Abstract

Mobocertinib (formerly known as TAK-788) is a targeted covalent tyrosine kinase inhibitor of epidermal growth factor receptor with exon 20 insertion mutations. This article describes the metabolism and excretion of mobocertinib in healthy male subjects after a single oral administration of [14C]mobocertinib. Mobocertinib-related materials were highly covalently bound to plasma proteins such as human serum albumin. The mean extraction recovery of total radioactivity was only 3.9% for six individual Hamilton pooled plasma samples. After extraction, mobocertinib was the most abundant component accounting for 7.7% of total extracted circulating radioactivity (TECRA) in the supernatant. Each of identified metabolites accounted for <10% of TECRA. Mobocertinib underwent extensive first-pass metabolism with the fraction of the dose absorbed estimated to be approximately 91.7%. Fecal excretion of mobocertinib metabolites was the major elimination route. Mobocertinib was mainly eliminated via oxidative metabolism with a fraction of approximately 88% metabolized by CYP3A4/5. The other minor elimination pathways included cysteine conjugation, metabolism by other cytochrome P450s, and renal excretion of unchanged mobocertinib.

SIGNIFICANCE STATEMENT This article describes the metabolism and excretion of a targeted covalent inhibitor mobocertinib in humans after a single oral administration of [14C]mobocertinib. Mobocertinib was highly covalently bound to human plasma proteins. No metabolite accounted for >10% of total extracted circulating radioactivity in human plasma. Mobocertinib was mainly eliminated via CYP3A4/5 mediated oxidative metabolism followed by fecal excretion after approximately 91.7% of the dose was absorbed.

Introduction

Epidermal growth factor receptor (EGFR) exon 20 insertion (ex20ins) mutations are typically located after the C-helix of the tyrosine kinase domain, which occur in approximately 2%–3% of all non–small cell lung cancer (NSCLC) patients, representing 4%–12% of observed EGFR mutations for NSCLC (Yasuda et al., 2012; Vyse and Huang, 2019; Remon et al., 2020; Meador et al., 2021; Riely et al., 2021; Malapelle et al., 2022). The EGFR ex20ins are the third most common EGFR mutation subtype after the exon 19 deletions and exon 21 L858R mutation (Arcila et al., 2013; Remon et al., 2020; Meador et al., 2021). Unlike the classical EGFR mutations in NSCLC, EGFR ex20ins are associated with de novo resistance to existing first-, second-, and third-generations of reversible and irreversible EGFR-directed tyrosine kinase inhibitors such as gefitinib, erlotinib, neratinib, afatinib, dacomitinib, and osimertinib (Yasuda et al., 2012; Vyse and Huang, 2019; Remon et al., 2020; Russell et al., 2023). As a result, platinum-based chemotherapy remains the preferred therapy for patients with EGFR ex20ins (Remon et al., 2020; Russell et al., 2023). There is an urgent need for the development of an efficient medicine for the treatment of NSCLC with this subset of mutations.

Mobocertinib (formerly known as TAK-788) is a targeted covalent inhibitor of EGFR ex20ins mutants in NSCLC (Gonzalvez et al., 2021; Han et al., 2021; Riely et al., 2021). Phase 1/2 dose-escalation trial of mobocertinib identified 160 mg/day as the recommended phase 2 dose and maximum tolerated dose (Riely et al., 2021). The data from phase 1 and 2 trials indicated that the overall response rate was 28% with a median duration of response of 17.5 months in 114 patients with locally advanced or metastatic NSCLC with EGFR ex20ins mutations receiving mobocertinib 160 mg orally once daily (Duke et al., 2023). Based on the data, mobocertinib was granted accelerated approval by US Food and Drug Administration for the treatment of patients with the same indication (whose disease has progressed on or after platinum-based chemotherapy) in 2021 (Kwon et al., 2022; Duke et al., 2023; Russell et al., 2023). In the phase 3 EXCLAIM-2 study, the primary endpoint was not met (Jänne et al., 2023), which led to a voluntary withdrawal of mobocertinib worldwide.

In the present study, the metabolism and excretion of mobocertinib were investigated in healthy male subjects after a single oral administration of [14C]mobocertinib. Metabolite profiling was conducted using high-performance liquid chromatography (HPLC)-accelerator mass spectrometry (AMS) or radiometric detection. Metabolites were identified with HPLC-high resolution mass spectrometry (HRMS). This study was conducted as a part of an open-label, 2-period, fixed-sequence phase 1 clinical trial that evaluated the absolute bioavailability, mass balance, pharmacokinetics, metabolism, and excretion of [14C]mobocertinib in healthy male subjects (Hanley et al., 2024).

