doses prepared in 5% dimethyl sulfoxide and 5% cremophor in water. † Doses prepared in 30% hydroxypropyl-ti-cyclodextrin.
‡ Doses prepared in 80% polyethylene glycol 400 in water.
§ Doses prepared in 0.5% methylcellulose with 0.2% Tween 80.

(V/F) were used to simulate the GDC-0980 plasma concentrations for the modeling of the xenograft efficacy study.
Xenograft efficacy study. An indirect response model, described by differ-

K ti
Kmax ti Cn KCn50tiCn
(2)

ential eq. 1, was fitted to the xenograft efficacy data. d(TV)
where Kmax is the maximum value of K (hti1), C is the plasma concentration of GDC-0980 (tiM), n is the Hill coefficient, and KC50 is the concentration of

dt
ti kng(TV) ti K (TV)
(1)
GDC-0980 where K is 50% of Kmax.
GDC-0980 plasma concentrations in mice were simulated based on the

where t is time (h), TV is the tumor volume (mm3), kng is the net growth rate constant (hti 1), and K is the rate constant (hti 1) associated with the reduction of tumor by GDC-0980. K is described further as eq. 2:
parameters determined in the pharmacokinetic studies, and mean TVs from each dose group were used in the modeling. The ED60 was determined by fixing PD parameter estimates and simulating the daily dose required for a 60% reduction of the TV relative to that of the vehicle control group at the end

A

B

100

80

60

40

20

0

100

80

60

40

20

0

Male

Male

Female

Female
of the study. The plasma concentration needed to achieve tumor stasis was determined when K was equal to kng.
Prediction of Human PK Parameters and Efficacious Doses. Human volume of distribution and plasma CL were predicted using simple allometric scaling and incorporating the “rule of exponents” (Mahmood and Balian, 1996) for CL prediction. In vivo CL and volume of distribution at steady-state (Vss) estimated in preclinical species were scaled as a function of body weight using the following power function (allometric equation): Y ti aWb, where Y is CL ti MLP, or volume of distribution, W is the body weight, and a and b are the allometric coefficient and exponent, respectively. MLP is the maximum lifespan potential, as described by Boxenbaum (1982). Body weights of 0.02, 0.25, 3, and 10 kg were used for mouse, rat, monkey, and dog, respectively. Predicted human CL and Vss were extrapolated assuming a body weight of 70 kg.

100

80

60

40

20

0
Male Female Male BDC

FIG. 3. Recovery of [14C]GDC-0980-related radioactivity after oral administration of 2 mg/kg (100 tiCi/kg) of GDC-0980 to male and female bile duct-intact (A) and BDC (B) Sprague-Dawley rats. Results presented as mean ti S.D. (n ti 3 rat/sex).
FIG. 4. Recovery of [14C]GDC-0980-related radioactivity after oral administration of 5 mg/kg (20 tiCi/kg) of GDC-0980 to male and female bile duct-intact and BDC beagle dogs. Results presented as mean of two animals.

FIG. 5. Whole-body autoradiogram of the radioactivity distribution in a male Sprague-Dawley rat at 1 h after a single oral administration of [14C]-GDC-0980 at a target dose of 2 mg/kg (200 tiCi/rat).

Human PK profile simulations using GastroPlus. The PBPKPlus module of the GastroPlus simulation software (Simulations Plus, Lancaster, CA) was used to predict the PK profile of GDC-0980 in human, based on the physico- chemical properties of the compound, its permeability (determined in MDCK cells and converted to a human Peff of 1.34 ti 10ti4 cm/s), and the predicted CL from allometric scaling incorporating the MLP correction. The solubility profile (at pHs ranging from 2.3 to 10), Log P (2.02), and pKa (3.1 and 4.8) were determined experimentally. The inputs for plasma protein binding (Fup ti 42%) and blood-plasma partitioning (1.1) were those presented in this article. For all of the other parameters, the default settings of GastroPlus were used. The plasma concentration-time profile and PK parameters simulated with a dose of 2 mg (starting dose in Phase I) were compared with those obtained in patients (Wagner et al., 2009; Bendell et al., 2010). The simulated profile was fitted subsequently with a one-compartment model with first-order oral ab- sorption using WinNonlin.
The human dose associated with 60% tumor growth inhibition compared with vehicle control was simulated based on the pharmacodynamic parameters estimated in the MCF7-neo/HER2 breast tumor xenograft model and substi- tuting the mouse PK parameters with the human parameters derived from the simulated human profile. Simulations were performed using SAAM II.
Results
Permeability and Transport in MDCK Cells. The apparent per- meability (P ) of GDC-0980 as assessed in parental MDCKI cells
app
was high (Table 1) and comparable with that determined for meto- prolol, the high P marker used in the same experiment (E. G. Plise,
app
unpublished observations). The ER (P /P ) was low (1.7),
app,B-A app,A-B
suggesting minor involvement of efflux transporter(s) in the B-A direction in this cell line. In contrast, when bidirectional transport was studied in MDR1-MDCK and Bcrp1-MDCKII cells, the ERs were 19 and 6.7, respectively. In the presence of the P-gp and Bcrp1 inhibitors, GF120918 and fumitremorgin C, these ERs were reduced to 1.1 and 0.74, respectively, confirming that GDC-0980 was a substrate of both transporters.
Metabolic Stability Study in Cryopreserved Hepatocytes. The hepatic CL predicted from the metabolic stability after a 3-h incuba- tion in cryopreserved hepatocytes is presented in Table 2. Predicted hepatic CL was low in all of the species tested (ti30% of hepatic blood flow) (Davies and Morris, 1993).
In Vitro Blood-to-Plasma Partitioning, Plasma Protein Binding, and Brain Tissue Binding. The mean blood-to-plasma partitioning of GDC-0980 in mouse, rat, cynomolgus monkey, dog, and human

