Vorolanib

Journal of Pharmaceutical and Biomedical Analysis

j ourna l h om epage: www.elsevier.com/locate/jpba
Journal of Pharmaceutical and Biomedical Analysis 199 (2021) 114034
Development of a rapid and sensitive UPLC–MS/MS assay for simultaneous quantitation of Vorolanib and its metabolite in human plasma and application to a pharmacokinetics study
Xin Zhenga,1, Huitao Gaoa,1, Yanbao Zhanga, Xinge Cuia, Ranran Jiaa, Junli Xueb,
Wenbo Tangb, Yang Wangc, Hua Lic, Xuefei Chenc, Hongyun Wanga,∗
a Clinical Pharmacology Research Center, Peking Union Medical College Hospital, Peking Union Medical College and Chinese Academy of Medical Sciences,
Beijing 100730, China
b Shanghai East Hospital, Tongji University, Shanghai 200123, China
c Betta Pharmaceuticals Co., Ltd, China

a r t i c l e i n f o a b s t r a c t

Article history:
Received 17 January 2021
Received in revised form 18 March 2021 Accepted 19 March 2021
Available online 22 March 2021

Keywords: Vorolanib UPLC–MS/MS
Method validation Pharmacokinetics Isomers
Vorolanib is an oral tyrosine kinase inhibitor that targets vascular endothelial growth factor receptor (VEGFR) and platelet-derived growth factor receptor (PDGFR). A sensitive and specific LC–MS/MS assay was developed and fully validated for simultaneous quantification of vorolanib and its main metabolite X297 in human plasma. The two analytes were extracted from K2-EDTA plasma samples by protein pre- cipitation (PP) with acetonitrile, and chromatographically separated on a C18 reverse-phase column using a gradient elution. A SCIEX 5500 QTRAP® mass spectrometer system was operated in multiple-reaction monitoring mode (MRM) and all components were detected using positive electrospray ionization (ESI). The results successfully demonstrated that the method had satisfactory linearity, sensitivity, and selec- tivity in the concentration ranges of vorolanib (1.00−1000 ng/mL) and X297 (0.500−500 ng/mL).
In this study, two concentration related peaks in the vorolanib and X297 detection channels were observed, which were speculated to be isomers of vorolanib and X297. In order to standardize the sam- ple pretreatment process, the effect of lamp light and pH on the isomer reconversion was evaluated. The results indicated, that the exposure of samples to lamp light during the handling procedures, did not cause the conversion of the isomers. For the first time a robust and specific ultra-performance liquid chromatography tandem mass spectrometry (UPLC–MS/MS) assay for the high-throughput quantifi- cation of vorolanib and X297 in human plasma was established and validated following bioanalytical validation guidelines. The proposed method was successfully applied to clinical trials evaluating the pharmacokinetics of vorolanib tablets in Chinese advanced solid tumor patients.
© 2021 Elsevier B.V. All rights reserved.

1. Introduction

Receptor tyrosine kinases (RTKs) are the second major receptors on the cell surface and key components of signal transduction path- ways that mediate intercellular communication. These single-pass transmembrane receptors, which bind polypeptide ligands-mainly growth factors, play key roles in processes such as cellular growth, differentiation, metabolism, and motility [1]. Because of their roles as growth factor receptors, many RTKs are involved in the

∗ Corresponding author at: No. 41 Damucnag, Xidan, Xicheng District, Beijing, 100032, China.
E-mail address: [email protected] (H. Wang).
1 These authors contributed equally to this work.
occurrence or progression of various cancers, either through recep- tor/ligand overexpression or receptor function mutations [1,2]. Therefore, receptor tyrosine kinases have become hot spots in the research and development of anti-tumor drugs, among which the important anticancer targets are vascular endothelial growth fac- tor receptor (VEGFR) [3], platelet-derived growth factor receptor (PDGFR) [4] and epidermal growth factor receptor (EGFR) [5].
Angiogenesis is necessary for the progression from benign to malignant tumors, as well as the growth and metastasis of malignant cells [6–8]. Many studies have shown that angiogen- esis is caused by the overexpression of VEGF, a growth factor that plays an important role in the occurrence and develop- ment of abnormal angiogenesis [9–12]. In addition to VEGF, PDGF also plays an important role in the angiogenesis by recruiting pericytes to the newly formed blood vessels and maintaining

https://doi.org/10.1016/j.jpba.2021.114034

0731-7085/© 2021 Elsevier B.V. All rights reserved.

