Rapid Quantitative Determination of Epirubicin and Its Metabolites in Plasma Using High Performance Liquid Chromatography and Fluorescence Detection T. Dine, C. Brunet, M. Luyckx and M. Cazin Laboratoire de Pharmacologie, Pharmacocinetique et Pharmacie Clinique (Professeur J . C. Cazin), FacultP de Pharmacie, Rue du Professeur Laguesse, 59045 Lille Cedex, France

P. Gosselin and J. L. Cazin Laboratoire de Radiopharmacie et Pharmacie Clinique, Centre Oscar Lambret (Professeur A. Demaille), BP309, 59020 Lille Cedex, France

A rapid sensitive and selective isocratic technique was developed for the analysis of plasma epirubicin and three of its known fluorescent metabolites epirubicinol, 4'-O-~-~-glucuronyl-4'-epidoxorubicin and 4'-O-fi-D-ghCUrOnyl 1,3-dihydro-4'-epidoxorubicin, with daunorubicin as an internal standard, by using high performance liquid chromatography (HPLC) with fluorescence detection and a 'Hypersil ODs' column. The drugs were easily and efficiently extracted with a Sep-Pak C,* cartridge and the mean recoveries were greater than 85%. Intraassay and Interassay coefficients of variation (plasma samples) were better than 8.25%. An example of pharmacokinetic study obtained in a cancer patient after intravenous injection of epirubicine is described.

INTRODUCTION Epirubicin (EpiDX, Farmorubicin) is an anthracycline with antitumor activity, synthesized in the Farmitalia Carlo Erba Laboratories. Clinical studies indicated activity against breast and ovarian cancer, nonHodgkin's lymphomas, soft-tissue sarcomas, pancreatic and gastric cancer and acute leukaemia (Cersosimo et al., 1986). Structurally, EpiDX (2) differs from doxorubicin (1) (DX, Adriamycin) in the epimerization of the OH group in position 4' of the daunosamine sugar.

H2m

HLOH

\ Me0

/

o

\ H

HO

'0

1

Me0

GLU-EPIDX

GLU-EPIDXOL

/

0

H

HO

H

o +'

'0

1

Scheme 1. Major metabolic pathways of EpiDX

d

NHp

In most studies, HPLC with fluorescence detection was adopted for the specific, sensitive and accurate assay Preclinical studies in animal models (Casazza, 1979; of EpiDX and its metabolites in biological fluids Yeung et al., 1988) and randomized clinical trials (Bon(Robert, 1980; Moro et al., 1982; Maessen et al., 1987; fante er d.,1979; Hurteloup et d.,1983; Ganzina, 1983; Camaggi et al., 1988). This paper describes a rapid and sensitive method for Young, 1984) have proved that this anthracycline induces less acute toxicities and is less cardiotoxic than the extraction and determination of EpiDX and its three DX when used in equimolar doses. The differences in metabolites in human plasma using high performance toxicity between DX and EpiDX might depend on a liquid chromatography with reversed-phase column and different metabolic pathway (Scheme 1). Indeed, EpiDX fluorescence detection. This method is of great interest and epirubicinol (EpiDXol), its main metabolite, are in pharmacokinetic studies with the determination of conjugated with glucuronid acid occurring at the pharmacokinetic parameters in cancer patients. equatorial-OH group in the 4' position. CorrespondWe report the experimental details and the validation data of the assay procedure, then data on plasma levels ing amounts of 4'-O-P-~-glucuronyl-4'-epiDX (GluEpiDX) and 4'- 0-P-~-glucuronyl-l,3-dihydro-4'-epiDX of EpiDX and its fluorescent metabolites in a cancer (Glu-EpiDXol) can be found in biological fluids patient, following a 50 mg/m2 intravenous dose of (Weenen et al., 1984). EpiDX. (2) EPlDX

CCC-0269-3879/90/0020-0023 $2.00 20 BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 1, 1990