Materials and MethodsTest Articles and Reagent

Good manufacturing practice grade [14C]mobocertinib succinic acid salt (specific activity 0.468 μCi/mg; radiochemical purity 99.6%; chemical purity 102.4%) used in the human absorption, distribution, metabolism, and excretion (ADME) study was supplied by Almac Sciences (Craigavon, UK). Mobocertinib, M65 (also known as AP32914), M67 (also known as AP32960), M70, and M108 (also known as M59) reference standards were provided by Takeda Pharmaceuticals (Cambridge, MA). The deuterium-labeled internal standards for mobocertinib, M65, and M67 were manufactured by Almac. cDNA-expressed recombinant human cytochrome P450 (Supersomes) (rhCYPs) were purchased from Corning (Woburn, MA). All other chemicals and reagents for this work were of analytical grade and purchased from commercial sources.

Study Design

This human ADME study is an open-label, 2-period, fixed-sequence phase 1 clinical trial (Hanley et al., 2024). In the study, six healthy male subjects received a single oral dose of 160 mg (approximately 100 μCi) [14C]mobocertinib solution. The whole blood, plasma, urine, and feces were collected for pharmacokinetics, mass balance, and metabolic profiling analysis. The study and related documents were approved by institutional review board. Additional information about the clinical study design can be found in the previous publication (Hanley et al., 2024).

Sample Collection for Metabolite Profiling

Blood samples were collected at 0 (predose), 1, 2, 4, 6, 12, 24, 48, 72, 96, 120, and 168 hours postdose. The plasma samples were obtained from the blood samples after centrifugation. Urine and fecal samples were collected at predose, 0–12 hours (urine only), 12–24 hours (urine only), 0–24 hours (feces only), and every 24-hour interval up to 312 hours until the discharge criteria were met for each subject. Fecal samples were homogenized with at least 4 volumes of ultrapure water after collection. Plasma, urine, and fecal homogenates were stored at –80°C until sample analysis.

Sample Preparation for Metabolite Profiling

Plasma samples were pooled from 0- to 168-hour plasma from each subject based on the Hamilton pooling method (Hamilton et al., 1981). An aliquot of each Hamilton pooled plasma sample was subject to AMS analysis for calculation of extraction, column, and overall recoveries. Separately, an aliquot of each pooled plasma sample (200 μl) was extracted with 600 μl of acetone containing mobocertinib, M70, M65, and M65 standards, methanol, and methanol containing 0.1% formic acid, respectively, and centrifuged. The combined supernatant was concentrated down to approximately 100 μl under a nitrogen stream at ambient temperature followed by addition of 300 μl of 60% aqueous methanol and centrifuged. The supernatant was subject to HPLC fractionation followed by AMS analysis. The recovery of radioactivity in the supernatant was assessed by AMS.

Urine samples were pooled proportionally based on the volumes collected in each interval for each subject. The radioactivity in the pooled urine samples represented a mean of 92.2% of the radioactivity recovered in all urine collected. Aliquots of individual pools were taken proportionally based on the volumes collected from the individuals to obtain a combined urine pool from all subjects. Aliquots of the combined urine pool were centrifuged and concentrated. Aliquots of the concentrated urine samples were subjected to HPLC-radiometric analysis, and the recovery of radioactivity following concentration was assessed.

Fecal homogenate samples were pooled proportionally based on the weight of feces collected in each interval for each subject. The radioactivity in the pooled fecal homogenate samples represented a mean of 89.0% of the radioactivity recovered in all feces collected. An aliquot of individual pooled fecal homogenate sample from each subject was extracted twice with acetonitrile containing 0.1% formic acid and then once with acetonitrile/methanol (1:1, v/v) containing 0.1% formic acid once. After each extraction, the mixture was centrifuged, and the pellet was used for next extraction. The supernatants were collected after each extraction. The combined extracts were dried under vacuum. The dried residues were reconstituted with acetonitrile/methanol (1:1, v/v) containing 0.1% formic acid and centrifuged prior to analysis. Aliquots of clean supernatant samples were subjected to HPLC-radiometric analysis. The recovery of radioactivity following extraction and reconstitution was assessed.

HPLC-Radiometric Analysis for Human Urine and Feces

Sample analysis was performed on an Agilent 1290 Infinity II ultra-high performance liquid chromatography (UHPLC) system (Agilent Technologies, Wilmington, DE) with a high-speed binary pump (G7120A), an autosampler (G7129A), a thermostatted column compartment (G7116B), a diode array detector (G7117A), and a fraction collector (G1364C). The chromatographic separations were achieved on a Gemini NX C18 column (250 × 4.6 mm, 5 μm, 110 Å; Phenomenex, Inc., Torrance, CA) with a C18 guard column (4 × 3 mm, Phenomenex, Inc.) at 40°C. Mobile phases A and B were 10 mM ammonium acetate (pH 5.6, adjusted with acetic acid) and acetonitrile, respectively. The HPLC flow rate was 0.75 ml/min. The UV detector was set at 288 nm. The HPLC gradient used for urine metabolite profiling was 0–5 minutes, 5%–10% B; 5–30 minutes, 10%–40% B; 30–40 minutes, 40%–45% B; 40–44 minutes, 45%–40% B; 44–45 minutes, 40%–60% B; 45–50 minutes, 60%–95% B; 50–54 minutes, 95% B; 54–55 minutes, 95%–5% B; 55–60 minutes, 5% B. The HPLC gradient used for fecal metabolite profiling was 0–5 minutes, 10%–28% B; 5–30 minutes, 28%–45% B; 30–40 minutes, 45%–40% B; 40–44 minutes, 40%–75% B; 44–45 minutes, 75%–95% B; 45–50 minutes, 95% B; 50–51 minutes, 95%–10% B; 51–60 minutes, 10% B. HPLC eluates were collected by the fraction collector carrying 4 × 96-well plates (IsoPlate, PerkinElmer, Waltham, MA) with a collection time of 9.375 seconds per well. The plates were dried under vacuum, mixed with scintillation cocktail (approximately 150 μl), and counted for 10 minutes per well by MicroBeta2 (PerkinElmer). The UHPLC system was controlled by Laura software, Version 5 (LabLogic System Inc., Tampa, FL). The radiometric data were then imported and analyzed using Laura to obtain the radiochromatograms.

