Orbitrap HRMS‐Based Comprehensive Metabolite Characterization of Duvelisib in Cross‐Species Hepatocytes

Balabhau Warkad; Swapnil Sharma; Prakash Niguram; Amol Chhatrapati Bisen; Sarvesh Paliwal; Anand Prakash

doi:10.1002/rcm.70141

. 2026 Jul 27;40(20):e70141. doi: 10.1002/rcm.70141

Orbitrap HRMS‐Based Comprehensive Metabolite Characterization of Duvelisib in Cross‐Species Hepatocytes

Balabhau Warkad ¹, Swapnil Sharma ¹, Prakash Niguram ², Amol Chhatrapati Bisen ³, Sarvesh Paliwal ¹, Anand Prakash ^4,^✉

PMCID: PMC13408335 PMID: 42515940

ABSTRACT

Rationale

Duvelisib (DVB), a selective PI3K‐δ/γ inhibitor approved for chronic lymphocytic leukemia, requires detailed metabolic characterization to support early drug development and selection of appropriate toxicology species. Because drug metabolism varies across species, an integrated system such as primary hepatocytes containing both Phase I and Phase II enzymes is critical for accurately defining metabolic pathways and identifying potential reactive metabolites.

Methods

DVB metabolism was investigated using primary hepatocytes from human, monkey, dog, rat, and mouse. Incubated samples were analyzed using high‐resolution mass spectrometry (HRMS) to detect and characterize metabolites. Structural elucidation was performed based on MS/MS fragmentation patterns. Mechanistic studies were conducted using a selective aldehyde oxidase (AO) inhibitor to assess the role of AO in DVB metabolism.

Results

DVB underwent diverse metabolic transformations, including oxidation, N‐dealkylation, N‐glucuronidation and glutathione (GSH) conjugation, leading to the identification of 18 putative metabolites. Thirteen metabolites, including several glucuronide and GSH conjugates were newly identified. AO‐mediated oxidation on the purine ring emerged as a major pathway and was significantly reduced in the presence of an AO inhibitor, confirming its involvement. Two previously unreported GSH conjugates were also characterized, suggesting potential sites of reactive metabolite formation and associated toxicity risk. Comparative analysis revealed distinct species‐dependent metabolic profiles, with both shared and species‐specific metabolites identified among human, mouse, rat, dog, and monkey hepatocytes.

Conclusions

The present study provides a comprehensive cross‐species metabolic profile of DVB, highlighting key pathways and novel metabolites, including those linked to potential toxicity. Mouse hepatocytes are recommended for future toxicological studies.

Keywords: aldehyde oxidase, characterization, GSH, hepatocytes, metabolites, Orbitrap

1. Introduction

DVB is an orally administered drug used to treat certain blood cancers, including chronic lymphocytic leukemia, small lymphocytic leukemia, and follicular lymphoma. Its therapeutic activity arises from the inhibition of the phosphoinositide 3‐kinase (PI3K) enzymes PI3K‐δ and PI3K‐γ, which play critical roles in cancer cell growth and survival. Phosphoinositide 3‐kinases are dual‐specificity lipid kinases involved in key signal transduction and protein trafficking pathways [1, 2, 3, 4, 5]. Early metabolite identification is a cornerstone of drug discovery and development, enabling recognition of pharmacologically active or potentially toxic biotransformation products [6]. Accordingly, metabolic stability studies using species‐specific hepatocytes are widely employed to characterize both Phase I and Phase II metabolism [7, 8]. These data guide structural optimization, predict in vivo performance, support structure metabolic stability relationships and enable compound triage prior to animal studies [9]. As reported in the literature, in vitro hepatocyte studies have indicated that the hepatic extraction ratio (> 0.3) of DVB is moderate to high in dogs and mice, whereas it is low to moderate (≤ 0.3) in rats, monkeys and humans. DVB is primarily metabolized by CYP3A4 with additional contributions from CYP1A2, CYP2B6 and CYP2C8. Among the UGT enzymes, UGT1A4 facilitates direct glucuronidation of DVB with smaller contributions reported from UGT1A3 and UGT1A9. Overall, DVB is slowly metabolized in rats, moderately metabolized in dogs and humans, extensively metabolized in mice and monkeys [10, 11, 12, 13, 14, 15, 16]. Sonawane et al. reported five metabolites in human and rat liver microsomes as well as in human S9 fractions, whereas the US FDA label for DVB also lists its major metabolites [17].

Despite these reports, the literature lacks a comprehensive cross‐species comparison of DVB metabolism, particularly regarding potentially toxic metabolites. A detailed understanding of metabolic pathways across species is critical because interspecies variability can significantly affect the translational relevance of preclinical safety data. This gap underscores the need for thorough investigations using hepatocytes from multiple species to identify species‐specific metabolites, assess their toxicological implications and define the enzymatic pathways responsible for DVB metabolism, thereby enabling better prediction of metabolite‐mediated toxicity risks. The objective of this study was to comprehensively identify and characterize DVB metabolites using mouse, rat, dog, monkey and human hepatocyte systems capable of capturing both Phase I and Phase II biotransformation pathways. Leveraging LC‐HRMS with high‐resolution Orbitrap detection, we characterized metabolite structures and delineated their biotransformation pathways using precise MS/MS fragmentation data. In total, 18 putative metabolites were identified and their corresponding metabolic pathways were proposed. The involvement of aldehyde oxidase (AO) was confirmed through experiments using a selective AO inhibitor, which produced a marked reduction in the major oxidative metabolite on the purine moiety, thereby establishing AO‐mediated oxidation on the aza‐aromatic ring. Additionally, previously unreported glucuronide and glutathione (GSH) conjugates were detected, the latter indicating the formation of reactive intermediates that may pose potential toxicity liabilities.

