1470
Chem. Res. Toxicol. 2007, 20, 1470–1476
Photoaddition of Fluphenazine to Nucleophiles in Peptides and Proteins. Possible Cause of Immune Side Effects Sergio Caffieri,*,† Giorgia Miolo,† Roberta Seraglia,‡ Daniele Dalzoppo,† Francesca M. Toma,† and Gerard M. J. Beyersbergen van Henegouwen† Department of Pharmaceutical Sciences, UniVersity of PadoVa, Via Francesco Marzolo 5, I-35131 PadoVa, Italy, and CNR-ISTM, Corso Stati Uniti 4, I-35127 PadoVa, Italy ReceiVed April 19, 2007
By the action of UVA light, fluphenazine reacted with nucleophiles through a mechanism involving defluorination of its trifluoromethyl group, giving rise to carboxylic acid derivatives that were easily detected by electrospray mass spectrometry. This photoreaction took place with alcohols, sulphydryls, and amines. When irradiation of fluphenazine was carried out in the presence of an amino acid at pH 7.4, the R-amino group was covalently bound to the drug. With amino acids possessing a further nucleophilic residue on the side chain, such as lysine, tyrosine, and cysteine—but not serine—both groups reacted, resulting in a fluphenazine–amino acid–fluphenazine diadduct. The same occurred with the physiological peptide glutathione (γ-glutamylcysteinylglycine). By means of MALDI mass spectrometry, it was shown that fluphenazine also covalently bound to peptides and proteins such as calmodulin. This binding may result in the formation of antibodies, ultimately leading to the destruction of the granulocytes and thus suggesting that photoactivation of this drug may play a role in its clinical side effects, such as agranulocytosis. Introduction Fluphenazine is a phenothiazine derivative with general properties similar to those of chlorpromazine. It has a piperazine side chain (Scheme 1) and is used in the treatment of a variety of psychiatric disorders, including schizophrenia, mania, severe anxiety, and behavioral disturbances (1). The two most frequently occurring side effects are extrapyramidal reactions and photosensitivity (2). The latter complex of adverse reactions, resulting from simultaneous exposure of the body to sunlight and the drug, may involve a photototoxic response, in fact an exaggerated sunburn, or photoallergy. Photoallergy resembles contact hypersensitivity; the only difference is that the drug requires activation by light, instead of, for example, enzymes, in order to provoke this unwanted immune response (3). The photoexcited drug or its reactive photoproduct, called hapten, irreversibly binds to proteins, resulting in the formation of an antigen; this is supposed to be the initial step in photoallergy (4). Subsequently, the antigen triggers immunologic reactions that eventually result in the symptoms of contact hypersensitivity, in this case, photoallergy (5, 6). Photosensitivity as reported in clinical literature concerns the light-exposed areas of the body, mostly the skin and sometimes the eyes. However, normal photobiological processes in man, such as the UVB-induced vitamin D3 production from 7-dehydrocholesterol in the skin or the conversion of bilirubin in the blood of, for example, premature babies with visible light, demonstrate that systemic effects occur as a result of the simultaneous exposure to light and an endogenous compound present in the skin or blood. Vitamin D3 is essential for proper * Phone: 39 049 827 5706. Fax: 39 049 827 5366. E-mail: sergio.caffieri@ unipd.it. † University of Padova. ‡ CNR-ISTM.
bone calcification and the photoconversion of bilirubin in the visible light therapy of neonatal jaundice results in a decrease in brain-damaging effects. Referring to this kind of natural photobiological processes, it has been demonstrated that the combination of drug and light may also cause systemic effects; in other words, after photoexcitation of the drug in, for example, the skin or blood, the eventual biological effect occurs in an organ that never sees any light (7). Although it has not yet received attention in clinical research, further investigations with experimental animals (8, 9) support the concept that simultaneous exposure to sunlight and a drug or another chemical may provoke systemic effects in man as well. A life-threatening systemic side effect of fluphenazine is agranulocytosis (2). Agranulocytosis is a severe reduction in the number of white blood cells (specifically granulocytes) in the circulation and is in fact an impairment of the immune system. From investigations of diverse drugs, circumstantial evidence has been obtained that agranulocytosis is caused by reactive intermediates (10–12). The latter may occur during enzymatic conversion of the causative drug by activated leukocytes in the blood, amongst others. One of the possible mechanisms is that these reactive intermediates (irreversibly) bind to plasma proteins, or biomacromolecules on a cellular membrane, and that the resulting antigens provoke an immune response via the formation of antibodies, ultimately leading to the destruction of the granulocytes (13). However, besides enzymatic activation, photodegradation of the drug in vivo should be considered as a causative factor. The fact is that extensive research with diverse drugs and the rat as animal model has demonstrated that photodegradation in vivo occurs. Irreversible binding to biomacromolecules in internal organs such as the liver and the spleen was shown and photoproducts resulting from reactive intermediates were also found in leukocytes of animals exposed to UVA (8, 9).
