Room temperature phosphorescence optosensor for anthracyclines

Analyst, January 1998, Vol. 123 (151–154)

151

Room temperature phosphorescence optosensor for anthracyclines Fausto Alava-Morenoa, Mar´ıa Jesus ´ Valencia-Gonzáleza and Marta Elena D´ıaz-Garc´ıab a Department of Analytical and Food Chemistry, Faculty of Sciences of Vigo, University of Vigo, As Lagoas-Marcosende, P.O. Box 874, E-36200 Vigo, Spain b Department of Physical and Analytical Chemistry, University of Oviedo, c/Juli´ an Claveria 8, E-33006 Oviedo, Spain. E-mail: [email protected] A flow-through optosensor for anthracyclines based on the anthracycline–europium chelate room temperature phosphorescence energy transfer is proposed. The sensor was developed in conjunction with a flow injection analysis system and is based on the transient immobilization on a non-ionic resin (packed in a flow-through cell) of the anthracycline–europium chelate. The analytical performance characteristics of the proposed sensor for semi-automated analysis and control of very low levels of anthracycline were as follows: the detection limits for daunorubicin, doxorubicin and epirubicin were 9.0, 5.8 and 5.8 ng ml21, respectively, with an RSD of 1% for the determination of 0.22 mg ml21 of each anthracycline (n = 10). Most of the common metal ions in biological samples did not interfere, except FeIII, which caused serious interference and should be masked with 1,10-phenanthroline. The recommended method was successfully tested for determination of anthracyclines in clinical samples (urine and pharmaceutical preparations). Keywords: Flow injection; phosphorimetry; sensors; optosensors; anthracyclines

The anthracycline (ATC) group of antibiotics are amongst the most clinically important anti-tumour agents in current cancer chemotherapy; doxorubicin in particular has the broadest spectrum of activity of any known antineoplastic agent.1 The 4Aepimer epirubicin was introduced into clinical medicine about 10 years ago2 and has been reported to have anti-tumour activity equivalent to that of doxorubicin but with lower cardiac and systemic toxicity. Only a few methods, such as fluorimetry,3 visible spectrophotometry,4 voltammetry5 and liquid chromatography6–8 have been reported for their determination in biological fluids and in dosage forms. The unique sensitized luminescence characteristics of certain lanthanide ions, notably europium(iii), were first exploited for the determination of tetracyclines by Hirschy et al.9 In aqueous solution, free europium ions do not luminesce efficiently, since any excited-state species are relaxed by vibrionic energy transfer to the aqueous solvent shell.10 In order to observe luminescence and to avoid radiationless relaxation mechanisms, the water should be removed or, at least, this solvent shell should be replaced by other ligands. Typically, luminescence is enhanced by the use of micelleforming reagents and water exclusion ligands such as trioctylphosphine oxide.11 Unlike molecular phosphorescence, sensitized lanthanide luminescence occurs readily in liquid solution at room temperature. Because energy transfer from the organic triplet state of the ligand to the emitting level of the lanthanide ion is an intramolecular process, the luminescence is not quenched by oxygen. In recent years, the combined use of flow injection analysis (FIA) and solid substrate room temperature phosphorescence

