Journal of Pharmaceutical and Biomedical Analysis 100 (2014) 254–261

Contents lists available at ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba

An UPLC–MS/MS method for separation and accurate quantification of tamoxifen and its metabolites isomers Cécile Arellano a,∗ , Ben Allal a,b , Anwar Goubaa c , Henri Roché d,e , Etienne Chatelut a,b a

EA4553, Université Paul Sabatier Toulouse III, Toulouse F-31000, France Institut Claudius Regaud, IUCT-O, Laboratoire de pharmacologie, 1 Avenue Joliot-Curie, Toulouse F-31059, France Universita Degli Studi Di Genova, Italy d Département d’Oncologie Médicale, IUCT-O, 1 Avenue Irène Joliot-Curie, 31059 Toulouse, France e Université Paul Sabatier, Toulouse III, Toulouse F-31000, France b c

a r t i c l e

i n f o

Article history: Received 19 May 2014 Received in revised form 24 July 2014 Accepted 26 July 2014 Available online 4 August 2014 Keywords: UPLC–MS/MS Tamoxifen Endoxifen 4-Hydroxytamoxifen Tamoxifen-N-oxide

a b s t r a c t A selective and accurate analytical method is needed to quantify tamoxifen and its phase I metabolites in a prospective clinical protocol, for evaluation of pharmacokinetic parameters of tamoxifen and its metabolites in adjuvant treatment of breast cancer. The selectivity of the analytical method is a fundamental criteria to allow the quantification of the main active metabolites (Z)-isomers from (Z) -isomers. An UPLC–MS/MS method was developed and validated for the quantification of (Z)-tamoxifen, (Z)-endoxifen, (E)-endoxifen, Z -endoxifen, (Z) -endoxifen, (Z)-4-hydroxytamoxifen, (Z)4 -hydroxytamoxifen, N-desmethyl tamoxifen, and tamoxifen-N-oxide. The validation range was set between 0.5 ng/mL and 125 ng/mL for 4-hydroxytamoxifen and endoxifen isomers, and between 12.5 ng/mL and 300 ng/mL for tamoxifen, tamoxifen N-desmethyl and tamoxifen-N-oxide. The application to patient plasma samples was performed. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Tamoxifen (Z-1-(p-dimethylaminoethoxyphenyl)-1,2diphenyl-1-butene; TAM) is a non-steroidal selective estrogen receptor modulator (SERM). TAM is a competitive antagonist of estrogen receptor activity, which inhibits the estrogen-dependent growth and proliferation pathway in epithelial breast cancer cells. TAM is the main drug used to treat women with estrogen receptor (ER)-positive tumors. As an adjuvant, TAM provides significant clinical benefits in pre- and postmenopausal patients with early-stage breast cancer, prolonging survival and significantly reducing the incidence of recurrences. The most commonly used administration schedule is 20 mg TAM daily for 5 years. However a significant number of patients (30–50%) experience disease recurrence or progression during TAM therapy and despite a good overall tolerability profile subsequently died of disease highlighting the individual differences in response to TAM. TAM is extensively metabolized by the human cytochrome P450 enzymes into several metabolites resulting from its N-demethylation by the cytochrome CYP3A4/5 enzyme and

∗ Corresponding author. Tel.: +33 5 31 15 55 74. E-mail address: [email protected] (C. Arellano). http://dx.doi.org/10.1016/j.jpba.2014.07.033 0731-7085/© 2014 Elsevier B.V. All rights reserved.

hydroxylation by the cytochrome CYP2D6 enzyme. The Z-endoxifen and Z-4-hydroxytamoxifen (4-HOTam) resulting metabolites have been shown to have antiestrogenic activities which are 30- to 100fold more potent than TAM [1–4]. Several publications report that poor functionality/activity of the human cytochrome CYP2D6 leads a decrease in the plasmatic concentration of TAM and its active metabolites (Z-endoxifen and Z-4-HOTam), and it was suggested that genetic polymorphisms of cytochrome CYP2D6 increase the risk of breast cancer recurrence in patients receiving TAM as an adjuvant therapy [5,6]. However, the relationship between CYP2D6 genotype and TAM treatment efficiency is still the subject of debate and will not be definitively established until results of a properly prospective clinical trial are obtained [7–9]. Methods for the quantification of TAM and its related metabolites during clinical studies and therapeutic monitoring have been largely developed in the past decade. Teunissen et al. [10] have reported an overview of the LC–MS, LC–MS/MS and UPLC–MS/MS methods developed for quantification of TAM and its metabolites in a biological matrix during metabolism studies or pharmacokinetic applications. Most of these studies focus on the quantification of the more therapeutically active metabolites: 4-HOTam and Z-endoxifen without verification of their method selectivity for the 4 -hydroxytamoxyfen (4 -HOTam) and Z -endoxifen isomers [11–13]. Therefore, this lack of selectivity could introduce a bias in