TABLE 6
Concentrations of radioactivity (microgram equivalents of GDC-0980 per gram of sample) in tissues of male rats after a single oral dose of [14C]GDC-0980 at
2 mg/kg
Time after Compound Administration
pooled whole blood ranged from 0.75 to 1.3 and appeared concentra- tion independent from 1 to 40 tiM (Table 3).
Protein binding for GDC-0980 was low in the five species tested, with the free fraction ranging from approximately 29% in mouse plasma to 52% in dog plasma (Table 4). Protein binding was inde-

Tissue

Blood (in heart) Bile (in duct) Kidney cortex Kidney medulla Liver
Urinary bladder Brain (cerebrum) Brain (cerebellum) Brain (medulla) Spinal cord
Cecum
Large intestine Small intestine
1 h
1.71
77.3
8.73
8.50
23.9
3.52
0.151
BQL
0.165
0.139
13.9
5.70
64.4
8 h
BQL
7.23
0.316
0.360
2.17
0.170
BQL
BQL
BQL
BQL
9.01
3.73
1.09
24 h
BQL
6.34
0.320
0.382
1.95
BQL
BQL
BQL
BQL
BQL
1.62
1.22
0.805
120 h
BQL
BQL
BQL
BQL
0.183
BQL
BQL
BQL
BQL
BQL
BQL
BQL
BQL
pendent of the concentration in the concentration range of 1 to 40 tiM. Binding to mouse brain tissue measured at 10 tiM GDC-0980 was 92.1 ti 2.4%.
Pharmacokinetics of GDC-0980 in Mouse, Rat, Monkey, and Dog. The semilog plots of GDC-0980 plasma concentration versus time for mouse, rat, monkey, and dog after intravenous and oral administrations are presented in Fig. 2. The pharmacokinetic param- eters are presented in Table 5. GDC-0980 had low plasma CL (ti30% of hepatic blood flow) in mice, rats, and dogs (6.30, 15.4, and 6.38 ml ti minti1 ti kgti1, respectively) and a moderate plasma CL (approx- imately 43% of hepatic blood flow) of 18.9 ml ti minti1 ti kgti1 in monkeys. Terminal half-life values ranged from 0.558 h in monkey to

Skeletal muscle
3.750
BQL
BQL
BQL
6.41 h in dog. The V
ss
was moderate to high in all of the species,

BQL, below quantitation limit (0.120 tig equivalent/g tissue). corresponding to approximately 1.1- to 5.6-fold total body water.

TABLE 7
Plasma and brain concentrations in wild-type and knockout mice after oral (20 mg/kg) administration of GDC-0980
Results presented as mean of three animals ti S.D.
Total Free
Mice Time
Brain Plasma Brain-to-Plasma Ratio Brain Plasma Brain-to-Plasma Ratio
h tiM tiM tiM tiM
FVBn 1 0.52 ti 0.055 6.4 ti 1.1 0.082 ti 0.0087 0.042 ti 0.0044 1.9 ti 0.32 0.022 ti 0.0023
Mdr1a/1b(ti/ti)/Bcrp1(ti/ti) 1 9.9 ti 5.2 9.5 ti 4.6 1.0 ti 0.20 0.78 ti 0.41 2.8 ti 1.3 0.27 ti 0.054
FVBn 6 0.83 ti 0.31 10 ti 2.9 0.081 ti 0.0071 0.065 ti 0.024 2.9 ti 0.84 0.021 ti 0.0020
Mdr1a/1b(ti/ti)/Bcrp1(ti/ti) 6 10 ti 2.8 10 ti NC 1.1 ti NC 0.79 ti 0.23 2.9 ti NC 0.30 ti NC
NC, not calculated; n ti 2.

