Development and validation of an UPLC-MS/MS method for the therapeutic drug monitoring of oral anti-hormonal drugs in oncology

A liquid chromatography-mass spectrometry assay was developed and validated for simultaneous quantification of anti-hormonal compounds abiraterone, anastrozole, bicalutamide, Δ(4)-abiraterone (D4A), N -desmethyl enzalutamide, enzalutamide, Z-endoxifen, exemestane and letrozole for the purpose of therapeutic drug mon- itoring (TDM). Plasma samples were prepared with protein precipitation. Analyses were performed with a triple quadrupole mass spectrometer operating in the positive and negative ion-mode. The validated assay ranges from 2 to 200ng/mL for abiraterone, 0.2–20ng/mL for D4A, 10–200ng/mL for anastrozole and letrozole, 1–20ng/ mL for Z-endoxifen, 1.88–37.5ng/mL for exemestane and 1500–30,000ng/mL for enzalutamide, N -desmethyl enzalutamide and bicalutamide. Due to low sensitivity for exemestane, the final extract of exemestane patient samples should be concentrated prior to injection and a larger sample volume should be prepared for exemestane patient samples and QC samples to obtain adequate sensitivity. Furthermore, we observed a batch-dependent stability for abiraterone in plasma at room temperature and therefore samples should be shipped on ice. This newly validated method has been successfully applied for routine TDM of anti-hormonal drugs in cancer patients.


Introduction
Breast cancer and prostate cancer are the most common malignancies in women and men in the Western world [1]. As these cancer types are highly dependent on growth-stimulating hormones, antihormonal therapy is a first-line treatment strategy. Anti-hormonal drugs are generally administered orally or subcutaneously. Oral drugs for treatment of breast cancer include tamoxifen, anastrozole, letrozole and exemestane and oral drugs for prostate cancer therapy include bicalutamide, abiraterone acetate and enzalutamide. The group of oral anticancer drugs is rapidly expanding [2][3][4][5], however, most oral antihormonal agents have been on the market for a longer period of time.
Although many patients benefit from anti-hormonal therapy in terms of progression-free survival, treatment outcome is variable. This may be attributed to variability in drug levels and exposure. For some anti-hormonal drugs, such as tamoxifen and abiraterone acetate, a clear exposure-response relationship has been described [6][7][8]. This relationship is the basis for therapeutic drug monitoring (TDM); individualized drug dosing by monitoring drug concentrations in patient blood, plasma or serum. In current practice, oral anti-hormonal drugs are administered at fixed doses, which could lead to suboptimal exposure or high blood concentrations and adverse-events. Recommendations for pharmacokinetic TDM are based on clinical studies and guidelines and proposed targets for anti-hormonal drugs for the treatment of breast cancer and prostate cancer can be found in literature [8]. Ultimately, implementation of individualized dosing with TDM may be an important tool to improve treatment outcome and efficacy in breast cancer and prostate cancer patients.
To facilitate TDM, there is a need for bioanalytical assays to quantify drugs of interest. Liquid chromatography-mass spectrometry (LC-MS/MS) is a useful and often applied analytical technique for determining drug concentrations. When developing an analytical method for TDM, it is important to choose a clinically relevant calibration range. This quantitation range should be built around the proposed target concentration, covering the majority of samples as seen in the clinic. Our lab has experience developing and validating methods for TDM of anticancer agents [9][10][11]. Previously published LC-MS/MS assays for quantification of abiraterone [11][12][13][14][15][16], anastrozole [17], bicalutamide [13,[18][19][20], Z-endoxifen [21][22][23][24], enzalutamide [9,11,13,[25][26][27][28], exemestane [29,30] and letrozole [31,32] are limited to measuring one to four analytes. Furthermore, there are no articles reporting steady-state concentrations of anastrozole, letrozole and exemestane in humans for TDM purpose, and no articles describing Zendoxifen analysis in plasma. To our knowledge this is the first bioanalytical assay for simultaneous quantification of six anti-hormonal drugs in oncology, including the active metabolites Z-endoxifen, Ndesmethyl enzalutamide and Δ(4)-abiraterone (D4A), which enables concurrent quantification of these analytes to efficiently determine plasma concentrations for TDM purpose. Although methods have been developed for the combined analysis of anti-hormonal drugs for either prostate cancer or breast cancer, the development of an assay for both types of anti-hormonal drugs is complicated due to different chemical drug properties. Furthermore, the development of such an assay is challenging because target concentration ranges span from 0.2-30,000 ng/mL, compounds show a variety in MS response and a highly selective chromatographic method is needed to separate isomers of abiraterone and Z-endoxifen.