C. Arellano et al. / Journal of Pharmaceutical and Biomedical Analysis 100 (2014) 254–261

the plasmatic concentrations obtained for these metabolites. The use of LC or UPLC tandem mass spectrometry theoretically offers a highly specific technique for quantification. Nevertheless, the resolution of TAM metabolites showing the same ionization and transition pattern has to be verified as done by Jaremko et al. [14]. Jager et al. [15] report that the lack of selectivity could induce large discrepancies in the reported concentrations average for endoxifen and 4-HOTam. And yet, even in the more recent LC–MS/MS reports [16], this selectivity criterion is not always completely satisfied. Murdter et al. [17] recently reported the concentration levels of a panel of phase I and II metabolites of TAM observed after 6 months of treatment determined with HPLC–MS/MS method but analytical results and validation were not reported in detail. Our aim was to combine plasmatic concentrations of TAM and its phase I metabolites with the pharmacogenetic characteristics of a large number of patients (i.e. 1000 inclusions). To this end, we developed and validated a rapid and selective UPLC–MS/MS method for the quantification of TAM, Z-endoxifen, 4-HOTam, Z -endoxifen, 4 -HOTam, N-demethyltamoxifen (N-DMTam), including also tamoxifen-Noxide (Tam-N-ox). The method was then applied to clinical samples. 2. Experimental 2.1. Chemicals Z -endoxifen, Z-4-HOTam, Z-4 -HOTam, Z-Tam-N-oxide, and Z-N-DMTam were obtained from Toronto Research chemicals (North York, Canada), d5 tamoxifen was purchased from Alsachim (Illkisch; France). Endoxifen (E:Z, 1:1 mixture), acetonitrile (HPLC grade), isopropanol and hexane (chromasolv for HPLC quality), dichloromethane and formic acid were purchased from Sigma–Aldrich (St-Quentin Fallavier, France) and methanol from Sharlau (Barcelona, Spain). Ultrapure water was prepared with a Milli-Q System (Millipore Corporation, Molsheim, France). Human plasma was obtained from “Etablissement Franc¸ais du Sang” (CHU Purpan, Toulouse, France). Minispike filters (EDGE, 13 mm, nylon 0.22 ␮m) were purchased from Waters (St-Quentin en Yvelines, France). 2.2. UPLC–MS/MS quantification 2.2.1. MS detection A Waters Acquity UPLC MS/MS composed of UPLC Sample Manager coupled to a Waters TQ Detector (Waters, St Quentin en Yvelines, France) was used for the quantitative analysis. The UPLC system consisted of an Acquity UPLC® separation module (Waters, Milford, Connecticut, USA) controlled by MassLynx 4.1 software and the QuantiLynx application was used for quantification. Detection was performed by the mass spectrometer Acquity detector with electrospray ionization (ESI) in positive ion mode. The mass spectrometer was used in the multiple-reaction monitoring (MRM) mode, MS collision parameters were summarized in Table 1. The temperature of the ESI source during the run was respectively set at 148 ◦ C (for the source) and 349 ◦ C (for the desolvation gas). The gas flow of the cone was set at 1 L/h and the gas flow of the desolvation was set at 649 L/h. 2.2.2. LC analysis The chromatographic separations were performed on an UPLC BEH C18 1.7 ␮m of 2.1 × 100 mm Column (Waters, Milford, MA, USA) thermostated at T = 50 ◦ C. LC eluent consisted in a gradient of phase A (2 mM ammonium formate acidified with formic acid (0.1%, v/v) and phase B (acetonitrile acidified with formic acid (0.1%, v/v)) at a flow rate of 0.3 mL/min. Phase B, initially set at 35% increased linearly to 65% over 3.5 min, then phase B was decreased to the

255

Table 1 MS collision parameters. Compounds

Parent (m/z)

Daughter (m/z)

Collision (V)

E-endoxifen Z-endoxifen Z -endoxifen 4-HOTam 4 -HOTam Tam-N-ox TAM N-DMTam d5 -TAM

374.22 374.22 374.22 388.22 388.22 388.22 372.22 358.22 377.22

57.99 57.99 57.99 71.97 71.97 71.97 71.97 57.99 72.04

32 32 32 38 38 38 50 30 32

Abbreviations: HOTam: hydroxytamoxifen, Tam-N-ox: tamoxifen-N-oxide, TAM: tamoxifen, N-DMTam: N-desmethyltamoxifen, d5 -TAM: deutered tamoxifen.

initial conditions over 0.5 min and the system was re-equilibrated for 2 min before the following injection. The autosampler was thermostated at 10 ◦ C, volumes of 5 ␮L were injected into the UPLC with a run time duration of 6 min. A needle wash solution containing a mix of methanol, isopropanol, water, acetonitrile (1/1/1/1, v/v/v/v) and 0.1% formic acid (v/v) was used to ensure needle cleanup in between each point of analysis; the seal wash solution was a combination of acetonitrile/water (1/3, v/v) solution. 2.3. Standard solution for calibration Standard solutions were prepared from stock solutions (stored at −20 ◦ C) of TAM or metabolites in methanol. The concentrations of stock solutions were 5 mg/mL for TAM, N-DMTam, Tam-N-ox, 4HOTam and endoxifen (Z/E); 2.5 ng/mL for 4 -HOTam and 1 ng/mL for Z -endoxifen and 5 mg/mL for d5 tamoxifen (d5 -TAM). These stock solutions were diluted to prepare pooled calibration solutions in plasma at 1250 ng/mL and 20 ng/mL for 4-, 4 -HOTam and E-, Z-, Z -endoxifen and 500 ng/mL for TAM, N-DMTam and Tam-N-ox. Calibration solutions were then diluted in blank plasma to prepare calibration points with concentrations ranging from 0.5 to 125 ng/mL for E-, Z-, Z -endoxifen and 4-, 4 -HOTam and from 12.5 to 300 ng/mL for N-DMTam, Tam-N-ox and TAM. A working solution of d5 -TAM at 500 ng/mL was prepared in methanol, and added at a final concentration of 25 ng/mL as an internal standard (IS) in each plasma sample before extraction for the quantification of all compounds. 2.4. Validation protocol Calibration samples and quality control (QC) points at low, medium, and high concentrations in the validation intervals were prepared in blank plasma from independent dilutions of TAM or metabolites in plasma and used for determination of linearity, precision and accuracy of the method. Calibration curves were constructed by correlating peak area ratio for each compound (versus d5 -TAM used as internal standard) as a function of the concentration of the spiked standard solutions. Calibration points were set to 0.5, 1, 5, 10, 20, 31.25, 62.5 and 125 ng/mL for E-, Z-, Z -endoxifen, 4HOTam, and 4 -HOTam and 12.5, 25, 50, 75, 150, 225 and 300 ng/mL for TAM, N-DMTam, and Tam-N-ox. Regression analysis was performed with weighting 1/X which gave the best fitting. Calibrations started from 0.5 ng/mL for endoxifen and HOTam isomers, which can be taken as the LLOQ, and 12.5 ng/mL for other compounds, which is much higher than the LLOQ (signal to noise ratio respectively higher than 5000 for TAM and NDM-Tam and higher than 1000 for Tam-N-ox), this range was chosen according to the levels of concentration expected in patients. The choice of the number of QC levels was determined by the range of the concentrations used for calibration for each analyte.

