Investigating the Role of 2-Phenylpropenal in Felbamate-Induced

1568

Chem. Res. Toxicol. 2004, 17, 1568-1576

Investigating the Role of 2-Phenylpropenal in Felbamate-Induced Idiosyncratic Drug Reactions M. Popovic´,† S. Nierkens,‡ R. Pieters,‡ and J. Uetrecht*,†,§ Faculties of Pharmacy and Medicine, University of Toronto, Toronto, Ontario M5S 2S2, Canada, and IRAS-IT, Universiteit Utrecht, Utrecht 3508 TD, Netherlands Received July 3, 2004

Felbamate (2-phenyl-1,3-propanediol dicarbamate, FBM) can cause aplastic anemia and hepatotoxicity. The mechanism of FBM-induced toxicities is unknown; however, it has been proposed that 2-phenylpropenal, a reactive metabolite of FBM, is responsible. The pathway leading to this metabolite involves hydrolysis of FBM to 2-phenyl-1,3-propandiol monocarbamate (MCF), oxidation to 3-carbamoyl-2-phenylpropionaldehyde (CBMA), and spontaneous loss of carbon dioxide and ammonia. We made a polyclonal antibody against 2-phenylpropenal bound to protein and confirmed its specificity using ELISA. We attempted to develop an animal model of FBM-induced aplastic anemia and/or hepatotoxicity, and we also used the antibody to try to detect covalent binding of 2-phenylpropenal using immunoblotting. However, none of the animals developed evidence of bone marrow or liver toxicity, and we were unable to detect covalent binding, possibly because significantly less 2-phenylpropenal is formed in rodents than in humans. As this type of idiosyncratic drug reaction is believed to be immune-mediated, we also studied the potential of FBM and its metabolites to stimulate an immune response using the reporter antigen popliteal lymph node assay in female Balb/c mice. We found that neither FBM nor MCF induced an immune response in popliteal lymph nodes (PLNs). However, CBMA treatment appeared immunogenic, causing footpad inflammation, hardening, scab formation, and an increase in thickness. The PLN cell count in CBMA-treated mice increased 8-fold as compared to control, FBM-, or MCF-treated mice. Immunohistochemical analysis of the CBMA-exposed PLNs revealed germinal center formation, indicating B cell proliferation, later confirmed by flow cytometry. Most of the cells expressing the activation surface marker CD54 were B cells. We also found that CBMA treatment caused an increase in the production of IgM and IgG1 antibodies as well as IL-4 and IFN-γ cytokines. Our findings indicate that 2-phenylpropenal is a very potent immunogen, supporting its possible involvement in the FBMinduced hepatotoxicity and aplastic anemia.

Introduction 1

Felbamate (FBM) is an antiepileptic drug used for monotherapy and adjunctive therapy in the treatment of partial and generalized seizures, including LennoxGastaut syndrome in children (1). Within a short period of time following the release of FBM in August of 1993, more than 100 000 patients were treated with it, and more than 50 cases of aplastic anemia or hepatotoxicity were reported in association with the treatment (2). This prompted the U.S. Food and Drug Administration to restrict the use of FBM to a handful of patients who were already on the drug and who were refractory to other epilepsy treatments (3). FBM-induced liver and bone marrow toxicities are unpredictable, patient-dependent events, which were not observed in phase III clinical trials. These reactions are referred to as idiosyncratic drug reactions (IDRs), hypersensitivity reactions, or type B reactions (4). * To whom correspondence should be addressed. Tel: 416-978-8939. E-mail: [email protected]. † Faculty of Pharmacy, University of Toronto. ‡ Universiteit Utrecht. § Faculty of Medicine, University of Toronto. 1 FBM, felbamate, 2-phenyl-1,3-propanediol dicarbamate; MCF, monocarbamate felbamate, 2-phenyl-1,3-propandiol monocarbamate; CBMA, aldehyde felbamate, 3-carbamoyl-2-phenylpropionaldehyde.

Previous research provided evidence that chemically reactive metabolites of drugs, rather than parent drugs themselves, are responsible for many IDRs (5). It has been previously proposed that the reactive metabolite responsible for FBM-induced IDRs is the R,β-unsaturated aldehyde, 2-phenylpropenal (6). The metabolic pathways leading to the reactive 2-phenylpropenal, as well as other pathways that inactivate FBM, are shown in Figure 1; FBM is initially hydrolyzed to an alcohol, monocarbamate felbamate (MCF), which undergoes oxidation to an aldehyde, aldehyde felbamate (CBMA), which spontaneously forms 2-phenylpropenal. Hydrolysis to MCF appears to be the rate-limiting step in the production of 2-phenylpropenal. In the study by Dieckhaus et al., an attempt to develop an animal model of FBM-induced toxicity was described (7). Dieckhaus treated both male SpragueDawley and UDP-glucuronyl transferase 1 deficient Gunn rats, previously depleted of GSH, with either FBM or MCF, and monitored them for an increase in 2-phenylpropenal production, neutrophil infiltration in the liver, and a decrease in the blood cell precursors in the bone marrow (7). Initially, we attempted to create an animal model of FBM-induced IDRs using various rat and mice strains by exposing them to FBM alone or concurrently with