2. Experimental

2.1. Chemicals and Reagents

DVB (98%) was obtained from Angene International Limited (England, UK). Ketoconazole (99%) and Raloxifene HCl (98%) were purchased from Sigma‐Aldrich (USA). Human, mouse, rat, dog and monkey hepatocytes were purchased from Gibco (USA). β‐nicotinamide adenine dinucleotide phosphate tetra sodium salt (NADPH) was purchased from Sisco research laboratory (India). Dimethyl sulfoxide (DMSO) and formic acid were obtained from Sigma‐Aldrich (USA). Ammonium formate was purchased from Honeywell (USA). Acetonitrile (ACN) and methanol were purchased from J. T. Baker (USA). All other chemicals and other reagents were used without further purification or modification.

2.2. In Silico SOM Prediction of DVB Through XenoSite Software

In silico metabolism prediction of DVB was conducted using the XenoSite web predictor, which employs neural network‐based algorithms to predict cytochrome P450‐mediated sites of metabolism (SOM) [18, 19] The XenoSite software predicted potential oxidative metabolic hot spots predominantly located on the aromatic and heterocyclic regions of DVB, consistent with sites susceptible to oxidation and N‐dealkylation reactions. Furthermore, the model predicted the formation of Phase II metabolites, including glucuronide conjugation on the purine ring and a glutathione adduct associated with the chlorinated site. Notably, the computationally predicted metabolites demonstrated strong concordance with the in vitro metabolite profile obtained in our study, indicating that the XenoSite model provides a reliable approximation of DVB's biotransformation pathways. These findings highlight the applicability of in silico metabolism prediction tools in supporting early‐stage drug metabolism assessments and mechanistic elucidation (refer to Figure S1).

2.3. Cryopreserved Hepatocytes Incubation and Sample Preparation

Cross‐species hepatocyte metabolism of DVB was carried out using mouse, rat, dog, monkey, and human hepatocytes [20, 21]. A 10‐μM solution of DVB was incubated with mouse, rat, dog, monkey and human cryopreserved hepatocytes. One vial of cryopreserved hepatocytes was quickly transferred from the liquid nitrogen storage system into a 37°C water bath. The immersed vial was gently shaken in the water bath until completely thawed. The thawing process took approximately 90–120 s. The vial was removed immediately and transferred to a centrifuge tube containing 45 mL of prewarmed (37°C) water, thawed from the water bath when the last ice crystal was about to disappear. The contents of the vial were added to the media, and the vial was rinsed twice with thawed media to recover any residual cells. The tube was inverted a few times gently for proper mixing, followed by centrifugation for 5 min at 50 × g at room temperature. After centrifugation, the supernatant was decanted and the cells were resuspended by adding 1.0 mL of incubation buffer to the cell pellet. The cell pellet was gently mixed with the incubation buffer until dispersed, and the cells were resuspended. For cell counting, 100 μL of trypan blue dye and 25 μL of cell suspension were added to 375 μL of incubation buffer. The cells were transferred to a hemocytometer and counted under an inverted microscope. The hepatocyte lots used in this study were sourced from BioIVT (human, mouse, rat and dog) and Biopredic International (monkey). The details are as follows: Human (Catalog No. M00995, Lot No. ZMV) with a yield of 6.5 million cells and 87% viability; Mouse (Catalog No. M005052, Lot No. COE) with a yield of 6.2 million cells and 86% viability; Rat (Catalog No. M00005, Lot No. QWD) with a yield of 5.53 million cells and 84% viability; Dog (Catalog No. M00205, Lot No. WRM) with a yield of 9.4 million cells and 88.5% viability; and Monkey (Catalog No. HEP133006, Lot No. HEP133) with a yield of 6.6 million cells and 86.4% viability. Based on the cell count, the cell density was adjusted to 2 × 10⁶ viable cells/mL by adding appropriate volume of incubation buffer. The cell suspension (2 × 10⁶ cells/mL) was placed on ice. The vibramax shaker at 425 rpm was placed inside the CO₂ incubator at 37°C after achieving the appropriate temperature and CO₂ level in the incubator. The cell suspension (200 μL) containing 2 × 10⁶ cells/mL and 300 μL of test compound in incubation buffer (20 μM) was added to separate wells in a 48‐well plate. The 48‐well plates were placed on Vibramax shaker at 425 rpm to allow the compound solution and cell suspension to warm for approximately 10 min with gentle shaking, and used thawing media asINVITROGRO HTMedium. The reaction was initiated by transferring 200 μL of incubation (INVITROGRO KHB, no phenol red) buffer containing the test compound to 200 μL of cell suspension containing 2 × 10⁶ cells/mL in the respective well (The density of cells in the final assay was 1 × 10⁶ cells/mL, 10 μM used as a final concentration of test compound from 0.1% DMSO concentration). At each time point (0, 60, 120, and 180 min), 80 μL of sample was withdrawn from the 48‐well plate and the reaction was terminated immediately by adding 160 μL of ice‐cold ACN. Two biological replicates were used for analysis; one replicate was used as the verapamil incubation control. The samples were vortex‐mixed for 10 min and centrifuged at 4000 rpm for 15 min. After centrifugation, 100 μL supernatant was transferred to the loading plate, and the samples were subjected to HRMS analysis.