the stabilization and maturation of blood vessels. What’s more, pericyte-derived VEGF and cell-cell contacts may participate in promoting endothelial survival and may guide migration. The previously established endothelial/pericyte cell connections and vascular stability are disrupted when PDGF/PDGFR signaling is inhibited [13]. Considering the synergistic effects of VEGF and PDGF signaling pathways in angiogenesis, therapeutic drugs that simul- taneously inhibit VEGF and PDGR pathways are under constant research [9].
The first generation of small-molecule VEGFR inhibitors repre- sented by sunitinib and sorafenib suffer from poor kinase selectivity and have medium drug resistance and high toxicity, such as dermal toxicity and confusional state [14–16]. Compared with sunitinib, the safety of the second generation VEGFR inhibitors, such as axi- tinib has been improved, but the effective dose is still the maximum tolerated dose (MTD) [6]. Vorolanib is a highly potent VEGFR/PDGFR tyrosine kinase inhibitor (TKI) with limited tissue accumulation and a small volume of distribution, which is designed to main- tain the target effects while reducing side effect [16]. Sharing the same chemical scaffold with sunitinib, the main reason why vorolanib itself could harbor a lower toxicity under the similar antitumor effect is the intermittent inhibition of VEGFR/PDGFR, specifically [17]. Betta Pharmaceutical Co., Ltd. (Hangzhou, China) has completed a phase I clinical study of vorolanib in combina- tion with everolimus in patients with advanced clear-cell renal cell carcinoma (RCC) and is currently conducting a further evalua- tion of this combination in advanced RCC patients (NCT03095040) [16]. The phase I clinical study of vorolanib in combination with everolimus (5 mg/day) in the treatment of advanced clear-cell RCC showed good efficacy with a disease control rate (DCR) of 100 % and an objective response rate (ORR) of 32 %. Meanwhile, the combination did not show any unexpected adverse events [16]. Preclinical studies have shown that X297 is the main active metabolite of vorolanib, and its exposure in male rats is up to
23.5 % of vorolanib exposure. Therefore, it is useful to measurethe metabolite in a simultaneous method from a clinical perspec- tive.

In this article, a method using LC–MS/MS for the simultaneous detection of vorolanib and X297 in human plasma within 4.0 min was fully validated and applied to the pharmacokinetic study of a phase I clinical trial in patients with solid tumors.

2. Experiment

2.1. Chemicals, reagents, standards and materials

Analytical standards of vorolanib (purity 99.5 %) and X297 (purity 98.4 %) were supplied by Kananji Pharmaceuticals Research Co., Ltd. (Shanghai, China), who was the sponsor of pharma- cokinetic study of vorolanib. Erlotinib-D6 hydrochloride reference material, to be used as internal standard (IS), was purchased from Toronto Research Chemicals (Canada). Methanol and acetonitrile purchased from Honeywell (USA) were chromatographic grade. Formic acid with analytical grade (98 %, pure) was obtained from Sinopharm Chemical Reagent Co., Ltd. (Beijing, China). Intralipid (20 %, emulsion) and dimethyl sulphoxide (DMSO) (analytical grade) were acquired from Sigma-Aldrich Corp. (St. Louis, MO, USA). All ultrapure water used in this study was prepared using a water purification Milli-Q® system (Millipore, MA, USA). Drug- free human plasma was obtained from the whole blood of healthy volunteers, which was collected using K2-EDTA vacuum collec- tion tubes. Blood collection took place after informed consent was obtained according to the guidelines of the Phase I Unit of the Peking Medical College.

2.2. Liquid chromatographic and mass spectrometric equipment and conditions

×All the plasma samples were analyzed by an UPLC–MS/MS system to determine their drug concentrations. Chromatographic separation of target compounds was carried out on a Waters ACQUITY UPLC system equipped with an Acquity BEH C18 column (2.1 mm 50 mm, 1.7 µm) (Waters Corporation, Milford, MA, USA) with an injection volume of 10 µL at a flow rate of 0.40 mL/min. Both the autosampler and the column compartment have individ- ual temperature control, which were maintained at 10 ◦C and 40⦁C, respectively. Mobile phase A (MP-A) was composed of 0.1 %
formic acid in deionized water, whereas mobile phase B (MP-B) was composed of 0.1 % formic acid in HPLC-grade acetonitrile. A
4.00 min gradient elution program was set as follows: 30 % of MP-B (0.00–1.0 min), from 30 % to 70 % of MP-B (1.00–1.51 min), from 70
% to 90 % of MP-B (2.00–2.01 min), from 90 % to 30 % (3.50–3.51 min) and 30 % of MP-B (3.51–4.00 min).
The SCIEX QTRAP® 5500 (SCIEX, Concord, ON, Canada) was oper-ated in positive ionization mode using the multiple reaction mode (MRM) with a dwell time of 150 ms per transition. All source param- eters, which were optimized to improve sensitivity, were source temperature, 500 ◦C; collision gas, medium level; gas1, 60 psi; gas2, 50 psi; and curtain gas, 20 psi.