10.1021/tx0498197 CCC: $27.50 © 2004 American Chemical Society Published on Web 11/06/2004

2-Phenylpropenal-Triggered Immune Response in Mice

Chem. Res. Toxicol., Vol. 17, No. 12, 2004 1569

Figure 1. Metabolic scheme of FBM.

various enzyme inhibitors. Unfortunately, as with most attempts to reproduce IDRs in an animal, we were unsuccessful. We also used an antibody against 2-phenylpropenal-modified protein to try to detect covalent binding of this metabolite in liver and bone marrow proteins. Finally, we performed the reporter antigen popliteal lymph node assay (RA-PLNA) using female Balb/c mice to investigate the immunogenicity of FBM and its nonreactive metabolites, MCF and CBMA, the latter being a direct precursor to a reactive metabolite 2-phenylpropenal. We hypothesized that once formed in vivo, the reactive metabolite 2-phenylpropenal could form novel antigenic entities in the host, and it could also act as an immunogen to induce an immune response in the popliteal lymph node (PLN) of Balb/c mice; both properties of this FBM metabolite could contribute to its capacity to induce IDRs.

Materials and Methods Chemicals and Reagents. FBM was kindly supplied by Carter-Wallace, Inc. (Princeton, NJ), and 2-phenylpropenal was a generous gift from Dr. Timothy Macdonald’s laboratory (Charlottesville, VA). The following antibodies were used for immunohistochemistry: rat anti-mouse B220, hamster antimouse CD3 (cultured in Dr. Pieters’s laboratory), rabbit antirat horseradish peroxidase (HRP; DAKO, Glostrup, DK), and goat anti-hamster biotin antibody (Jackson Laboratories, Maine); for ELISPOT: alkaline phosphatase (AP)-conjugated goat antimouse, human adsorbed IgG1, IgG2a, and IgM (Southern Biotechnology Associates, Birmingham, AL); and for ELISA analysis: IL-4 and IFN-γ capture and detecting antibodies (BD PharMingen, Hamburg, DE), TNF-R (BioSource, Camarillo, CA), and streptavidin-HRP (CLB, Amsterdam, NL). The following antibodies, used for a three-color flow cytometry staining, were purchased from BD Pharmingen: cychrome-conjugated CD3, fluorescein isothiocyanate (FITC)-conjugated CD3, FITC-conjugated CD4, phycoerythrin (PE)-conjugated CD8a, PE-conjugated CD19, FITC-conjugated CD80, FITC-conjugated CD86, FITC-conjugated CD11c, FITC-conjugated CD54, FITC-conjugate CD8a, PE-conjugated NKT, streptavidin-cychrome, and biotin-conjugated rat-anti-mouse MHC-II, while PE-conjugated anti-macrophage marker (F4/80) was obtained from Caltag Laboratories (Burlingame, CA). Synthesis of 2-Phenylpropenal-Keyhole Limpet Hemocyanin (KLH). 2-Phenylpropenal (5 mg in 0.25 mL acetonitrile)

and KLH (2.5 mg in 1.2 mL PBS, pH 8) were incubated at room temperature for 20 h. The sample was then dialyzed against distilled water for 11 h and lyophilized overnight. This procedure was repeated except that 2-phenylpropenal (100 µL of 20 mg/ mL in acetonitrile) was incubated with bovine serum albumin (BSA; 1 mL of 1 mg/mL in PBS, pH 8). Because KLH is too large to analyze by mass spectrometry, the extent of 2-phenylpropenal adduct formation was followed in the reaction with BSA using matrix-assisted laser desorption/ionization (MALDI), and it was assumed that the efficiency of the reaction with KLH was similar. On the basis of the change in molecular mass following the incubation of 2-phenylpropenal with BSA protein, on average about 42 amino acids out of 607 present in the BSA molecule were modified by 2-phenylpropenal (Figure 2). Production of anti-2-Phenylpropenal-KLH Antiserum. Polyclonal anti-2-phenylpropenal-KLH antibody was raised in a pathogen-free New Zealand White rabbit (Hazleton, MD) housed in the animal care facility at the National Institutes of Health (Bethesda, MD). 2-Phenylpropenal-KLH antigen (2 mg) was diluted in a mixture of saline and Freund’s complete adjuvant (1:1, v/v), and 3 mL of this antigen was injected into 10 subcutaneous sites along the spine and one intramuscular site in a hind limb. Secondary immunization took place 2 weeks after primary immunization, and booster injection took place 2 weeks after the secondary injection, both using Freund’s incomplete adjuvant. Three weeks after the final immunization, the serum was collected and stored at -20 °C. ELISA. 2-Phenylpropenal-BSA or BSA alone (100 µL of 10 µg/mL in PBS) was incubated overnight at 4 °C in flat bottom 96 well MaxiSorp plates (San Diego, CA). The plates were washed eight times with PBS containing 0.05% Tween-20 and then blocked in 5% fetal calf serum (FCS; w/v) in PBS for 2.5 h at room temperature. The plates were washed, coated with anti2-phenylpropenal antiserum (1/100 up to 1/100 000 dilution), and incubated for 2 h at room temperature. The serum was removed, and the plates were washed and then incubated with secondary AP-conjugated goat anti-rabbit IgG [1:10 000 in 2% FCS (w/v) in PBS; 100 µL/well] at room temperature for 1 h. P-nitrophenyl phosphate (100 µL/well) was added to the plates, and the colorimetric reaction was monitored at 405 nm using a SpectraMax Plus plate reader (Sunnyvale, CA). To check for serum cross-reactivity, competitive ELISA was performed by preincubating 2-phenylpropenal antiserum in 1% BSA/PBS with increasing concentrations (0.38-380 µM) of 2-phenylpropenal-BSA, atropic acid, phenylalanine, or clozapine for 2 h at 37 °C prior to its addition to coated ELISA plates (100 µL/well).