2.4. LC‐HRMS Instrumentation and Conditions

Metabolite identification was carried out using Ultra‐high performance liquid chromatography coupled with Q‐Exactive Orbitrap high‐resolution mass spectrometer (UHPLC‐HRMS; Thermo Fisher Scientific, Bremen, Germany). The chromatographic separation was achieved on a Zorbax XDB C18 (150 × 4.6 mm, 3.5 μm) column by using a gradient mobile phase of 10 mM ammonium formate in water containing 0.1% formic acid (mobile phase A) and ACN (mobile phase B). The following gradient program was set as time (min) versus % mobile phase B: 0–2/10, 2–4/35, 4–7/35, 7–15/75, 15–22/75, 22–24/10, and 24–28/10. The column temperature and flow rate were set to 45°C and 0.25 mL/min, respectively and the injection volume was 10 μL. Data were acquired in data‐dependent full MS and MS/MS modes with mass accuracy of less than 5 ppm. Mass spectrometry (MS) analyses were performed using an H‐ESI atmospheric‐pressure ionization interface, with data acquisition in positive ion mode and both full MS and data‐dependent MS/MS scanning. The ion transfer tube temperature was set to 320°C and the vaporizer temperature was maintained at 300°C. Sheath and auxiliary gases were controlled at 35°C and eight arbitrary units, respectively, with a sweep gas of zero. For HRMS, data acquisition was carried out for 28 min with a full scan range of 100–1100 Da. The automatic gain control (AGC) target was set at 100% for MS and 80% for MS/MS, with injection times of 100 msec (MS) and auto settings for MS/MS. RF lens voltage was set at 70, and resolution was set to 90 000 (MS) and 35 000 (MS/MS). Higher‐energy collisional dissociation (HCD) was performed using a normalized collision energy of 40%. Dynamic exclusion was set at 2.5 s to minimize repeated analysis of the same ion. Data acquisition and processing were done using Xcalibur software (Version 4.2, Thermo Fisher Scientific) and Compound Discoverer software (Version 3.0). Expected Metabolite Identification with FISh Scoring and Background Removal workflow. The workflow included retention time alignment, detection of expected metabolites and biotransformation products, isotope pattern matching, fragment ion annotation using FISh scoring and background subtraction using corresponding blank samples. A mass tolerance of ±5 ppm was applied for precursor and fragment ion matching. Chromatographic peak detection was performed using a signal threshold of 3, and background features were excluded by applying a minimum fivefold sample‐to‐blank intensity ratio.

3. Results and Discussion

This study comprehensively investigates Phase І and Phase II metabolism of DVB across hepatocyte species, viz. human, monkey, dog, rat and mouse. A total of 18 metabolites were putatively identified and characterized, revealing significant species‐specific variations and novel metabolic pathways. Figure 1 and Table 1 depict the proposed metabolite pathway and HRMS data of observed metabolites, respectively. However, Table 2 exemplifies identified metabolites across species.

Proposed plausible structures and metabolic pathways of DVB, including its metabolites.

TABLE 1.

Orbitrap MS data for putatively identified metabolites of DVB.

Peak ID	Putative metabolite identification	RT (min)	Molecular formula	m/z [M + H]+	Molecular formula change	Mass difference from DVB	Fragment ions	PPM error
DVB	Parent	19.49	C₂₂H₁₈ClN₆O⁺	417.1208	—	—	282, 256, 206, 162, 136, 119, 109, 94, 92, 69, 53	−4.2
M1	Oxidation, GSH on ‐Cl	10.4	C₃₂H₃₃N₉O₈S⁺	704.2264	‐(Cl) + (C₁₀ H₁₆ N₃ O₇ S)	+287.1012	686, 575.1803, 279	−3.1
M2	Oxidation, GSH on ‐Cl	10.6	C₃₂H₃₃N₉O₈S⁺	704.2222	‐(Cl) + (C₁₀ H₁₆ N₃ O₇ S)	+287.1012	686, 575.1803, 279	−3.4
M3	N‐Glucuronidation	14.8	C₂₈H₂₆ClN₆O₇ ⁺	593.1519	+(C₆H₈O₆)	+176.0306	417, 312	−3.3
M4	Oxidation glucuronidation	15.1	C₂₈H₂₆ClN₆O₈ ⁺	609.1475	+(C₆H₈O₇)	+192.0264	433, 355, 152 282	−4.6
M5	N‐Dealkylation	15.2	C₁₇H₁₆ClN₂O⁺	299.0933	‐(C5H2N₄)	−118.0276	266, 220, 191, 162, 282	−4.1
M6	Oxidation	15.3	C₂₂H₁₈ClN₆O₂ ⁺	433.1177	+(O)	+15.9941	298, 272, 217, 136	−3.7
M7	Oxidation	16.0	C₂₂H₁₈ClN₆O₂ ⁺	433.1161	+(O)	+15.9941	298, 272, 196, 136	−3.4
M8	N‐Glucuronidation	16.2	C₂₈H₂₆ClN₆O₇ ⁺	593.1516	+(C₆H₈O₆)	+176.0306	417, 312, 282	−3.4
M9	Oxidation glucuronidation	16.6	C₂₈H₂₆ClN₆O₈ ⁺	609.1475	+(C₆H₈O₇)	+192.0264	433, 355, 328, 152, 125, 282	−4.0
M10	Oxidation	16.8	C₂₂H₁₈ClN₆O₂ ⁺	433.1177	+(O)	+15.9941	298, 283, 248, 136	−3.4
M11	Oxidation	17.3	C₂₂H₁₈ClN₆O₂ ⁺	433.1154	+(O)	+15.9941	152, 108, 282	−3.6
M12	Hydration	17.5	C₂₂H₂₀ClN₆O₂ ⁺	435.1264	+(H₂O)	+18.0107	152, 108, 282	−3.1
M13	Oxidation glucuronidation	18.2	C₂₈H₂₆ClN₆O₈ ⁺	609.1478	+(C₆H₈O₇)	+192.0264	355, 299, 95, 71	−3.9
M14	Oxidation	19.1	C₂₂H₁₈ClN₆O₂ ⁺	433.1176	+(O)	+15.9941	340, 205, 152, 125, 282	−3.7
M15	Oxidation	19.9	C₂₂H₁₈ClN₆O₂ ⁺	433.1148	+(O)	+15.9941	340, 152, 108, 282	−3.7
M16	Hydration	20.0	C₂₂H₂₀ClN₆O₂ ⁺	435.1264	+(H₂O)	+18.0107	233, 154, 152,	−3.0
M17	Oxidative dechlorination	20.2	C₂₂H₁₈N₆O₂ ⁺	399.1491	‐(Cl), +(H O)	−17.9719	328, 264, 136	−3.7
M18	Oxidation	21.9	C₂₂H₁₈ClN₆O₂ ⁺	433.1158	+(O)	+15.9941	298, 283, 248, 136	−3.8

