An integrated assessment of the ADME properties of the CDK4/6 Inhibitor ribociclib utilizing preclinical in vitro, in vivo, and human ADME data

Abstract Ribociclib (LEE011, Kisqali ®) is a highly selective small molecule inhibitor of cyclin‐dependent kinases 4 and 6 (CDK4/6), which has been approved for the treatment of advanced or metastatic breast cancer. A human ADME study was conducted in healthy male volunteers following a single oral dose of 600 mg [14C]‐ribociclib. Mass balance, blood and plasma radioactivity, and plasma ribociclib concentrations were measured. Metabolite profiling and identification was conducted in plasma, urine, and feces. An assessment integrating the human ADME results with relevant in vitro and in vivo non‐clinical data was conducted to provide an estimate of the relative contributions of various clearance pathways of the compound. Ribociclib is moderately to highly absorbed across species (approx. 59% in human), and is extensively metabolized in vivo, predominantly by oxidative pathways mediated by CYP3A4 (ultimately forming N‐demethylated metabolite M4) and, to a lesser extent, by FMO3 (N‐hydroxylated metabolite M13). It is extensively distributed in rats, based on QWBA data, and is eliminated rapidly from most tissues with the exception of melanin‐containing structures. Ribociclib passed the placental barrier in rats and rabbits and into milk of lactating rats. In human, 69.1% and 22.6% of the radiolabeled dose were excreted in feces and urine, respectively, with 17.3% and 6.75% of the 14C dose attributable to ribociclib, respectively. The remainder was attributed to numerous metabolites. Taking into account all available data, ribociclib is estimated to be eliminated by hepatic metabolism (approx. 84% of total), renal excretion (7%), intestinal excretion (8%), and biliary elimination (1%).

The compound has been approved by a number of health authorities, including the United States Food and Drug Administration (US FDA) and the European Medicines Agency, for the treatment of women with hormone receptor (HR)-positive, human epidermal growth factor receptor 2 (HER2)-negative advanced or metastatic breast cancer in combination with an aromatase inhibitor (AI) or fulvestrant. 1,2,3 Additional marketing authorizations are under review by health authorities worldwide.
During drug development, a human ADME study, in which six male healthy volunteers were administered 600 mg of [ 14 C]-labeled ribociclib was conducted. The human ADME study is essential to identify the major circulating drug-related components in order to assess any potential quantitative and/or qualitative differences in metabolism between humans and the animal species used in non-clinical safety assessment.
The importance of the study is discussed extensively in various regulatory guidances 4,5 as well as in the literature. 6,7,8 Furthermore, the study provides key data that can be used to estimate the main elimination pathways. Correlation of these data with in vitro phenotyping experiments allows a quantitative assessment of the enzymes responsible for the majority of metabolic elimination which, together with the identification of the major circulating metabolites, has important consequences for the understanding of possible drug drug interaction (DDI) liabilities. 9,10 The objectives of the human ADME study of ribociclib were (a) to determine the rates and routes of excretion of [ 14 C]-ribociclib-related radioactivity (mass balance) following a single oral dose of 600 mg [ 14 C]ribociclib to six healthy male subjects, (b) to determine the pharmacokinetics (PK) of total radioactivity in blood and plasma, (c) to characterize the plasma PK of ribociclib and N-desmethyl metabolite M4 (LEQ803), (d) to characterize the urine concentrations of ribociclib and LEQ803, (e) to identify and quantify ribociclib and its metabolites in excreta in order to elucidate key biotransformation pathways and clearance mechanisms, (f) to characterize the plasma PK of ribociclib and metabolites based on radiometry data, and (g) to assess the safety of a single 600 mg oral dose of [ 14 C]-ribociclib administered to healthy male subjects.
The purpose of this article is to describe the design and results of the human ADME study for ribociclib. In addition, the results of relevant in vitro studies and in vivo radiolabeled animal ADME studies are briefly described. Finally, an integrated assessment of all relevant data was performed in order to estimate the relative contributions of the various clearance pathways of ribociclib in humans.