2.3. Preparation of stock solutions, quality controls (QC) and calibration standards (CS)

All stock solutions of vorolanib and X297 for calibration standards and quality controls were prepared separately at a con- centration of 1.00 mg/mL in DMSO. These stock solutions were then diluted appropriately with acetonitrile-water (1/2, v/v) to obtain a serial of working solution at several concentration levels. CS samples were prepared by spiking an appropriate amount of cor- responding working solution into mixed blank plasma which was prepared by plasma anticoagulated by K2-EDTA from six different healthy volunteers with equal volumes. The final concentrations of calibrators were 1.00, 2.00, 5.00, 10.0, 50.0, 100, 500, 1000 ng/mL for vorolanib and 0.500, 1.00, 2.50, 5.00, 25.0, 50.0, 250, 500 ng/mL for X297, respectively. Using similar preparation procedures, the lower limit of quantitation (LLOQ), low-QC (LQC), medium-QC (MQC), high-QC (HQC) and dilution QC (DQC) plasma samples were 1.00 3.00, 60.0, 800, and 2000 ng/mL for vorolanib, and 0.500, 1.50, 30.0, 400, and 1000 ng/mL for X297, respectively. Internal stan- dard (Erlotinib-D6) stock solution (1 mg/mL) was prepared by 100
% DMSO and was further diluted with acetonitrile-water (4/1, v/v)
to obtain IS working solution (50.0 ng/mL).
Subsequently, the CS samples, QC samples and IS working solu- tion were sub-packaged into appropriate volumes and stored at
−80 ◦C.

2.4. Preparation of plasma samples

A total volume of 50 µL of each of the CS, QC or bioanalysis samples was pipetted and 50 µL of a 50.0 ng/mL IS solution was added before the analytes were extracted by protein precipitation using 900 µL acetonitrile. The samples were subjected to vortex mixing for 60 s, then were centrifuged at 17,000 g for 10 min. A 200 µL aliquot of supernatant of each sample was collected and evaporated to dryness under a stream of nitrogen at room temper- ature. Finally, the residues obtained were dissolved with 400 µL acetonitrile-water (1/2, v/v), and the volume of injection was 10 µL.

2.5. Analysis of data

Analyst software with version 1.7.1 (AB Sciex, Concord, Canada) was used for data collection and processing. Calibration curves of the peak area ratio of the analyte to its corresponding IS (y) versus the nominal concentration (x), were calculated using least-squares linear regression with a weighting factor of 1/x2.

2.6. Method validation

The UPLC–MS/MS method was validated for the items of carry- over, integrity of dilutions, stability, extraction recovery, matrix effect, precision and accuracy, LLOQ, linearity and selectivity according to National Medical Products Administration (NMPA) of China, European Medicines Agency and US Food and Drug Adminis- tration (FDA) guidelines for the validation of bioanalytical methods [18–20].

2.6.1. Selectivity
In selectivity test, blank plasma from six different volunteers were utilized to demonstrate the selectivity of the method. One DB (blank plasma sample) and one LLOQ sample were prepared from each lot of matrix, with a total of 12 samples. If there was a certain interference in the blank sample, the response at the cor- responding retention time should be assessed via comparison with the response of LLOQ samples. The peak area of interference should not be greater than 20 % of the average value of analyte peak area of LLOQ samples and no more than 5 % of the average peak area of internal standard in the standard curve samples and quality control samples.

2.6.2. Linearity of calibration standards

A linear calibration curve was obtained using eight concentra- tion levels spanning the analytical range of vorolanib and X297. The coefficient of determination R2 0.980 meant the standard curve had a good linearity. The calculated concentration of calibration standard samples should be within 85 %–115 % of the nominal value, and LLOQ should be within 80 %–120 %. For the determination accu- racy, at least 75 % of the calibration standard samples (minimum 6 effective concentrations) should meet the above requirements and at least 50 % of the calibration standards with the same concentra- tion should meet the requirements.

2.6.3. LLOQ, accuracy and precision
The precision and accuracy were assessed by analyzing QC sam- ples at four concentration levels in sixfold over two days using the calibration curve as described in Section 2.6.2. The accuracies were calculated by comparing the average result at each concentration level with their corresponding nominal value. The intra-batch and inter-batch relative error (RE%) and the relative standard deviation (RSD%) should be within 15 % for LQC, MQC and HQC samples, which should be within 20 % for LLOQ samples.

2.6.4. Matrix effect
The matrix effect for vorolanib, X297 and internal standard were estimated by calculating the matrix factors (MFs), obtained by using the ratio of the peak area in the presence of matrix to the peak area in pure solution of the analyte at three concentration levels (LQC, MQC and HQC). The matrix factor normalized by internal standard was calculated by the ratio of the matrix factor of analyte to that of internal standard. Within-subject variability (defined as the varia- tion CV %) of matrix effects at three QC levels should be no more than 15 % across the three concentration levels for all compounds.

2.6.5. Hyperlipidemic plasma matrix effect
The hyperlipidemic plasma matrix effect was evaluated by the six-fold analysis of QC samples prepared in hyperlipidemic plasma at three analyte concentration levels. Hyperlipidemic plasma matrix was prepared as below: 60 µL of intralipid (20 %, emul- sion) was pipetted into an Eppendorf tube after the addition of 3940 µL blank human plasma, then vortex mixing for 10 s. The final concentration of triglyceride was 300 mg/dL (3.40 mM). At least four quality control samples at each concentration level should be within the range of 85 %–115 % of the theoretical concentration, and the RSD% should not exceed 15 %.