Bioavailability after oral administration ranged from 22.9% in mon- key to 125% in dog. The renal CL of GDC-0980 was assessed in monkey and was negligible, representing ti3% of the plasma CL.
Mass Balance and Routes of Elimination in Rat. The excretion of radioactivity was determined after the administration of a single oral dose of [14C]GDC-0980 (2 mg/kg) to bile duct-intact and BDC male and female Sprague-Dawley rats. [14C]GDC-0980-derived radioactiv- ity was excreted rapidly after oral administration to bile duct-intact male and female rats, primarily within the first 48 h after dosing (87.1 to 88.9% of the administered dose in males and females, respectively; L. Salphati, unpublished observations). Routes and rates of excretion were similar in both males and females. A high percentage of radio- activity was recovered in feces (87.5% in males and 89.1% in females; Fig. 3A). Approximately 42.6 to 41.1% of the administered radioac- tivity was recovered in the bile of male and female BDC rats through 120 h postdose (Fig. 3B). The combined recovery in urine and bile indicated that a minimum of 57 and 50% of the oral dose was absorbed, in males and females, respectively.
Mass Balance and Routes of Elimination in Dog. [14C]GDC- 0980-derived radioactivity recovered in feces accounted for 68.6 and 58.7% of the administered dose in male and female bile duct-intact dogs, respectively. Totals of 12.2 and 14.1% of the dose were recov- ered in urine from bile duct-intact male and female dogs, respectively (Fig. 4). After oral administration of [14C]GDC-0980 to male BDC dogs, the mean fecal, urinary, and biliary recoveries were 17.8, 21.6, and 41.7%, respectively, of the administered dose. The combined biliary and urinary radioactivity recoveries suggested that a minimum of approximately 63% of the oral dose was absorbed.
QWBA in Rats. The tissue distribution of 14[C]GDC-0980 was determined in male Sprague-Dawley rats using whole-body autora- diography at 1, 8, 24, and 120 h after oral administration. Compound- derived radioactivity was absorbed rapidly, with concentrations de- tected in all of the tissues at 1 h (Fig. 5; Table 6). Radioactivity levels were below the limit of quantitation at 24 h postdose for all of the tissues except the gastrointestinal tract and excretory tissues. Tissues with the highest radioactivity included the liver, kidney, and intestine.

A FVBn Control Brain Mdr1a/b(-/-)/Bcrp(-/-) Brain
GDC-0980 (mg/kg) 0 20 0 20
1 hr 6 hr 1 hr 6 hr

pAktSer473
tAkt

pS6Ser235/236
tS6
Actin

FIG. 6. Modulation of pAkt and pS6 in the brains of wild-type and Mdr1a/b(ti/ti)/
Bcrp1(ti/ti) mice after oral administration of GDC-0980 at 20 mg/kg. A, Western blot of mouse brains probed with antibodies against pAkt, total Akt, pS6, total S6, and

B
FVBn Control Brain
Mdr1a/b(-/-)/Bcrp(-/-) Brain
actin. B, quantitation of mean pAkt and pS6 inhibition at 1 and 6 h postdose in wild-type

100
pAkt
pS6
pAkt
pS6
and Mdr1a/b(ti/ti)/Bcrp1(ti/ti) triple- knockout mice. Results are presented as the

mean of three animals ti S.D.

80

60

40

20

01 hr 6 hr

1hr 6 hr

1 hr 6 hr 1 hr 6 hr

High levels of radioactivity measured in the bile suggested that this was a major route of elimination. The lowest levels of radioactivity were measured in the brain and spinal cord (Table 6).
Studies in Mdr1a/b(ti/ti)/Bcrp1(ti/ti) and FVBn Mice. GDC- 0980 was administered orally (20 mg/kg) to FVBn and Mdr1a/b(ti /ti )/Bcrp1(ti /ti ) mice. Concentrations measured in plasma and brains were comparable between 1 and 6 h postdose in each strain of mice. Plasma concentrations were also similar be- tween FVBn and Mdr1a/b(ti /ti )/Bcrp1(ti /ti ) mice at the two time points (Table 7). In contrast, brain concentrations were 12- to 20-fold higher in Mdr1a/b(ti /ti )/Bcrp1(ti /ti ) mice than those in FVBn mice. Likewise, brain-to-plasma ratios were 0.08 in FVBn mice and close to 1 in Mdr1a/b(ti /ti )/Bcrp1(ti /ti ) mice (Table 7). However, the ratio of free brain-to-free plasma concentration was approximately 0.3 (using binding data presented here).
A marked inhibition (ti80%) of the PI3K pathway markers pAkt and pS6 was observed in the brains of triple-knockout mice (Fig. 6). Suppression of pS6 was similar at 1 and 6 h, whereas pAkt levels started to recover by 6 h postdose. Inhibition (20–70%) of the path- way also was observed in the brains of wild-type mice, with the suppression of pS6 lasting up to 6 h (Fig. 6).
PK-PD Modeling. Pharmacokinetic studies. The parameters (% coefficient of variation) k , k , and V/F were estimated for GDC-0980
a e
after single oral administrations to mice of 1, 5, and 10 mg/kg (Fig. 7A) GDC-0980 in MCT and were 0.55 (42) hti1, 0.44 (23) hti1, and 1.18 (32) l/kg, respectively. These estimated parameters were used to simulate plasma concentration-time profiles when modeling the xeno- graft efficacy data, because serial blood samples adequate for mod- eling were not collected from tumor-bearing mice during the study. The comparison between observed plasma concentrations and model predictions is presented in Fig. 7B and indicates that the PK model used described these data appropriately.