Open in a new tab

TABLE 2.

Identified metabolites for DVB across species of hepatocytes.

Metabolites	HH	MkH	DH	RH	MH
M1	x	x	√	√	√
M2	x	x	√	√	√
M3	√	x	x	x	x
M4	√	√	x	x	x
M5	√	√	√	√	√
M6	√	√	√	√	√
M7	√	√	√	√	√
M8	√	x	x	x	x
M9	√	√	x	x	√
M10	√	√	√	√	√
M11	√	√	x	x	√
M12	x	x	x	x	√
M13	√	√	x	x	√
M14	√	√	√	√	√
M15	√	√	x	√	x
M16	x	x	x	x	√
M17	√	√	√	√	√
M18	√	x	√	x	√

Open in a new tab

Abbreviations: DH, dog hepatocytes; HH, human hepatocytes; MH, mouse hepatocytes; MkH, monkey hepatocytes; RH, rat hepatocytes.

3.1. Parent Depletion and Formation of Metabolites Across the Species of Hepatocytes

It was observed that DVB exhibited high stability in monkey (91%), dog (92%) and rat (95%) hepatocytes, moderate stability in human hepatocytes (64%) and unstable in mouse hepatocytes (25%) (Table S1). Tentatively, 18 metabolites were identified in this study, comprising 11 Phase I and 7 Phase II metabolites across mouse, rat, dog, monkey and human hepatocytes. The predominant Phase I metabolic pathways included oxidation, N‐dealkylation, oxidative dechlorination and hydration. These pathways were detected in all five species except hydration, which was observed only in mouse hepatocytes. Phase II metabolic pathways involved glucuronidation, oxidation followed by glucuronidation, and oxidation, dechlorination followed by glutathione conjugation observed in rat, dog and mouse hepatocytes, respectively. Seven oxidative metabolites were identified. Among them, M14 and M15 were formed on the purine moiety, likely through aldehyde oxidase metabolism, whereas M6, M7, M10, M11 and M18 originated from oxidation on the phenyl ring of the isoquinolinone moiety. Metabolites M7, M10 and M14 were detected across all species. Of these, M14 was the predominant oxidative metabolite, showing a significantly higher response, particularly in mouse hepatocytes. The other oxidative metabolites were present at comparatively lower levels, with M7 and M10 being minor but consistently observed in all species. Metabolites M12 and M16 were formed via hydration and were unique to mouse hepatocytes. The N‐dealkylation metabolite (M5) was detected in all five species. Phase II metabolites were also characterized. N‐glucuronide conjugates (M3 and M8) were identified exclusively in human hepatocytes. Oxidation followed by glucuronidation (M4, M9 and M13) occurred in human and monkey hepatocytes, with M9 additionally detected in mouse hepatocytes. Oxidation, dechlorination followed by glutathione conjugated metabolites such as M1 and M2 were observed in dog, rat, and mouse hepatocytes. The identification of GSH conjugates (M1 and M2) suggests DVB's potential to form covalent adducts with nucleophilic proteins.

3.2. Role of Aldehyde Oxidase on Oxidation of DVB

It has been reported that CYP3A4 is the primary cytochrome P450 enzyme responsible for DVB metabolism. In addition, CYP1A2, CYP2B6, and CYP2C8 may be involved to a lesser extent in the formation of oxidative metabolites [21]. Interestingly, DVB consists of electron‐deficient centers at the carbon adjacent to the nitrogen atom associated with the purine ring. To explore the role of aldehyde oxidase in the oxidation at these sites and their relevance in humans, DVB was incubated in human hepatocytes in the presence of raloxifene (selective aldehyde oxidase inhibitor) and ketoconazole (CYP3A4 inhibitor). Figure 2 shows that the formation of the oxidative metabolite M14 increased from 0 to 180 min during DVB incubation in human hepatocytes. However, in the presence of raloxifene and ketoconazole, the metabolite formation dropped significantly. Results revealed that in the presence of ketoconazole and raloxifene, the responses to oxidative metabolites were reduced by 84% and 78%, respectively. Generally, aldehyde oxidase plays a substantial role in the oxidation of nitrogen‐containing heterocyclic compounds. These data indicate that both CYP3A4 and AO contribute significantly to the oxidative metabolism of DVB in human hepatocytes. The observed decrease in metabolite formation upon inhibition suggests that the metabolic stability of DVB is enhanced by attenuating both CYP3A4 and AO pathways [22]. The proposed mechanism was presented in Figure 3. This finding emphasizes the importance of considering multiple metabolic pathways when evaluating drug metabolism and potential drug–drug interactions.