Calibration standards, quality control samples
Calibration standards and QC samples were prepared by spiking 100 μL working solution to 900 μL K 2 EDTA plasma. Independent working solutions were used for the preparation of calibration standards and QC samples. Combined QC samples were prepared for abiraterone, anastrozole, bicalutamide, D4A, N-desmethyl enzalutamide, enzalutamide and letrozole, while separate QC samples were prepared for exemestane. Final concentrations of calibration standards and quality control samples are depicted in Table 1. Calibration standards and QC samples were stored at −20°C.
autosampler vials with insert.

Sample preparation of exemestane patient samples and QC samples
Directly after sample collection in the clinic, whole blood samples were centrifuged for 10 min at 2000 ×g at 4°C and plasma was stored at −20°C. For exemestane patient-and QC samples, 500 μL of plasma was aliquoted and a volume of 20 μL of IS working solution was added to each sample. Proteins were precipitated using 1000 μL of acetonitrile. Samples were vortex-mixed for 10 s and centrifuged for 5 min at 23,000 ×g. The supernatant was transferred to 2 mL containers and the samples were dried under a gentle stream of nitrogen at 40°C. The residue was reconstituted in 50 μL water-methanol (1:1 v/v), vortexmixed for 10 s and centrifuged for 5 min at 23,000 ×g. The supernatant was transferred to amber-colored autosampler vials with insert. To correct for the difference in sample preparation of exemestane calibration samples and QC samples, a dilution factor of 0.03 and an internal standard concentration of 3.4 was used for quantification of patient samples and quality control samples.

Analytical equipment and conditions
Analytes were separated chromatographically using a Shimadzu LC system with a binary pump, a degasser, an autosampler and a valco valve (Nexera 2 series, Shimadzu corporation, Kyoto, Japan). The temperature of the autosampler was kept at 4°C and the column oven at 50°C. Mobile phase A consisted of 0.1% formic acid in water and mobile phase B consisted of acetonitrile-methanol (50:50, v/v). Gradient elution was applied at a flow rate of 0.6 mL/min through an Acquity BEH C 18 column (100 Å, 2.1 × 15 mm, 1.8 μm) with an additional Acquity BEH C 18 Vanguard pre-column (100 Å, 2.1 × 5 mm, 1.8 μm) (Waters, Milford, MA, USA). The following gradient was applied: 45% B (0.0-4.0 min), 100% B (4.0-5.0 min), 45% B (5.0-6.0 min). The divert valve directed the flow to the mass spectrometer between 0.5 and 5 min and the remainder to the waste container.
A triple quadrupole mass spectrometer 6500 (Sciex, Framingham, MA, USA) with a turbo ion spray (TIS) interface operating in the positive and negative mode was used as a detector. Bicalutamide was determined in negative ion mode to obtain adequate assay sensitivity, while all the other compounds were measured in positive ion mode. For quantification, multiple reaction monitoring (MRM) chromatograms were acquired and processed using Analyst® 1.6.2 software (AB Sciex). General and analyte specific mass spectrometric parameters are listed in Table 2 and the structures and the proposed fragmentation patterns of the analytes are depicted in Fig. 1.

Validation procedures
The assay was validated for calibration model, accuracy and precision, LLOQ, sensitivity and selectivity, dilution integrity, carry-over and stability. Adjustments were made to typical validation practices to fit TDM purposes; four instead of six to eight calibrators were investigated, QC concentrations were prepared at three levels (LLOQ, medium, and high concentrations) and no matrix effects were evaluated. A reduced number of calibration standards increases the turnaround of the assay. We choose not to evaluate matrix effects as poor reproducibility due to the use of different matrices will also be reflected in the sensitivity experiments and because we use isotopically labeled internal standards to correct for matrix related effects. Accuracy and precision were calculated as described previously [9].

Clinical application
This assay was developed to support pharmacokinetic monitoring of abiraterone, anastrozole bicalutamide, D4A, N-desmethyl enzalutamide, Z-endoxifen, enzalutamide, exemestane and letrozole. As part of routine clinical care, K 2 EDTA blood samples (4 mL) were collected from patients who were treated with one of these drugs at the Antoni van Leeuwenhoek -The Netherlands Cancer Institute. Plasma samples were collected and processed as described in this report.