| Pharmacokinetic studies
ADME studies (including QWBA in pigmented and non-pigmented rats) using radiolabeled ribociclib were conducted in rat and dog.
Relevant institutional and national guides for the care and use of laboratory animals were followed. Exposure of organs to total radioactivity was measured by QWBA of tissue sections. 11 Placental transfer in rat and rabbit was assessed by comparison of ribociclib concentrations in maternal and fetal plasma. Details of the dosing routes, formulations used, and the sampling schedules are provided in Table S1. The synthetic route of [ 3 H] -and [ 14 C]-ribociclib, which were used in selected ADME studies, is described in Table S4.
In these studies, PK of ribociclib in plasma, and total radioactivity analysis in blood, plasma, urine, and feces were measured by a combination of validated bioanalytical assays and LSC measurements.
Metabolite profiling was conducted on plasma, urine, bile, and feces samples in the rat and dog ADME studies, and further metabolite profiling on plasma and milk was conducted in a dedicated rat milk excretion study. QWBA studies were conducted with the tritium label in Hanover Wistar and partially pigmented Long-Evans rats. A further QWBA study was conducted with the carbon-14 label in male Long Evans rats in order to support the dosimetry calculation for the human ADME study. Briefly, 40-µm thick lengthwise dehydrated whole body sections were exposed for 1 day to Fuji BAS III imaging plates (Fuji Photo Film Co., Ltd., J-Tokyo) in a lead-shielded box and room temperature, and scanned in a Fuji BAS 5000 phosphor imager at a 50-µm scanning step. Concentrations of total radiolabeled components in the tissues were determined by comparative densitometry and digital analysis of the autoradiogram; blood samples of known radioactivity concentrations processed under the same conditions as the samples to analyze were used as calibrators.
Metabolite profiling in the ADME studies was conducted using liquid chromatography-mass spectrometry. was taken from the fetuses and ribociclib was measured using a validated bioanalytical assay. Resulting concentrations were compared with maternal plasma concentrations. In vitro blood distribution of [ 3 H]-ribociclib was investigated at the nominal blood concentrations of 100, 1000, and 10 000 ng/mL (rat and dog) and at 10, 100, 1000, and 10 000 ng/mL (human). Heparinized blood was spiked with [ 3 H]-ribociclib and incubated for 1 h at 37°C with constant agitation. After the incubation, blood cells and plasma were separated by centrifugation (1500 g, 10 min, 37°C). Total radioactivity was measured by LSC on triplicate aliquots, taken before (blood) and after (plasma) centrifugation. The hematocrit of the whole blood was determined in triplicate after centrifugation in micro hematocrit capillaries (13 000 g, 5 min). Calculation of f p and C bc /C p is described in Table S5.
In vitro plasma protein binding of [ 3 H]-ribociclib was investigated at nominal plasma concentrations of 100 and 1000 ng/mL (rat and dog) and at 10, 100, 1000, and 10 000 ng/mL (human). Stock solutions were spiked into plasma to achieve the intended concentrations.
After incubation for 1 hour at 37°C under constant gentle agitation, the spiked plasma samples (n = 3) were centrifuged (2000 g, 10 min, 37°C) in pre-warmed Centrifree devices. Total radioactivity was determined in the ultrafiltrate (C u , concentration of unbound compound) and in the sample introduced into the reservoir before ultrafiltration (C p ). The unbound (f u ) and the bound (f b ) fraction in plasma were calculated as follows: f u (%) = C u /C p × 100; f b (%) = 100-F u (%).