2.6.6. Hemolyzed plasma matrix effect
Six different lots of fresh blank whole blood were lysed by freez- ing at -80 ◦C for at least 1 h, then thawing under ultrasound at room temperature for 30 min, followed by the individual vortex mixing of each lot for 1 min. Six sources of hemolyzed plasma were prepared by adding 2 µL of each of the lysed whole blood lots to six 98 µL aliquots of pooled blank plasma. Then QC samples were prepared at three concentration levels for each source of hemolyze plasma. The acceptance standard of hemolytic matrix effect wase same as that of hyperlipidemic plasma matrix effect.

2.6.7. Extraction recovery
The recoveries of vorolanib, its metabolite and internal standard were assessed by comparing peak areas obtained from extracted spiked samples with those from extracted blanks to which an equivalent amount of analyte was added during reconstitution. The guidance does not mention an acceptance criterion for recovery, but the extent of the recovery of the analytes and IS should be pre- cise, reproducible and consistent along the tested concentration range.

2.6.8. Stability
Stock solutions of vorolanib, X297 and IS were evaluated for stability after storage at -80 ◦C for 6 months and after being kept at room temperature for 8 h by the ratio of the peak areas of the stored solution to those of the freshly prepared solution.
The stabilities of vorolanib and X297 in human plasma (anti- coagulant: K2-EDTA) under different conditions were investigated. These conditions included the stability of the samples after 5 freeze-
thaw cycles, short-term stability after storage at room temperature for 24 h, long-term stability after storage at −80 ◦C for 6 months. The auto-sampler stability (10 ◦C for 48 h) of processed samples
was also evaluated. Six replicates of LQC, MQC and HQC were ana- lyzed for each stability assessment, of which, the average RE% and RSD% should be less than 15 %.For stability of whole blood samples, LQC, MQC and HQC’ (1/4 of HQC concentration) samples were prepared with fresh whole blood. Initially the samples were homogenized by gently inverting
them, then they were placed on a water bath of 37 ◦C for 15 min to
achieve an even distribution of analytes among the components of whole blood. The above whole blood quality control samples were divided into two groups (A and B). For group A, plasma samples were obtained by centrifugation (1900 g, 4 ◦C) for 10 min imme- diately after the water bath to obtain the corresponding plasma sample. For group B, whole blood QC samples were stored at room temperature for 2 h after water bath, and then centrifuged to obtain plasma samples. All whole blood QC samples of the two groups were analyzed and determined in the same analytical batch. The mean value of the stability sample of each concentration in group A was regarded as a reference of the concentration (coefficient of variation should be no more than 15 %). The mean concentration of group B should be within 85 %–115 % of that of group A, and the coefficient of variation should be also no more than 15 %.

2.6.9. Dilution integrity
The concentrations of DQC for vorolanib and X297 were 2000 ng/mL and 1000 ng/mL, respectively. Six replicates of DQC samples for vorolanib and X297 were diluted 10-fold and prepared for analysis. The accuracy and precision for the six replicates after dilution should be no more than 15 %.

2.6.10. Carry-over
To evaluate the carry-over, two double blank samples were ana- lyzed following ULOQ samples. The peak response of carry-over effect should be below 20 % and 5 % of the LLOQ for analytes and internal standard, respectively.

2.7. Application of the method

The validated LC–MS/MS method was utilized for a pharma- cokinetic study of a phase I dose-escalation clinical trial, which was conducted in accordance with Declaration of Helsinki and the principles of Good Clinical Practice (GCP).
The phase I clinical trial was an open, single-center dose esca- lation study, which was approved by the Ethics Committee of Shanghai East Hospital. In this clinical trial, two solid tumor patient cohorts received vorolanib capsule of 2 dose levels, namely 200 mg and 400 mg respectively. Each patient was administered single dose and multiple doses of vorolanib sequentially. K2-EDTA tubes were used to collect blood samples. For QD (once-daily) single dose, sam- ples were collected at the following time points: within 0.5 h before dosing on day 1, 0.5, 1.0, 2.0, 3.0, 4.0, 6.0, 8.0, 12.0, 24.0, 36.0 and
48.0 h after dosing. For QD multiple dose, samples were collected within 0.5 h before dosing on day 8, day 15 and day 22 as well as within 0.5 h before dosing on day 28, 0.5, 1, 2, 3, 4, 6, 8, 12 and 24 h after dosing.
All blood samples were then centrifuged for 15 min at 1900 g (4 ◦C) to obtain the corresponding plasma. All clinical samples were frozen at −80 ◦C until analysis.