Sample preparation
Previous validation procedures showed that abiraterone is not stable in acetonitrile [34], therefore, working solutions were prepared in K 2 EDTA plasma. Protein precipitation was chosen as highthroughput method for sample preparation with a sample:acetonitrile ratio of 50:100 (v/v). With this composition of the final extract, no further dilution was necessary prior to injection. However, with this simple sample preparation we were unable to quantify exemestane in patient samples due to low sensitivity. Therefore, we developed a method for the quantification of exemestane in patient samples and QC samples, using a 10-fold larger volume of plasma (500 μL instead of 50 μL). The final extracts of these samples were evaporated to dryness and reconstituted in 50 μL of reconstitution solvent. To preserve a fast turn-around, we prepared combined calibration standards containing all analytes including exemestane at a higher concentration range (62.5 to 1250 ng/mL) these calibration standards were prepared with simple protein precipitation, without the need for concentrating the final extract. During development and validation it was shown that we could easily correct for the difference in sample preparation of exemestane calibration samples and QC samples by applying a dilution factor in the processing software.

Mass spectrometry and chromatography
The analytical setup was developed for simultaneous quantification of anti-hormonal drugs to monitor drug exposure. Chromatographic separation was pivotal and challenging for Z-endoxifen and abiraterone, as both analytes show extensive metabolism, including the formation of isomers. Therefore, baseline separation of these isomers was required. This was achieved by using an ultra-pressure liquidchromatography (UPLC) column. Orbitrap MS (Thermo Fischer) spectra were obtained of the abiraterone metabolites to determine the accurate mass. These spectra confirm that both metabolites and abiraterone have the same accurate mass (349.24 g/mol) and are therefore considered isomers. Representative chromatograms of QC LLOQ and blank samples are presented in Fig. 2 for each analyte. Furthermore, Fig. 3 depicts the MRM chromatograms of Z-endoxifen and abiraterone of a patient sample, showing that the chromatographic system is capable of separating the isomers of these drugs. Calibration ranges were chosen so that analyte concentrations in patient samples were within this range. Reported C trough concentrations of enzalutamide and N-desmethyl enzalutamide are 11.4 mg/L and 13.0 mg/L, respectively [35]. A calibration range around these high plasma concentrations, however, caused saturation of the MS detector resulting in non-linearity of the calibration model. To overcome this, the MRM channel was adjusted (+2) to monitor m/z values of naturally occurring isotopes of both parent and product ions [36]. With this modification, enzalutamide and N-desmethyl enzalutamide could both be measured in a clinically relevant concentration range without the need for sample dilution.

Calibration model
Four non-zero calibration standards were prepared and analyzed in three separate runs. Linearity of the calibration model was determined by plotting the peak area ratio of the analyte/IS against the corresponding concentration (x) of the calibration standard. The reciprocal of the squared concentrations (1/x 2 ) was used as a weighting factor for all analytes. For each calibration curve the calibration concentrations were back-calculated from the response ratios. The deviations of the nominal concentrations should be within ± 15%. At the LLOQ level a deviation of ± 20% was permitted. All calibration curves (n = 3) of all analytes met these criteria. The assay was linear for the validated concentration ranges of 2-200 ng/mL voor abiraterone, 0.2-20 ng/mL for D4A, 10-200 ng/mL for anastrozole and letrozole, 1-20 ng/mL for Z-endoxifen, 62.5-1250 ng/mL for exemestane and 1500-30,000 ng/ mL for enzalutamide, N-desmethyl enzalutamide and bicalutamide.

Accuracy and precision
Intra-and inter-assay bias and precisions of the method were determined by analyzing five replicate QC samples in three consecutive runs at LLOQ, mid and upper limit of quantification (ULOQ)   concentration levels. The intra-and inter-assay biases and precisions should be within ± 20% and ≤20%, respectively, for the LLOQ concentration and within ± 15% and ≤15%, respectively, for other concentrations. Table 3 summarizes the intra-and inter-assay biases and precisions of the assay. All values were within the acceptance criteria.

Carry-over
Carry-over was investigated by injecting two double blank samples subsequently after an ULOQ sample in three independent runs. The peak area in the blank processed samples should be ≤20% of the peak area in the LLOQ sample and ≤5% of the internal standard area. There were no peaks observed in the first blank processed sample for any analyte, which means that there was no carry-over.

Specificity and selectivity
Six individual batches of K 2 EDTA plasma were used to assess the specificity and selectivity of the method. A double blank sample and a sample spiked at the LLOQ were processed of each batch. The samples were prepared to determine whether endogenous compounds interfere at the mass transitions chosen for the analytes and internal standards. Samples were processed and analyzed according to the described procedures. Interferences co-eluting with the analytes or internal standards in the blanks were all ≤20% of the peak area of the analytes at LLOQ or ≤5% of the internal standard areas. Deviations of the nominal concentrations were within ± 20% for at least 4 out of 6 batches for all analytes. Selectivity was therefore considered acceptable.