Graph depicting the role of aldehyde oxidase in the oxidation of DVB.

Proposed plausible mechanism for aldehyde oxidase and CYP‐mediated metabolism of DVB.

3.3. Characterization of the DVB and Its Phase I and II Metabolites

3.3.1. MS/MS Fragmentation Pattern of DVB

The full mass spectrum of DVB has manifested a precursor ion at m/z 417.1208 [M + H]⁺ with a retention time (RT) of 19.49 min (Figure S17). Chemically DVB is (S)‐3‐(1‐((7H‐purin‐6‐yl)amino)ethyl)‐8‐chloro‐2‐phenylisoquinolin‐1(2H)‐one. The MS/MS fragmentation pattern of DVB indicated a prominent bond cleavage at the amino bridge connecting the isoquinolinone and purine moieties, resulting in major product ions at m/z 417.1208 → 282.0668 and 136.0613, respectively. Additional product ions observed at m/z 417.1208 → 324.0611, 267.0432, 246.0903, 232.0745, 207.0671, 203.0716, 180.0211, 132.0804, and 119.0352 are considered structure indicative. These fragments are significant for elucidating plausible metabolite structures using MS/MS fragmentation data and accurate mass measurements. Figure 4 and Table 1 exemplify the MS/MS fragmentation data and proposed structures.

MS/MS fragmentation pattern of DVB depicting putative structures (*m/z* 417.1208).

3.3.2. M1 and M2 (Oxidation, Dechlorination Followed by glutathione Conjugation)

M1 and M2 metabolites were eluted at RT of 10.4 and 10.60 min (Figure S17) and showed a protonated ion at m/z 704.2264 [M + H] ⁺ with chemical formula of C₃₂H₃₃N₉O₈S⁺, respectively. Both M1 and M2 had identical fragmentation spectra, indicating the addition of 287 Da to the parent molecule and suggesting oxidation, dechlorination and subsequent GSH conjugation. The characteristic product ions, viz., m/z 279 and 264, corresponded to the oxygenated portion of the isoquinoline ring. In addition, the product ion m/z 136.0616 was unchanged, indicating no modification on the purine moiety. The fragmentation pattern suggested that the metabolites were products of multiple biotransformations, including oxidation (+16 Da), dechlorination (−35 Da) and glutathione conjugation (+305 Da) of the parent. The GSH adduct formation was confirmed with the identification of neutral loss of 129 Da from m/z 704.2264 → 575.1804. Other characteristic fragments were also identified, which indicate GSH conjugation, such as m/z 704.2264 → 440.1273, 429.1117, 337.0994. 294.0576. However, no such direct fragments were observed, indicating GSH conjugation, which was confirmed by the loss of −129 Da (Figure 5 and Figure S2).

MS/MS fragmentation pattern of M1 depicting proposed structures (*m/z* 704.2264).

3.3.3. M3 and M8 (N‐Glucuronidation at Purine Ring of DVB)

M3 and M8 metabolites were eluted at an RT of 14.8 and 16.2 min (Figure S17) and showed protonated ion at m/z 593.1519 [M + H] ⁺ with chemical formula of C₂₈H₂₆ClN₆O₇ ⁺ respectively. Both M3 and M8 had identical fragmentation spectra, indicating the addition of +176 mass units to the parent molecule and suggesting glucuronidation on the purine moiety (Figures S2 and S3). The product ions like m/z 417.1209 → 282.0670 and 267.0433 were observed intact, which represent the unmodified isoquinolinone moiety. The characteristic fragment, like 136.0613, was converted to 136.0613 → 312.0940 with the addition of a glucuronide conjugate. Interestingly, the glucuronide conjugate was confirmed by a neutral loss of 176 Da from 593.1519 → 417.1209. Using MS/MS data, it was determined that glucuronidation could be occurring on the nitrogen atom in the purine ring; however, no such direct fragment was identified. Moreover, these metabolites were observed only in humans and they were designated as human‐specific metabolites (Figures S3 and S7).

3.3.4. M4, M9 and M13 (Oxidation Followed by Glucuronidation on the Purine Ring of DVB)

The M4, M9 and M13 metabolites were eluted at an RT of 15.1, 16.6, and 18.2 min (Figure S18) and showed protonated ion at m/z 609.1475 [M + H] ⁺ with chemical formula of C₂₈H₂₆ClN₆O₈ ⁺, respectively. The identical MS/MS fragmentation pattern was observed for M4, M9 and M13 indicating the addition of +192 Da mass units to the parent molecule and suggesting oxidation followed by glucuronidation of the purine moiety. The characteristic fragments like m/z 355.0687 and 152.0562 indicate the oxidation on the parent molecule and the m/z 152 fragment has manifested oxidation on the purine ring. Subsequent glucuronidation was confirmed with a fragment of interest m/z 152.0562 → 328.0870. Notably, glucuronide adduct formation was confirmed by the neutral loss of 176 Da from m/z 609.1474 → 433.1154. Other fragments at m/z 282.0670 and 267.0432 were observed intact, and no modification on the isoquinolinone moiety (Figure 6 and Figures S8 and S12).

MS/MS fragmentation pattern of M4 depicting proposed structures (*m/z* 609.1475).

3.3.5. M5 (N‐Dealkylation of DVB)

M5 metabolite was eluted at an RT of 15.2 min (Figure S18) and showed a protonated ion at m/z 299.0933 [M + H]⁺ with a chemical formula of C₁₇H₁₆ClN₂O⁺. The MS/MS spectra of M5 has manifested several product ions at m/z 282.0671, 267.0435, 256.0513, 232.0757, 206.0358, and 180.0205. The fragment ion m/z 282.0671 was formed by the loss of NH₃ from m/z 299.0933 and m/z 267.0435 was found to be retained as a parent. The MS/MS spectra of M5 exhibited cleavage of the purine moiety from the parent (Figure S4).