256

C. Arellano et al. / Journal of Pharmaceutical and Biomedical Analysis 100 (2014) 254–261

The concentrations of QC were respectively set at 1.5 ng/ml (3 times LLOQ), 2.5 ng/mL (low level), 25 ng/mL (medium level), and 100 ng/mL (high level) for E-, Z-, Z -endoxifen, 4-HOTam, and 4 HOTam; and at 20 ng/mL (low level), 125 ng/mL (medium level), and 250 ng/mL (high level) for TAM, N-DMTam, and Tam-N-ox. Intra-day variability was assessed by analysing six replicates on the same day and inter-day validation was based on two replicates per day on 6 different days for QCs and the lowest calibration points.

2.7. Stability Freeze–thaw stability was performed in the laboratory by preparation of a pool of QC samples at each level of concentration used, stored at −20 ◦ C. Two series of QC (low, medium and high concentrations) were analyzed the day they were made and each time after the three consecutive freeze–thaw cycles for a period of 8 days. 2.8. Patients

2.5. Liquid–liquid extraction procedure Two hundred microliters of plasmatic sample were added by 5 ␮L of 1 M sodium hydroxide, and vortex for 30 s. Samples were then added to 1000 ␮L of extraction solvent (dichloromethane/isopropanol/hexane; 1/1.3/2.5, v/v/v), placed on a rotary shaker for 5 min, and centrifuged for 2 min at 1000 × g to separate the aqueous phase from the organic one, the latter was then removed and dried at 37 ◦ C under air stream. The dry residue was then dissolved in a 200 ␮L of a mixture of acetonitrile/0.02 M ammonium formate buffer solution (v/v).

2.6. Matrix effect and recoveries The extraction recovery was calculated by the comparison of the peak area for extracted samples (prepared in plasma) with non-extracted standards prepared in acetonitrile/0.02 M ammonium formate buffer solution (v/v) at the same concentration representing 100% recovery. The recovery was calculated at QC concentrations and at the median concentration in the calibration range: 62.5 ng/mL for endoxifen and HOTam isomers (a calibration solution corresponding to 150 ng/mL of TAM, NDN-Tam and TamN-oxide) and 125 ng/mL for TAM, NDN-Tam and Tam-N-oxide (a calibration solution corresponding to 75 ng/ml for endoxifen and HOTam isomers). The median concentrations of the calibration curves and a lowest QC were also used to determine the matrix effect by comparing the area of each compound spiked in plasma extracted as described just before versus the area of compound in solution in acetonitrile/0.02 M ammonium formate buffer solution (v/v) without extraction; at least four independent plasmas were tested (test in duplicate in three plasmas and a single in the fourth one).

Clinical samples were collected as specified in the prospective multicenter protocol “Pharmacology of hormonotherapy in adjuvant treatment of breast cancer” (Institut Claudius-Regaud, Toulouse, France), INCA French registry: RECF1586. All the patients gave written informed consent. The study was approved by the Ethics Committee of Midi-Pyrénnées Region (Comité de Protection des Personnes Sud Ouest et Outre Mer III). All patients took 20 mg of TAM daily per os. Plasma samples were collected in heparincontaining tubes before the beginning of the treatment, and then every 6 months during 3 years (7 samples collected). The blood sample was taken just before the morning intake of TAM in order to determine the trough concentrations of TAM and its metabolites. Heparin plasma was separated and immediately stored at −20 ◦ C until analysis. Plasma TAM and its metabolites were determined in 23 patients corresponding to the first included patients for whom the collected plasmas were all available (from 6 to 36 months after beginning of their treatment). Patient blood samples were taken 30 min before patient took the drug and plasma was prepared within 30 min after blood collection. Plasmas were then aliquoted (500 ␮L in special frozen tube) and stored at −20 ◦ C. For the analysis, samples were defrosted, and 200 ␮L of plasma were collected and 10 ␮L of IS (d5 -TAM, 500 ng/mL) were added before extraction. 3. Results and discussion The UPLC–MS method was performed and validated in accordance with the requirements defined by the regulatory guidelines [18]: recovery, linearity, sensitivity, accuracy, and precision. Extraction recoveries resulting from the peak ratio of compounds in the sample after extraction versus the peak ratio without

Table 2 Matrix effect and recoveries. Conc.a (ng/mL)

1.5

62.5

Matrix effect (%) (n = 7) CV(%)c Mean ± SDb

Mean ± SDb

E-endoxifen Z-endoxifen 4-HOTam 4 -HOTam Z -endoxifen

99.7 ± 4.9 107.2 ± 3.5 102.5 ± 7.3 106.6 ± 4.0 109.7 ± 4.3

102.9 ± 4.2 106.2 ± 3.4 107.0 ± 5.4 104.0 ± 4.9 103.1 ± 4.3

Conc.a (ng/mL)

20

N-DMTam TAM Tam-N-ox

102.8 ± 1.8 104.5 ± 1.9 75.4 ± 14.5

Conc.a (ng/mL)

25

d5-TAM (IS)