(SS-RTP) optosensing detection in an aqueous carrier used as a continuous flow stream has been demonstrated for ultra-trace aluminium12 and tetracycline13 determination. In the same vein, we report here the combined use of FIA and SS-RTP europium optosensing detection to determine anthracyclines at very low concentrations. The method is based on the transient immobilization (on to a non-ionic resin packed in a flow-through cell) of the anthracycline–EuIII complex, which exhibits RTP. Experimental Reagents Hydrochlorides of daunorubicin, doxorubicin (Sigma, St. Louis, MO, USA) and epirubicin (Pharmacia Farmitalia, Madrid, Spain) and europium chloride hexahydrate (Fluka, Buchs, Switzerland) were used as received, unless stated otherwise. The non-ionic resin Amberlite XAD-2 (Sigma) was packed in a column and thoroughly cleaned by passing 2 m HCl until no atomic absorption of iron could be measured in the effluents. Analytical-reagent grade chemicals were employed for the preparation of all solutions. Freshly prepared ultra-pure water (Milli-Q/Milli-Q2 system, Millipore, Bedford, MA, USA) was used in all experiments for both sample and standard solutions. The carrier solution used in the FIA experiments consisted of 0.07 m N,N,NA,NA-tetramethylethylenediamine (TEMED) (Merck, Darmstadt, Germany)–0.1 m HCl (pH 7.5, m = 0.25 m, adjusted with NaCl). Instrumentation Phosphorescence emission measurements were carried out with a Perkin-Elmer (Norwalk, CT, USA) LS-5 luminescence spectrometer, which employs a xenon-pulsed (10 ms half-width, 50 Hz) excitation source and is equipped with a Perkin-Elmer Model 3600 data station. The delay times used were typically 0.03 ms and the gate time was 2 ms. Instrument excitation and emission slits were set at 10 and 20 nm, respectively. Steadystate phosphorescent signals were recorded with a Perkin-Elmer Model 560 chart recorder. Fig. 1 illustrates the simple optosensing FIA manifold used. A conventional Hellma (Mulheim, Germany) Model 176.52 flow cell of 25 ml volume was used. At the end of the flow cell, a small piece of nylon was placed to prevent particle displacement by the carrier. The resin was loaded into the flow cell with the aid of a syringe and the other end of the flow cell was kept free. The cell was then connected to the flow system and 10 min were allowed for the particles to settle. In order to ensure that the complex first retained by the packing solid material was in the light path, the resin level was maintained 1 mm lower than that of the cell windows. The resin packed in this way could be used for 4 months or even longer with satisfactory RTP readings. A four-channel Gilson Minipuls-3 peristaltic pump (Teknokroma, Barcelona, Spain) was used to generate the flowing

152

Analyst, January 1998, Vol. 123

streams. Omnifit 1106 rotary valves (Teknokroma) were used for sample introduction (valve A in Fig. 1) and for elution of the retained ATC–EuIII chelate (valve B). PTFE tubing (0.8 mm id) and fittings were used for connecting the flow-through cell, the rotary valves and the carrier solution reservoirs. pH measurements were made with a MicropH 2002 pH meter (Crison, Barcelona, Spain). General procedure Samples or standards (1 ml) were injected via valve A and pumped through the flow system. The ATC–EuIII complex formed (see below) reached the flow cell where it was retained on the resin. The high SS-RTP of the complex on the solid support was measured at the spectral maxima, lex = 393 nm and lem = 615 nm. Once the RTP measurement had been made, 2 ml of 0.5 m HCl were injected via valve B (to strip the ATC– EuIII chelate retained on the solid phase), before proceeding with the next sample injection. For injection in this flow-through sensing approach, standard and sample solutions were prepared as follows. An appropriate aliquot of the ATC standard solution (or the sample) was transferred into a 10 ml calibrated flask, 25 ml of EuIII solution (6.6 3 1022 m) were added and the solution was diluted to volume with the carrier solution. For urine analysis, to 5 ml of sample were added 25 ml of EuIII solution (6.6 3 1022 m) and 0.1 ml of aqueous 1,10-phenanthroline solution (5 x 1023 m) (in order to prevent interference from the possible presence of iron in the samples). The solution was diluted to 10 ml with 0.15 m TEMED solution of pH 7.5. For pharmaceutical liquid preparations, 1 ml of the sample was transferred into a 100 ml calibrated flask and diluted to volume with ultra-pure water. An appropriate aliquot of this solution was further diluted by placing it in a 10 ml calibrated flask, and 25 ml of EuIII solution (6.6 3 1022 m) and 0.1 ml of 1,10-phenanthroline solution (5 3 1023 m) and diluting to volume with the carrier solution. Reagent blanks were prepared and measured following the same procedure. Results and discussion Selection of the sensor phase Anthracyclines are known to form luminescent metal chelates with doubly and triply charged metal cations;14 it has been suggested that the complexation takes place through the bdiketone oxygens on rings B and C (Fig. 2) readily forming sixmembered rings with the metal ions. The benzoyl moiety is believed to be involved in possible energy transfer processes.