3. Results and discussion

3.1. Development of methods

A method using UPLC–MS/MS for the detection of vorolanib and its metabolite X297 in plasma was investigated. Both positive and negative ionization modes were investigated for the detection of all analytes, and better response was observed in the positive ion- ization mode. Suitable MRM transitions for vorolanib, X297 and IS were m/z 440.2 283.2, m/z 426.2 283.1 and m/z 400.2 339.2, respectively. The electrospray sources and optimized ionization conditions are depicted in Supplementary Table 1. Fig. 1 illustrated the product ion spectra of vorolanib, X297 and IS.
To improve the reproducibility and sensitivity, the LC condi- tions were optimized. Acetonitrile was chosen as the organic phase since it could stimulate a higher mass spectrometric response and a lower background compared with methanol. What’s more, a slight amount of formic acid could increase the signal-to-noise ratios and improved peak shape. It turned out that an applicable chromato- graphic profile and sensitivity were achieved when acetonitrile containing 0.1 % formic and water containing 0.1 % formic acid
pumped at 0.4 mL min−1 under 40 ◦C was used as eluent. Finally,
the retention times of the peaks of vorolanib, X297 and internal standard for quantitative analysis in MRM chromatogram were
1.72 min, 1.50 min and 0.98 min, respectively. Protein precipitation, solid-phase extraction and liquid-liquid extraction were commonly involved in the early stage of method development. After test, the protein precipitation method was eventually adopted as it repre- sented a very simple and rapid means of sample preparation with the low matrix effects.
To sum up, we established a novel and sensitive method involv- ing UPLC–MS/MS for the determination of vorolanib and X297 in human plasma.

3.2. Method validation

3.2.1. Linearity
Linearities over certain range for vorolanib and X297 were obtained by least-squares regression with a weighting factor of 1/x2 (the reciprocal of the squared concentration) in each valida- tion batch. Consequently, in each validation run at least 75 % of the non-zero standards were within 15 % of the actual value (or 20 % for the LLOQ) with correlation coefficient (r2) of all standard curves greater than 0.99.

3.2.2. Selectivity
Vorolanib, X297 and IS were separated on a BEH C18 col- umn at the corresponding retention time of each analytes. As shown in Fig. 2, the typical chromatograms of blank sample, LLOQ, and a plasma sample from a patient at 2 h after single dose of 200 mg vorolanib were illustrated. No endogenous interference was observed at the retention times of vorolanib (1.72 min), X297 (1.50 min) and the internal standard (0.98 min) in the MRM chro- matograms of double blank samples from human plasma. The concentration of LLOQ samples were 1.00 ng/mL and 0.500 ng/mL for vorolanib and X297 with accuracies (RE%) 20 %. A satisfac- tory selectivity of analytes and IS was concerned acceptable in this method.

3.2.3. Precision, accuracy and LLOQ
The LLOQ was 1.00 ng/mL for vorolanib and 0.500 ng/mL for X297, respectively. The analyte response at the LLOQ was > 5 times of the response compared to the blank sample, which indicated enough sensitivity of the method.
The intra- and inter-day accuracy (RE%) and the precision (RSD%) of LLOQ, LQC, MQC and HQC samples all met the acceptance criteria (Table 1 ).

3.2.4. Extraction recovery and matrix effect
The observed recoveries (Table 2) of vorolanib from plasma sam- ples of three QC levels were 100.1 %, 102.9 % and 96.3 %, respectively. The observed recoveries of X297 from three concentration levels were 96.2 %, 102.6 % and 94.8 %, respectively. The extraction recov- ery of IS was 97.3 %, which implied that the extraction efficiency was applicable for further quantification.
The IS normalized matrix factors were 102.2 %–105.8 % for vorolanib, and 101.0 %–105.0 % for X297, respectively (Table 2). These indicated that there was insignificant matrix effect for vorolanib and X297.

3.2.5. Matrix effect of hyperlipidemic and hemolyzed plasma
Samples at all tested QC levels prepared with hyperlipidemic plasma and hemolyzed plasma matrixes were described as previ- ous in 2.6.5 and 2.6.6 Sections. As shown in Supplementary Table 2, the accuracies of hyperlipidemic plasma and hemolyzed plasma QCs were within 15 % of their nominal concentrations, and so were the coefficients of precision.

3.2.6. Stability
Vorolanib, X297 and IS in the stock solutions were stable after 6 months storage at -80 ◦C or 8 h at room temperature (Supple- mentary Table 3). All analytes in whole blood remained stable after being placed at room temperature for 2 h (Supplementary Table 4). The results of stability of analytes in plasma samples were present in Table 3 . It revealed that vorolanib and X297 kept stable in plasma

Product ion spectra: (A) Vorolanib; (B) X297; (C) Erlotinib-D6 (IS).
Image

Fig. 2. Typical MRM chromatograms of plasma samples: (A) blank plasma sample; (B) LLOQ plasma sample; (C) plasma sample obtained 2 h after 200 mg single oral administration of vorolanib.at room temperature for 24 h and at 80 ◦C for 6 months. After sub- jecting to five freeze-thaw cycles, plasma samples remained stable. Processed samples were proved to be stable after placing in the auto-sampler at 10 ◦C for 2 days. The results of stabilities supported the conditions of clinical trial executive and ensured the reliability of the quantitation results.
3.2.7. Dilution integrity
The average accuracy (RE%) and precision (RSD%) of the both analytes were less than 15 % for DQC samples after 10-fold dilution, which expanded the quantification ranges to 1.00−2000 ng/mL for vorolanib and 0.500−1000 ng/mL for X297.