Fig. 7C. The inhibition of the MCF7-neo/HER2 tumor growth ap- peared to be dose dependent.
An indirect response model was fitted to the tumor data from all of the doses (eq. 1) and appeared to describe them adequately. The comparison between the observed TVs and the model predictions is presented in Fig. 7D. The estimates of the pharmacodynamic param- eters describing MCF7-neo/HER2 tumor growth and reduction effect by GDC-0980 are presented in Table 8.
Prediction of Human PK Parameters and Efficacious Doses. Human CL and volume of distribution of GDC-0980 were predicted by
p
allometric scaling (Fig. 8) using the estimates of CL and volume of
p
distribution from mouse, rat, monkey, and dog. For CL , the MLP
p
correction was applied as proposed by Mahmood and Balian (1996), based on the value of the allometric exponent in the simple allometry plot (1; L. Salphati, unpublished observations). The predicted CL and vol-
p
ume of distribution were 5.1 ml ti minti1 ti kgti1 and 1.8 l/kg, respectively. The plasma concentration-time profile after oral administration of
the starting dose (2 mg) was simulated using GastroPlus. Parameters used as inputs are presented in this article or corresponded to the default values in the software. Plasma CL predicted by allometry was assumed to be equivalent to hepatic CL; thus, hepatic CL was set to 5.1 ml ti minti1 ti kgti1 in GastroPlus (21.4 l/h for 70 kg human). The simulated and the average concentrations from patients (Wagner et al., 2009; Bendell et al., 2010) are presented in Fig. 9. The AUC
0 –24, Cmax, and T1/2 estimated from the GastroPlus simulation (0.13 tiMtih, 0.017 tiM, and 4.4 h, respectively) are presented in Table 9, along with the actual values obtained in patients in Phase I (Wagner et al., 2009). The fraction absorbed (F ) and the bioavailability (F) predicted
a
by GastroPlus were 99 and 74%, respectively.
The simulated profile was fitted using a one-compartment model with first-order absorption. The estimated parameters from that fit were substituted for the mouse parameters in the PK-PD model described here and used to simulate human doses expected to achieve

Xenograft efficacy studies. GDC-0980 was administered orally,
efficacy. The human daily dose and AUC
0–24
associated with 60%

daily, to MCF7-neo/HER2 tumor-bearing mice at doses ranging from 0.25 to 7.5 mg/kg. Tumor growth inhibition results are presented in
tumor growth inhibition compared with vehicle on day 21 were 55 mg and 3.8 tiMtih, respectively. This dose was determined by iteration,

A

100

10

1

0.1

0.01

1mg/kg 5 mg/kg 10 mg/kg

B

12

10

8

6

4

y = 0.83x R² = 0.92

0.001 0.0001
0

4

8

12 16 20 24
Time (hr)

2

0

0 2 4 6 8 10 12
Observed GDC-0980
FIG. 7. Plasma concentration-time profiles in nude mice and xenograft studies in MCF7-neo/HER2 tumor-bearing mice. A, concentration-time profiles after oral ad- ministration of GDC-0980 at 1, 5, and 10

C

3000

2500

2000

1500

1000

500

0

0 96 192 288 384 480 576
D

2100
1800
1500
1200
900
600
300
0
Plasma Concentration (µM)

y = 1.03x R² = 0.93

0 300 600 900 1200 1500 1800 2100
mg/kg. B, observed versus predicted plasma concentrations obtained by fitting a one- compartment model with first-order absorp- tion. C, TVs versus time after oral admin- istration of GDC-0980 daily at various doses. D, observed versus predicted TVs obtained by fitting of an indirect response model.

Time (hr) Observed Tumor Volume (mm3)

TABLE 8
Pharmacodynamic parameters estimated from MCF7-neo/HER2 tumor xenograft studies

ti40% of the radioactivity recovered (Fig. 4). Combined with the recovery in urine (21.5%), these results indicated that approximately 63% of the oral dose was absorbed. In both species, the calculated

Parameter

k , hti1 ng
K , hti1 max
KC , tiM
50
n
ED , mg/kg
60
C , tiM
stasis
Estimate

0.0042 0.020 2.0
1(fixed) 1.1
0.52
Coefficient of Variation
%
2.7
11
25
bioavailability (F) was approximately 100%, whereas the fraction absorbed did not appear to exceed 63% based on the recovery of radioactivity in bile and urine. Biliary secretion is a major route of elimination for GDC-0980, and it is possible that it undergoes enterohepatic cycling, after elimination either as the parent com- pound or as a conjugate cleaved in the intestine, allowing GDC- 0980 to be recycled. The reabsorption of the parent compound could contribute to increased AUC and thus F. Indeed, a secondary “hump” can be observed after oral administration in rats and dogs