@ Heyden & Son Limited, 1990

DETERMINATION OF EPIRUBlClN

EXPERIMENTAL Chemicals. EpiDX was obtained from Farmitalia Carlo Erba Laboratories (Milan, Italy). EpiDXol, Glu-EpiDX and GluEpiDXol were gifts from Dr Robert of Fondation BergoniC (Bordeaux, France). Daunorubicin (DNR), used as internal standard was supplied by Roger Bellon Laboratories (Neuilly sur Seine, France). Methanol (analytical grade) and acetonitrile (HPLC grade, Lichrosolv) were obtained from Merck (Darmstadt, FRG). Na2HP0, and NaH,PO, (Rectapur) were from Prolabo (Paris, France). Used for buffers, injectable preparation water (sterile and apyrogenic) was obtained from Maco-Pharma Laboratories (Tourcoing, France). Human plasma was obtained from the Oscar Lambret Centre (Lille, France). Preparation of standard solutions and plasma samples. Standard stock solutions of EpiDX, EpiDXol, Glu-EpiDX and GluEpiDXol were prepared in water at concentrations of SO pg/mL, 12.5 pg/mL and 700 ng/mL in polypropylene tubes, to prevent loss due to the adsorption to glass (Bots et a/., 1983). They were kept frozen at -20 "C for 4 weeks. The internal standard stock solution ( 5 mg/mL) was diluted with water to give a final concentration of 2.5 pg/mL. From standard stock solution of EpiDX, working solutions ranging between 2 and 1250 ng/mL were prepared weekly by suitable dilution with water in polypropylene tube and stored at 4 ° C when not in use. Known amounts of EpiDX were added to aliquots of 1 mL of human plasma in a polypropylene tube. Calibration curves were obtained by analysing 1 mL plasma sample spiked with 2.44; 4.88; 19.54; 78.125; 312.55 and 1250 ng of EpiDX (five samples for each concentration) and 7.32; 14.65; 29.30; 117.19; 468.75 and 937.5 ng of EpiDXol (five samples for each concentration).

Quantitative determinations. For drug quantification (EpiIIX, EpiDXol, Glu-EpiDX and Glu-EpiDXol), the peak ratio (drug peak area/DNR peak area) was calculated for each plasma sample and the amount of drug determined by reference to the calibration curve. Fluorescence of Glu-EpiDX and GluEpiDXol were assumed to be equal to EpiDX and EpiDXol, based on the similarity of their fluorescence spectra and the proposed molecular structure of their chromophores.

RESULTS AND DISCUSSION HPLC Figure I s h o w s t w o chromatograms: (a) w a s obtained after extraction of blank plasma. No p e a k that could interfere with t h e determination of E p i D X , its metabolites ( E p i D X o l , G l u - E p i D X a n d G l u - E p i D X o l ) or the internal s t a n d a r d ( D N R ) w a s s e e n ; ( b ) w a s obtained a f t e r extraction of plasma s p i k e d with k n o w n a m o u n t s of E p i D X , EpiDXol, Glu-EpiDX and G l u - E p i D X o l . Identification of free d r u g s ( E p i D X , D N R ) a n d metabolites in plasma samples was achieved by comparison with t h e authentic specimens obtained respectively from t h e Farmitalia Laboratories, Roger Bellon Laboratories and Dr Robert. E p i D X a n d its metabolites ( E p i D X o l , G l u - E p i D X and C l u - E p i D X o l ) , as well as D N R (internal s t a n d a r d ) were simultaneously well separated, identified and quantified by this HPLC proc e d u r e with retention times of 6.54, 3.62, 1.45, 1.05 and 13.90 min, respectively.