10.1021/tx700123u CCC: $37.00 2007 American Chemical Society Published on Web 09/21/2007
Photoaddition of Fluphenazine to Nucleophiles
Chem. Res. Toxicol., Vol. 20, No. 10, 2007 1471
Scheme 1. Molecular Structure of Fluphenazine and Its Main Photoproducts Described in the Text
Referring to fluphenazine, photoreactivity resulting in irreversible binding to biomacromolecules should also be considered as a possible initial step eventually leading to agranulocytosis: (a) Fluphenazine is photolabile upon exposure to UVB and UVA (14); the latter also penetrates the network of capillary blood vessels. (b) The possibility of light exposure during the presence of fluphenazine in the body is high because the plasma half-life of the drug is between 15 h and several days; besides, it is administered daily over a long period (1). (c) The extent of complex binding of fluphenazine to proteins is high, about 99% in plasma (15). This facilitates photobinding because fluphenazine is already in close contact with the target molecule so that not only the photoexcited molecule in its long living triplet state but also the singlet state with a lifetime of 1 × 10-8 and shorter contributes to the covalent binding. (d) Complex binding is not limited to plasma but also involves proteins ubiquitously present in cellular membranes, including those of leukocytes. One of these proteins, essential for proper functioning of almost all kind of cells, is calmodulin, of which fluphenazine is an antagonist (16). The foregoing prompted us to investigate the possibility that fluphenazine photobinds to nucleophilic compounds present in the human body, peptides such as glutathione, and proteins such as calmodulin, which has been taken as a model because of its well-known strong noncovalent binding with the drug, that could favor the subsequent photobinding. Besides, photoaffinity labeling of calmodulin has already been studied with trifluoperazine, (17) which has a molecular structure very close to that of fluphenazine (trifluoperazine has a methyl group linked to the nitrogen of the piperazine ring and fluphenazine a hydroxyethyl group, see Scheme 1). Special attention was given to the trifluoromethyl group because it has recently been shown that the main photoproduct of fluphenazine results from hydrolytic photocleavage of this group (14). Such a reaction had already been observed with other trifluoromethyl derivatives. For example, in the case of 2-hydroxy-4-trifluoromethylbenzoic acid, the authors proposed the attack of nucleophiles at the trifluoromethyl moiety through the involvement of the triplet excited state and showed photobinding of the compound to bovine serum albumin (18). The same mechanism was found in the photochemical transformation of fluoxetine (19). Defluorination of 3,5-diaminotrifluoromethylbenzene to 3,5-diaminobenzoic acid by a nucleophilic substitution of the fluoride by water was also reported (20). These last authors also proposed a mechanism for this photoreaction.
In the present investigation, the photoreactivity of the CF3 moiety of fluphenazine was studied, with the aim of verifying the possible interaction with nucleophilic groups present in solvents different from water and in biomolecules such as amino acids, peptides, and in particular, proteins. For our investigations, we made use of mass spectroscopy as an analytical tool for qualitative identification of photoadducts.
Experimental Procedures Materials. Fluphenazine dihydrochloride, amino acids, glutathione, and calmodulin from bovine brain were purchased from Sigma-Aldrich Italia, Milano, Italy. The 14-peptide Val-Ala-Ser-Phe-Lys-Gln-Ala-Phe-Asp-Ala-ValGly Val-Lys (fragment 303–316 of thermolysin) was a kind gift of Prof. V. De Filippis of the Department of Pharmaceutical Sciences, University of Padova. Irradiation. Two 1 × 10-3 M solutions of the drug were irradiated in pyrex test tubes (10 mm in diameter) by means of two Philips HPW125 lamps, mainly emitting at 365 nm. The total energy on the sample was monitored by means of a radiometer (model 97503, Cole-Parmer Instrument Company, Niles, IL), equipped with a 365-CX sensor. The samples were cooled by circulating water during irradiation. Mass Spectrometric Measurements. To monitor the photobinding of fluphenazine to amino acids and glutathione, we performed the MS analyses on API-TOF mass spectrometer MARINER (PerSeptive Biosystems, Stafford, TX) connected to an infusion pump (PUMP 11, Harvard Apparatus, Holliston, MA, USA). All MS experiments were performed in the positive-ion mode. Full-scan mass spectra were recorded between 120 and 2500 mass units with a scan rate of 4 s per scan in MS mode. The source temperature was 25 °C and the desolvation temperature was 140 °C. The ESI probe voltage was 4.0 kV. The ESI drying and nebulizing gas was nitrogen. Nozzle potential was 90 V. Samples were dissolved in 50 % acetonitrile/water containing 1% formic acid and infused at a flow rate of 10 µL/min. Data were acquired by Mariner Workstation 4.0 and processed by Data Explorer 4.0 (PerSeptive Biosystems, Framingham, MA). MALDI mass measurements were performed on an Ultraflex II TOF-instrument (Bruker Daltonik, Bremen, Germany), able to work in linear, reflectron, and post-source decay (PSD) modes. The instrumental conditions for the analysis of the 14-peptide in positive reflectron mode were IS1 ) 25 kV; IS2 ) 21.65 kV; reflectron potential, 26.3 kV; delay time ) 0 ns. In this case, R-cyano-4hydroxycinnamic acid (Sigma-Aldrich) was used as the matrix (saturated solution in H2O/Acetonitrile, 50/50 v/v). The peptide solutions were diluted 10 times with a 0.1% trifluoroacetic acid
1472 Chem. Res. Toxicol., Vol. 20, No. 10, 2007
Caffieri et al.