108.8 ± 4.9

4.9 3.2 7.1 3.7 3.9

125

2.5

CV(%)c

Mean ± SDb

CV(%)c

Recovery (%) (n = 6) Mean ± SDb

4.1 3.2 5.1 4.7 4.2

99.5 ± 6.6 101.3 ± 7.1 99.8 ± 8.0 97.9 ± 7.2 97.7 ± 7.1

6.6 7.0 8.0 7.4 7.2

58.5 ± 9.9 58.0 ± 7.6 78.8 ± 13.5 73.2 ± 14.3 59.5 ± 8.9

75 1.7 1.8 19.9

95.8 ± 3.9 93.4 ± 6.8 96.5 ± 4.7

150 4.1 7.3 4.8

91.6 ± 4.3 88.7 ± 3.9 98.4 ± 6.0

4.7 4.3 6.1

25

62.5

100

125

56.4 ± 6.1 56.1 ± 7.7 79.8 ± 15.5 66.8 ± 10.6 63.2 ± 8.2

70.4 ± 11.3 71.4 ± 13.8 59.6 ± 7.3 74.0 ± 12.5 58.9 ± 13.8

61.2 ± 4.2 64.2 ± 5.9 63.3 ± 4.7 70.7 ± 8.1 55.8 ± 7.3

60.7 ± 8.4 61.3 ± 5.8 84.7 ± 14.7 71.6 ± 11.6 67.5 ± 9.0

20

75

125

150

250

81.6 ± 14.9 85.4 ± 15.2 88.1 ± 8.5

71.9 ± 10.4 68.0 ± 3.5 70.0 ± 9.9

73.3 ± 10.3 72.9 ± 11.6 61.4 ± 9.7

74.7 ± 8.3 71.6 ± 9.3 73.6 ± 6.2

74.5 ± 9.7 70.5 ± 12.4 62.9 ± 11.5

25 4.5

82.1 ± 6.7d

Abbreviations: HOTam: hydroxytamoxifen, Tam-N-ox: tamoxifen-N-oxide, TAM: tamoxifen, N-DMTam: N-desmethyltamoxifen, d5 -TAM: deutered tamoxifen. a Concentrations. b Mean (n = 5) and SD: standard deviation. c CV (%): coefficient of variation. d IS recovery, n = 18.

Table 3 Intra-day and inter-day precision and accuracy and coefficient of variation of the calibration slope for assay of tamoxifen and metabolites.

E-endoxifen

8.9

Z-endoxifen

5.5

Z -endoxifen

11.3

4-HOTam

16.0

4 -HOTam

9.4

Calibration slope CV (%) (n = 10) N-DMTam

9.3

Tam-N-ox

15.7

TAM

6.5

Nominal conc. (ng/mL)

Mean CV (%) Bias (%) Mean CV (%) Bias (%) Mean CV (%) Bias (%) Mean CV (%) Bias (%) Mean CV (%) Bias (%)

Intra-daya

Inter-dayb

0.5 LLOQ

1.5

2.5

25

100

0.5 LLOQ

1.5

2.5

25

100

0.52 ± 0.08 14.7 3.3 0.48 ± 0.04 7.8 −3.3 0.48 ± 0.06 12.9 −4.4 0.48 ± 0.11 22.07 −4.5 0.44 ± 0.07 15.3 −12.5

1.3 ± 0.1 11.1 −12.5 1.4 ± 0.1 10.4 −5.7 1.4 ± 0.2 14.0 −8.6 1.5 ± 0.1 10.5 −1.9 1.4 ± 0.2 14.1 −8.6

2.3 ± 0.4 15.2 −4.9 2.3 ± 0.3 12.4 −8.5 2.5 ± 0.3 13.3 1.2 2.5 ± 0.2 8.1 0.5 2.6 ± 0.2 9.2 4.6

23.4 ± 1.7 7.4 −6.5 23.5 ± 1.9 8.1 −5.9 24.2 ± 2.1 8.8 −3.4 23.7 ± 1.2 5.2 −5.0 23.7 ± 1.2 5.2 −5.0

99.0 ± 5.1 5.1 −1.0 96.3 ± 4.9 5.1 −3.7 103.3 ± 4.2 4.1 3.3 92.7 ± 4.5 4.8 −7.5 110.1 ± 5.5 5.0 10.1

0.48 ± 0.10 20.1 −4.5 0.45 ± 0.07 16.1 −9.3 0.41 ± 0.08 20.1 −17.8 0.49 ± 0.07 15.0 −2.5 0.45 ± 0.07 15.5 −10.8

1.42 ± 0.16 11.3 −5.3 1.52 ± 0.17 11.5 1.1 1.51 ± 0.18 11.9 0.6 1.52 ± 0.19 12.4 1.07 1.49 11.9 −0.71

2.5 ± 0.26 10.7 −1.9 2.4 ± 0.3 13.1 −4.3 2.4 ± 0.3 13.9 −5.7 2.4 ± 0.3 13.3 −2.0 2.5 ± 0.3 13.5 −0.6

23.3 ± 4.0 17.2 −6.83 23.1 ± 3.3 14.3 −7.5 21.7 ± 2.7 12.5 −13.2 23.69 ± 2.4 10.2 −5.2 23.6 ± 2.4 10.1 −5.4

102.6 ± 16.5 16.1 2.6 100.1 ± 13.6 13.6 0.1 96.8 ± 13.6 13.9 −3.2 98.5 ± 12.2 12.4 −1.5 102.4 ± 12.0 11.8 2.4

Nominal conc. (ng/mL)

12.5

20

125

250

12.5

20

125

250

Mean CV (%) Bias (%) Mean CV (%) Bias (%) Mean CV (%) Bias (%)