Stability
Stability of the analytes was tested under various conditions. All stability experiments were performed in triplicate. The analytes were considered stable in the plasma or processed sample when 85%-115% of the initial concentration was recovered. Furthermore, analytes were considered stable in the stock solution when 95%-105% of the original concentration was recovered.
All analytes were stable at −20°C in plasma for at least 21 weeks. Short-term stability in plasma was determined after five days at room temperature (20-25°C) and at 4°C in dark and exposed to light. Analytes were stable under these short-term storage conditions, except for abiraterone, which was unstable at room temperature in both light and dark. Additional stability experiments showed that abiraterone was stable for only 4 h in plasma at room temperature. However, when the experiment was repeated in a two-year old batch of plasma, abiraterone was proven stable at room temperature up to 48 h. Fig. 4 shows the stability of abiraterone, given as the recovery (%) of the original concentration up to 48 h in two different batches of plasma. The underlying mechanism for this batch-dependent stability remains to be elucidated but could possibly be caused by enzymes, which are active in fresh plasma and less active in older plasma.
The effect of three freeze (−20°C)/thaw cycles on the stability of each compound was investigated after thawing samples to room temperature with a minimum interval of 12 h on three separate occasions and comparison with freshly prepared calibration samples. All analytes were stable for three freeze/thaw cycles. Five-day stability was proven for all analytes in final extract at 4°C. Furthermore, exemestane was stable in dried extract at 4°C for at least five days. Stability in stock solution was demonstrated at 124 days at −20°C.

Clinical application
This analytical assay was used to determine plasma concentrations of abiraterone, anastrozole bicalutamide, D4A, N-desmethyl enzalutamide, Zendoxifen, enzalutamide, exemestane and letrozole in samples from patients using these drugs. The chromatograms of abiraterone and Z-endoxifen show additional peaks with identical transitions, belonging to isomeric metabolites. The presence of these isomeric metabolites has been previously described [12,14,34,37]. Applicability of the assay was shown in samples from patients treated with these drugs and the results are listed in Table 4. Ten patients were included for each drug and one sample was drawn from each patient. All values were within the validated range, except for two exemestane samples being below the LLOQ and one letrozole sample being above the ULOQ. The quantitation ranges of previously published methods (exemestane 0.2-0.4 ng/mL, [29,30]; letrozole 6-430 ng/mL [32]) might be sufficient to determine exemestane and letrozole concentrations within the validated range. However, our method was developed for the purpose of therapeutic drug monitoring and therefore the quantitation range was chosen to measure the majority of samples from the clinic. These results demonstrate the applicability of this method for quantification of the selected oral anti-hormonal drugs and three active metabolites for therapeutic drug monitoring.

Conclusion
The development and validation of a combined assay for the quantification of abiraterone, anastrozole bicalutamide, D4A, N-desmethyl enzalutamide, Z-endoxifen, enzalutamide, exemestane and letrozole in plasma is described. The validated assay ranges from 2 to 200 ng/mL voor abiraterone, 0.2-20 ng/mL for D4A, 10-200 ng/mL for anastrozole and letrozole, 1-20 ng/mL for Z-endoxifen, 1.88-37.5 ng/ mL for exemestane and 1500-30,000 ng/mL for enzalutamide, N-desmethyl enzalutamide and bicalutamide. Exemestane patient samples and QC samples should be concentrated to increase the sensitivity of the assay, and enzalutamide and N-desmethyl enzalutamide should be monitored at +2 m/z values to prevent detector saturation and therefore the need for sample dilution. Furthermore, the chromatographic method of this assay is highly selective and capable of separating isomers of abiraterone and Z-endoxifen Due to instability of abiraterone in Table 4 Plasma concentrations of the analytes in patient samples of patients treated with these drugs (n = 10). Z-Endoxifen was measured in plasma from patients using tamoxifen and abiraterone and Δ(4)-abiraterone (D4A) were determined in plasma from patients using abiraterone acetate. Abbreviation: o.d. = once daily. Two exemestane samples were below the lower limit of quantification (LLOQ). c One letrozole samples was above the upper limit of quantification (ULOQ).
plasma at room temperature, abiraterone patient samples should be shipped on dry ice. In conclusion, the presented assay is considered suitable to support therapeutic drug monitoring of oral anti-hormonal drugs in clinical daily oncology practice.

Funding
Conflicts of interest and sources of funding: this research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.