through successive simulations of different doses; the target of 60% tumor growth inhibition was reached at 55 mg.
Discussion
The PI3K pathway, which is commonly altered in numerous can- cers (Liu et al., 2009a), has been identified as a promising target for the treatment of malignancies. Components of this signaling pathway, such as PI3K or mTOR, can be targeted by small molecules, and several inhibitors specific for one of these two kinases are being evaluated in patients (Engelman, 2009). Inhibitors of mTOR (mTORC1), referred to as rapalogs, have been approved for the treatment of cancers, and numerous dual PI3K/mTOR inhibitors are currently in clinical trials (Benjamin et al., 2011).
Preclinical pharmacokinetic studies provide key parameters during the discovery and the early stages of development of a compound. These parameters help to assess whether sufficient exposure can be achieved in further animal studies (efficacy and toxicology) and also are used to predict pharmacokinetic properties in humans. In addition, when integrated with efficacy data in PK-PD models, they can be used to estimate efficacious and target doses in humans (Gibbs, 2010).
(Fig. 2, A and D), which may be indicative of enterohepatic cycling.
The QWBA study in rats also indicated that GDC-0980 distributed rapidly and extensively into tissues, which was consistent with the moderate volume of distribution measured in the rat PK study. How- ever, among all of the tissues, the levels of radioactivity in the central nervous system were the lowest. This can be attributed to the effect of efflux transporters limiting access of GDC-0980 to the brain. Indeed, despite its high permeability in parental MDCK cells, studies in transfected cells indicated that GDC-0980 was a substrate for P-gp and Bcrp1 (Table 1). These results also were consistent with the poor brain penetration observed in wild-type mice and the marked increase (ti20-fold) in brain concentration achieved in the P-gp- and Bcrp1- deficient mice (Table 7). Although the high levels of GDC-0980 in the brains of the knockout mice led to pronounced PI3K pathway sup- pression, for up to 6 h, inhibition of the pathway was also detected in the brains of wild-type animals (Fig. 6) despite much lower concen- trations. The effects observed were consistent with the free brain concentrations determined in both strains (ti0.8 tiM in the knockout and ti0.06 tiM in the wild-type mice), which were greater than the

GDC-0980, a PI3K/mTOR inhibitor being developed as an anticancer
IC
50
(or K
i
) for each of the PI3K isoforms and mTOR (0.005–0.027

agent, had low plasma CL (and blood CL based on blood-to-plasma ratio) in the mouse, rat, and dog and moderate CL in monkey, compared with hepatic blood flow (Davies and Morris, 1993). Renal CL was a minor route of elimination in monkeys (Table 4), dogs, and rats (Figs. 3 and 4). In contrast, the [14C]GDC-0980-associated radio- activity measured in the bile of BDC rats represented approximately 42% of the dose administered (Fig. 3B), whereas almost all of the dose administered to intact rats was recovered in the feces (Fig. 3A). The combined recovery in urine and bile indicated that 50 to 57% of the oral dose was absorbed. The large contribution of the biliary route in the elimination of GDC-0980 was supported by the results from the rat QWBA study, where high levels of radioactivity were measured in the bile. Likewise, in the dog, biliary elimination was prominent, with
tiM) (Sutherlin et al., 2011). Similar findings with pAkt were reported with another PI3K inhibitor, 2-(1H-indazol-4-yl)-6-(4-methanesulfo- nyl-piperazin-1-ylmethyl)-4-morpholin-4-yl-thieno[3,2-d]pyrimidine (GDC-0941) (Salphati et al., 2010a). However, in addition to this marker, potent inhibition of mTOR by GDC-0980 (Wallin et al., 2011) led to the prolonged suppression of pS6, a downstream effector of mTOR kinase (Fig. 6). Inhibition of PI3K in the brain, through the administration of inhibitors of this pathway in combination with P-gp/BCRP inhibitors or the design of compounds circumventing these transporters, could have implications for the treatment of central nervous system tumors relying on the activation of this pathway, such as glioblastoma (Holand et al., 2011). The fairly constant brain-to- plasma ratio observed between 1 and 6 h suggested that by 1 h a

A B
100000 CL x MLP = 95•W1.38 Human (predicted) 1000 Vss = 1.85•W0.99

R2 = 0.97 R2 = 0.93
Human (predicted)

10000
100

Dog
1000 Monkey Dog
10 FIG. 8. Predicted human clearance (A) and
100 volume of distribution (B) using allometric

Rat 1 Rat Monkey
scaling.

10

1
0.1

Mouse Mouse

0.1
0.01
0.1
1
10
100
0.01
0.01
0.1
1
10
100

Body Weight (kg)
Body Weight (kg)

10

1

0.1

0.01
FIG. 9. GastroPlus simulation of plasma concentration-time profile at the starting dose (2 mg) and actual concentrations (tiS.D.) measured in patients (n ti 3) in Phase 1 (Bendell et al., 2010) and simulation at the dose predicted to be associated with 60% tumor growth inhibition.

0.001

0.0001
GastroPlus Simulation (55 mg)
Mean concentration (n=3 patients; 2 mg) GastroPlus Simulation (2 mg)

0 4 8 12 16 20 24
Time (hr)