HPLC-AMS Analysis for Human Plasma

HPLC-AMS analysis and radiochromatogram generation for plasma samples were conducted at Accium BioSciences, Inc. (Seattle, WA). All HPLC separation and fraction collection procedures were performed using a Prominence HPLC system (Shimadzu Scientific Instruments Inc., Columbia, MD) equipped with a binary pump with gradient control (LC-20AD), an autosampler (SIL-20ACHT), a column oven (CTO-20AC), UV-visible detector (SPD-20AV), a communication BUS module (CBM-20 A), and a fraction collector (FRC-10 A). The HPLC fraction collection was the same as HPLC-radiometric analysis for urine analysis except that HPLC eluates were collected with two 96-well fraction collection plates with a collection time of 18 seconds per well. The collected fractions or fraction pools were subjected to AMS analysis following graphitization procedure. AMS measurements were performed on a 1.5SDH Compact AMS System (National Electrostatics Corporation, Middleton, WI). The data acquisition software developed by National Electrostatics Corporation provides 14C counts, 13C, and 12C currents, as well as the isotopic ratios 14C/13C and 13C/12C. Further calculations including normalization, corrections for fractionation, and machine and chemical background, were performed using in-house developed and validated software at Accium BioSciences, Inc. The radiometric data were analyzed using Excel (Microsoft Corporation, Redmond, WA) to obtain the radiochromatograms.

HPLC-HRMS Analysis

Liquid chromatography was performed using a 1290 Infinity UHPLC system (Agilent Technologies) with a binary pump (G4220A), an autosampler (G4226A), a thermostatted column compartment (G1316C), and a diode array detector (G4212A). The HPLC analysis conditions were the same as those used in the HPLC-radiometric analysis. The HPLC gradient for plasma analysis was the same as that for urine analysis. The HPLC eluate was introduced via electrospray ionization source directly into the TripleTOF 5600 high resolution mass spectrometer (SCIEX, Framingham, MA) using a positive ion mode with the following conditions: ionization voltage (IS) = 5000 V, gas 1 = 50 (arbitrary units), gas 2 = 50 (arbitrary units), curtain gas = 30 (arbitrary units), source temperature = 550°C, de-clustering potential = 80 V, and collision energy = 64 ± 15 V. The mass defect tolerance was set at 50 mDa with the mass of mobocertinib. The data were acquired using Analyst TF 1.7.1 (SCIEX) and processed with PeakView 2.2 (SCIEX).

Incubation of [14C]Mobocertinib in Human Plasma for Determination of Extraction Recovery of Total Radioactivity (TRA)

Blank human plasma spiked with [14C]mobocertinib solution with a final concentration of 0.48 μM (23,532 dpm/ml) was incubated for 0 and 10 minutes at room temperature or for various periods of time (0.5, 1, 3, 8, 24, 48, and 168 hours) at 37°C in a shaking water bath. For the 0-minute sample, duplicate aliquots of 100 μl of plasma were subject to liquid scintillation counting (LSC) measurement. For other incubation time, aliquots (300 μl) of the incubated plasma were removed and treated with acetone (0.9 ml) to precipitate the protein, followed by vortexing and centrifugation to collect the supernatant. The protein pellet was subsequently extracted with methanol (0.9 ml) and methanol containing 0.1% formic acid (0.9 ml), respectively, followed by vortexing and centrifugation. The supernatants were combined, and duplicate aliquots were taken for LSC measurement. The extraction recovery of TRA at time t was determined by the following equation.

Incubation of Mobocertinib in Human Plasma and Human Serum Albumin (HSA) for Detection of HSA-Mobocertinib Conjugate

Blank human plasma or HSA (approximately 0.6 μM) in pH 7.4 phosphate buffer (495 μl) was spiked with 5 μl of 20 mM mobocertinib (final concentration 200 μM). Additionally, blank human plasma (495 μl) was spiked with either 5 μl water or 20 mM dextromethorphan as negative control (final concentration 200 μM). All samples were incubated at 37°C for 0 hours, 2 hours, and overnight (>16 hours). At each time point, 33 μl aliquots were diluted 30-fold with the addition of 957 μl water. The diluted samples were analyzed directly via microflow HPLC-HRMS using an M5 Eksigent system (SCIEX) coupled with a TripleTOF 6600 Q-TOF mass spectrometer (SCIEX) in the trap and elute mode.