Chromatographic conditions and instrumentation. Chromatographic analysis was performed with an HP1090 High Performance Liquid Chromatograph (Hewlett Packard, Orsay, France) equipped with a variable volume injector, an automatic sampling system and a Hewlett Packard HP1046 A fluorescence detector operating at excitation and emission wavelengths 254 nm and 565 nm, respectively. The output from the detector was connected to a Hewlett Packard 9000 model 300 integrator and the data recorded on a terminal printer HP Thinkjet. Analyses were performed on a 5 pm Hypersil ODS C , 8 column (100 mm x 4 . 6 mm ID; Hewlett Packard, Orsay, France) operating at 25 "C. The analytical column was protected by a small pre-column (20 mm x 4.6 mm ID) packed wth Hypersil C I S5 pm (Hewlett Packard). During assay development components were eluted isocratically with a mobile phase consisting of acetonitrile+ formate buffer mixtures (35:65, v/v) at a flow rate of 1.5-2.5 mL/min. The formate buffer was prepared in water with ammonia adjusted to pH 4 with pure formic acid. Plasma extraction procedure. We added 50 p L internal standard stock solution (Daunorubicine, 2.5 pg/mL in water) to plasma samples (1 mL). This solution was filtered through a Sep-Pak C,8 cartridge (Waters Associates, Milford, MA, USA) previously washed with methanol (3 mL), then 3 x 3 mL 0.05 M phosphate buffer (Na,HPO,/NaH,PO,) pH 7.00. The cartridge was eluted first with 2 x 3 mL phosphate buffer (discharged) and then with 3 mL methanol. The last organic eluate was evaporated to dryness in a warm water-bath under a stream of nitrogen. The residue was reconstituted with 50 p L of the solvent used for the HPLC separation. Centrifugation in small conical tubes (at 30.5 x g for 10 mins) eliminates as a precipitate the lipid material extracted. The clear supernatant (20 pL) was then injected into the chromatograph.

@ Heyden & Son Limited, 1990

2

4

6

8

10

12

I4

(Time mi".)

Figure 1. (a) Chromatogram of extracted blank plasma. (b) Chromatogram of plasma spiked with amounts of EpiDX, its metabolites and DNR (internal standard). Peaks: 1, EpiDX; 2, EpiDXol; 3, Glu-EpiDX; 4, Glu-EpiDXol; 5, DNR. Extraction procedure and chromatographic conditions are described in the text.

BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 1, 1990 21

T. D I N E , C. BRUNET, M. LUYCKX, M. C A Z I N , P. GOSSELIN A N D J. L. C A Z I N ~~

Table 1. Precision and accuracy of EpiDX determinations Amounts

Average amounts

added

found

(ns/mL)

2.44 4.88 19.53 78.13 312.5 1250

(nslmL)

2.38*0.15 4.79* 0.35 19.1311.85 77.04f 3.60 314.20* 8.10 1252.90* 3.79

CV interassay

cv Accuracy

("/I

(%I

97.54 98.15 97.90 98.61 100.54 100.23

6.25

(%)

8.24 7.55 6.32 6.88 4.35 3.79

4 !I

1

intraassay

3.78 2.54

CV, coefficient of variation.

2

4

6

8

Tlma (mln

10

14

12

1

Figure 2. HPLC chrornatograrn of a plasma from a cancer patient obtained 20 min after intravenous injection of EpiDX at the dose of 50 rng/m2.Peaks: 1. EpiDX; 2,EpiDXol; 3,Glu-EpiDX; 4,Glu-EpiDxol; 5 , Daunorubicin (internal standard); 6, unknown metabolite.