steady-state condition had been reached between brain and plasma. In addition, the similar plasma concentrations measured in the wild-type and triple-knockout mice implied that the two transporters had min- imal impact on systemic exposure to GDC-0980, as has been reported for other substrates of both P-gp and Bcrp1 (Polli et al., 2009; Salphati et al., 2010a). It is worth noting that although GDC-0980 brain penetration was improved markedly in Mdr1a/b(ti/ti)/Bcrp1(ti/ti) mice, the ratio of free brain-to-free plasma concentrations was ap- proximately 0.3, not approaching 1 as would be expected if free drug was in equilibrium between brain and plasma. Efflux at the blood-brain barrier by transporters other than P-gp and Bcrp1 may explain this result; experimental variability in the estimation of the free fraction in brain or plasma also could contribute to the calculation of this lower than expected free brain-to-plasma ratio. The in vitro determination of binding to brain homogenate nevertheless has been proposed as a reliable method to estimate free brain concentration (Liu et al., 2009b).
Plasma protein binding was low, not exceeding 71% in all of the species over the range of concentrations tested (Table 4), which appeared consistent with the moderate volume of distribution across species (ti1–5.6-fold total body water) (Davies and Morris, 1993).
well as strong in vitro-in vivo correlations in preclinical species, and convergent predictions with different methods provide more confi- dence. GDC-0980 was generated from the same chemical series as GDC-0941 (Folkes et al., 2008), a compound that had previously entered clinical trials. The evaluation of the different human predic- tion methods used with that compound (Salphati et al., 2010b) was informative and helped build confidence in the approach selected with GDC-0980. Although the allometry predicted plasma CL and the extrapolation of in vitro hepatocyte stability data predicted hepatic blood CL, these could be compared directly because the blood-to- plasma ratio was 1.1.
The plasma concentration-time profile of GDC-0980 for the start- ing dose in human (2 mg) was predicted using GastroPlus and the PBPKPlus module. In this simulation, human CL predicted by inter- species scaling (5.1 ml ti minti1 ti kgti ) was conservatively chosen as an input for hepatic CL (21.4 l/h; 70 kg human). The PK parameters (AUC, Cmax, and T1/2) derived from the simulation were similar to the values obtained in patients in Phase I (Table 9) (Wagner et al., 2009), and the simulated profile (Fig. 9) described the actual plasma con- centrations very closely (Wagner et al., 2009; Bendell et al., 2010).

The predicted human CL of GDC-0980 was 5.1 ml ti minti1 ti kgti1 The Vss estimated by PBPKPlus based on the GDC-0980 physico-

using MLP as a correction factor. This predicted human plasma CL was comparable with the hepatic CL of 3.3 ml ti minti1 ti kgti 1 projected from hepatocyte stability data (Table 2). Plasma CL mea- sured in preclinical species was overall consistent with in vitro pro- jections (Tables 2 and 5), and the narrow range of predictions obtained based on in vitro results and allometry provided some confidence that the human plasma CL would be low. Prediction of human CL remains challenging however, and among the numerous methods evaluated recently (Ring et al., 2011), it did not appear that one was notably more accurate. The use of physiology-based pharmacokinetic models for the prediction of plasma concentration-time oral profiles also resulted in fairly poor performance (Poulin et al., 2011). In general, a thorough characterization and understanding of the preclinical PK, as
chemical properties and input from in vitro studies also was compa- rable with that extrapolated from preclinical data by allometry (1.9 versus 1.8 l/kg). To estimate efficacious doses and exposures in human, the parameters obtained from the fit (one-compartment model, first-order absorption) of this simulated profile were substituted for the mouse parameters in the PK-PD model described here. This approach assumes a similar relationship between PD marker response and efficacy in human and mouse xenograft models as well as com- parable drug distributions in xenograft and human tumor. Despite these caveats, the prediction of efficacy from PK-PD models of xenograft data can be useful in the design of clinical trials and can help establish PK and PD targets. This method has been applied successfully in predicting targets for human efficacy (Simeoni et al.,

TABLE 9

Predicted (GastroPlus) and observed AUC, C
max
, and T
1/2
in human after administration of a 2-mg oral dose and predicted human dose and parameters associated

with 60% tumor growth inhibition
Observed data reported as mean ti S.D. (n ti 3) (Wagner et al., 2009).
AUC0–24 Cmax T1/2
Dose
Predicted Observed Predicted Observed Predicted Observed
tiMtih tiM h
2mg 0.13 0.17 (0.06) 0.017 0.019 (0.005) 4.4 6.4 (2.6)
55 mg 3.8 0.43 4.4

2004; Tanaka et al., 2008). The GDC-0980 human daily dose and exposure predicted to be associated with 60% tumor growth inhibition were 55 mg and 3.8 tiMtih, respectively, based on the MCF7-neo/
HER2 xenograft model (PI3K mutant). This target tumor growth inhibition was selected, as proposed by Wong et al. (2011), as the minimum likely to correspond to clinical response. The predicted efficacious dose appeared to be consistent with the results reported by Bendell et al. (2010), showing some clinical response with 50 mg (AUC ti5 tiMtih) GDC-0980 dosed daily for 21 days in a 28-day cycle of treatment. Although such consistency is promising, the validity of the preclinical efficacy models also depends on their relevance; the work and predictions presented here were based on the MCF7-neo/
HER2 xenograft model (PI3K mutant). Efficacy studies and predic- tions using models with a different mutational status (e.g., PI3K wild-type, phosphatase and tensin homolog null) also may help to define a range of potential efficacious doses.
In summary, the preclinical absorption, distribution, metabolism, and excretion properties of GDC-0980 provided the basis for further clinical evaluation, and the use of PK and PK-PD simulations in early development helped to assess its potential for clinical efficacy. GDC- 0980 is currently in Phase II clinical trials.