Determination of Relative Contribution of Mobocertinib Metabolism in rhCYPs

Mobocertinib (0–100 μM) was incubated with rhCYP1A2, 2B6, 2C8, 2C9, 2C19, 2D6, 3A4, and 3A5 (20 pmol/ml). The final concentration of proteins in the incubation was adjusted to 0.5 mg/ml with appropriate control Supersomes with either P450 oxidoreductase or with both P450 oxidoreductase and cytochrome b5. The incubations were conducted in triplicate in 0.1 M potassium phosphate buffer supplemented with 2 mM NADPH and 3 mM MgCl2, pH 7.4, at 37°C for 5 minutes. The total incubation volume was 100 μl and at the end of the incubation time, the reactions were terminated by addition of 100 μl of acetonitrile containing 0.1 μM of each deuterated mobocertinib, M65, and M67 internal standards. The plates were kept in a refrigerator set at 4°C to maximize protein precipitation and then centrifuged at 1800 × g at 0°C for 10 minutes. The supernatants were analyzed by HPLC with tandem mass spectrometry (MS/MS). The formation of two major in vitro metabolites of mobocertinib, M65 and M67, was monitored to estimate the metabolic clearance (CLint,rhCYP) of mobocertinib. The intrinsic clearance in HLMs by each rhCYP (CLint,HLM,rhCYP) was calculated from the mean CLint,rhCYP using relative activity factor. The effect of individual rhCYP3A4 and 3A5 mediated clearance was extrapolated for a combined rhCYP3A4/5-mediated clearance based upon their relative expression in human liver microsomes. The contribution of each rhCYP was then obtained by dividing each CLint,HLM,rhCYP to the sum of CLint,HLM,rhCYP of all seven rhCYPs described above.

Radioactivity Determination

The radioactivity contained in samples was determined by LSC in duplicate (when possible). Liquid scintillation analysis was automatically corrected for chemiluminescence and chemical quench using an external standard method. Background radioactivity was automatically subtracted. In general, samples were counted for 2 minutes per well. Aliquots of plasma, plasma extracts, urine, and fecal homogenate extracts were mixed with scintillation cocktail and analyzed by LSC. For urine and fecal metabolite profiling analysis, Isoplate-96 well plates were counted for 10 minutes per well by MicroBeta2 (PerkinElmer).

Metabolite Identification

Identification and characterization of metabolites were conducted using HPLC-HRMS. The metabolite structures were elucidated based on the mass shift from mobocertinib, elemental analysis using HRMS data, and product ions obtained from accurate MS/MS spectra. Identification of metabolites M65, M67, M70, and M108 were also confirmed by comparative analysis of retention time and MS/MS spectra with respective authentic standards.

ResultsAbsolute Bioavailability, Pharmacokinetics, and Mass Balance

Absolute oral bioavailability, pharmacokinetics of total radioactivity in blood and plasma, pharmacokinetics of mobocertinib and its two active metabolites M65 (AP32914) and M67 (AP32960) in plasma, and cumulative excretion of radioactivity in urine and feces from this human ADME study were reported in the previous publication (Hanley et al., 2024). In the current manuscript, metabolite profiling and identification of [14C]mobocertinib in plasma, urine, and feces from the same study along with some in vitro studies were investigated and presented.

Excretion of Administered Radioactive Dose

Following a single oral dose of 160 mg (approximately 100 μCi) [14C]mobocertinib to six healthy male subjects, a geometric mean of 3.6%, 76.0%, and 79.6% of the administered dose was recovered in urine, feces, and both excreta combined, respectively (Hanley et al., 2024).

Effect of [14C]Mobocertinib Incubation Time in Human Plasma on the Extraction Recovery of TRA

An in vitro experiment was conducted to determine the effect of incubation time on extraction recovery of human plasma incubated with [14C]mobocertinib. The recovery of TRA after 10 minutes at room temperature and after 0.5, 1, 3, 8, 24, 48, and 168 hours at 37°C was shown in Table 1. TRA recovery was dependent on incubation time. Longer incubation resulted in lower extracted TRA. After 10-minute incubation at room temperature, the TRA was almost completely recovered. The TRA recovery was 89.3% after 0.5-hour incubation at 37°C and it decreased to 4.0% after 168-hour incubation.