Figure 1. ESI mass spectrum of the (a) water and (b) methanol solution of fluphenazine irradiated with 5 J/m2 UVA. (a) m/z 438, intact fluphenazine; 414, photoproduct 3; 454, photoproduct 2. (b) m/z 438, intact fluphenazine; 428, photoproduct 4; 454, photoproduct 2.
(TFA) aqueous solution. Five microliters of peptide final solution were mixed with the same volume of the matrix solution. The resulting solution was deposited (1 µL) on the stainless steel sample holder and allowed to dry before being introduced into the mass spectrometer. External mass calibration was done using the Peptide Calibration Standard (Bruker Daltonics), on the basis of the monoisotopic values of [M + H]+ of angiotensin II, angiotensin I, substance P, bombesin, ACTH clip (1–17), ACTH clip (18–39), somatostatin 28 at m/z 1046.5420, 1296.6853, 1347.7361, 1619.8230, 2093.0868, 2465.1990, and 3147.4714, respectively. The instrumental conditions for the analysis of calmodulin in the positive linear mode were as follows: IS1 ) 25 kV; IS2 ) 23.40 kV; delay time ) 70 ns. Sinapinic acid (from Sigma-Aldrich, saturated solution in acetonitrile/water, 50/50 v/v) was used as the matrix. The initial calmodulin solutions were diluted 10 times with 0.1% TFA aqueous solution. Five microliters of calmodulin final solutions were mixed with the same volume of matrix solution. The resulting solution was deposited (1 µL) on the stainless steel sample holder and allowed to dry before being introduced into the mass spectrometer. External mass calibration was carried out using the protein mixture “Protein Calibration Standard 1” (Bruker Daltonics), which allows instrumental calibration in a mass range 3000–25000 Da, on the basis of [M+ H]+ ions of insulin (m/z 5734.51), ubiquitin (m/z 8565.76), cytochrom C (m/z 12360.97), and horse myoglobin (m/z 16952.30). Three independent measurements were carried out for each sample.
Figure 2. ESI mass spectra of a water solution (pH 7.4) of fluphenazine irradiated in the presence of (a) glycine or (b) cysteine. (a) m/z 438, intact fluphenazine; 471, glycine–fluphenazine monoadduct. (b) m/z 438, intact fluphenazine; 517, cysteine–fluphenazine monoadduct 6 (z ) 1); 259, cysteine–fluphenazine monoadduct 6 (z ) 2); 912, cysteine–fluphenazine diadduct 7 (z ) 1); 456, cysteine-fluphenazine diadduct 7 (z ) 2).
Chromatographic Analysis. TLC was carried out on silica gel plates F254 with a 2 mm thickness (Merck KGaA, Darmstadt, D) and eluted with 1:1 ethyl acetate:methanol. After elution, the bands of interest were scraped off, extracted with ethanol, and taken to dryness. NMR Spectrometry. NMR measurements were carried out on a Bruker AMX 300 spectrometer.
Results and Discussion Fluphenazine dihydrochloride was dissolved (2 × 10-3 M) in water and irradiated with 5 J/m2 UVA; the crude mixture was diluted with an equal volume of acetonitrile and analyzed by mass spectrometry, using the ESI technique, which leaves the small molecules in their [M + H]+ form. The mass spectrum (Figure 1a and Table 1) consists of a few peaks only, showing that, at least for relatively low UVA doses, photolysis of fluphenazine (m/z 438) is quite simple. The peaks at m/z 454 and 414 belong to the already characterized (14) fluphenazine N-oxide 1 and carboxylic acid 2, respectively (Scheme 1).
Table 1. Mass Spectral Data (relative intensity) for the Fluphenazine Solutions Irradiated in the Presence of Various Nucleophilic Solvents, Amino Acids, and Glutathionea substrate
fluphenazine
monoadduct z ) 1
water methanol ethanol aq. ammonia Gly Cys Glu Trp Arg His Pro Lys Tyr Ser GSH
438 (79) 438 (100) 438 (100) 438 (100) 438 (100) 438 (100) 438 (100) 438 (14) 438 (27) 438 (100) 438 (100) 438 (58) 438 (74) 438 (100) 438 (75)
414 (100) 428 (66) 442 (17) 413 (20) 471 (77) 517 (3) 543 (2) 600 (9) 570 (3) 551 (