Acknowledgments
We thank the Drug Metabolism and Pharmacokinetics, Chemistry, and Small Molecule Pharmaceutical Sciences Departments and the In Vivo Studies Group at Genentech for their contributions to the results presented.

Authorship Contributions
Participated in research design: Salphati, Pang, Plise, Lee, Olivero, Prior, Sampath, and Wong.
Conducted experiments: Plise, Lee, Prior, and Wong. Contributed new reagents or analytic tools: Zhang.
Performed data analysis: Salphati, Pang, Plise, Lee, and Wong.
Wrote or contributed to the writing of the manuscript: Salphati and Pang.

References
Bendell JC, Wagner AJ, Dolly S, Morgan JA, Papadatos-Pastos D, Ware JA, Mazina KE, Lauchle J, Burris H, and De Bono J (2010) A first-in-human phase I study to evaluate the dual PI3K/mTOR inhibitor GDC-0980 administered QD in patients with advanced solid tumors or non-Hodgkin’s lymphoma (NHL). European Society for Medical Oncology (ESMO) Meeting; 2010 Oct 8–12, 2010; Milan, Italy. Abstract 4960. ESMO, Viganello-Lugano, Switzerland.
Benjamin D, Colombi M, Moroni C, and Hall MN (2011) Rapamycin passes the torch: a new generation of mTOR inhibitors. Nat Rev Drug Discov 10:868–880.
Boxenbaum H (1982) Interspecies scaling, allometry, physiological time, and the ground plan of pharmacokinetics. J Pharmacokinet Biopharm 10:201–227.
Castanedo G, Dotson J, Goldsmith R, Gunzner J, Heffron T, Mathieu S, Olivero A, Staben S, Sutherlin DP, Tsui V, et al. (2008) Phosphoinositide 3-kinase inhibitor compounds and methods of use. World patent WO/2008/073785. 2008 June 19.
Chalhoub N and Baker SJ (2009) PTEN and the PI3-kinase pathway in cancer. Annu Rev Pathol 4:127–150.
Ciraolo E, Morello F, and Hirsch E (2011) Present and future of PI3K pathway inhibition in cancer: perspectives and limitations. Curr Med Chem 18:2674–2685.
Davies B and Morris T (1993) Physiological parameters in laboratory animals and humans. Pharm Res 10:1093–1095.
Engelman JA (2009) Targeting PI3K signalling in cancer: opportunities, challenges and limita- tions. Nat Rev Cancer 9:550–562.
Engelman JA, Luo J, and Cantley LC (2006) The evolution of phosphatidylinositol 3-kinases as regulators of growth and metabolism. Nat Rev Genet 7:606–619.
Folkes AJ, Ahmadi K, Alderton WK, Alix S, Baker SJ, Box G, Chuckowree IS, Clarke PA, Depledge P, Eccles SA, et al. (2008) The identification of 2-(1H-indazol-4-yl)-6-(4- methanesulfonyl-piperazin-1-ylmethyl)-4-morpholin-4-yl-thieno[3,2-d]pyrimidine (GDC- 0941) as a potent, selective, orally bioavailable inhibitor of class I PI3 kinase for the treatment of cancer. J Med Chem 51:5522–5532.