TABLE 1

Effect of incubation time of [14C]mobocertinib in human plasma on the extraction recovery of TRA

Detection of HSA-Mobocertinib Conjugate in In Vitro Human Plasma or HSA Incubation

The deconvoluted mass spectra for the incubated human plasma or HSA with water, dextromethorphan, or mobocertinib at different time points are shown in Fig. 1. There were primarily four HSA related peaks observed in blank human plasma (Fig. 1A) and human plasma incubated with dextromethorphan (Fig. 1E) overnight and human plasma (Fig. 1B) and HSA (Fig. 1F) incubated with mobocertinib for 0 hour. The most abundant two peaks were native HSA (m/z 66,438) and HSA-cysteine conjugate (HSA-Cys) (m/z 66,557), respectively (Fig. 1E). The less abundant two peaks were glycosylated HSA and HSA-Cys, respectively. In comparison, four new peaks emerged in human plasma (Fig. 1C) and HSA (Fig. 1G) incubated with mobocertinib for 2 hours. Those peaks were also observed in samples incubated with mobocertinib overnight with higher abundance (Figs. 1, D and H). The two most abundant new peaks, m/z 67,024 and m/z 67,143 (Fig. 1D), were 586 Da higher than HSA and the HSA-Cys peaks, respectively. The mass difference of 586 Da was consistent with the molecular weight of mobocertinib (585.7 Da). Therefore, peaks at m/z 67,024 and m/z 67,143 represented HSA-mobocertinib and HSA-Cys-mobocertinib conjugates, respectively. The two minor new peaks, m/z 67,191 and m/z 67,306 (Fig. 1D), were mobocertinib conjugated to glycosylated HSA and glycosylated HSA-Cys, respectively.

Fig. 1.

Deconvoluted mass spectra for blank human plasma incubated overnight (A), human plasma incubated with mobocertinib after 0 (B) and 2 hours (C), and overnight (D), human plasma incubated with dextromethorphan overnight (E), HSA incubated with mobocertinib after 0 (F) and 2 hours (G), and overnight (H). Mobo, mobocertinib.

Metabolite Profiling in Human Plasma

Human plasma samples were pooled across 0 to 168 hours for each of the six subjects using the Hamilton pooling method (Hamilton et al., 1981). As the plasma extraction recovery of TRA was extremely low, the metabolite profiling was conducted using HPLC-AMS analysis. After extraction of each individual Hamilton pooled plasma, the supernatant was concentrated to a small volume and then diluted with methanol/water (v/v = 60/40) prior to TRA measurement and HPLC fractionation. The extraction recovery of TRA in the plasma was 3.2%, 3.3%, 3.5%, 3.5%, 4.5%, and 5.1% for the six subjects, respectively, with a mean recovery of 3.9%. The column recovery ranged from 83.4% to 97.9% with a mean recovery of 91.0%. The overall recovery, which was the cumulative radioactivity recovered across all collected HPLC fractions as a percentage of the radioactivity present in an equivalent amount of pooled plasma sample prior to extraction as measured by AMS, ranged from 2.9% to 4.5% with a mean recovery of 3.5%.

A representative radiochromatogram from a Hamilton pooled plasma sample is shown in Fig. 2A. There were nine circulating components identified in human plasma including M70, M53, M66, M65, M106, M67, M107, mobocertinib, and M108. Mobocertinib was the most abundant extracted circulating component and accounted for 7.65% of total extracted circulating radioactivity (TECRA). M108, M70, M107, M106, and M67 accounted for 7.14%, 6.76%, 5.94%, 5.30%, and 5.17% of TECRA, respectively. Other observed metabolites each accounted for <4% of TECRA. There was no single metabolite accounting for >10% of TECRA. The mean TECRA of each of the identified components is shown in Table 2.

Fig. 2.

Representative radiochromatograms of (A) plasma, (B) urine, and (C) feces after a single oral administration of [14C]mobocertinib to healthy male subjects. Mobo, mobocertinib. *Denotes unidentified metabolites.

TABLE 2

Mean distribution of mobocertinib and its metabolites in pooled plasma (percent TECRA), urine (percent of the dose), and feces (percent of the dose) from humans

Metabolite Profiling in Human Urine

The recovered dose was low in urine of all subjects with a geometric mean recovery of 3.6% of the administered dose. As a result, the urinary metabolite profiling was conducted in a pooled urine sample across all 6 subjects. Metabolite profiling of the pooled urine sample revealed the presence of 5 metabolites, M53, M66, M67, M69, and M70 (Fig. 2B), all of which were identified by HPLC-HRMS analysis except for M69. Percent of the dose excreted as mobocertinib and its metabolites in urine are summarized in Table 2. The recovery from processing of urine samples was 102.6%, and the HPLC column recovery was 102.2% for urine analysis.

Unchanged mobocertinib accounted for approximately 0.9% of the dose in urine. M67 was the most abundant metabolite, which represented 0.8% of the dose. Metabolites M53, M66, and M70 represented approximately 0.4%, 0.5%, and 0.4% of the dose, respectively. Unknown metabolite M69 accounted for 0.3% of the dose.

Metabolite Profiling in Human Feces

A geometric mean of 76.0% of the dose was recovered in feces from all six subjects. Metabolite profiling of the reconstituted extracts from the pooled fecal homogenate of individual subjects exhibited the presence of many metabolites as shown by a representative radiochromatogram (Fig. 2C). The radioprofiles were qualitatively similar across six individual subjects. The metabolites with relatively high intensity as shown by the radiochromatogram were identified by HPLC-HRMS analysis, which included M35, M42, M44, M46, M53, M55, M60, M63, M66, M67, M68, M70, and M71. Percent of the dose excreted as mobocertinib and its identified metabolites are summarized in Table 2. The mean extraction recovery was 88.2%, and the mean reconstitution recovery was 92.8%. The HPLC column recovery was 101.0% for fecal analysis.