3.3.6. M6, M7, M10, M11 and M18 (Oxidation at the Isoquinolinone Ring of DVB)

The oxidative metabolites such as M6, M7, M10, M11 and M18 eluted at RT of 15.3, 16.0, 16.8, 17.3, and 21.9 min (Figures S18 and S19), respectively. These showed protonated molecular ions at m/z 433.1177 [M + H]⁺ with the formula C₂₂H₁₈ClN₆O₂ ⁺. The identical MS/MS fragmentation pattern was observed for these five metabolites, which exhibited the addition of +16 Da mass units to the parent molecule, suggesting oxidation on the isoquinolinone ring. Subjecting all these metabolites to HCD resulted in several product ions m/z 298.0616, 283.0382, 248.0692 and 219.0679, which were characteristic for oxidation taking place on the isoquinolinone moiety. The other m/z 136.0613, 132.0808 and 119.0351 were observed intact, especially the purine ring (Figures S5, S6, S9, S10, and S16).

3.3.7. M14 and M15 (Oxidation at the Purine Ring of DVB)

The metabolites M14 and M15 eluted at RT of 19.1 and 19.9 min, respectively (Figure S19). The distinct molecular ion of m/z 433.1176 [M + H]⁺observed with C₂₂H₁₈ClN₆O₂ ⁺. The identical MS/MS fragmentation pattern was observed for M14 and M15, indicating the addition of 16 Da to the parent molecule and suggesting oxidation of the purine moiety. Subjecting these metabolites to HCD resulted in several product ions, m/z 282.0670, 267.0434, 232.0757 and 207.0684, which were observed intact. The characteristic fragment ions m/z 152.0562 and 125.0459 represented oxidation taking place on the purine moiety, indicating an addition of 15.9941 Da (Figure 7 and Figure S13).

MS/MS fragmentation pattern of M14 depicting proposed structures (*m/z* 433.1176).

3.3.8. M12 and M16 (Hydration on the Purine Ring of DVB)

The metabolites M12 and M16 eluted at RT of 17.5 and 20.0 min, respectively (Figure S20). The distinct molecular ion of m/z 435.1264 with C₂₂H₂₀ClN₆O₂ ⁺. The identical MS/MS fragmentation pattern was observed for M12 and M16, indicating the addition of 18 Da to the parent molecule and suggesting hydration on the purine moiety. The characteristic product ion was m/z 154.0723, which was the addition of 18 Da mass units to m/z 136 Da (purine ring). Other product ions m/z 282.0670, 267.0434, 232.0757 and 207.0684 were observed intact. These metabolites were observed exclusively in mouse hepatocytes and were identified as mouse‐specific (Figures S11 and S14).

3.3.9. M17 (Oxidative Dechlorination on the Isoquinolinone Ring of DVB)

M17 was found to be formed by oxidative dechlorination m/z 399.1491 [M + H] ⁺ and observed at RT of 20.3 min (Figure S20). M17 formation was predicted as the [M + H]⁺ decreased 17.97 Da as compared to DVB. The MS/MS pattern results revealed product ions at m/z 264.1029 and 249.0790, indicating oxidative dechlorination at the isoquinolinone moiety (Figure S15).

Foremost, in this study, the initial sites of metabolism were assessed using the XenoSite software and the predictions were largely consistent with the experimentally observed data. The predicted metabolic hotspots suggested a likelihood of glucuronide conjugation on the purine ring, as well as dechlorination followed by GSH conjugation. Oxidation was also predicted on both the purine and isoquinolinone moieties. These predictions were supported by the results obtained from incubated hepatocyte samples, indicating strong concordance between the in silico and in vitro findings. This alignment underscores the significance of in silico tools in identifying metabolic sites and increases confidence in the predicted metabolic pathways. The identified pathways, such as glucuronidation and GSH conjugation, enhance the molecule's hydrophilicity, thereby promoting its elimination from the body. Additionally, the formation of chlorine‐derived electrophiles triggers major cellular detoxification mechanisms, wherein the thiol group of glutathione reacts with reactive chlorine species or chlorinated compounds. This reaction neutralizes these electrophiles, prevents potential damage to proteins or DNA and facilitates their excretion [23].

4. Conclusions

This study comprehensively reports putatively identified 18 metabolites of DVB across multiple species, including human, monkey, dog, rat and mouse hepatocytes, revealing that metabolism predominantly occurred on the purine ring via oxidation, N‐glucuronidation and oxidation followed by glucuronide conjugation. Unprecedented glutathione conjugates were also identified, their formation via dechlorination and subsequent GSH addition indicates a propensity for DVB to neutralize electrophiles and potentially limit toxicity. The contribution of AO to DVB oxidation was confirmed using a selective AO inhibitor, which demonstrated that oxidation was restricted to the purine ring in its presence. A complementary CYP3A4 inhibition study was also conducted to clarify its involvement in oxidative metabolism. Together, these results offer actionable insights for designing DVB analogues with enhanced efficacy and metabolic stability, while species‐specific features such as mouse‐exclusive hydration and human‐specific glucuronidation highlight the need for caution when extrapolating across species. Overall, the detailed metabolic profile established here advances the understanding of DVB disposition and supports future safety and efficacy evaluations.

