Erdafitinib and bosutinib (purity of >98%) were purchased from Beijing Sunflower and Technology Development Co., Ltd. (Beijing, China). Posaconazole (purity over 98%) was purchased from Jiangsu Aikang Biomedicine Research and Development Co., Ltd.; isavuconazole (purity over 98%) was purchased from Sigma (St Louis, MO, United States). Methanol and acetonitrile in this study were HPLC grade and were purchased from Tianjin Kermel Chemical Reagent Co., Ltd. (Tianjin, China). Analytically pure formic acid was purchased from Tianjin Kermel Chemical Reagent Co., Ltd. (Tianjin, China).

The ultra-high liquid chromatography instrument was Waters ACQUITY UPLC I-Class, including quaternary solvent manager, sample manager-flow through needle, and high-temperature column heater with active pre-heating (Waters, United States). The mass spectrometer was Waters XEVO TQD triple quadrupoles mass spectrometer, and the electro-spray ionization (ESI) source was used (Waters, United States). Other instruments included electronic analytical balance and vortex mixer, and ultra-pure water equipment.

Six beagle dogs (male, weight 7.5 ∼ 9.5 kg) were purchased from Hubei Yizhicheng Biotechnology Co., Ltd. (Shiyan, Hubei). The animal license number was SCXK (Hubei) 2016–0020. The six beagle dogs were raised in the Laboratory Animal Center of Henan University of Science and Technology (Luoyang, China). All the experimental behaviors and operations were approved by the Institutional Ethics Committee of Henan University of Science and Technology (Luoyang, China). The experimental operation was carried out according to the Laboratory Animals Guidelines for Ethical Review of Welfare (GB/T 35,892–2018).

This experiment adopted a three-period self-control experimental design. All the beagle dogs fasted for 12 h before the experiment but were free to drink water. In the first period (group A), erdafitinib was dissolved in 0.5% carboxymethyl cellulose sodium solution and was orally administered to beagle dogs at a dose of 4 mg/kg. The blank blood was collected before erdafitinib was given, and then, at the different time points of 0, 0.5, 1, 1.5, 2, 3, 4, 6, 9, 12, 24, and 36 h, blood samples (about 1.0 ml) were collected from the veins of anterior and posterior limbs and taken into 1.5-ml heparinized EP tubes. The plasma were separated by centrifugation for 10 min at 10,000 rpm and frozen at −80°C for later analysis.

After a week of drug washout period, the second period (group B) of experiment was carried out. On the day of the experiment, the same six beagle dogs were orally given posaconazole at a dose of 7 mg/kg, and after 30 min, erdafitinib (4 mg/kg) was orally administered to beagle dogs. The time point of blood collection was the same as that of the first period.

Then, after a week of the second drug washout period, the third period (group C) of experiment was carried out. On the day of the experiment, the same six beagle dogs were orally given isavuconazole at a dose of 7 mg/kg, and after 30 min, erdafitinib was orally administered at a dose of 4 mg/kg. The time point of blood collection was the same as that of the first period.

Erdafitinib (10 mg) was accurately weighed and dissolved with methanol in a volumetric flask, and then, the volume was fixed with methanol to obtain a standard stock solution (1 mg/ml). IS stock solution (1 mg/ml) was prepared by the same method. The stock solution of erdafitinib (1 mg/ml) was diluted 10 times with methanol to obtain the standard solution of 100, 10, and 1 μg/ml for calibration curve and quality control (QC) samples. The stock solutions of erdafitinib and IS were stored in the refrigerator at −20°C.

The calibration curves with eight different concentration levels of erdafitinib were prepared by adding different concentration levels and volumes of standard application solution to blank beagle dog plasma, and the different concentration levels of calibration curve in beagle dog plasma were as follows: 1, 5, 10, 25, 50, 100, 250, and 500 ng/ml. QC samples of low, medium, and high concentration levels (2.5, 50, and 375 ng/ml) were prepared in the same way. The IS working solution at the concentration of 200 ng/ml was obtained by dilution of its stock solution with methanol.