Unchanged mobocertinib accounted for a mean of 5.9% of the dose. Metabolite M67 was the most abundant component contributing to 11.9% of the dose. Other identified metabolites M35, M42, M44, M46, M53, M55, M60, M63, M66, M68, M70, and M71 each represented from 1% to 6% of the dose. There were a few low-level metabolites which were not able to be identified by HPLC-HRMS analysis. The mean percent of the dose was less than 3% for each unidentified metabolite.

Overall, mobocertinib, identified metabolites, and unidentified metabolites accounted for approximately 7%, 50%, and 12% of the dose, respectively, in urine and feces combined.

Metabolite Identification

Mobocertinib metabolites identified by HPLC-HRMS analysis are summarized in Table 3. The HRMS and MS/MS of mobocertinib and identified metabolites and proposed structural fragmentations are shown in Supplemental Figs. 1–18. The proposed structures of metabolites and the metabolic pathways of mobocertinib in humans are shown in Fig. 3.

TABLE 3

Mobocertinib metabolites detected and identified in healthy male subjects

Fig. 3.

Metabolic pathways of mobocertinib in humans. P, U, F means that a metabolite was observed in the respective matrix. P, plasma; U, urine; F, feces.

Reaction Phenotyping in rhCYPs

Mobocertinib was primarily metabolized by cytochrome P450 (CYP)3A4/5 based on the intrinsic clearance determined via the formation rate of two major in vitro metabolites M65 and M67 using rhCYPs. The relative contribution of CYP3A4/5 to hepatic metabolism of mobocertinib was 93.5%. Each of the other major CYPs including CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP2C19, and CYP2D6 contributed to less than 3% of mobocertinib metabolism (Table 4).

TABLE 4

% contribution of rhCYPs to mobocertinib metabolism in vitro

Fraction of the Dose Absorbed

Fraction of the dose absorbed was estimated based on the mass balance and metabolite profiling data after a single oral administration of [14C]mobocertinib. A geometric mean of 3.6% and 76.0% of the dose was recovered in urine and feces of the six subjects, respectively. To process the samples for metabolite profiling, a combined urine pool across six subjects and individual fecal homogenate pool from each subject were prepared. A mean value of 92.2% and 89.0% of the radioactivity recovered in each matrix was pooled for urine and fecal homogenates, respectively, for metabolite profiling and identification. As a result, a mean of 3.3%, 67.7%, and 71.0% of the dose for urine, feces, and total excreta, respectively, was processed for the radio-profiling. A mean of 5.9% of the dose was identified as unchanged mobocertinib in fecal metabolic profile. All other identified components were either oxidative (major) or conjugative (minor) metabolites in feces. Those types of metabolites are generally not produced by gut microbiota (Pant et al., 2023). Assuming that mobocertinib found in feces was due to the unabsorbed dose and all fecal metabolites were formed after absorption followed by biliary excretion, the fraction of oral absorption for mobocertinib was estimated to be approximately 91.7% [(71.0–5.9)/71.0 × 100%] when normalized to the administered dose. It should be noted that when mobocertinib was incubated in human fecal homogenate for 24 hours at room temperature, no metabolism or degradation of mobocertinib was observed. Because unchanged mobocertinib found in feces might be partially due to biliary excretion, the 5.9% of the dose recovered in feces as unchanged mobocertinib is a conservative estimate of the maximum unabsorbed mobocertinib.

Discussion

Metabolism and excretion of mobocertinib were evaluated in healthy male volunteers after a single oral administration of [14C]mobocertinib. The geometric mean of total recovery of the administered dose was 79.6% in urine and feces with the fecal excretion as the major elimination route. The relatively low total recovery was not surprising as the covalent binding nature of mobocertinib and its metabolites retaining the intact acrylamide can slow down the elimination rate in humans, while the sample collection was conducted in a limited time after administration. Similar recovery was also observed for other targeted covalent inhibitors administered to humans (Dickinson et al., 2016; Vuu et al., 2022).