| Human enzyme phenotyping studies
In vitro incubations were carried out in 100-mM potassium phosphate buffer (pH 7.4) and 5-mM MgCl 2 with a total volume of 200 µL. Substrate and either human liver microsomes (HLM, BD Biosciences, mixed gender pool of n = 50) or recombinant human enzymes (BD Biosciences) were added and pre-incubated for 3 minutes at 37°C. The reaction was started by addition of a fresh solution of NADPH (1 mmol/L final concentration). The samples were incubated at 37°C with agitation of 500 rpm. For incubations with CYP2A6, CYP2C9, CYP2C18, and CYP4A11, the phosphate buffer was replaced by TRIS buffer (100 mmol/L, pH 7.4) and for flavin-containing monooxygenases (FMOs), a glycine buffer (50 mM, pH 9.5) was used.
The enzymatic reaction was stopped and the protein was precipitated by addition of an equal volume of methanol. After 30 min at −80°C, the samples were centrifuged at 30'000 x g for 15 min.
The supernatant was withdrawn. Aliquots were analyzed by LSC and the supernatant was diluted with water to obtain a final solution containing less than 20% of the organic solvent. For samples of low substrate concentration, supernatants were evaporated to dryness under nitrogen, then re-suspended in water containing less than 20% of methanol. Samples were analyzed by HPLC combined with radiodetection.
FMO inactivation in HLM: The FMO in HLM was inactivated by heating up the microsomal fraction at 50°C for 1 minute as described below: tubes with the appropriate volume of phosphate buffer and 2 µL 1M MgCl 2 were pre-warmed in a water bath at 50°C for at least 5 minutes. 8 µL HLM (20 mg/mL stock solution) were added to each tube (n = 2). All tubes were mixed quickly and incubated for exactly 1 minute at 50°C and immediately cooled down to 0°C in an iced

| Study Drug
The radiolabeled drug [ 14 C]-ribociclib succinate salt was synthesized by the Isotope Laboratory of Novartis, Basel, Switzerland. The synthetic route is described in Table S4. The final drug product was ana-

| Study Volunteers
This single-center, open-label, single oral dose study enrolled six healthy, non-smoking, male Caucasian volunteers who were determined as being in good health according to their medical history, physical examination, vital signs, electrocardiogram, laboratory tests, and urinalysis. Healthy male volunteers were selected as the foundation of the extensive human ADME data should be based on a small cohort (6) of young healthy male volunteers with subsequent extension/bridging to actual patients as needed for the investigation of variables such as age, gender, ethnicity, and health on the metabolic profile. Subjects with relevant radiation exposure of > 0.2 mSv in the 12 months prior to the initiation of the study were excluded. The subjects were exposed to a radia-

| Safety Assessments
Safety analysis included monitoring and recording of all adverse events, laboratory tests (ophthalmologic exam, hematology, blood chemistry, and urinalysis), vital signs, electrocardiogram, and physical examination.

| Dose Administration and Pharmacokinetic Sampling
After an overnight fasting of approximately 10 hours, each subject re-

| PK of ribociclib and Metabolite LEQ803 in Plasma and Urine
Ribociclib and LEQ803 were measured in plasma and urine using a validated bioanalytical assay. Plasma aliquots at each time point were subjected to protein precipitation with three volumes of acetonitrile containing 0.1% (v:v) formic acid, followed by dilution and analysis by liquid chromatography-tandem mass spectrometry in selected reaction monitor-positive ion mode using heated electrospray ionization as the ionization technique. For urine aliquots, six volumes of acetonitrile containing 0.1% (v:v) formic acid were added, followed by dilution and analysis in the same way as for plasma. Components were separated using a YMC-Triart C18 (2.0 x 30 mm, 1.9 µm) column (YMC Co. Ltd.). Mobile phase A was held at 95% for 0.5 min, then reduced to 80% at 0.8 min, 60% at 2 min, and 5% at 2.5 min, where it was held until 4 min. Finally, it was increased to 95% at 4.1 min. Mobile phase A was 0.1% formic acid in