13.1 ± 1.1 8.3 4.4 13.9 ± 1.0 7.5 11.2 13.6 ± 0.8 6.2 9.1

17.5 ± 1.3 7.0 −10.3 19.7 ± 1.7 8.9 −1.65 18.5 ± 1.0 5.4 −7.3

104.2 ± 6.7 6.3 −16.7 113.2 ± 2.5 2.2 −9.5 108.1 ± 5.3 4.9 −13.5

283.0 ± 38.0 13.7 13.2 286.5 ± 38.1 13.3 14.6 260.2 ± 21.2 8.1 4.1

13.4 ± 0.7 5.4 7.1 14.0 ± 1.0 7.4 11.9 13.4 ± 1.5 11.1 7.2

20.1 ± 1.9 9.6 0.7 20.6 ± 2.5 12.0 3.1 21.76 ± 3.2 14.8 8.8

124.8 ± 8.9 7.2 3.8 126.1 ± 18.8 14.9 0.9 129.70 ± 15.0 11.6 3.7

263.5 ± 28.4 10.7 5.4 258.9 ± 28.3 10.9 3.6 276.35 ± 29.8 10.8 10.5

Abbreviations: HOTam: hydroxytamoxifen, Tam-N-ox: tamoxifen-N-oxide, TAM: tamoxifen, N-DMTam: N-desmethyltamoxifen, d5 -TAM: deutered tamoxifen. CV (%): coefficient of variation. a n = 6. b Mean n = 12

C. Arellano et al. / Journal of Pharmaceutical and Biomedical Analysis 100 (2014) 254–261

Calibration slope CV (%) (n = 10)

257

258

C. Arellano et al. / Journal of Pharmaceutical and Biomedical Analysis 100 (2014) 254–261

Fig. 1. MRM chromatogrames of standard compounds at the lowest calibration concentration: 0.5 ng/mL (A, B); 12.5 ng/mL (D, E, F) and 25 ng/mL (EI, C).

Fig. 2. TIC chromatogram of blank plasma (A) and standard compounds (B): with respectively from left to the right E-endoxifen, Z-endoxifen, Z -endoxifen, 4hydroxytamoxifen and 4 -hydroxytamoxifen (10 ng/mL) and tamoxifen, N-desmethyltamoxifen and tamoxifen-N-oxide (12.5 ng/mL) and numbers representing their retention time in minutes.

Fig. 3. Example of chromatogram from patient plasma at the steady state of 18-month daily taken 20 mg PO.

C. Arellano et al. / Journal of Pharmaceutical and Biomedical Analysis 100 (2014) 254–261

extraction ranged between 55.8% and 88% depending on the chemical properties of compounds and were constant with an acceptable RSD within 15%, including IS extraction recovery (Table 2). LC or UPLC–MS/MS methods should exclude any matrix effects during analysis in between the matrix component and the quantified compounds during elution [19]. This was evidenced by analysis of blank plasma spiked with target compounds after extraction (peak A) compared with the analysis of a solution of target compounds in solvent (peak B); for each compound, the peak ratio A/B showed that there was no matrix effect under the chromatographic conditions used [20,21]: main ratios being between 88.7 and 107.7% with variability (CV) under 10% (Table 2). The higher CV (15.4%) for the lowest ratio obtained with Tam-N-ox at 20 ng/mL could not be considered as a matrix effect. This point is very important to have reliable concentrations of analytes in further plasma analyses. No memory effect was detected when a blank sample (solvent) was analyzed after the highest QC sample or the highest calibration point. Calibration curves were linear in the analytical range (r > 0.995), the accuracy (bias) was set at 15% for calibrators and the coefficient of variation of the slope of the calibration line was less than 16% (n = 10, Table 3). The concentration for the validation range was selected according to the plasma concentrations previously reported in the literature for TAM and its metabolites. This range enabled us to quantify concentrations that were at least ten times smaller than the ones found in patient plasma samples for Z-endoxifen, or three times smaller than those for HOTam; the method reported here was twice as much sensitive for endoxifen quantification than the LC–MS/MS method reported by Teunissen et al. (LOQ = 1 ng/mL) [16]. UPLC–MS/MS technology compared to LC–MS/MS has also the advantage of a higher resolution for the separation of a set of related compounds within a short run time. Compared to previous reports, only two other UPLC–MS methods have been put forward for the quantification of TAM and its metabolites but they did not allow the separation of Z- and Z -isomers [11] or E- and Z-endoxifen [12] as the method we developed does and our analysis time was about half that of the HPLC–MS/MS methods reported by Jaremko et al. [14] and Teunissen et al. [16], respectively 11.6 min and 10 min. In the case of TAM, the absence of selectivity for quantification of its metabolites could give wrong plasma concentrations, particularly for the active metabolites endoxifen and 4-HOTam [15]. Specificity of this method was demonstrated by the MRM chromatographic profiles of analytes reported in Fig. 1 showing a good resolution between the different metabolites and particularly between the 4- and 4 isomers of endoxifen and HOTam. In addition, the absence of peak interference with TAM or metabolites was demonstrated by the analysis of 5 batches of blank plasma and was evidenced in Fig. 2 by the chromatographic profile from blank plasma and the total ion recording (TIC) chromatogram of a mixture of analytes. Moreover, therapeutically active TAM is the (trans) Z-isomer which represents the only Z-configuration produced by enzymatic metabolism [22] and the conversion to cis isomers is considered negligible [23]; therefore the quantification of all E-isomers of TAM and metabolites are not required. Nevertheless the most commonly available commercial standard of endoxifen is a mixture of Z- and E-isomers, it is therefore very important to have the ability to separate them since the response is different for each isomer. We have not studied the question of a possible Z/E isomerization, this point was previously documented, and it was shown that the E-isomer ratio is negligible in vivo [3,24]; analysis of patient plasma samples in the laboratory confirmed the absence of E-endoxifen. Because exposure to light could induce Z/E isomerization, light sensitivity is an important point to consider; it has been shown that TAM and its metabolites are stable in daylight [12] during sample preparation time. In agreement with studies of freeze–thaw stability of TAM and its metabolites with sample storage at −70 ◦ C/−80 ◦ C