The incomplete recovery of TRA from in vitro human plasma incubated with [14C]mobocertinib was postulated to result from covalent binding of mobocertinib to plasma proteins, which increased in terms of time. Mobocertinib carries an acrylamide moiety that is electrophilic and capable of covalent binding to biological nucleophiles such as proteins. It has been reported that targeted covalent inhibitors could covalently bind to plasma proteins, particularly the most abundant protein HSA (Chandrasekaran et al., 2010; Liu et al., 2020; Vuu et al., 2022). HSA-mobocertinib conjugates were observed in both human plasma and HSA solution incubated with mobocertinib after 2 hours. The abundance of conjugates was greater in the samples incubated overnight compared with the 2-hour incubation, suggesting that more mobocertinib was bound to HSA after a longer incubation time. This was consistent with the decreasing extraction recovery of TRA at a longer incubation time. In contrast, HSA-dextromethorphan conjugates were not observed in human plasma incubated with dextromethorphan overnight. Dextromethorphan does not have a reactive functional group such as acrylamide and is not expected to covalently bind to HSA. The data clearly suggests that covalent binding of mobocertinib to HSA contributed to the low TRA recovery in human plasma incubated with [14C]mobocertinib. Further characterization of the specific residue(s) conjugated to mobocertinib in HSA was not pursued. HSA-Cys conjugate was the HSA protein with the only free cysteine-34 forming a disulfide bond with a cysteine. Interestingly, HSA-Cys-mobocertinib conjugate was detected in both human plasma and HSA solution incubated with mobocertinib. The observation suggested that mobocertinib was able to bind to noncysteine residue(s) of the HSA-Cys conjugate. That was not uncommon as several covalent inhibitors bearing an acrylamide warhead have been reported to covalently bind to lysine residues such as lysine-190 and lysine-137 of HSA (Liu et al., 2020; Vuu et al., 2022). Similar to other covalent drugs such as afatinib, alflutinib (furmonertinib), ibrutinib, neratinib, olmutinib, and pyrotinib, the covalent binding of mobocertinib to the noncysteine residue(s) of HSA was potentially reversible (Liu et al., 2020). This was evidenced by the absence of any metabolite derived from conjugation between noncysteine amino acid and mobocertinib in human plasma, urine, and feces. In contrast, metabolites (i.e., M70 and M71) related to cysteine mobocertinib conjugate were observed in the study, which suggested that the covalent binding of mobocertinib to cysteine might be either irreversible or much more stable than noncysteine mobocertinib conjugate. Conjugation of mobocertinib to other plasma proteins was not evaluated, and their contribution to covalent binding could not be excluded.

The plasma extraction recovery of TRA for targeted covalent inhibitors is usually lower than that of noncovalent small molecules as the covalent inhibitor and its metabolites that retain the intact warhead can covalently bind to plasma proteins, which cannot be extracted. The target plasma extraction recovery of TRA is usually >90% for noncovalent small molecules while it can range from high recovery down to a single digit recovery for covalent inhibitors. For example, targeted covalent inhibitors acalabrutinib, sotorasib, and osimertinib had an extraction recovery of 85%, 69%, and 8%, respectively, in human plasma in vivo (Dickinson et al., 2016; Podoll et al., 2019; Vuu et al., 2022). Similarly, the mean extraction recovery of TRA from in vivo human plasma samples was only 3.9% for mobocertinib in this study. The majority of mobocertinib-related materials (approximately 96%) were covalently bound to the plasma proteins in the circulatory system. Nevertheless, those covalently bound mobocertinib-related materials were inactive toward the target protein as their warhead was already neutralized by the plasma proteins. As a result, the metabolite profiling was focused on the supernatant after extraction.

Because the radioactive concentration was extremely low in the supernatant, the extracted components of the individual Hamilton pooled human plasma from 0- to 168-hour samples were profiled by HPLC-AMS, an ultrasensitive technique for detection of 14C concentration. The metabolites were characterized by HPLC-HRMS thereafter. Mobocertinib was the most abundant component observed in the supernatant. All other metabolites each accounted for <10% of TECRA. All eight identified metabolites are potentially active except M70 as they retain an intact acrylamide moiety. Although mobocertinib accounted for 7.7% of TECRA, it only accounted for a mean of 0.27% (7.7% × 3.5%; 3.5% was the mean overall recovery) of TRA present in the Hamilton pooled plasma. This is close to the molar area under the concentration–time curve from 0 to infinity ratio (0.172%) of mobocertinib measured with a validated HPLC-MS/MS method to plasma TRA (Hanley et al., 2024). It is worth noting that there were numerous metabolites accounting for a total of 53% of TECRA that were not identified due to low abundance in the plasma supernatant. Each of those uncharacterized metabolites accounted for <4% of TECRA.