| Total radioactivity measurements
Total radioactivity in blood and plasma was analyzed by accelerator mass spectrometry (AMS) on a National 5.3.2. Electrostatics Corporation 1.5SDH Compact AMS System (Middleton) in the bioanalytical laboratory of Accium BioSciences, Seattle, WA, USA. A known aliquot of each blood and plasma specimen was transferred to a prebaked quartz tube.
Sample volumes were selected to achieve approximately 1-2 mg total carbon. An AMS batch consisted of unknown specimens and chemical blank(s) to monitor for any in-process contamination. The samples were then dried using vacuum centrifugation and submitted to graphitization. Graphitization consisted of combustion of samples followed by reduction to graphite according to published methods. 12  For feces homogenates, quadruplicate, accurately weighed (500 mg) aliquots were dried in a stove at 50ºC for 3 hours. After the addition of 100-µL combustion aid (Perkin Elmer) to the dry homogenates, the samples were combusted in a sample oxidizer model 307 (Perkin Elmer). The absorber agent for CO 2 was 7-mL Carbo-Sorb® E (Perkin Elmer). At the end of the combustion cycle, the absorber was mixed with 13 mL of the scintillant Permafluor® E (Perkin Elmer). The samples were placed in the liquid scintillation analyzer for 60 minutes before counting. The total [ 14 C]-radioactivity of the samples was determined by counting until a statistical error (2 s) of 0.5% was obtained with a counting time of 10 or 30 minutes, depending on the level of radioactivity. Quantitative metabolite profiling in plasma was accomplished using a Shimadzu Prominence high-performance liquid chromatography (HPLC) system (Shimadzu, Columbia, MD, USA), coupled with fraction collection and analysis of total radioactivity in individual fractions using AMS (as described above). A second injection of each sample, conducted immediately after the first, was used to provide HPLC fractions for metabolite identification. The fractions identified as containing radioactivity from the AMS analysis were analyzed by HPLC coupled with high-resolution mass spectrometry to identify the metabolites. The HPLC-MS/MS methodology used is described below. Urine pools were analyzed directly for metabolite profiling (column recovery of a representative sample was measured), whereas feces pools were subjected to extraction. Specifically, to each aliquot of feces homogenate, two volumes of acetonitrile were added. After centrifugation, the pellet was washed further with an additional two volumes of acetonitrile. Additional pellet wash steps were done with methanol, dimethyl sulfoxide, acetone, and dichloromethane.

| Quantitative metabolite profiling in urine and feces
All washes were combined, evaporated under nitrogen, and reconstituted with a solution of 15-mM ammonium formate (pH 3.5). This feces extract reconstitute was analyzed for metabolite profiling.

| Absorption
Pharmacokinetic parameters in rat and dog are listed in Table S2.
Plasma clearance was high in rats (3.1 L/h/kg in males and 7.8 L/h/ kg in females) and dogs (1.9 L/h/kg), and volume of distribution was large (7.9 L/kg in rat and 27.9 L/kg in dog). Elimination half-life was moderate in rats (1.9-3.2 h) and long in dog (18.1 h). Bioavailability was 37%-55% in rat and 64%-87% in dog. Ribociclib showed a moderate first-pass effect in rats, with 66% absorption and 37% bioavailability.
Tissue distribution of ribociclib was studied in rats. Based on quantitative whole-body autoradiography, total radioactivity showed marked distribution into the extravascular compartment, except for brain, after intravenous or oral dose administration,

| Metabolism
Metabolism of ribociclib was investigated in rat and dog. Predominant metabolic pathways in the rat were direct conjugation to the sulfate metabolite M8 and N-dealkylation to LEQ803 with subsequent Phase II reactions. In dog, metabolism was dominated by oxidative pathways like dealkylation, C-and N-oxygenation, oxidations, and combinations of these reactions. In both species, the major plasma component was unchanged ribociclib. Metabolite LEQ803 was found in the circulation of both species and represented between 3% (dog) and 38% (rat) of plasma exposure to ribociclib. A higher ratio of LEQ803 to ribociclib AUC after p.o. compared to iv dosing in rat and dog indicated a first pass effect in generation of this metabolite. The higher clearance in female rats compared to males was attributed to a more pronounced metabolism to the sulfate metabolite M8 in females. 13 The formation of M8 was a minor metabolic pathway in dog. Representative metabolite profiles from rat and dog ADME studies are provided in Figure S3 and Figure S4.

| Excretion
In rat and dog, the predominant route of elimination was fecal.