Author Contributions

Balabhau Warkad: methodology, investigation, validation, writing – review and editing. Swapnil Sharma: supervision. Prakash Niguram: validation, methodology. Amol Chhatrapati Bisen: investigation, validation. Sarvesh Paliwal: supervision, validation. Anand Prakash: validation, supervision, investigation.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: In silico predicted site of metabolism of DVB with Phase I and Phase II isoforms by XenoSite web predictor tool.

Figure S2: MS/MS fragmentation pattern of M2 depicting proposed structures (m/z 704.2264).

Figure S3: MS/MS fragmentation pattern of M3 depicting proposed structures (m/z 593.1519).

Figure S4: MS/MS fragmentation pattern of M5 depicting proposed structures (m/z 299.0933).

Figure S5: MS/MS fragmentation pattern of M6 depicting proposed structures (m/z 433.1177).

Figure S6: MS/MS fragmentation pattern of M7 depicting proposed structures (m/z 433.1174).

Figure S7: MS/MS fragmentation pattern of M8 depicting proposed structures (m/z 593.1523).

Figure S8: MS/MS fragmentation pattern of M9 depicting proposed structures (m/z 609.1475).

Figure S9: MS/MS fragmentation pattern of M10 depicting proposed structures (m/z 433.1177).

Figure S10: MS/MS fragmentation pattern of M11 depicting proposed structures (m/z 433.1174).

Figure S11: MS/MS fragmentation pattern of M12 depicting proposed structures (m/z 435.1331).

Figure S12: MS/MS fragmentation pattern of M13 depicting proposed structures (m/z 609.1478).

Figure S13: MS/MS fragmentation pattern of M15 depicting proposed structures (m/z 433.1148).

Figure S14: MS/MS fragmentation pattern of M16 depicting proposed structures (m/z 435.1331).

Figure S15: MS/MS fragmentation pattern of M17 depicting proposed structures (m/z 399.1564).

Figure S16: MS/MS fragmentation pattern of M18 depicting proposed structures (m/z 433.1158).

Figure S17: EIC of DVB, M1, M2, M3, and M8.

Figure S18: EIC of M4, M5, M9, M10, M13, M14, and M18.

Figure S19: EIC of M6, M7, M11, M14, M15, and M18.

Figure S20: EIC of M12, M16, and M17.

Table S1: Percentage remaining of DVB in cross‐species hepatocytes over time.

RCM-40-e70141-s001.docx^{(4.7MB, docx)}

Acknowledgments

The authors gratefully acknowledge the Vice Chancellor of Banasthali Vidyapith in Rajasthan for providing research infrastructure. The authors also acknowledge the Bioinformatics Center at Banasthali Vidyapith, which is funded by DBT and DST and supports the FIST and CURIE programs at the Department of Bioscience and Biotechnology. All the authors read and approved the content of the manuscript for publication.

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this paper.