The metabolic profiles in urine and feces indicated that mobocertinib was extensively metabolized with oxidative metabolism the major elimination pathway in humans. Mobocertinib accounted for only 0.9% and 5.9% of the dose in urine and feces, respectively. Oxidative metabolites identified including M35, M42, M44, M46, M53, M55, M60, M63, M66, M67, and M68 accounted for 43.5% of the dose, cysteine conjugate M70 accounted for 3.8% of the dose, and oxidative cysteine conjugate M71 accounted for 2.7% of the dose in both urine and feces combined. The fraction of the dose absorbed was estimated to be approximately 91.7% after a single oral administration of mobocertinib to humans. This is based on the hypothesis that unchanged mobocertinib (5.9% of the dose) was the unabsorbed dose in feces. The assumption was supported by an ADME study in bile duct cannulated rats after a single dose of [14C]mobocertinib. In that study, unchanged mobocertinib only accounted for 0.2% of the administered dose in the rat bile in which 55.8% of the dose was recovered (data on file). That being said, if any portion of mobocertinib observed in human feces was due to biliary excretion, the fraction of the dose absorbed would be >91.7%. Although mobocertinib was well absorbed in humans, the majority of mobocertinib was eliminated by extensive first-pass metabolism, resulting in moderate oral bioavailability (36.7%) (Hanley et al., 2024). The mobocertinib escaping from first-pass metabolism entered the systemic circulation and was then cleared mainly by metabolism and minimally by renal excretion (Fig. 4). An in vitro reaction phenotyping study indicated that CYP3A4/5 contributed to 93.5% of mobocertinib metabolism among the major CYPs. This was consistent with the in vivo human data. In a clinical drug–drug interaction study, the area under the concentration–time curve from 0 to infinity of mobocertinib increased 743% after coadministration of itraconazole with mobocertinib compared with administration of mobocertinib alone in healthy volunteers (Zhang et al., 2021). This suggests that CYP3A4/5 was responsible for approximately 88.1% [(1 – 1/8.43) × 100%] of the elimination of mobocertinib in humans. Assuming there was no biliary excretion of unchanged mobocertinib, renal excretion of unchanged parent accounted for approximately 1.4% [0.9/(71.0 – 5.9) × 100%] of the elimination of mobocertinib, and the remaining elimination pathways accounted for a total of approximately 10.5% (100% –88.1% – 1.4%) including cysteine conjugation and metabolism by other CYPs.

Fig. 4.

Mass balance and metabolism of [14C]mobocertinib in healthy male subjects. Fa, fraction of the dose absorbed; F, absolute bioavailability (Hanley et al., 2024); fm,CYP3A4/5, fraction metabolized by CYP3A4/5; fm,others, fraction metabolized by cysteine conjugation and other CYPs; fe,r, fraction of renal excretion; CLr, renal clearance (Hanley et al., 2024). aAssuming there was no biliary excretion of unchanged mobocertinib in humans. bThe value was from the samples pooled for metabolite profiling; a total of 8.2% of the dose in feces and 0.3% of the dose in urine were not pooled for metabolite profiling.

In summary, after a single oral administration of [14C]mobocertinib to healthy male subjects, approximately 91.7% of the dose was absorbed followed by significant first-pass metabolism. Mobocertinib related components were highly covalently bound to plasma proteins such as HSA, leaving only approximately 3.9% of noncovalently bound mobocertinib related TRA to circulate in plasma. Mobocertinib was the most abundant component in plasma supernatant after extraction, and no metabolite accounted for >10% of TECRA. Mobocertinib was mainly eliminated via oxidative metabolism with a fraction of approximately 88% metabolized by CYP3A4/5 followed by fecal excretion. Minor elimination pathways included cysteine conjugation, metabolism by other CYPs, and renal excretion of unchanged mobocertinib.

Acknowledgments

The authors thank scientists Minna Zheng and her team at Accium BioSciences, Inc for generating metabolite profile in human plasma using AMS.

Data Availability

The authors declare that all the data supporting the findings of this study are contained within the manuscript.

Authorship Contributions

Participated in research design: Chen, Shah, Kato, Griffin, Zhang, Pusalkar, Cohen, Li, Chowdhury, Zhu.

Conducted experiments: Chen, Shah, Kato, Cohen.

Performed data analysis: Chen, Shah, Kato, Pusalkar, Cohen, Zhu.

Wrote or contributed to the writing of the manuscript: Chen, Shah, Griffin, Zhang, Pusalkar, Cohen, Zhu.

FootnotesReceived June 12, 2024.Accepted July 15, 2024.

↵1Current affiliation: PepGen Therapeutics, Boston, Massachusetts.

↵2Current affiliation: Chugai Pharmaceutical Co., Ltd., Kita City, Tokyo, Japan.

↵3Current affiliation: Novo Nordisk Pharmaceuticals, Lexington, Massachusetts.

↵4Current affiliation: Servier Bioinnovation LLC, Cambridge, Massachusetts.

↵5Current affiliation: Boston Pharmaceuticals, Inc., Cambridge, Massachusetts.

This work was supported by Takeda Development Center Americas, Inc. All authors were Takeda employees when the work was performed. Some of the authors are the shareholders in Takeda. All authors declare no competing interest. This work received no external funding.

dx.doi.org/10.1124/dmd.124.001841.

↵This article has supplemental material available at dmd.aspetjournals.org.

AbbreviationsADMEabsorption, distribution, metabolism, and excretionAMSaccelerator mass spectrometryCYPcytochrome P450EGFRepidermal growth factor receptorex20insexon 20 insertionHPLChigh-performance liquid chromatographyHRMShigh resolution mass spectrometryHSAhuman serum albuminHSA-CysHSA-cysteine conjugateLSCliquid scintillation countingMS/MStandem mass spectrometryNSCLCnon–small cell lung cancerrhCYPcDNA-expressed recombinant human cytochrome P450 (Supersomes)TECRAtotal extracted circulating radioactivityTRAtotal radioactivityUHPLCultra-high performance liquid chromatographyCopyright © 2024 by The Author(s)