The beagle dog plasma (100 µl) to be tested was taken into 1.5-ml EP tube, and 20 µl of IS working solution was added and mixed well. In addition, 200 µl of 1 mol/L sodium hydroxide solution was added and mixed well. Then, 1 ml of ethyl acetate was added and vortexed for 1 min. The upper organic phase was transferred in another 1.5-ml EP tube and blow-dried with nitrogen flow. The residue was dissolved with 100 µl of mobile phase, take 50 µl into the sample of automatic injector, and 5 µl of supernatant was injected into UPLC-MS/MS for detection.

The chromatographic column was waters Acquity UPLC BEH C18 column (2.1 × 50 mm, 1.7 μm), the mobile phase was acetonitrile (A) and 0.1% formic acid in water (B), the gradient elution procedure was used, and the flow rate was 0.40 ml/min. The gradient elution procedure was as follows: 0–0.5 min A 10%, 0.5–1.0 min A 10%→90%, 1.0–2.0 min A 90%, and the termination time was 2 min. The volume of each injection was 5.0 µl, the column was set at 40°C, and the temperature of sample tray was set at 4°C.

In the positive ion mode, erdafitinib and IS were monitored by multiple reaction monitoring. The parent ion and daughter ion used for quantification were as follows: m/z 447.0 → 361.9 for erdafitinib and m/z 529.8 → 141.0 for IS, respectively. The cone voltage and collision energy of erdafitinib were 30 and 20 V, respectively, and the collision energy and cone voltage of IS were 25 and 20 V, respectively. The desolvation temperature was 1,000°C, and the capillary voltage was 2.0 kV. Masslynx 4.1 (Waters, United States) was used for data acquisition and instrument control.

Using batch processing method, the concentration levels of erdafitinib in groups A, B, and C were detected by the developed UPLC-MS/MS technique in this study. Then, DAS (Drug and Statistics, version 2.0) was used to calculate the important pharmacokinetic parameters of erdafitinib through the statistical moment method. The main pharmacokinetic parameters of erdafitinib were as follows: T_max, C_max, t_1/2, CL, Vd, and AUC. Then, the mean concentration–time curve of erdafitinib was drawn.

The data were processed using SPSS 18.0 statistical software. Taking group A as the control group, the differences of pharmacokinetic parameters between group B and group A and between group C and group A were compared, respectively. The p-value was calculated by independent sample t-test, p < 0.05, which was statistically significant.

Six blank beagle dog plasma samples from different sources, standard solution of erdafitinib in blank beagle dog plasma, and samples after administration of erdafitinib were taken. According to the plasma sample processing method, the specificity of the UPLC-MS/MS method was investigated by comparing the chromatograms.

Under the above chromatographic and mass spectrometric conditions, erdafitinib and IS were separated completely, and the endogenous substances did not interfere with the detection. The representative chromatograms were shown in Supplementary Figure S2; A was the chromatogram of a blank beagle dog plasma sample, B was the chromatogram of plasma standard solution—a beagle dog plasma sample spiked with erdafitinib and IS, and C was the chromatogram of a beagle dog sample. The retention times of erdafitinib and IS were 0.92 and 0.81 min, respectively. The total running time for each sample was 2.0 min.

Chromatographic separation was based on the difference of adsorption and solubility of analytes by stationary phase. The components with strong adsorption or solubility had large partition coefficient and long retention time; on the contrary, the components with weak adsorption or solubility had small partition coefficient and short retention time. It could be seen from the chromatogram that the retention time of erdafitinib was slightly longer than that of IS. Therefore, under this chromatographic condition, the partition coefficient of erdafitinib was slightly larger than that of IS.

After treatment with plasma sample treatment method, the plasma standard solutions of erdafitinib with the concentration levels of 1, 5, 10, 25, 50, 100, 250, and 500 ng/ml were prepared and detected. The peak areas of erdafitinib and IS were recorded, respectively. The concentration of erdafitinib was taken as the abscissa x, the ratio of the peak area of erdafitinib and IS was taken as the ordinate y, and the standard curve was drawn by the least square method with a weighted (1/x ²) to ensure the accuracy of low concentration point calculation. The lowest concentration of standard curve was the LLOQ.

At the concentration range of 1–500 ng/ml for erdafitinib, the typical regression equations of erdafitinib were y = 0.9009 *x + 0.3741 (r ² = 0.999 3), which exhibited an excellent linearity. The LLOQ was 1.0 ng/ml, and the precision was below 8.67%, whereas the accuracy ranged from −2.33% to 4.29% (Table 1).