Table 1
Inter-run and intra-run accuracy and precision.
Analyte Vorolanib X297

1.00 ng/mL 3.00 ng/mL 60.0 ng/mL 800 ng/mL 0.500 ng/mL 1.50 ng/mL 30.0 ng/mL 400 ng/mL
Mean 0.906 2.96 60.8 774 0.500 1.49 31.1 402
Day 1 (n = 6) RSD% 10.3 5.5 4.3 2.1 10.3 12.5 4.6 1.5
RE%
Mean −9.4
0.957 −1.3
2.95 1.3
63.1 −3.3
776 0.0
0.533 −0.7
1.44 3.7
31.3 0.5
386
Day 2 (n = 6) RSD% RE%
Mean 7.9
−4.3
0.905 3.5
−1.7
2.85 1.3
5.2
58.4 4.5
−3.0
740 15.9
6.6
0.467 6.6
−4.0
1.44 2.1
4.3
30.1 3.4
−3.5
398
Day 3 (n = 6) RSD% RE%
Mean 6.9
−9.5
0.923 2.3
−5.0
2.92 2.4
−2.7
60.8 2.1
−7.5
763 16.1
−6.6
0.500 11.5
−4.0
1.46 1.6
0.3
30.8 2.0
−0.5
396
Inter-day (n = 18) RSD% RE% 8.4
−7.7 4.1
−2.7 4.2
1.3 3.7
−4.6 14.6
0.0 10.1
−2.7 3.4
2.7 2.9
−1.0

Day Item LLOQ LQC MQC HQC LLOQ LQC MQC HQC

Table 2
Recovery and matrix effect of analytes in plasma samples.

Analytes Conc. (ng/mL) Recovery% (n = 6) Mean recovery% (RSD %) Matrix effect %a (n = 6) (RSD %)
3.00 100.1 102.2(2.3)
Vorolanib 60.0 102.9 99.8(3.3) 105.8(2.0)
800 96.3 102.4(3.6)
1.50 96.2 101.0(7.4)
X297 30.0 102.6 97.9(4.3) 105.0(1.9)
400 94.8 103.4(3.1)
Erlotinib-D6 50.0 97.3 NA NA
NA: not applicable.
a The matrix effect was represented by the IS normalized matrix factor.

Table 3
Stability of vorolanib and X297 in human plasma samples and processed samples (n = 6).
Analytes Stability types Nominal concentration (ng/mL) Mean RSD% RE%
Vorolanib Short-term (Room 3.00 3.05 1.6 1.7
Temperature for 24 h) 60.0
800 59.1
762 4.6
3.4 −1.5
−4.8
Auto-sampler (10◦C for 3.00
48 h) 60.0 3.06
64.0 6.5
4.3 2.0
6.7
800
3.00
Freeze-thaw (5 cycles) 60.0
800 778
3.11
61.3
798 1.3
5.7
1.9
4.1 −2.8
3.7
2.2
−0.3

Long term (6 months)

X297 Short-term (Room Temperature for 24 h)
400 431 2.3 7.8
1.50 1.47 5.0 −2.0
Freeze-thaw (5 cycles) 30.0 31.1 2.2 3.7
400 412 3.4 3.0
1.50 1.48 9.3 −1.3
Long term (6 months) 30.0 28.3 3.3 −5.7
400 376 8.3 −6.0

Auto-sampler (10◦C for 48 h)
3.00 2.75 9.8 −8.3
60.0 56.6 4.5 −5.7
800 710 9.2 −11.3
1.50 1.60 4.0 6.7
30.0 30.7 3.0 2.3
400 401 2.3 0.3
1.50 1.60 10.3 6.7
30.0 30.1 8.1 0.3

3.2.8. Carry-over
No residues were observed at the retention time of analytes and the internal standard in the DB samples analyzed following the ULOQ samples throughout the entire validation process, which showed that carry-over was acceptable in the condition of the present method.

3.3. Application in the pharmacokinetic study of a phase I clinical trail

Analysis of vorolanib and X297 in plasma samples from a phase I dose-escalation study was performed according to this full- validated method. During the phase I investigation, 428 plasma
samples from 19 enrolled subjects were collected and analyzed to obtain the plasma concentrations of vorolanib and X297. The aver- age concentration-time curves of vorolanib and X297 in plasma after a single dose of 400 mg (n = 3) in the phase I clinical trial are shown in Fig. 3. And the pharmacokinetic parameters of the phase I clinical trial are shown in Table 4.

3.4. Incurred sample reanalysis (ISR)

A total of 60 plasma samples from the phase I clinical trial were reanalyzed for vorolanib and X297 to measure the reproducibil- ity, robustness and accuracy of the present method. The ISR result illustrated that 98.3 % of the differences of percentage concentra-

Image

Fig. 3. The concentration-time profiles of vorolanib and X297 in plasma after a single oral dose of 400 mg of vorolanib in Chinese solid tumor patients (n = 3): (A) vorolanib (mean + SD); (B) X297 (mean + SD).