Fig. 1 FIA manifold for the proposed flow-through sensing system. A and B, injection valves; carrier, 0.07 m TEMED–0.1 m HCl (pH 7.5, m = 0.25 m, adjusted with NaCl).

The solid support for flow-injection RTP optosensing should ideally be translucent so that reflected or absorbed light entering or leaving the probe would be negligible. Taking into account our previous experience with solid support selection using a similar system,13 Amberlite XAD-2 was finally selected as the best sensor phase for the transient immobilization of preformed ATC–EuIII chelate. The luminescence properties of the ATC–EuIII chelate retained on the Amberlite XAD-2 are particularly well suited for RTP. The phosphorescence spectra of the ATC–EuIII chelate are shown in Fig. 3 As can be seen, the ATC molecule absorbs the excitation radiation and transfers the excited energy through its triplet state to the lowest resonance level of Eu3+, whereafter the ion undergoes a radiative transition resulting in a line-type emission characteristic of the ion. The ion-specific emission appears at a long wavelength (615 nm) with a large Stokes’ shift (around 200 nm). The most important feature in this context is the long luminescence lifetime, which makes it possible to apply time-resolved detection for effective background elimination and to enhance sensitivity. Although the luminescence emission originates from the europium ion, the surroundings have proved to play a very important role in the process. This could explain the shorter wavelength emission and lower intensity observed for the ATC–EuIII chelate in solution compared with the complex immobilized on the solid support. The ‘rigidly held’ mechanism hvexc

Eu3+

Amberlite XAD-2 NON-IONIC RESIN

O

O

B

C

R2 +

A

Eu3+ O

O

B

C

OCH3 O

OH

D H

OH R1

R2 A

OCH3 O

Fig. 2

OH

D H

OH R1

RTP emission

Probable response mechanism for the proposed sensor.

Fig. 3 Luminescence spectra for the ATC–EuIII complex retained on Amberlite XAD-2. (a) Excitation spectrum and (b) phosphorescence spectrum.


should account for the higher intensity observed when the ATC–EuIII complex is retained on the solid.15 Chemical requirements and experimental conditions The study of the influence of pH on the luminescence of the different chelates with three anthracyclines adsorbed on the Amberlite XAD-2 resin showed that the optimum pH for retention/detection occurred at pH 7.5 in all cases. The luminescence signal was not appreciably quenched by oxygen, probably because the emission process involves intra- rather than intermolecular energy transfer. The influence of EuIII concentration on optimum complex formation was studied using different ATC concentrations in the range 0–1026 m. Concentrations of EuIII > 1.3 3 1024 m had no influence on the luminescence signal. Therefore, in order to ensure complete ATC–EuIII complex formation, 1.7 3 1024 m EuIII was used in further studies. Elution of the ATC–EuIII chelate from the solid support was studied using HCl, H2SO4 and Na2SO4 at different concentrations as leaching agents. Acids were preferred to Na2SO4 and HCl and H2SO4 provided similar results. HCl was finally selected in order to keep resin shrinkage problems to the minimum (the same anion composition in the carrier and in the eluent). It was verified that 2 ml of HCl completely washed the complex out of the resin and allowed its re-use. As expected, the analytical signal decreased slightly as the carrier flow rate increased. A flow rate of 0.50 ml min21 was selected for subsequent work as a compromise between sensitivity and sample throughput. Selectivity of the RTP sensor The selectivity study for this RTP flow-through sensor for anthracyclines was focused first on those inorganic components potentially present in the samples to be analysed (e.g., Na+, K+, Mg2+, Ca2+ and Cl2). The results for the determination of 0.22 mg of the different anthracyclines studied following the recommended procedure are summarized in Table 1. Na+, K+, Mg2+, Ca2+ and Cl2 at twice the concentrations levels normally present in serum did not interfere in the determination of the anthracyclines. The effect was also checked for other cations such as Cu2+ and Fe3+ that could be present in these samples as impurities and displace EuIII from the ATC chelate. As shown in Table 1, the only interference was produced by FeIII. It was found, however, that the quenching influence of FeIII could be eliminated by masking it with 1,10-phenanthroline. Analytical performance characteristics The analytical figures of merit of the proposed sensing system were evaluated. Calibration graphs were prepared from the results of triplicate 1 ml injections of the corresponding ATC– EuIII chelate standard solution. The detection limit, using the 3sB criterion (sB being the standard deviation of the blank), and the observed relative standard deviation were evaluated for the