References

1. Gabani B. B., Dittakavi S., Kurakula P. K., et al., “Validated LC‐MS/MS Method for Simultaneous Quantitation of Three PI3K Inhibitors, Copanlisib, Duvelisib and Idelalisib in Mouse Plasma: Application to a Pharmacokinetic Study in Mice,” Analytical Chemistry Letters 11, no. 2 (2021): 140–152, 10.1080/22297928.2021.1895885. [DOI] [Google Scholar]
2. Blair H. A., “Duvelisib: First Global Approval,” Drugs 78, no. 17 (2018): 1847–1853, 10.1007/s40265-018-1013-4. [DOI] [PubMed] [Google Scholar]
3. Weis T. M., Gutierrez J., Kabel C. C., King A. C., Daley R. J., and Stump S. E., “Real‐World Management of Targeted Therapies in Chronic Lymphocytic Leukemia,” Journal of Oncology Pharmacy Practice 28, no. 6 (2022): 1411–1433, 10.1177/10781552221090869. [DOI] [PubMed] [Google Scholar]
4. Rodrigues D. A., Sagrillo F. S., and Fraga C. A. M., “Duvelisib: A 2018 Novel FDA‐Approved Small Molecule Inhibiting Phosphoinositide 3‐Kinases,” Pharmaceuticals 12, no. 2 (2018): 69, 10.3390/ph12020069. [DOI] [PMC free article] [PubMed] [Google Scholar]
5. Vangapandu H. V., Jain N., and Gandhi V., “Duvelisib: A Phosphoinositide‐3 Kinase δ/γ Inhibitor for Chronic Lymphocytic Leukemia,” Expert Opinion on Investigational Drugs 26, no. 5 (2017): 625–632, 10.1080/13543784.2017.1312338. [DOI] [PMC free article] [PubMed] [Google Scholar]
6. Bisen A. C., Sanap S. N., Agrawal S., Biswas A., and Bhatta R. S., “Chemical Metabolite Synthesis and Profiling: Mimicking In Vivo Biotransformation Reactions,” Bioorganic Chemistry 139 (2023): 106722, 10.1016/j.bioorg.2023.106722. [DOI] [PubMed] [Google Scholar]
7. Cai R., Liu J., Wang X., An T., and Zhang L., “Identification of Daurisoline Metabolites in Rats via the UHPLC‐Q‐Exactive Orbitrap Mass Spectrometer,” Journal of Pharmaceutical and Biomedical Analysis 252 (2025): 116482, 10.1016/j.jpba.2024.116482. [DOI] [PubMed] [Google Scholar]
8. Jaiswal A., Biradar R., Deshmukh V., Deshpande R., and Nandi S., “Comprehensive Identification and Characterization of In Vitro and In Vivo Metabolites of the Novel GLP‐1 Receptor Agonist Danuglipron Using UHPLC‐QToF‐MS/MS,” Journal of Pharmaceutical and Biomedical Analysis 267 (2026): 117128, 10.1016/j.jpba.2025.117128. [DOI] [PubMed] [Google Scholar]
9. Shao Y., Xie S., Zhu H., Du X., and Xu R.‐a., “Development of a Novel and Quick UPLC‐MS/MS Method for the Pharmacokinetic Analysis of Duvelisib in Beagle Dogs,” Journal of Pharmaceutical and Biomedical Analysis 187 (2020): 113355, 10.1016/j.jpba.2020.113355. [DOI] [PubMed] [Google Scholar]
10. Bowman C. M. and Benet L. Z., “In Vitro‐In Vivo Extrapolation and Hepatic Clearance‐Dependent Underprediction,” Journal of Pharmaceutical Sciences 108, no. 7 (2019): 2500–2504, 10.1016/j.xphs.2019.02.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
11. Flinn I. W., Patel M., Oki Y., et al., “Duvelisib, an Oral Dual PI3K‐δ, γ Inhibitor, Shows Clinical Activity in Indolent Non‐Hodgkin Lymphoma in a Phase 1 Study,” American Journal of Hematology 93, no. 11 (2018): 1311–1317, 10.1002/ajh.25228. [DOI] [PMC free article] [PubMed] [Google Scholar]
12. Horwitz S. M., Koch R., Porcu P., et al., “Activity of the PI3K‐δ,γ inhibitor duvelisib in a phase 1 trial and preclinical models of T‐cell lymphoma,” Blood 131, no. 8 (2018): 888–898, 10.1182/blood-2017-08-802470. [DOI] [PMC free article] [PubMed] [Google Scholar]
13. Mostafa G. A. E., Kadi A. A., AlMasoud N., Attwa M. W., Al‐Shakliah N. S., and AlRabiah H., “LC‐MS/MS Method for the Quantification of the Anti‐Cancer Agent Infigratinib: Application for Estimation of Metabolic Stability in Human Liver Microsomes,” Journal of Chromatography B 1179 (2021): 122806, 10.1016/j.jchromb.2021.122806. [DOI] [PubMed] [Google Scholar]
14. O'Brien S., Patel M., Kahl B. S., et al., “Duvelisib, an Oral Dual PI3K‐δ,γ Inhibitor, Shows Clinical and Pharmacodynamic Activity in Chronic Lymphocytic Leukemia and Small Lymphocytic Lymphoma in a Phase 1 Study,” American Journal of Hematology 93, no. 11 (2018): 1318–1326, 10.1002/ajh.25243. [DOI] [PMC free article] [PubMed] [Google Scholar]
15. Zhu J., Wang P., Shehu A. I., Lu J., Bi H., and Ma X., “Identification of Novel Pathways in Idelalisib Metabolism and Bioactivation,” Chemical Research in Toxicology 31, no. 7 (2018): 548–555, 10.1021/acs.chemrestox.8b00023. [DOI] [PMC free article] [PubMed] [Google Scholar]
16. Bisen A. C., Rawat P., Sharma G., et al., “Hesperidin: Enrichment, Forced Degradation, and Structural Elucidation of Potential Degradation Products Using Spectral Techniques,” Rapid Communications in Mass Spectrometry 37, no. 20 (2023): e9615, 10.1002/rcm.9615. [DOI] [PubMed] [Google Scholar]
17. Sonawane D., Sahu A. K., Jadav T., and Sengupta P., “UHPLC–Q‐TOF–MS/MS‐Based Metabolite Profiling of Duvelisib and Establishment of Its Metabolism Mechanisms,” Biomedical Chromatography 36, no. 4 (2022): e5314, 10.1002/bmc.5314. [DOI] [PubMed] [Google Scholar]
18. Matlock M. K., Hughes T. B., and Swamidass S. J., “XenoSite Server: A Web‐Available Site of Metabolism Prediction Tool,” Bioinformatics 31, no. 7 (2015): 1136–1137, 10.1093/bioinformatics/btu761. [DOI] [PubMed] [Google Scholar]
19. Niguram P. and Kate A. S., “Comprehensive Metabolite Identification Study of Arterolane Using Hydrophilic Interaction Liquid Chromatography With Quadrupole‐Time‐of‐Flight Mass Spectrometry,” Rapid Communications in Mass Spectrometry 36, no. 16 (2022): e9335, 10.1002/rcm.9335. [DOI] [PubMed] [Google Scholar]
20. Black S. R., Nichols J. W., Fay K. A., Matten S. R., and Lynn S. G., “Evaluation and Comparison of In Vitro Intrinsic Clearance Rates Measured Using Cryopreserved Hepatocytes From Humans, Rats, and Rainbow Trout,” Toxicology 457 (2021): 152819, 10.1016/j.tox.2021.152819. [DOI] [PMC free article] [PubMed] [Google Scholar]
21. Dick R. A., “Refinement of In Vitro Methods for Identification of Aldehyde Oxidase Substrates Reveals Metabolites of Kinase Inhibitors,” Drug Metabolism and Disposition 46, no. 6 (2018): 846–859, 10.1124/dmd.118.080960. [DOI] [PubMed] [Google Scholar]
22. Dalvie D. and Di L., “Aldehyde Oxidase and Its Role as a Drug Metabolizing Enzyme,” Pharmacology & Therapeutics 201 (2019): 137–180, 10.1016/j.pharmthera.2019.05.011. [DOI] [PubMed] [Google Scholar]
23. Lash L. H., “Renal Glutathione: Dual Roles as Antioxidant Protector and Bioactivation Promoter,” Biochemical Pharmacology 228 (2024): 116181, 10.1016/j.bcp.2024.116181. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials