Ecofriendly, Simple, Fast and Sensitive UPLC-MS/MS Method for Determination of Erdafitinib, a Novel Tyrosine Kinase Inhibitor, in Plasma and Its Application to Metabolic Stability

Abstract

Erdafitinib is an oral pan-fibroblast growth factor receptor (FGFR) inhibitor and has a potent antitumor activity against FGFR-aberrant malignancies. Erdafitinib has a narrow therapeutic index, and its pharmacokinetics are influenced by genetic variability and interacting medication. Routine therapeutic drug monitoring and dose adjustment are recommended. This study aims at developing a new UPLC-MS/MS method for determination and quantitation of erdafitinib in human plasma using ibrutinib as an internal standard. The sample ionization was performed by using electrospray ionization in positive mode, and multiple reaction monitoring was used for the quantification of target analytes. The chromatographic separation of erdafitinib and IS was achieved by an UPLC BEH C₁₈ column (2.1 mm × 100 mm, 1.7 μm). Erdafitinib metabolic stability was studied using intrinsic clearance and in vitro half-life. The greenness of the developed method was evaluated using appropriate, analytical Eco-Scale and AGREE software. The linearity of the established UPLC-MS/MS assay ranged from 0.5 to 1000 ng/mL with r > 0.99 with a limit of quantitation of 0.5 ng/mL. The accuracy and precision were within acceptable limits and the average recovery and matrix effects were 86.11% and 90.51%, respectively. Erdafitinib metabolic stability was studied and its in vitro half-life was 7.28 min and intrinsic clearance was 95.11 µL/min/mg. The assessment of the greenness profile of the method indicated that the method is eco-friendly. The proposed method can be utilized for therapeutic drug monitoring and pharmacokinetic studies.

1. Introduction

Urothelial carcinoma or transitional cell carcinoma is the most common type of cancer, affecting the transitional cells of the urinary system. It accounts for the majority of bladder cancer diagnosis with more than 90% of bladder cancer cases. It is more common in males than females [1,2,3] and one out of five advanced urothelial carcinoma patients has a mutation in the fibroblast growth factor receptor (FGFR) gene, which is responsible for increased tumor cell growth and survival and accompanied with lower sensitivity to immune interventions [4]. Erdafitinib (ERDA) is a potent selective FGFR tyrosine kinase inhibitor [5]. It has been recently approved by the FDA for the treatment of advanced cancer of the urinary bladder [6]. ERDA binds to and blocks the enzymatic activity of cell proteins, FGFRs [7]. It also binds to other proteins, including tyrosine kinase 4 (FLT4), vascular endothelial growth factor receptor 2 (VEGFR2) and platelet-derived growth factor receptor alpha (PDGFRA) [3]. It is currently being evaluated as monotherapy or combination therapy with anti-programmed cell death protein-(PD-1) in a mouse model with PD-1 blocked in an autochthonous lung cancer; erdafitinib monotherapy was effective but without survival benefit and anti-PD-1 treatment alone was ineffective. However, erdafitinib and an anti-PD-1 combination induced significant tumor regression and improved survival rate [8].

It was reported that ERDA was well tolerated in Japanese patients with advanced or refractory solid tumors [9,10]. It was absorbed rapidly following single-dose administration with the maximum concentration achieved after 2–3 h and 2–6 h, following multiple daily dosing and the steady state was reached after 8 days of treatment [9]. ERDA pharmacokinetics can be affected by co-administration with other cytochrome P450 inhibitors/inducers, such as some antifungal drugs. The co-administration of ERDA with fluconazole, posaconazole and isavuconazole (CYP3A4 inhibitor) or itraconazole (CYP2C9 inhibitor) increases ERDA blood levels [11,12]. Race, age and mild hepatic impairment have no significant effect on ERDA blood concentration (11).

The toxicity profile of ERDA has shown ocular toxicity even after a short treatment duration [13,14,15]. ERDA has some adverse effects, including cases of elevated aspartate transaminase (AST), alanine aminotransferase (ALT) and alkaline phosphatase [16,17]. ERDA most common adverse effects are dysgeusia, stomatitis, diarrhea, hyperphosphatemia and dry mouth [7]. Moreover, hyperphosphatemia was a common side effect in patients treated with ERDA [15].

There is a large inter-individual difference in ERDA pharmacokinetics, leading to variability in doses required to achieve target concentration [9]. To achieve this balance and to individualize drug dosing, therapeutic drug monitoring (TDM) of ERDA is required. ERDA dose is considered based on the presence of susceptible FGFR genetic alteration in patient tumor. However, the source of variability is multifactorial, including pathophysiology, environment, diet, drug-drug interaction. All these factors have an impact on ERDA pharmacokinetics and pharmacodynamics. In order to achieve this goal, it is necessary to develop bio-analytical assays that are reliable, very sensitive and fast. Few methods were published using HPLC-UV, but they are characterized by high limits of quantitation (50 ng/mL) [18] with others using LC-MS/MS [19,20]. The present work was formulated to develop a novel, simple, sensitive, accurate and validated method for the determination and quantification of ERDA in human plasma that could be applied to the TDM, toxicokinetic and pharmacokinetic studies.

2. Materials and Methods

2.1. Chemicals

ERDA (Figure 1) and IS were procured from Toronto Research Chemicals (North York, ON, Canada) and Beijing Mesochem (Beijing, China), respectively. Ammonium acetate, ammonium format, tert-butyl methyl ether, formic acid and acetonitrile (HPLC grade) were purchased from Sigma Aldrich (St. Louis, MO, USA) and DMSO (Dimethyl sulfoxide) (Sigma Aldrich; Darmstadt, Germany). Human blank plasma was kindly delivered by “Blood Bank of King Khalid University Hospital” (Riyadh, Saudi Arabia).

2.2. Instrumentations and Analytical Conditions

ERDA analysis was performed using an H-class ultra-performance liquid chromatography (UPLC) system (Waters, Milford, MA, USA). An Acquity BEH C₁₈ column (2.1 mm × 100 mm, 1.7 µm) “Waters, Milford, Massachusetts, USA” was used for separation and determination of EDRA. A mixture of 0.1% formic acid in acetonitrile and water (0.1% formic acid) (65:35) was used as a mobile phase and eluted at flow rate 0.25 mL/min. The outlet of the column was coupled to a Waters TQD Class Mass Spectrometer (Waters, Milford, MA, USA) with an electrospray ionization (ESI) source operated in the positive ion mode. Mass spectrometry parameters were adjusted at optimized conditions for separation of ERDA and IS Multiple reaction monitoring (MRM) mode was selected and used, in which the ion transitions for ERDA and IS were m/z 447.2 → 86.0 and m/z 441.1 → 84.0, respectively (Figure 2). The details of compound and instrumental parameters were illustrated in

Table 1. Mass optimization parameters for ERDA and IS.

Parameters	ERDA	IS
I. Compound Parameters
Precursor ion (m/z)	447.2	441.1
Product ion (m/z)	86.0	84.0
Dwell time (s)	0.05	0.05
Cone Voltage (V)	38	50
Collision energy (eV)	30	40
II. Instrument Parameters
Collision gas flow rate (mL/min)	0.1	0.1
Nitrogen flow rate (L/h)	600	600

. Each of ERDA and IS was separated at about 1.23 and 1.02 min, respectively, with a total run time of 2.0 min.

2.3. Standard Solutions, Calibration Standards and Quality Control (QC) Samples

Stock solutions of ERDA were prepared by dissolving 5.0 mg of ERDA in DMSO to yield a concentration of 2.0 mg/mL. Separate working solutions (100 µg/mL) for calibration curves (CCs) and quality control (QC) samples were prepared by further dilution of the stock solution with methanol. IS stock solution (1.0 mg/mL) and its working solution (100 µg/mL) were prepared with the same method as ERDA. Then, the working standards of CCs and QC samples were spiked to blank human plasma samples to attain the CCs range of 0.5–1000 ng/mL and QC samples (1.5, 150 and 750). Spiked plasma samples were kept in a deep freezer at −80 °C until further analysis while the solutions were stored in a refrigerator.

2.4. Sample Preparation

To 100 µL CCs and QC plasma samples, 10 µL IS working solution and 50.0 µL acetonitrile were added. After mixing by vortex for 1 min, 1.0 mL tert-butyl methyl ether was added. Samples were then mixed by shacking for 20 min. After centrifugation at 10,500× g for 10 min at 8 °C, the supernatant was transferred to a clean tube and evaporated until dryness using vacuum speed concentrators at 40 °C. The dried extracts were dissolved using 100 µL of the mobile phase and was injected (5 µL) into the UPLC-MS/MS system for analysis.

2.5. Method Validation

According to the FDA and EMA bioanalytical method validation guidelines [21,22], the proposed method was validated. The validation was conducted in terms of the selectivity, sensitivity, calibration curve, lower limit of quantification (LLOQ), accuracy and precision, matrix effect, extraction efficacy (recovery) and stability at various processing and storage conditions.

2.5.1. Selectivity

To evaluate the method selectivity, chromatograms of blank human plasma from six different sources were compared to the plasma spiked with ERDA at low limit of quantitation (LLOQ) and IS. The precision and accuracy of LLOQ are required to be ≤20% and ±20% of the nominal concentration, respectively.

2.5.2. Accuracy and Precision

In addition to LLOQ, three different levels of quality control samples (QCs) of the ERDA were analyzed in six replicates on the same day and in three successive days to evaluate intra-day and inter-days’ accuracy and precision, respectively. According to the guidelines, precision (in the term of RSD%) and accuracy of the method should be within ±15% for the all QCs and ±20% for the LLOQ.

2.5.3. Matrix Effect and Recovery

Enhancement or suppression of signal on ESI-MS/MS response due to matrix components was evaluated by comparing the response of the analyte with respect to the response in the solvent. Extracted drug-free plasma samples obtained from six different sources were spiked with the analyte at concentrations equivalent to three QC levels and IS (100 ng/mL). Post extracted blank plasma and the pure standard solution containing the analytes and IS was also prepared at the same level of QC concentrations. To quantify for the loss of the compounds during the sample treatment process, the extraction efficiency (recovery) was estimated by comparing the analysis response (peak areas) of pre-extraction spiked serum with those obtained from post-extraction samples at the three QC concentrations.

2.5.4. Linearity

Calibration curves of ERDA in human plasma were constructed using concentrations ranging from 0.5 (LLOQ) to 1000.0 ng/mL. They were obtained by weighted 1/X2 linear regression analysis. The peak area ratio of ERDA to IS was plotted against the ERDA concentration. Precision and accuracy for the back calculated concentrations of the calibration points should be within ≤15% and ±15% of their nominal values, respectively. However, for LLOQ the precision and accuracy should be within ≤20% and ±20% of their nominal values. The minimum concentration of ERDA in CC was considered as LLOQ, which had a signal-to-noise ratio of ≥5 in comparison with blank plasma [14,19].

2.5.5. Stability

Two levels of QC samples (LQC and HQC) under different processing and storage conditions were analyzed in six replicates to evaluate the stability of the method. Processing conditions and the short-term stability of ERDA were evaluated after storage for 10 h at room temperature (approximately 24 °C). The freeze–thaw stability was determined using QC plasma samples after three cycles of freeze at −80 °C and thaw at room temperature. The long-term stability was evaluated by analyzing QC plasma samples after storage at −80 °C for 4 weeks. The auto-sampler stability was evaluated by re-injecting the reconstituted QC samples after 24 h of its storage in autosampler. The acceptable limits for QC samples should be ±15% for accuracy and ≤15% for precision.

2.6. In Vitro Metabolic Stability of ERDA

A stock solution of studied ERDA was prepared at a concentration of 1 mg/mL in DMSO and then diluted in phosphate buffer (pH = 7.4) to obtain a concentration of 800 ng/mL. A 20 mM NADPH (25 µL) solution in phosphate buffer was added to 10 µL of ERDA (800 ng/mL) following addition of 450 µL of 0.1 M phosphate buffer (pre-warmed, 37 °C). Incubations were initiated by 10 µL of microsomes and reactions were started. The tubes were shaken and kept at 37 °C for 1 h. Then 250 µL of acetonitrile containing IS (100 ng/mL) was added at different time intervals (0.0, 0.5, 10.0, 20.0, 30.0 and 40.0 min) to terminate the reaction. After centrifugation at 10,500× g for 5 min at 10 °C, the supernatant was transferred to the Eppendorf tube. Then, 5 µL of each sample was injected into the UPLC-MS/MS. ERDA remaining concentration was calculated at each time interval for evaluation of ERDA metabolic stability. Different concentrations of ERDA were prepared in the mobile phase to construct a calibration curve of peak area ratio (ERDA/IS) against incubation time. The range of the calibration curve was 0.5–1000 ng/mL. From this curve, another curve was plotted between logarithm (Ln) concentrations that exhibited linearity and incubation times. Metabolic half-time (t_1/2) was calculated from the slope of the linear regression [23].

2.7. Assessment of the Method Greenness

Three different methods were used to assess the greenness of the method. The first method is according to national environmental method index (NEMI) protocol [24]. The profiles and acceptance criteria of greenness are developed and optimized by interpreting the data from an analytical procedure into a greenness report that includes chemicals used, pH, and waste generated. The report patterns are distinguished by four features or parameters: hazardous; corrosive; persistent, bio-accumulative and toxic (PBT); and applying the profile criteria of the emergency planning and community right-to-know act (EPCRA) [25].

The second method is analytical Eco-Scale tool [26,27], and it is based on penalty points subtracted from a base of 100. The higher the score, the greener, as well as economical, the analytical procedure is. However, the disadvantage of this method is the lack of information about the structure of the hazards obtained and the lack of information about the impact of the analytical procedure on the environment, such as the use of solvents, other reagents, occupational hazard or generation of waste. The third method is the Analytical GREEnness calculator. It is a new assessment approach proposed by Pena-Pereira et al. [28]. The evaluation criteria of Analytical GREEnness Metric Approach and Software (AGREE) were taken from the twelve principles of green analytical chemistry and transformed into 0–1 range. The higher average scores the method receives, the greener it is.

3. Results

For MS/MS optimization, ERDA (500 ng/mL) was directly infused into a mass spectrometer instrument using Electrospray ionization (ESI) to produce ions. It was observed that ERDA is more sensitive in ESI positive mode and produced the protonated molecular ion of m/z 447.2. Protonated precursor ion fragmentation produces the highest products ion of m/z 86.0. Similarly, the internal standard (ibrutinib) also produced the protonated precursor ion of m/z 441.1, and, after fragmentation, the highest product ion produced was m/z 84.0. Figure 2 represents a full scan for ion spectra of protonated parent (precursor) ions for [M + H] + for ERDA and IS. The optimized parameters chosen for separation are illustrated in Table 1.

Several types of columns with different length and particle size, including Waters BEH C₁₈, Waters BEH HILIC and Waters CSH C₁₈, were tested. Finally, Waters BEH C₁₈ was selected. Chromatographic conditions were optimized to achieve the highest sensitivity, optimum peak shape and shorter retention time. Different compositions of mobile phase were tested. Mobile phase mixtures of different ratios of acetonitrile and water, ammonium format or ammonium acetate with or without formic acid (0.1%) were investigated, concerning their effects on the chromatographic behavior. As a result, a mixture of 0.1% formic acid in acetonitrile and water (0.1% formic acid) (65:35) was used as a mobile phase and eluted at flow rate 0.25 mL/min. It was selected at a flow rate of 0.25 mL/min as it provided the best results regarding the peak shape and response.

3.1. Sample Preparation

Sample preparation is a critical step that affects the sensitivity of the analytical method. Protein precipitation was tested using acetonitrile and methanol as a precipitating agent. It was fast and easy. However, the extraction efficiency was low. Moreover, different solvents, such as ethyl acetate, dichloromethane, diethyl ether and tert-butyl methyl ether, were investigating for their efficiency to extract ERDA from plasma. The best extraction recovery was obtained using tert-butyl methyl ether. More efficient recovery was achieved using protein precipitation followed by extraction with tert-butyl methyl ether. The advantage of using acetonitrile for protein perception prior to liquid–liquid extraction is to eliminate the effects of matrix interference [29]. The use of acetonitrile as precipitating agent reduces the solution layer around proteins causing them to aggregate and drop out the solution.

3.2. Method Validation

According to the FDA and EMA bioanalytical method validation guidelines [23,24], the proposed method was validated. Method selectivity and sensitivity, linearity, lower limits of quantification, extraction efficiency (recovery), linearity, lower limits of quantification, matrix effect, accuracy and precision and stability studies were evaluated for the validation of the proposed method. In addition to LLOQ, three concentrations of QC samples representing the entire range of the standard curve were used for method validation purposes; LLOQ, low QC sample (within three times of LLOQ), middle QC sample (near the center of the calibration range), and high QC sample (near the upper limit of the calibration range).

3.2.1. Specificity

No interference was observed at the retention times of both ERDA and IS. Figure 3 represents the chromatograms of blank plasma and plasma samples spiked with LLOQ level of ERDA. The response (peak area) of ERDA at its LLOQ was at least five times higher than that of the blank. These results indicate the specificity of the current method.

3.2.2. Linearity and Lower Limit of Quantification (LLOQ)

Standard calibration curves were constructed in the range of 0.5–1000.0 ng/mL by plotting peak area ratios (ERDA/IS) against ERDA-spiked concentrations. Linearity was calculated using the least square method (1/_X2). The linearity of the method was >0.997.

LLOQ for the determination of ERDA in plasma was found to be 0.5 ng/mL using the proposed method. The LLOQ in the proposed method was very low, which reflects its efficiency for application in the quantitative detection of ERDA plasma for pharmacokinetic studies and TDM. Figure 3 shows the MRM chromatograms of plasma samples spiked with ERDA at its LLOQ, along with blank plasma samples.

3.2.3. Precision and Accuracy

Precision and accuracy were evaluated in the term of relative standard deviation (RSD %) for precision and the term of recovery percentage of nominated concentrations for accuracy. Both precision and accuracy were evaluated by analysis of three concentrations of QC samples (LLQC, MQC, HQC) and LLOQ, six times in one day (intra-day) and for three successive days (inter-days). The results of accuracy and precision for ERDA are presented in

Table 2. Inter- and intra-day precision and accuracy of ERDA in plasma.

Nominal Conc. (ng/mL)	Inter-Day			Intra-Day
	Actual Conc. (Mean ± SD)	Precision (CV %)	Accuracy (%)	Actual Conc. (Mean ± SD)	Precision (CV %)	Accuracy (%)
0.5	0.41 ± 0.04	9.76	82.00	0.41 ±0.046	11.22	82.00
1.5	1.32 ± 0.18	13.50	88.00	1.30 ± 0.11	8.46	86.66
150	130.75 ±15.62	11.95	87.16	130.23 ± 6.87	5.28	86.82
750	644.15 ±37.10	5.76	85.88	643.21 ± 77.17	11.99	85.76

. The low values of % RSD indicated that the precision of the developed method was within the acceptable range required by the concerning guidelines for the determination of EDRA in human plasma.

3.2.4. Recovery and Matrix Effect

Extraction efficiency or recovery and matrix effects of ERDA and IS from plasma samples were evaluated using QC plasma samples of ERDA at three different concentration levels: low (1.5 ng/mL), medium (150 ng/mL) and high (750 ng/mL). The recovery value ranged from 85.06% to 86.93% for ERDA and was 86.25% for IS, which indicates an excellent extraction recovery of the optimized method for ERDA from plasma samples. Matrix effects ranged from 89.89% to 91.60% for ERDA and were 89.49 for IS (

Table 3. Recovery % and matrix effects ERDA and IS in plasma.

Compound	Nominal Conc. (ng/mL)	Extraction Recovery		Matrix Effects
		Mean (%)	RSD (%)	Mean (%)	RSD (%)
ERDA	1.5	86.34	6.56	91.6	7.81
	150	86.93	10.39	90.03	5.16
	750	85.06	2.90	89.89	4.69
IS	100	86.25	6.50	89.49	5.61

). These results reflect that the matrix effect of the current method is negligible.

3.2.5. Stability

ERDA stability under different conditions of processing and storage was evaluated using plasma samples spiked at two different ERDA concentrations, LQC (15 ng/mL) and HQC (750 ng/mL).

Table 4. Stability of ERDA in plasma under different processing and storage conditions.

Stability	Nominal Con. (ng/mL)	Measured Con. (ng/mL)	Precision (%)	Accuracy (%)
Short-term	15	13.09 ± 1.56	11.92	87.27
	750	621.15 ± 59.82	9.63	82.82
Long-term	15	12.76 ± 1.57	12.30	85.07
	750	648.58 ± 35.38	5.45	86.48
Thaw and freeze	15	12.91 ± 0.92	7.13	86.07
	750	639.32 ± 61.90	9.68	85.24
Auto-sampler (24 h)	15	12.93 ± 0.83	6.41	86.20
	750	671.88 ± 54.78	8.15	89.58

represents the calculated recovery values for ERDA, which were more than 82.82% with CV % less than 12.30%, indicating that the analyt is stable under various conditions of processing and storage (Table 4).

3.3. In Vitro Metabolic Stability of ERDA

The liver is the most important site of drug metabolism in the body. Liver microsomes are subcellular fractions that contain membrane-bound drug-metabolizing enzymes. Microsomes can be used to determine the in vitro intrinsic clearance of a compound. The curve for ERDA metabolic stability (the ERDA percent remaining is plotted versus time) was established following sample analysis. The time points of the linear range of this curve (0–10 min) was used to plot another curve of time versus the natural logarithm (Ln) of ERDA remaining. The elimination rate constant or slope of the linear part was used for the calculation of ERDA in vitro half-life (t_1/2). The equation of the linear part of the curve was y = −0.0951x + 4.6188 with R² = 0.9961, that was used for the calculation of in vitro t_1/2 and in vitro t_1/2, was found to be 7.28 min and the inference clearance (CL int) was 95.11 µL/min/mg (Figure 4).

3.4. Greenness of the Method

The proposed method was evaluated in regard to its greenness. It provides an acceptable waste and corrosive profile. Applying the profile criteria of the emergency planning and community right-to-know act (EPCRA) [25] to evaluate the greenness of the proposed method showed that the method is excellent green. The pH of the solvents used is not less than two, so the corrosive profile is acceptable. In addition, the waste is less than 50 g and none of the solvent used is toxic; bioaccumulate of persistence indicates that this method agrees with the greenness profile criteria according to the emergency planning and community right-to-know act (EPCRA) [25]. Figure 5 represents the greenness of the method obtained by the Eco-Scale tool, which indicates the method is an excellent green analysis. The results obtained from applying the third method (AGREE software) also showed that the method is green, as the AGREE pictogram score was found to be 0.78. Therefore, the current method is shown in (Figure 5).

3.5. Comparison with the Previously Published Methods

The LLOQ values of the proposed method is less than the other previously methods [2,12,18], suggesting that this method is more sensitive than the other methods. Moreover, the proposed method is simpler, economic and fast

Table 5. Comparison of the proposed method with previous reported methods for the determination of ERDA in plasma samples.

Analytical Method	Linearity Range	MRM Transition	LLOQ (ng/mL)	Sample Volume µL	Sample Preparation	Separation Retention Time (min)	References
HPLC-UV	50–2000	____	50	100.0	Sep-pack column	7.7	[18]
UPLC-MS/MS	1.0–500	447.0 → 361.9	1.0	100.0	LLE (Ethyl acetate +Na OH)	0.92	[12]
LC-MS/MS	3.0–800	447.10 → 362.0	3.0	150.0	SPE	2.6	[20]
UPLC-MS/MS	0.5–1000	447.0 → 86.0	0.5	100	PPT +LLE	1.2	Current method

LLE: Liquid–liquid extraction; SPE: Solid phase extraction; PPT: Protein precipitation.

.

4. Conclusions

A specific and precise UPLC–MS/MS method was developed and validated for the determination of ERDA in plasma samples. The calibration curve was linear between the concentration range of 0.5–1000 ng/mL. The total run times was 2.0 min. Therefore, it is suitable for high throughput analysis. The stability of the analyte was not impacted by the expected sample handling, storage, preparation and analysis conditions. The greenness of the methods was evaluated using NEMI protocol, AGREE Software and the analytical Eco-Scale tool, and the results indicated that the excellent greenness of the method. ERDA metabolic stability in the human liver microsomes was successfully estimated.