At low, medium, and high concentration levels (2.5, 50, and 375 ng/ml), the precision and accuracy were estimated by sextuple detection of QC samples over three consecutive days. The precision was expressed by relative standard deviation (RSD, %), and the accuracy was expressed by relative error (RE, %).

Table 1 showed the results of intra-day and inter-day precision and accuracy of erdafitinib. The precision (% RSD) did not exceed 8.93%. Accuracy (% RE) was in the range from −3.08% to 4.29% at low, medium, and high concentration levels (2.5, 50, and 375 ng/ml).

At low, medium, and high concentration levels (2.5, 50, and 375 ng/ml), by comparing the ratio of the peak area of erdafitinib before and after the extraction, the extraction recovery was calculated. ME was also calculated in six replicates by comparing the response of erdafitinib in plasma matrix after extraction with that in neat solution, and six batches of blank matrix were obtained from blank plasma of different beagle dogs.

At low, medium, and high concentration levels (2.5, 50, and 375 ng/ml), the extraction recovery of erdafitinib was within the range of 79.22% ∼ 82.25% (Table 2), and the ME values for erdafitinib were 98.71% ∼ 102.53% (Table 2). ME did not affect the detection of samples.

At low, medium, and high concentration levels (2.5, 50, and 375 ng/ml), the stability of plasma samples was investigated under four different storage conditions: processed samples at 4°C in auto-sampler tray for 6 h, room temperature for 3 h, −80°C for 4 weeks, and three freeze-thaw cycles (−80°C to 25°C).

All results of the stability were summarized in Table 3, and it was found that erdafitinib was stable under the conditions described above.

At similar operational and environmental conditions, the plasma standard solutions of erdafitinib with the concentration level of 250 ng/ml were prepared and analyzed by two different analysts and repeated six times, respectively. The peak area of erdafitinib was recorded and RSD was calculated. The method was found to be rugged (RSD, 2.84%).

The stock solution stability of erdafitinib and IS was assessed individually after 6 weeks of storage at −20°C, respectively. This was done by comparing the signal response for six replicates of erdafitinib or IS (10 μg/ml) made using the freshly prepared stock solution with the signal response for six replicates of erdafitinib or IS (10 μg/ml) made using the stock solution prepared 6 weeks earlier.

Stock solution stability was tested after 6 weeks for erdafitinib and IS, and it was found that the stock solution of erdafitinib and IS was stable for up to 6 weeks.

Figure 2A displayed the average plasma drug concentration–time curves of erdafitinib in groups A and B, and the average plasma concentration–time curves of erdafitinib in groups A and C were displayed in Figure 2B. The main parameters of pharmacokinetic of erdafitinib in groups A, B, and C were summarized in Table 4.

The results showed that, after erdafitinib was combined with posaconazole, the C_max and AUC_0→t of erdafitinib increased by 27.19% and 47.62%, respectively. After erdafitinib was combined with isavuconazole, the C_max and AUC_0→t of erdafitinib increased by 23.13 %and 54.46%, respectively.

UPLC-MS/MS has the advantages of high sensitivity, strong specificity, short analysis time, and good reproducibility. Therefore, it is often used in the detection of biological samples and the study of pharmacokinetics and DDI (Wang et al., 2020b; Yang et al., 2021).

In this study, we explored ESI positive and negative to find the most sensitive ionization modes of erdafitinib and IS. It demonstrated that positive ion mode exhibited higher mass response than negative ion mode. As indicated in Figure 1, erdafitinib and IS showed protonated molecular ion [M + H]⁺ at m/z 447.0 and 529.8, respectively, and the most abundant product ions were m/z 361.9 and 141.0, respectively. As a result, the parent ion–to–daughter ion ratio for erdafitinib was m/z 447.0 → 361.9 and for IS was m/z 529.8 → 141.0.

The mobile phase should meet the requirements of HPLC and LC/MS. Volatile salts should be added to the mobile phase as far as possible, and surfactants and other non-volatile buffers should not be used as far as possible, such as phosphoric acid buffer. Phosphate and other non-volatile buffer salts will precipitate in the ion source and plug the capillary. In this experiment, formic acid, water, and ammonium formate were first selected to investigate the influence of mobile phase on peak pattern and peak intensity. Finally, 0.1% formic acid and acetonitrile were used as mobile phase, which could increase the ionization degree of the sample, increase the signal intensity and improve the peak pattern. The mobile phase gradient elution method could complete a detection in 2 min. The detection was faster than that reported (Elawady et al., 2020).