Table 4
Mean ± SD Median (Min,Max) Mean ± SD Mean ± SD Mean ± SD
Vorolanib X297 400 mg 400 mg 10.6 ± 6.00
12.6 ± 9.42 2.97(2.95−4.02)
4.00(2.95−4.02) 852 ± 419
180 ± 62.5 7050 ± 4830
1940 ± 1170 35.6 ± 18.7

Pharmacokinetic parameters of vorolanib and X297 in Chinese advanced solid tumor patients after a single oral dose of 400 mg vorolanib in the phase I clinical trial (n = 3). Analytes Dose T1/2 (h) Tmax (h) Cmax (ng/mL) AUClast (h*ng/mL) CL/F (L/h)

tions between original and reanalyzed samples were within 20 % of their mean for both vorolanib and for X297.

4. Discussion

It was reported that the isomerization of the Z-sunitinib to the E-sunitinib could happen in presence of light and low pH during sample pre-treatment. It was proved that the maximum percent- age of sunitinib as E-isomer was reached (44 % of E-sunitinib; the percentage was calculated with respect to the sum of E + Z) after a half hour exposure of light, and the transformation of Z to E-isomer increased up to 20 % after lowering the pH of the solution [21].

Vorolanib shares the same chemical scaffold with sunitinib. In our study, we also observed two concentration related peaks in the vorolanib and X297 detection channels, which were speculated to be the isomers of vorolanib and X297. The retention times of the two isomers of voronib were 0.82 min and 1.72 min, respectively, and those of X297 were 0.52 min and 1.50 min, respectively.
According to the experience of sunitinib research, in order to standardize the sample pretreatment process and to evaluate the effect of lamp light on isomer transformation, the appropriate aliquots of plasma samples and pure solutions at the three QC con- centrations of vorolanib and X297 were exposed to the lamp light for 6 h. The results listed in Supplementary Tables 5 and 6 showed that the RSD% and RE % of vorolanib and X297 in both plasma sam- ples and working solutions at three quality control levels were all within 15 % after 6 h of illumination. Vorolanib and X297 in both plasma samples and working solutions remained stable after being placed under the room temperature with lamp light for 6 h. Besides, both the first peak and the second peak were stable under light, indicating that no isomer transformation similar to sunitinib was observed. It meant that the exposure of samples to lamp light dur- ing the handling procedures did not caused the conversion of the isomers.

In addition, to explore the effect of pH on the conversion
extent of isomers, the solution of vorolanib and X297 were pro- cessed using methanol-H2O (1/1, v/v) containing different portions of hydrochloric acid (HCl). The results in Supplementary Table 7 showed that the isomer of vorolanib at 0.82 min was gradu-
ally transformed into the isomer of vorolanib at 1.72 min with the decrease of pH, reaching the maximum transformation when pH = 0.37, and the conversion rate was 99 %. For X297, with the decrease of pH, the isomer at 0.52 min was gradually transformed into the isomer of at 1.50 min, reaching the maximum transforma- tion when pH = 0.07, and the conversion rate was 100 %.
The results indicated that the isomers observed in the vorolanib and X297 detection channels were different from the reported iso- mers of sunitinib, and were not Z-E configurational isomers. The isomers of vorolanib and X297 observed in our method remained stable during the preparation and illumination. According to the results of the pH experiment, the transformation of the isomer at
0.82 min of vorolanib and the isomer at 0.50 min of X297 were closely related to pH, and the former isomers were presumed to be related to the formation of hydrogen bonds.

5. Conclusion

For the first time, an UPLC–MS/MS method was developed and fully validated for the simultaneous quantification of the novel anticancer agent vorolanib and its main metabolite in plasma samples. The method showed a good linearity over the range of
1.00 1000 ng/mL for vorolanib and 0.500 500 ng/mL for X297. The matrix effect was not significant, and the recovery was high, which indicated reliability and accuracy of this method. The LC–MS/MS method described in this paper fully meets the quantification requirements and has been successfully used for the research of the pharmacokinetics of vorolanib tablet in Chinese patients.

Future perspective

A few studies indicated that the third-generation RTKs could achieve a curative effect in solid tumors patients. Therefore, a highly specific and sensitive LC–MS/MS method based on the protein pre- cipitation procedure was developed and fully validated to support clinical studies. It can help to evaluate the PK of vorolanib after administration and to understand the relationship between PK and pharmacodynamics, and then provide an instructive clinical devel- opment strategy for vorolanib. Besides, the present method is alsoapplicable to provide an instructive development strategy for other RTKs with similar structure to vorolanib.

Author contributions

Xin Zheng: Writing-Original Draft and Methodology; Huitao Gao: Formal analysis and Validation; Yanbao Zhang: Valida- tion; Xinge Cui: Writing-Reviewing and Editing; Ranran Jia: Writing-Reviewing and Editing; Junli Xue: Clinical investigator; Wenbo Tang: Clinical investigator; Yang Wang: Resources; Hua Li: Resources; Xuefei Chen: Resources; Hongyun Wang: Supervision and Methodology.