Linearity, recovery and precision of the extraction technique

We obtained good linearity between peak area ratios and concentrations in plasma samples spiked with 02500 ng/mL EpiDx and 125 ng DNR. The method shows a sensitivity of 2 ng/mL for each compound when present in human plasma, taking as the limit of detection a value three times higher than the baseline noise. The sensitivity appears to be good enough for therapeutic monitoring and pharmacokinetics studies. The calibration curves were fitted by the least squares method for the peak-area ratio of the sample substance and the internal standard ( y ) versus the concentration of the analysed product (x). The linear regression equation was: y=O.O136x-O.O75 for EpiDX and y = 0.0048~- 0.029 for EpiDXol. The correlation coefficients were above 0.998 and no significant differences were observed between the equation parameters. To assess reproductibility, the same plasma was extracted 5 times for each point of the calibration curves. Replicate analysis of plasma samples to which known amounts of EpiDX and EpiDXol were added demonstrate that this method had acceptable accuracy and precision (Table 1 for EpiDX; Table 2 for EpiDXol). The coefficients of variation were all below 8.25%; recoveries for the drug and its metabolites were all better than 85%. Application to pharmacokinetic studies

This HPLC technique is well adapted for pharmacokinetic studies in human. Figure 2 shows a plasma HPLC chromatogram from a cancer patient, obtained

1000.

I

CONCENTRATION (ng/mlJ

4

1

0

5

10

15

20

25

30

35

40

50

45

TIME (hours)

Figure 3. Plasma levels of EpiDX after i.v. injection (50rng/m2).

30 min after intravenous injection of 50 mg EpiDX/m2. Blood samples were made at 5, 10, 20, 30 and 40min and 1, 2, 4, 8, 24, 32 and 48 h. Blood was collected on EDTA-coated tubes and immediately centrifuged. The plasma was frozen at -20 "C until analysis. The plasma disposition kinetics of EpiDX after i.v. administration in this patient was characterized by a triphasic decay pattern, and its pharmacokinetics were compatible with a three-compartments open model (Figure 3). During the first phase, there was a rapid decrease of plasma levels with r,/> (Y of about 4 min, indicating a rapid distribution. A second phase was then established until 8 h after administration, followed by a very slow terminal elimination phase with plasma t,,*y

Table 2. Precision and accuracy of EpiDXol determinations (mean of five injections in each case) Table 3. EpiDX plasma pharmacokinetics

added

Average amounts found

interassay

(ng/mL)

(na/mL)

("/I

Amounts

7.32 14.65 29.3 117.19 468.75 937.5

7.11*0.31 14.18*0.64 28.77 1.56 116.57k4.25 470.34*9.44 939.82* 19.23

*

cv

cv Accuracy

7.86 7.05 6.43 5.18 4.98 4.36

CV, coefficient of variation.

22 BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 1, 1990

("/I 97.13 96.79 98.19 99.47 100.34 100.24

intraassay

(%I 6.38

Parameters of the triexponentlal equatjonD

A

a

8

P

C

Y

(cg/mL)

(h-'l

(pg/mLl

(h-')

(ca/mL)

(h-')

0.0172

0.73

0.00876

0.036

2.46

1 1 92

5.02 2.11

Pharmacokinetic parameters of EptDX

tllla

hI2P

fl,2Y

AUC cghmL'

0.06 h

0.95 h

20 h

0.473

a : C(t ) = A e-u'

"d

C IT

(7)

105.6L h m-2 2938.5L/m2

+ 6 e-@'+ C e-y'

@ Heyden & Son Limited, 1990

DETERMINATION OF EPIRUBICIN

*- Glu EpiDX O- EpiDXol

1

1 0

; 5

: 10

: 15

: 20

: 25

:

30

: 35

: 40

: 45

I

50

TIME (hours)

Figure 4. Plasma levels of metabolites of EpiDX

of about 20 h. With a software created in our laboratory (Cazin et al., 1987), we computed pharmacokinetic parameters presented in Table 3. In this patient, we observed pharmacokinetic parameters different to those described in the literature (Weenen et al., 1983; Camaggi et al., 1985; Robert et al., 1985; Vrignaud et al., 1985; Eksborg et al., 1986). The volume of distribution and the plasma clearance were higher and f,,2y shorter. This example