In the field of pharmaceutical analysis, sample pretreatment and purification was an important step. Improving the pretreatment method could not only protect the detection instrument from pollution and prolong its service life but also reduce the detection matrix and improve the sensitivity and selectivity. It was reported that the plasma was pretreated by solid phase extraction (SPE) in the process of determination of erdafitinib in the human plasma with LC-MS/MS (Elawady et al., 2020). Because SPE needed special extraction column, the treatment process was relatively cumbersome. In this study, through the exploration of a series of pretreatment methods, the ethyl acetate liquid-liquid extraction method was selected to pretreat the plasma samples, and the sodium hydroxide solution was added to adjust the pH value of the samples, which could significantly improve the extraction recovery rate of erdafitinib. The influence of matrix interference on the determination was reduced.

DDI refers to the phenomenon that one drug changes the pharmacological effect of the other drug when taking more than two drugs at the same time or successively. Pharmacokinetic DDIs are those that affect the absorption, distribution, metabolism, and excretion of a concomitantly administered drug (Rabba et al., 2020). This change can lead to the change of the concentration of drug at the action site, thus affecting the duration and magnitude of the effect (Vázquez et al., 2020).

DDIs are one of the important factors leading to adverse reactions and are often associated with toxicity or therapeutic failure (Wang et al., 2020a), but sometimes, DDIs are beneficial to patients thought improving the bioavailability of drugs and producing synergistic or additive effects (Gerber et al., 2018). In any case, clinicians must be familiar with DDIs to improve prescribing tools.

Erdafitinib was a potent oral pan-FGFR inhibitor being developed as oncology drug for patients with alterations in the FGFR pathway. The geometric mean C_max of erdafitinib was 136 ng/ml after a single dose of erdafitinib in healthy adults (Poggesi et al., 2020). Patients with advanced or refractory solid tumors after treatment with different doses of erdafitinib, the exposure of erdafitinib increased in a dose-dependent manner, the median T_max ranged from 2–3 h after the initial dose to 2–6 h after multiple daily dosing (Nishina et al., 2018). After mice were given erdafitinib (30 mg/kg), C_max was about 810 ng/ml and t_1/2 was about 1.73 h, and after mice were given erdafitinib (10 mg/kg), C_max was about 110 ng/ml and t_1/2 was about 2.36 h (Elawady et al., 2021). The PK results of this study showed that, after beagle dogs were given erdafitinib (4 mg/kg), C_max was about 282 ng/ml, which was close to human C_max when converted according to dose (Poggesi et al., 2020).

Our previous research results show that posaconazole significantly increased the concentration of selinexor (a selective nuclear export inhibition) in rats, fluconazole, itraconazole, and isavuconazole that have significant effects on pharmacokinetics of selinexor and increased plasma exposure of selinexor in rats (Li et al., 2020; Zhou et al., 2021).

The results of the study showed that, after erdafitinib was combined with posaconazole, the C_max and AUC_0→t of erdafitinib increased by 27.19 and 47.62%, respectively, and the t_1/2 was prolonged to 6.33 h; after erdafitinib was combined with isavuconazole, the C_max and AUC_0→t of erdafitinib increased by 23.13 and 54.46%, respectively, and the t_1/2 was prolonged to 6.31 h. It is suggested that posaconazole and isavuconazole could affect the pharmacokinetics of erdafitinib in beagle dogs and increase the plasma exposure of erdafitinib. Because posaconazole was a strong inhibitor of CYP3A4, isaconazole was a moderate inhibitor of CYP3A4, so posaconazole and isaconazole could slow down the metabolism of erdafitinib by inhibiting CYP3A4. Therefore, when erdafitinib was combined with posaconazole and isavuconazole, possible pharmacokinetics DDIs should be considered. However, after all, human drug metabolism enzymes were different from beagle dogs, so the experimental results could only be used as a theoretical reference. DDIs in humans would need further validation. Therefore, the results of this study still had some limitations.