Funding sources

The funding sources of this study were National Natural Science Foundation of China (No: 81903726) and “13th Five-YearN¨ ational Major New Drug Projects (No: 2019ZX09734001).

Declaration of Competing Interest

The authors report no declarations of interest.

Acknowledgements

The authors greatly appreciate Kananji Pharmaceuticals Research Co., Ltd. (Shanghai, China) for funding this study. The authors thank all the staffs in Shanghai East Hospital, and others who participated in the clinical trial.

Appendix A. Supplementary data

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jpba.2021. 114034.

References

[1] S.R. Hubbard, W. Todd Miller, Receptor tyrosine kinases: mechanisms of Vorolanib activation and signaling, Curr. Opin. Cell Biol. 19 (2007) 117–123.
[2] P. Blume-Jensen, T. Hunter, Oncogenic kinase signalling, Nature 411 (2001) 355–365.
[3] R. Robert, Vascular endothelial growth factor (VEGF) and VEGF receptor inhibitors in the treatment of renal cell carcinomas, J. Pharmacol. Res. 120 (2017) 116–132.
[4] J. Brugarolas, Molecular genetics of clear-cell renal cell carcinoma, J. Clin. Oncol. 32 (2014) 1968–1976.
[5] S. Chowdhury, J.M. Larkin, M.E. Gore, Recent advances in the treatment of renal cell carcinoma and the role of targeted therapies, Eur. J. Cancer 44 (2008) 2152–2161.
[6] R.S. Hope, H.S. Roy, S. Herbst, L. Glenn, et al., Phase I trial of the oral antiangiogenesis agent AG-013736 in patients with advanced solid tumors: pharmacokinetic and clinical results, J. Clin. Oncol. 23 (2005) 5474–5483.
[7] J. Folkman, Fundamental concepts of the angiogenic process, Curr. Mol. Med. 3 (2003) 643–651.
[8] D. Hanahan, J. Folkman, Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis, Cell 86 (1996) 353–364.
[9] C.D. Ren, H. Shi, J.J. Jiang, et al., The effect of CM082, an oral tyrosine kinase inhibitor, on experimental choroidal neovascularization in rats, J. Ophthalmol. 2017 (2017) 13.
[10] S. Grisanti, O. Tatar, The role of vascular endothelial growth factor and other endogenous interplayers in agerelated macular degeneration, Prog. Retin. Eye Res. 27 (2008) 372–390.
[11] N. Hay, N. Sonenberg, Upstream and downstream of mTOR, Genes Dev. 18 (2004) 1926–1945.
[12] K. Watari, S. Nakao, A. Fotovati, et al., Role of macrophages in inflammatory lymphangiogenesis: enhanced production of vascular endothelial growth factor C and D through NF-nB activation, Biochem. Biophys. Res. Commun. 377 (2008) 826–831.
[13] G.J. Jaffe, D. Eliott, J.A. Wells, et al., A phase 1 study of intravitreous E10030 in combination with ranibizumab in neovascular age-related macular degeneration, Ophthalmology 123 (2016) 78–85.
[14] S. Bracarda, R. Iacovelli, L. Boni, et al., Sunitinib administered on 2/1 schedule in patients with metastatic renal cell carcinoma: the RAINBOW analysis, Ann. Oncol. 27 (2016) 366.
[15] S. Oudard, L. Geoffrois, A. Guillot, et al., Clinical activity of sunitinib rechallenge in metastatic renal cell carcinoma-results of the REchallenge with Sunitinib in MEtastatic RCC (RESUME) study, Eur. J. Cancer 62 (2016) 28–35.
[16] X.N. Sheng, X.Q. Yan, Z.H. Chi, et al., Phase 1 trial of vorolanib (CM082) in combination with everolimus in patients with advanced clear-cell renal cell carcinoma, EBioMedicine 55 (2020), 102755.
[17] J.C. Bendell, M.R. Patel, K.N. Moore, et al., Phase I, first-in-human, dose-escalation study to evaluate the safety, tolerability, and
pharmacokinetics of Vorolanib in patients with advanced solid tumors, Oncologist 24 (2019), 455–e121.
[18] China Food and Drug Administration, Guideline on Bioanalytical Method Validation, China Pharmacopeia (version 2015), Part 2, 2015, pp. 363–368.
[19] European Medicines Agency, Guideline on Bioanalytical Method Validation, 2011 http://www.ema.europa.eu/docs/en GB/documentlibrary/ Scientificguideline/2011/08/WC500109686.pdf.
[20] US Food and Drug Administration, Guidance for Industry: Bioanalytical Method Validation, 2018 https://www.fda.gov/downloads/Drugs/ GuidanceComplianceRegulatoryInformation/Guidances/UCM070107.pdf.
[21] B. Posocco, M. Buzzo, L. Giodini, et al., Analytical aspects of sunitinib and its geometric isomerism towards therapeutic drug monitoring in clinical routine, J. Pharm. Biomed. Anal. 25 (2018) 360–367.