well showed important individual variations of metabolism of EpiDX and pharmacokinetic parameters could be significantly modified in patients with liver metastases or abnormal liver function tests (Camaggi et al., 1982). Figure 4 shows the kinetics of the metabolitaes: EpiDXol, Glu-EpiDX and Glu-EpiDXol; t l l 2values for the elimination phase were 18, 27 and 14 h respectively. In conclusion, we have demonstrated that EpiDX and its metabolites in human were determined with satisfactory accuracy with our methodology. The drug and its free major metabolites could be simultaneously separated and detected by a sensitive and simple HPLC method. An example of a pharmacokinetic study shows that the proposed method is well adapted for such studies in the human patient and may be of great help for the clinical monitoring of this drug. Acknowledgements This work was supported by grants o f The Comites du Nord et du Pas de Calais de la Ligue Nationale FranGaise Contre le Cancer. We thank very truly Doctor Robert, who gave us metabolites o f epirubicine used in this study.

REFERENCES Bonfante, V., Bonnadona, G. and Villani, F. (1979). Canc. Treat. Rep. 63, 915. Bots, A. M. B., Van Oort, W. J., Moordhoek, J., Van Dijk, A., Klein, S . W. and Van Haesel, Q. G. C. M. (1983). J. Chromatogr. 272, 421. Camaggi, C. M., Comparsi, R., Strocchi, E., Testoni, F. and Pannuti. F. (1988). Canc. Chemother. Pharmacol. 21, 216. Camaggi, C. M.. Strocchi, E., Martoni, A,, Angelelli, B., Comparsi. R. and Pannuti, F. (1985). Drugs Exp. Clin. Res. 11, 285. Camaggi, C. M., Strocchi, E. and Tamassia. V. (1982). Canc. Treat. Rep. 66, 1819. Casazza, A. M. (1979). Canc. Treat. Rep. 63, 835. Cazin, J. L., Luyckx, M., Brunet, C., Gosselin, P. and Demaille, M. C. (1987). J. Pharm. Clin. 6, 585. Cerosimo, R. J. and Ki Hong, W . J. (1986). Clin. Oncol. 4, 425. Eksborg, S., Stendhal, U. and Lonroth, U. (1986). Eur. J. Clin. Pharmacol. 30, 629. Ganzina, F. (1983). Canc. Treat. Rep. 10, 1. Hurteloup, P., Cappelaere, P. and Armand, J. P. (1983). Canc. Treat. Rep. 67. 337.

@ Heyden & Son Limited, 1990

Maessen, P. A., Mross, K. €3.. Pinedo, H. M. and Van der Vijdh. W. J. F. (1987). J. Chromatogr. 417, 339. Moro, E., Jannuzzo. M. G., Ranghieri, M., Stegnjaich, S. and Valzelli. G. (1982). J. Chromatogr. 230, 207. Robert, J. (1980). J. Liq. Chromatogr. 3, 1561. Robert, J., Vrignaud, P., Nguyen-Ngoc, T.. Iliadis, A., Mauriac, L. and Hurteloup, P. (1985). Canc. Treat. Rep. 69, 633. Vrignaud, P., Eghbali, H., Hoerni, B., Iliadis, A. and Robert, J. (1985). fur. J. Canc. Clin. Oncol. 21, 1307. Weenen, H., Lankelma, J., Penders, P. G. M., Mc Vie, J. G., Ten Bokkel Huinink, W . W., De Planque, M. M. and Pinedo, H. M. (1983). Invest. New Drugs 1, 59. Weenen, H., Van Maanen, J. M. S., De Planque, M. M., Mc Vie, J. G. and Pinedo, H. M. (1984). Eur. J. Canc. Clin. Oncol. 20. 919. Yeung, T. K., Simmons, R. H. and Hopewell. J. W . (1988). Radioth. Oncol. 11, 263. Young, C. W. (1984). In: Bonnadonna, G. (ed) Advances in Anthracycline Chemotherapy: Epirubicin. p. 183. Masson, Milan.

BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 1. 1990 23