259

[11,16], we found that three freeze/thaw cycles of sample stored at −20 ◦ C have no impact on plasmatic concentrations of TAM and its metabolites. Moreover, the long term freezing effects at −20 ◦ C on TAM and its metabolites have also been documented in a previous study where Jaremko et al. [14] showed that the concentrations of Z- and Z -endoxifen and Z-, Z -4-OHTam were stable. After 6 months at −20 ◦ C, the mean concentrations of plasma Z-endoxifen, Z -endoxifen, Z-4-HOTam, and Z -4-HOTam were found to be 94.5%, 103.1%, 110.5%, and 98.1%, respectively, of the original fresh frozen plasma concentrations. TAM and N-DMTam were also stable for at least 6 months. However, long term stability for a longer freezing period has not yet been investigated and we plan to complete the stability study by re-analyzing a group of samples (e.g. every 3 months) over the period necessary for the completion of the study, which is to date for a minimum period of 36 months (Fig. 3). Murdter et al. [17] studied the separation of 3-hydroxytamoxifen and concluded that it was present in plasma samples in very small quantities compared to 4-HOTam and the assay of the latter was not validated. In fact, their method was validated for the assay of the main metabolites (phase I): endoxifen and 4HOTam and no information was reported on sample preparations, calibration range and matrix effect which is of critical importance in UPLC/LC–MS/MS analysis as mentioned above. The validated method which is presented here is able to properly separate Z- endoxifen and Z-4-HOTam (active metabolites) from the Z -isomers, and allowed accurate quantification of each isomer within 6 min; this short run time was an important criteria considering the overall number of plasma samples included in the ongoing clinical protocol. The precision and the accuracy of the method were demonstrated: for each compound, the bias (%) and the coefficient of variance (CV, %) were calculated for the measure of the intraand inter-day accuracy and precision respectively (Table 3). They ranged within 15% of the target concentrations and within 20% at LLOQ as usually expected by the regulatory guidelines. This method was applied for the determination of plasma concentration of TAM and its metabolites in 23 patients with a daily dose of 20 mg TAM per os. Plasma concentrations were monitored during 3 years for quantification every 6 months. Concentrations were assayed after patient plasma samples were thawed. Table 4 summarized the intra-individual variation of the plasma concentration of each metabolite and TAM for 23 patients. For each patient, the mean of the seven plasmatic concentrations taken every six months during the follow-up period of three years was calculated and the corresponding coefficient of variation was reported to reflect the intra-individual variability. In general, the order of magnitude of TAM and the metabolites’ concentrations found in our study were similar to those reported in the literature [12,14,24]. We observed that the intra-individual variability of the plasma concentration of TAM and its metabolites was lower than the interpatient variability, except for Tam-N-ox. Table 4 also reports the steady state plasmatic concentrations after 36 months of treatment, measured for each patient. In particular, CV (%) for inter-individual variability of Z-endoxifen was 56.7% whereas the mean CV (%) for intra-individual variability was only 23.2% which could results from CYP2D6 polymorphism. The largest intra-individual variability was found for TamN-ox (mean CV (%) of 61.9%). Moreover, contrary to TAM and other metabolites, a significant decreasing trend in concentrations was observed during the 3-year period: mean concentration was 23.1 ± 23.0 ng/mL at 6 months versus 7.43 ± 3.68 ng/mL at 3 years. Although degradation of Tam-N-ox during the long freezing time period (36 months) cannot be excluded, this degradation is not likely to occur in view of its stability after freeze/thaw cycles. We will be able to establish whether this decreasing trend in

260

C. Arellano et al. / Journal of Pharmaceutical and Biomedical Analysis 100 (2014) 254–261

Table 4 Intra-individual variation of concentrations (mean of individual CV (%) values) at steady-state between 6 and 36 months for each patients (n = 23), and inter-individual (CV (%)) from steady-state concentration (at 36 months). CompoundsCV (%)

Z-endoxifen

Z -endoxifen

4-HOTam

4 -HOTam

NDM-tam

Tam-N-ox

TAM

Intra-patient variability Mean (min; max)

23.2 (8.9; 49.7)

17.6 (6.9; 31.4)

22.3 (11.7; 52.3)

17.3 (7.6; 24.6)

20.6 (7.6; 47.8)

61.9 (12.1; 100.5)

22.1 (9.8; 59.7)

Inter-patient variability Mean conc. CV (%)

9.90 56.7

6.57 50.72

8.08 46.68

4.71 41.36

149.38 36.45

7.43 49.56

114.78 31.32

Abbreviations: TAM: tamoxifen, HOTam: hydroxytamoxifen, N-DMTam: N-desmethyltamoxifen, Tam-N-ox: tamoxifen-N-oxide, min: minimal mean, max: maximal mean, n = 23 patients.

Tam-N-ox concentrations is due to time-dependent pharmacokinetics or a stability issue by re-analyzing both quality controls and patient samples after a 3-year period of storage at −20 ◦ C. In most previous publications, Tam-N-ox concentrations were not determined. However, the role of Tam-N-ox has previously been explored and it was found to have a similar estrogen activity as TAM and could contribute to effect or adverse effect during TAM therapy [25–27]. For example, the N-DMTam/Tam-N-ox ratio was found to be highly significantly related to the prediction of CP2C19 activity [25]. These results agree with those of Jordan [26] who showed that Tam-N-ox represents an intermediate metabolic step between TAM and N-DMTam. Parte et al. [27] report that TAM may undergo oxidation and Tam-N-ox reduction in the liver in a cyclic fashion. Hepatic flavin monooxygenase oxidize TAM to Tam-N-ox and, in turn, a portion of the Tam-N-ox formed is reduced back to TAM by cytochrome P450s; Tam-N-ox could also be reduced while circulating in body fluids by heme proteins (e.g. hemoglobin) and may act as a reservoir for tamoxifen in tissues, yielding TAM as and when required [28]. This accurate and quick method of determining Tam-N-ox plasmatic concentrations will help to increase the understanding of the role of this metabolite. This is of particular importance in the light of the current debate concerning the exact implication of cytochrome CYP2D6 in TAM treatment failure.