| Enzyme phenotyping
In vitro oxidative metabolism of  As FMOs are more thermally labile than CYPs in the absence of NADPH, 14 heat treatment of HLM was used to estimate the contribution of FMO. HLM was heated to 50°C for 1 min to inactivate the FMO activity. Figure S1 shows the effect of heat treatment.  (Table S3)

| Mass balance
Following oral administration of [ 14 C]-ribociclib 600 mg, most of the radioactive dose was excreted in feces (mean 69.1 ± 4.72%) after 504 hours. In urine, mean 22.6 ± 5.39% of the dose was excreted after 504 hours. Mean mass balance in the six volunteers was 91.7 ± 1.01%, which is almost complete. A graphical representation of the cumulative excretion is provided in Figure 3.

| Absorption
Oral absorption of ribociclib was estimated to be moderate (ie, approximately 58.8% based on the mean recovery of radioactivity in urine (22.6%) and mean radiolabeled metabolites excreted in feces (36.2%) [assuming all metabolites detected in feces were formed systemically]) Further information can be found in Table S7 and   Table S11.

| Total Radioactivity (Drug-Related Material) PK in Blood and Plasma and Ribociclib PK in Plasma
Concentrations of total radioactivity in blood and plasma were measured by AMS, and concentrations of ribociclib in plasma were measured by validated bioanalytical assay, as described above. Results 8700 ng-eq·h/mL, respectively. The overall contribution of ribociclib to the total radioactivity in plasma, based on mean AUCinf was approximately 23%.

| Renal clearance of ribociclib
The median CL/F of ribociclib was 70.2 L/h. The median renal clearance of ribociclib was 5.55 L/h, which was more than 10-fold lower than the non-renal clearance (64.7 L/h).

| Metabolite Profiles in Plasma
Metabolite profiles, generated from plasma pooled across subjects at 1, 3, 24, and 48 hours post dose, are shown in Figure 4. The  Table S6).
M4 was found to co-elute with metabolites M19 and M62, and F I G U R E 3 Mean cumulative excretion of radioactivity in urine and feces in the human absorption, distribution, metabolism, and elimination study conducted in healthy male volunteers (N = 6) who received a single oral dose of 14 C-ribociclib 600 mg. Source data for Figure 3 are shown in Table S11

| Metabolite profiles in Excreta
A representative metabolite profile of urine and feces samples for the subject pool (N = 6) is shown in Figure 6. Tabulated data showing the percentage of dose attributable to the metabolites is shown in Table S7.
The urine pool contained an average of 22.1% of the administered dose (covering 98% of the radioactivity eliminated in urine). The largest drugrelated component identified in urine from radio profiling was ribociclib, representing 12.1% of the dose. Ribociclib in urine was estimated to be 6.75% of the dose by validated bioanalytical assay. The most likely reason for the discrepancy is that minor metabolites in urine under the detection limit of the radio profile led to an overestimation of the ribociclib amount. M4 was also significant, although it was found to co-elute with metabolites M61 and M62. Assuming that these components were negligible in abundance (based on a comparison of LC-MS peak areas), M4 represented approximately 3.74% of the administered dose excreted in this matrix. Numerous other metabolites were also detected but these were minor, with each representing ≤ 1.5% of the dose. The feces pool contained an average of 66.8% of the administered dose (covering 97% of the radioactivity eliminated in feces). The largest drug-related component in feces was ribociclib, representing 17.3% of the administered dose. M4 was also significant, although was found to co-elute with metabolite M19. Assuming that this metabolite was negligible in abundance not be reliably estimated due to co-elutions, but it is < 5.21% of the administered dose. A simplified metabolism scheme is shown in Figure 7, with a summary of MS data and further details on metabolite structures provided in Table S8, Table S9, and Table S10.

| Absorption
Ribociclib exhibited moderate to high absorption. In the rat ADME study, absorption (66%) was 1.8-fold higher than bioavailability (37%), indicating a moderate first pass effect. Bioavailability was higher in dogs (64%-87%). In the human ADME study, absorption was estimated to be 58.8% based on the mean percentage of the total radiolabeled dose in urine (22.6%) and the amount of dose attributable to metabolites in feces (36.2%), assuming all metabolites in feces were formed systemically and assuming that no ribociclib was excreted into feces via either hepatobiliary export or intestinal secretion. Given that intestinal secretion is suspected based on rat ADME data, this absorption value in human should be considered with caution. Oral bioavailability could not be calculated in the human ADME study as co-administration of a 13 C-iv microdose was not conducted.