153

three anthracyclines under investigation. The results obtained are summarized in Table 2. Adsorption on the resin for the three ATC–EuIII complexes studied was found to be quantitative in the working range: 99.8 ± 0.5% for the daunorubicin, 99.9 ± 0.3% for the doxorubicin and 99.9 ± 0.4% for the epirubicin complex. The concentrations of anthracyclines used in the adsorption studies were 8 3 1027 m for the respective anthracyclines. Analytical applications Following the procedure detailed under Experimental, the proposed RTP flow-through sensing approach was applied to the determination of anthracyclines in a pharmaceutical preparation and in spiked urine. Three different urine samples were spiked with a known amount of epirubicin and analysed using direct aqueous calibration and the standard additions technique. No pre-treatment of samples was necessary. As shown in Table 3, the results obtained by the two techniques compared favourably with those expected. The epirubicin recovery was within ±3% of the spiked amount at nanomolar levels. A pharmaceutical preparation (Farmiblastina from Pharmacia Farmitalia) was analysed according to the proposed procedure. The result obtained was 2.00 ± 0.02 mg ml21 (n = 6) of doxorubicin, which compared excellently with the concentration reported by the manufacturer (2 mg ml21) Conclusions The RTP-based optosensor developed for the anthracycline group has proved to be accurate and suitable for the assay of anthracyclines added to urine and present in a pharmaceutical formulation. It does not suffer from interferences due to the inorganic compounds, proteins (present in urine) or excipients. The sensor is characterized by a reproducible and repetitive

Table 1 Effect of diverse ions on the determination of 0.22 mg of the three anthracyclines Recovery (%)* Concentration/mg ml21 Daunorubicin Doxorubicin Epirubicin — 100.0 100.0 100.0 7000 99.2 100.3 100.0 400 101.1 100.8 100.3 200 98.5 99.2 99.2 50 99.2 99.7 99.7 2 49.2 50.3 50.5 2† 99.2 99.5 100.5 2 98.9 99.2 98.9 Cu2+ Cl2 7000 100.4 100.5 100.3 5 98.5 100.3 99.5 SO422 PO432 10 99.2 98.9 99.2 1,10-Phen 20 99.6 100.3 100.5 * Mean of three determinations. † 20 mg ml21 of 1,10-phenanthroline was added.

Ion — Na+ K+ Ca2+ Mg2+ Fe3+

Table 2 Analytical characteristics of the anthracyclines Anthracycline Daunorubicin Doxorubicin Epirubicin *

(m).

Ten measurements of 0.22 mg ml21.

Detection limit/m Linear range/m Calibration fit† 1.3 1.6 3 1028 3 3 1028–8 3 1027 y = 0.10 + 2.1 3 107 x, r = 0.9999 0.9 1.0 3 1028 2 3 1028–8 3 1027 y = 0.01 + 3.0 3 107 x, r = 0.9999 0.9 1.0 3 1028 2 3 1028–8 3 1027 y = 0.02 + 3.1 3 107 x, r = 0.9999 Calibration line fitted by least-squares regression; y = RTP intensity and x = anthracycline concentration