Gibaldi M and Perrier D (1982) Pharmacokinetics, Marcel Dekker, New York.
Gibbs JP (2010) Prediction of exposure-response relationships to support first-in-human study design. AAPS J 12:750–758.
Guertin DA and Sabatini DM (2009) The pharmacology of mTOR inhibition. Sci Signal 2:pe24. Ho¨land K, Salm F, and Arcaro A (2011) The phosphoinositide 3-kinase signaling pathway as a
therapeutic target in grade IV brain tumors. Curr Cancer Drug Targets 11:894–918. Kalvass JC, Maurer TS, and Pollack GM (2007) Use of plasma and brain unbound fractions to
assess the extent of brain distribution of 34 drugs: comparison of unbound concentration ratios to in vivo p-glycoprotein efflux ratios. Drug Metab Dispos 35:660–666.
Liu P, Cheng H, Roberts TM, and Zhao JJ (2009a) Targeting the phosphoinositide 3-kinase pathway in cancer. Nature Rev Drug Discov 8:627–644.
Liu X, Van Natta K, Yeo H, Vilenski O, Weller PE, Worboys PD, and Monshouwer M (2009b) Unbound drug concentration in brain homogenate and cerebral spinal fluid at steady state as a surrogate for unbound concentration in brain interstitial fluid. Drug Metab Dispos 37:787– 793.
Mahmood I and Balian JD (1996) Interspecies scaling: predicting clearance of drugs in humans. Three different approaches. Xenobiotica 26:887–895.
O’Reilly KE, Rojo F, She QB, Solit D, Mills GB, Smith D, Lane H, Hofmann F, Hicklin DJ, Ludwig DL, et al. (2006) mTOR inhibition induces upstream receptor tyrosine kinase signal- ing and activates Akt. Cancer Res 66:1500–1508.
Obach RS, Baxter JG, Liston TE, Silber BM, Jones BC, MacIntyre F, Rance DJ, and Wastall P (1997) The prediction of human pharmacokinetic parameters from preclinical and in vitro metabolism data. J Pharmacol Exp Ther 283:46–58.
Polli JW, Olson KL, Chism JP, John-Williams LS, Yeager RL, Woodard SM, Otto V, Castellino S, and Demby VE (2009) An unexpected synergist role of P-glycoprotein and breast cancer resistance protein on the central nervous system penetration of the tyrosine kinase inhibitor lapatinib (N-{3-chloro-4-[(3-fluorobenzyl)oxy]phenyl}-6-[5-({[2-(methylsulfonyl)ethyl]- amino}methyl)-2-furyl]-4-quinazolinamine; GW572016). Drug Metab Dispos 37:439–442.
Poulin P, Jones RD, Jones HM, Gibson CR, Rowland M, Chien JY, Ring BJ, Adkison KK, Ku MS, He H, et al. (2011) PHRMA CPCDC initiative on predictive models of human pharma- cokinetics, part 5: prediction of plasma concentration-time profiles in human by using the physiologically-based pharmacokinetic modeling approach. J Pharm Sci 100:4127–4157.
Ring BJ, Chien JY, Adkison KK, Jones HM, Rowland M, Jones RD, Yates JW, Ku MS, Gibson CR, He H, et al. (2011) PhRMA CPCDC initiative on predictive models of human pharma- cokinetics, part 3: comparative assessment of prediction methods of human clearance. J Pharm Sci 100: 4090–4110.
Sabbah DA, Brattain MG, and Zhong H (2011) Dual inhibitors of PI3K/mTOR or mTOR- selective inhibitors: which way shall we go? Curr Med Chem 18:5528–5544.
Salphati L, Lee LB, Pang J, Plise EG, and Zhang X (2010a) Role of P-glycoprotein and breast cancer resistance protein-1 in the brain penetration and brain pharmacodynamic activity of the novel phosphatidylinositol 3-kinase inhibitor GDC-0941. Drug Metab Dispos 38:1422–1426.
Salphati L, Wong H, Belvin M, Bradford D, Edgar KA, Prior WW, Sampath D, and Wallin JJ (2010b) Pharmacokinetic-pharmacodynamic modeling of tumor growth inhibition and bio- marker modulation by the novel phosphatidylinositol 3-kinase inhibitor GDC-0941. Drug Metab Dispos 38:1436–1442.
Simeoni M, Magni P, Cammia C, De Nicolao G, Croci V, Pesenti E, Germani M, Poggesi I, and Rocchetti M (2004) Predictive pharmacokinetic-pharmacodynamic modeling of tumor growth kinetics in xenograft models after administration of anticancer agents. Cancer Res 64:1094– 1101.
Sutherlin DP, Bao L, Berry M, Castanedo G, Chuckowree I, Dotson J, Folks A, Friedman L, Goldsmith R, Gunzner J, et al. (2011) Discovery of a potent, selective, and orally available class I phosphatidylinositol 3-kinase (PI3K)/mammalian target of rapamycin (mTOR) kinase inhibitor (GDC-0980) for the treatment of cancer. J Med Chem 54:7579–7587.
Tanaka C, O’Reilly T, Kovarik JM, Shand N, Hazell K, Judson I, Raymond E, Zumstein-Mecker S, Stephan C, Boulay A, et al. (2008) Identifying optimal biologic doses of everolimus (RAD001) in patients with cancer based on the modeling of preclinical and clinical pharma- cokinetic and pharmacodynamic data. J Clin Oncol 26:1596–1602.
Wagner AJ, Burris HA, De Bono JS, Jayson GC, Bendell JC, Gomez-Roca C, Dolly S, Zee YK, Ware JA, Yan Y, et al. (2009) Pharmacokinetics and pharmacodynamic biomarkers for the dual PI3K/mTOR inhibitor GDC-0980: Initial phase I evaluation. Mol Cancer Ther 8 (Suppl 1):B137.
Wallin JJ, Edgar KA, Guan J, Berry M, Prior WW, Lee L, Lesnick JD, Lewis C, Nonomiya J, Pang J, et al. (2011) GDC-0980 is a novel class I PI3K/mTOR kinase inhibitor with robust activity in cancer models driven by the PI3K pathway. Mol Cancer Ther 10:2426–2436.
Wong H, Choo EF, Alicke B, Ding X, La H, McNamara E, Theil F-P, Tibbitts J, Friedman LS, Hop CECA, et al. (2011) Antitumor activity of targeted and cytotoxic agents in xenograft models correlates with clinical response: A pharmacokinetic-pharmacodynamic analysis. Mol Cancer Ther 10 (Suppl 1):A11.
Wong KK, Engelman JA, and Cantley LC (2010) Targeting the PI3K signaling pathway in cancer. Curr Opin Genet Dev 20:87–90.

Address correspondence to: Dr. Laurent Salphati, Department of Drug Me- tabolism and Pharmacokinetics, Genentech, Inc., 1 DNA Way, MS 412A, South San Francisco, CA 94080. E-mail: [email protected]