[2]

[3]

[4]

[5]

[6]

[7]

4. Conclusion [8]

A quick UPLC method to accurately determine the concentration of TAM and its metabolites was developed. The method was used to determine plasmatic concentrations in patients, showing an inter-patient variability of endoxifen and 4-hydroxytamoxifen which could be due to CYP2D6 polymorphism. Further pharmacogenetic analyses will enable us to determine and quantify the impact of pharmacogenetics and other factors on TAM pharmacokinetics. Because of the significant inter-individual variability of TAM pharmacokinetics, the relevance and importance of Therapeutic Drug Monitoring for TAM are starting to be recognized [29]. In order to properly apply TDM in daily practice, further PK studies are necessary to increase our knowledge of this important drug. The analysis of TAM and its isomeric metabolites in such studies will be made easier by the use of the method developed in this paper.

[9]

[10]

[11]

[12]

[13]

Acknowledgement We thank Melanie White-Koning for her editorial assistance with the English.

[14]

[15]

References [1] M.D. Johnson, H. Zuo, K.H. Lee, J.P. Trebley, J.M. Rae, R.V. Weatherman, Z. Desta, D.A. Flockhart, T.C. Skaar, Pharmacological characterization of

[16]

4-hydroxy-N-desmethyl tamoxifen, a novel active metabolite of tamoxifen, Breast Cancer Res. Treat. 85 (2004) 151–159. V. Stearns, M.D. Johnson, J.M. Rae, A. Morocho, A. Novielli, P. Bhargava, D.F. Hayes, Z. Desta, D.A. Flockhart, Active tamoxifen metabolite plasma concentrations after coadministration of tamoxifen and the selective serotonin reuptake inhibitor paroxetine, J. Natl. Cancer Inst. 95 (2003) 1758– 1764. B.S. Katzenellenbogen, M.J. Norman, R.L. Eckert, S.W. Peltz, W.F. Mangel, Bioactivities, estrogen receptor interactions, and plasminogen activator-inducing activities of tamoxifen and hydroxy-tamoxifen isomers in MCF-7 human breast cancer cells, Cancer Res. 44 (1984) 112–119. Y.C. Lim, Z. Desta, D.A. Flockhart, T.C. Skaar, Endoxifen (4-hydroxy-Ndesmethyl-tamoxifen) has anti-estrogenic effects in breast cancer cells with potency similar to 4-hydroxy-tamoxifen, Cancer Chemother. Pharmacol. 55 (2005) 471–478. M.P. Goetz, S.K. Knox, V.J. Suman, J.M. Rae, S.L. Safgren, M.M. Ames, D.W. Visscher, C. Reynolds, F.J. Couch, W.L. Lingle, R.M. Weinshilboum, E.G. Fritcher, A.M. Nibbe, Z. Desta, A. Nguyen, D.A. Flockhart, E.A. Perez, J.N. Ingle, The impact of cytochrome P450 2D6 metabolism in women receiving adjuvant tamoxifen, Breast Cancer Res. Treat. 101 (2007) 113–121. K. Kiyotani, T. Mushiroda, Y. Nakamura, H. Zembutsu, Pharmacogenomics of tamoxifen: roles of drug metabolizing enzymes and transporters, Drug Metab. Pharmacokinet. 27 (2012) 122–131. M.A. Province, M.P. Goetz, H. Brauch, D.A. Flockhart, J.M. Hebert, R. Whaley, V.J. Suman, W. Schroth, S. Winter, H. Zembutsu, T. Mushiroda, W.G. Newman, M.T. Lee, C.B. Ambrosone, M.W. Beckmann, J.Y. Choi, A.S. Dieudonne, P.A. Fasching, R. Ferraldeschi, L. Gong, E. Haschke-Becher, A. Howell, L.B. Jordan, U. Hamann, K. Kiyotani, P. Krippl, D. Lambrechts, A. Latif, U. Langsenlehner, W. Lorizio, P. Neven, A.T. Nguyen, B.W. Park, C.A. Purdie, P. Quinlan, W. Renner, M. Schmidt, M. Schwab, J.G. Shin, J.C. Stingl, P. Wegman, S. Wingren, A.H. Wu, E. Ziv, G. Zirpoli, A.M. Thompson, V.C. Jordan, Y. Nakamura, R.B. Altman, M.M. Ames, R.M. Weinshilboum, M. Eichelbaum, J.N. Ingle, T.E. Klein, CYP2D6 genotype and adjuvant tamoxifen: meta-analysis of heterogeneous study populations, Clin. Pharmacol. Ther. 95 (2014) 216–227. H.B. Brauch, W. Schroth, J.N. Ingle, M.P. Goetz, CYP2D6 and tamoxifen: awaiting the denouement, J. Clin. Oncol. 29 (2011) 4589–4590. H.B. Brauch, W. Schroth, J.N. Ingle, M.P. Goetz, H. Brauch, W. Schroth, M.P. Goetz, T.E. Murdter, S. Winter, J.N. Ingle, M. Schwab, M. Eichelbaum, Tamoxifen use in postmenopausal breast cancer: CYP2D6 matters, J. Clin. Oncol. 31 (2013) 176–180. S.F. Teunissen, H. Rosing, A.H. Schinkel, J.H. Schellens, J.H. Beijnen, Bioanalytical methods for determination of tamoxifen and its phase I metabolites: a review, Anal. Chim. Acta 683 (2010) 21–37. E. Dahmane, J. Boccard, C. Csajka, S. Rudaz, L. Decosterd, E. Genin, B. Duretz, M. Bromirski, K. Zaman, B. Testa, B. Rochat, E. Dahmane, T. Mercier, B. Zanolari, S. Cruchon, N. Guignard, T. Buclin, S. Leyvraz, K. Zaman, C. Csajka, L.A. Decosterd, Quantitative monitoring of tamoxifen in human plasma extended to 40 metabolites using liquid-chromatography high-resolution mass spectrometry: new investigation capabilities for clinical pharmacology, Anal. Bioanal. Chem. 406 (2010) 2627–2740. L. Binkhorst, R.H. Mathijssen, I.M. Ghobadi Moghaddam-Helmantel, P. de Bruijn, T. van Gelder, E.A. Wiemer, W.J. Loos, Quantification of tamoxifen and three of its phase-I metabolites in human plasma by liquid chromatography/triple-quadrupole mass spectrometry, J. Pharm. Biomed. Anal. 56 (2011) 1016–1023. L. Madlensky, L. Natarajan, S. Tchu, M. Pu, J. Mortimer, S.W. Flatt, D.M. Nikoloff, G. Hillman, M.R. Fontecha, H.J. Lawrence, B.A. Parker, A.H. Wu, J.P. Pierce, Tamoxifen metabolite concentrations, CYP2D6 genotype, and breast cancer outcomes, Clin. Pharmacol. Ther. 89 (2011) 718–725. M. Jaremko, Y. Kasai, M.F. Barginear, G. Raptis, R.J. Desnick, C. Yu, Tamoxifen metabolite isomer separation and quantification by liquid chromatographytandem mass spectrometry, Anal. Chem. 82 (2010) 10186–10193. N.G. Jager, H. Rosing, S.C. Linn, J.H. Schellens, J.H. Beijnen, Importance of highly selective LC–MS/MS analysis for the accurate quantification of tamoxifen and its metabolites: focus on endoxifen and 4-hydroxytamoxifen, Breast Cancer Res. Treat. 133 (2012) 793–798. S.F. Teunissen, N.G. Jager, H. Rosing, A.H. Schinkel, J.H. Schellens, J.H. Beijnen, Development and validation of a quantitative assay for the determination of