| Distribution
In tissue distribution studies in male rats, total radiolabeled components were markedly distributed into the extravascular compartment except for brain and were eliminated rapidly from most tissues. In pigmented rats, specific distribution of radioactivity to the melanin-containing structures was observed. Ribociclib was found to pass the placental barrier in both rats and rabbits, and was excreted into the milk of lactating rats. Although placental transfer and milk excretion was observed in animals, it should be acknowledged that translatability of the data to human is complex and not well understood due to anatomical and functional differences of the placenta between species, and wide species differences in the protein and lipid content of milk. 19 (Table S6), and 17.9, 19.8, and 21.6% of ribociclib exposure, respectively. Twentynine other metabolites were identified in plasma, but these were minor (≤4.15%). Subsequent investigations, by bioanalytical assay or by relative exposure comparison across species, 23,24 confirmed that metabolites M4 and M13 were covered by rat, dog, and/or rabbit (data on file, Novartis Pharmaceuticals). Furthermore, M4 and M13 were found not to have a relevant contribution to total pharmacological activity in human considering both in vitro measurements and their in vivo exposure.

| Excretion
Ribociclib is mainly eliminated via metabolic clearance, with renal clearance playing a lesser role. Median renal clearance (CLr) of  Figure 3). This is similar to preclinical species, where 68.8%-84% and 5.9%-18.5% of administered radioactivity were found in feces and urine, respectively. Elimination occurred mainly by hepatic metabolism with a limited contribution of renal excretion of unchanged ribociclib, which represented 6.75% of the dose in urine, measured by validated bioanalytical assay (17.3% of dose in feces). A large number of metabolites were identified in excreta ( Figure 6, Table S7), but the most abundant was M4 (LEQ803), which represented 13.9 and 3.74% of the dose in feces and urine, respectively.

| Enzyme phenotyping
Using recombinant human enzymes, metabolites M4 (LEQ803) and M15 were efficiently formed by CYP3A4, but were almost not detectable in incubations of the other hepatic CYPs investigated. A correlation analysis (Table S3) also strongly suggested that the enzymatic formation of M4, M15, and minor metabolites is catalyzed by CYP3A.
The hydroxylamine metabolite M13 (CCI284) was formed readily by FMO3 and CYP2J2, but not by the other recombinant hepatic CYPs. Inactivation of FMO3 by heat treatment, correlation analysis (Table S3), and incubation with recombinant enzymes consistently indicated that hydroxylamine M13 (CCI284) is formed by FMO3.
Based on the kinetics of metabolite formation (Figure 2), it was estimated that 74% of the oxidative metabolism in HLM results from CYP3A4, whereas 26% is due to the hepatic FMO contribution. Incubation with CYP-selective inhibitors suggests a dominant contribution of CYP3A4/5 to oxidative metabolism in HLM (Table 2).
Collectively, all in vitro enzyme phenotyping approaches consistently indicate that oxidative hepatic metabolism of ribociclib is dominated by CYP3A4/5 (74%) with partial contribution by FMO3 (15%-26%). It is likely that inhibitors and/or inducers of CYP3A4 will influence the oxidative metabolic clearance of ribociclib in humans.

| Estimation of ribociclib elimination pathways
Integrating the human ADME results obtained in healthy male volunteers, the metabolic enzyme phenotyping, and rat biliary and intestinal excretion data, our data indicate that ribociclib is eliminated mainly by hepatic metabolism in humans, primarily via CYP3A with a lesser contribution by FMO3 and direct phase II metabolism (approx. 84% of total). The remainder is accounted for by renal excretion (7%), intestinal excretion (8%), biliary elimination (1%), and an unknown, but likely low, contribution from extrahepatic metabolism (FMO1) (Figure 8). The intestinal excretion and biliary elimination components are based on rat ADME data, and assume that these data translate to human.