RSD (%)*

†

154


return of the readout to the baseline after each single measurement cycle. The RTP sensor responds almost instantly to the presence of the anthracycline–EuIII complex, enabling a large number of samples to be measured in sequence without deterioration. A sampling frequency of 10–12 h21 can be achieved. The analytical future of RTP optosensing appears bright both from a practical and from a fundamental standpoint: from a practical point of view the intrinsically high spectral selectivity of phosphorescence can be substantially improved by measuring other related optical properties of the complexes (lifetime, light polarization, etc.) for multivariate analysis. In this way new sensing schemes could be exploited and their application to real sample multicomponent analysis should be easier, particularly when aided by chemometrics. Work in this direction is in progress as significant differences were observed in the lifetimes of the three ATC-Eu immobilized complexes studied. As few compounds can bind to the solid support used and transfer excitation energy to Eu3+, the use of time-resolved luminescence would increase even further the selectivity for a single or group of structurally analogous compounds (e.g., anthracyclines). From a basic point of view, time-resolved luminescence optosensing offers great potential, as yet unexploited, for fundamental studies on new sensing mechanisms and for the development and design of flow-through and probe-

Table 3 Analysis of urine samples Epirubicin concentration/mm Direct Sample No. Spiked calibration* 1 200 200 ± 6 2 250 246 ± 3 3 300 300 ± 4 * Each result is the mean ± s of three determinations.

Standard additions* 201 ± 3 247 ± 2 302 ± 3

type fibre-optical sensors for (bio)chemical species using delayed fluorescence and RTP measurements. Financial support from CICYT (Projects AMB 95/0476 and SAF 96/1484) is gratefully acknowledged. The authors also acknowledge Dr. I. Pelaez of the Oncology Service of the Hospital of Cabueñes, Gijón (Spain), for providing anthracycline pharmaceutical preparations. References 1 Black, D. J., and Livingston, R. B., Drugs, 1990, 39, 652. 2 Mross, K., Maessen, P., van der Vijgh, W. J. F., Gall, H., Boven, E., and Pinedo, H. M., J. Clin. Oncol., 1988, 6, 517. 3 Alykov, N. M., Nekrest Yanova, T. V., and Yakovleva, L. V., Zh. Anal. Khim, 1991, 46, 1642. 4 Sastry, C. S. P., and Lingeswara-Rao, J. S. V. M., Talanta, 1996, 43, 1827. 5 Ploegmakers, H. H. J. L., Moritz, P. A., Toll, P. J. M. M., and Van Oort, W. J., J. Autom. Chem., 1989, 11, 106. 6 Van der Vlis, E., Irth, H., Tjaden, U. R., and Van der Greef, J., Anal. Chim. Acta, 1993, 271, 69. 7 Leca, D., and Leca, F. R., Chromatographia, 1993, 35, 435. 8 De Jong, J., Vermorken, J. B., and Van der Vijgh, W. J. F., J. Chromatogr., 1992, 574, 309. 9 Hirschy, L. M., Dove, E. V., and Winefordner, J. D., Anal. Chim. Acta, 1983, 147, 311. 10 De, W., Horrocks, W., Jr., and Sidnik, D. R., Acc. Chem. Res., 1981, 14, 384. 11 Hemmilä, J., Dakubu, S., Mukkala, V. M., Siitari, H., and Lovgren, T., Anal. Biochem., 1984, 137, 335. 12 Pereiro Garc´ıa, R., Liu, Y. M., D´ıaz Garc´ıa, M. E., and Sanz- Medel, A., Anal. Chem., 1991, 63, 1759. 13 Alava-Moreno, F., D´ıaz-Garc´ıa, M. E., and Sanz-Medel, A., Anal. Chim. Acta, 1993, 281, 637. 14 Ming, L. J., and Wei, X. D., Inorg. Chem., 1994, 33, 4617. 15 Sanz-Medel, A., Anal. Chim. Acta, 1993, 283, 367.

Paper 7/03798H Received June 2, 1997 Accepted September 11, 1997