C. Arellano et al. / Journal of Pharmaceutical and Biomedical Analysis 100 (2014) 254–261

[17]

[18] [19]

[20]

[21]

[22]

tamoxifen and its five main phase I metabolites in human serum using liquid chromatography coupled with tandem mass spectrometry, J. Chromatogr. B 879 (2011) 1677–1685. T.E. Murdter, W. Schroth, L. Bacchus-Gerybadze, S. Winter, G. Heinkele, W. Simon, P.A. Fasching, T. Fehm, M. Eichelbaum, M. Schwab, H. Brauch, Activity levels of tamoxifen metabolites at the estrogen receptor and the impact of genetic polymorphisms of phase I and II enzymes on their concentration levels in plasma, Clin. Pharmacol. Ther. 89 (2011) 708–717, Erratum in Clin. Pharmacol. Ther. 91 (2012) 1087. FDA, Food and Drug Administration, Rockville, MD, 2002. A. Furey, M. Moriarty, V. Bane, B. Kinsella, M. Lehane, Ion suppression: a critical review on causes, evaluation, prevention and applications, Talanta 115 (2013) 104–122. F. Gosetti, E. Mazzucco, D. Zampieri, M.C. Gennaro, Signal suppression/enhancement in high-performance liquid chromatography tandem mass spectrometry, J. Chromatogr. A 1217 (2010) 3929–3937. K.H. Matuszweski, M.L. Constanzer, C.M. Chavez-Eng, Strategies for assessment of matrix effect in quantitative bioanalytical methods based on HPLC–MS/MS, Anal. Chem. 75 (2003) 3019–3030. J.K. Coller, N. Krebsfaenger, K. Klein, K. Endrizzi, R. Wolbold, T. Lang, A. Nüssler, P. Neuhaus, U.M. Zanger, M. Eichelbaum, T.E. Mürdter, The influence of CYP2B6, CYP2C9 and CYP2D6 genotypes on the formation of the potent antioestrogen Z-4-hydroxy-tamoxifen in human liver, Br. J. Clin. Pharmacol. 54 (2002) 157–167.

261

[23] Y. Zheng, D. Sun, A.K. Sharma, G. Chen, S. Amin, P. Lazarus, Elimination of antiestrogenic effects of active tamoxifen metabolites by glucuronidation, Drug Metab. Dispos. 35 (2007) 1942–1948. [24] M.F. Barginear, M. Jaremko, I. Peter, C. Yu, Y. Kasai, M. Kemeny, G. Raptis, R. Desnick, Increasing tamoxifen dose in breast cancer patients based on CYP2D6 genotypes and endoxifen levels: effect on active metabolite isomers and the antiestrogenic activity score, Clin. Pharmacol. Ther. 90 (2011) 605–611. [25] J. Gerde, J. Geisler, S. Lundgren, D. Ekse, J.E. Varhaug, G. Mellgren, V.M. Steen, E.A. Lien, Associations between tamoxifen, estrogens, and FSH serum levels during steady-state tamoxifen treatment of postmenopausal women with breast cancer, BMC Cancer 10 (2010) 313–326. [26] V.C. Jordan, New insights into the metabolism of tamoxifen and its role in the treatment and prevention of breast cancer, Steroids 72 (2007) 829–842. [27] P. Parte, D. Kupfer, Oxidation of tamoxifen by human flavin-containing monooxygenase (FMO) 1 and FMO3 to tamoxifen-N-oxide and its novel reduction back to tamoxifen by human cytochromes P450 and hemoglobin, Drug Metab. Dispos. 33 (2005) 1446–1452. [28] J. Gjerde, S. Gandini, A. Guerrieri-Gonzaga, L.L. Haugan Moi, V. Aristarco, G. Mellgren, A. Decensi, E.A. Lien, Tissue distribution of 4-hydroxy-Ndesmethyltamoxifen and tamoxifen-N-oxide, Breast Cancer Res. Treat. 134 (2012) 693–700. [29] N.G.L. Jager, H. Rosing, J.H.M. Schellens, S.C. Linn, J.H. Beijnen, Tamoxifen dose and serum concentrations of tamoxifen and six of its metabolites in routine clinical outpatient care, Breast Cancer Res. Treat. 143 (2014) 477–483.