1570

Chem. Res. Toxicol., Vol. 17, No. 12, 2004

Figure 2. MALDI spectrum of BSA before (A) and after (B) incubation with 2-phenylpropenal at room temperature for 20 h. Animals and Housing. Rats (150-175 g) and mice (6-8 weeks old) used for the development of an animal model of FBMinduced IDRs were obtained from Charles River (Montreal, QC) and housed in groups of two and four, respectively, in standard cages with free access to water and powdered lab chow. All of the animals were allowed 1 week to acclimatize. Afterward, animals were either left on a powdered lab chow or subjected to the outlined treatment. Upon conclusion of the study, the rats were sacrificed by ip injection (0.2 mL/kg) of an anesthetic mixture, 5:3 (v/v) of ketamine (100 mg/mL) to xylazine (20 mg/ mL), while the mice were sacrificed by cervical dislocation. Rat and mouse liver and bone marrow tissues were collected and stored at -70 °C before analysis. Female Balb/c mice used for the RA-PLNA assay were pathogen-free, 6-12 weeks old, and obtained from Harlan (Horst, NL). The mice were housed in filter-topped Macrolon cages with wood chip bedding at room temperature (23 ( 2 °C), 50-55% relative humidity, and a 12 h light/dark cycle. They were provided drinking water and food ad libitum and acclimatized for 1 week before the start of the study. All of the experiments were conducted according to the guidelines of the Veterinary Faculty at Utrecht University.

Popovic´ et al. Attempts to Create an Animal Model of FBM-Induced IDRs. We performed eight studies in an attempt to either develop an animal model of FBM-induced IDRs or to show the presence of 2-phenylpropenal-modified proteins in animals undergoing the FBM treatment. In these studies, a variety of cotreatments were used to inhibit protective metabolic pathways and to stimulate the immune system. In the first study, outbred female Sprague Dawley rats (one control and two treated) were treated with 800 mg/kg/day of FBM mixed with the diet for 1 week. In the second study, female inbred Brown Norway rats were treated for 5 weeks with either vehicle alone (n ) 1), with FBM (mixed with the diet) and poly IC (n ) 2), or with a combination of FBM, poly IC, and ketoprofen (n ) 2). Poly IC is an analogue of a double-stranded viral RNA and, as such, is recognized and taken up by the Toll-like receptor 3 on the cell surface, inducing a cascade of events, which can lead to an immune response (8). In addition, poly IC is a known inhibitor of P450 synthesis (9). Ketoprofen, on the other hand, is a nonsteroidal antiinflammatory drug, which has been earlier shown to increase the incidence of IDRs in D-penicillaminetreated Brown Norway rats (10). Ketoprofen (10 mg/mL/kg in NaHCO3) was ip injected 2 h before the poly IC (10 mg/kg in PBS). Animals were put on a diet containing 800 mg/kg/day FBM; ketoprofen (6 mg/kg/day) was mixed with the drinking water, and poly IC (5 mg/kg/day) was ip injected three times per week. In the third study, female Lewis rats (n ) 16) and female C57BL/6 mice (n ) 16) were divided into control, FBM only, FBM and poly IC, and FBM, poly IC, and ketoprofen treatment groups, each including four animals, which underwent a 2 week treatment regimen as outlined above. In the fourth study, male Brown Norway rats were divided into control (n ) 3), FBM only (n ) 2), and FBM with poly IC (n ) 5) treatment groups. Most of the animals were gavaged with FBM (2.3 g/kg/day in 0.5% methylcellulose) for a week, while two FBM- and poly IC-treated animals received drug administration for a month. The fifth study included male Brown Norway rats cotreated with FBM (gavaged 2.3 g/kg/day in 0.5% methylcellulose) and poly IC (5 mg/kg/day ip, three times a week) for 1, 2, 4, 7, 14, 23, and 28 days. We used one animal per treatment group per time point, with control animals being cotreated with poly IC as well. In the sixth study, we gavaged male Balb/c mice with FBM alone (2.3 g/kg/day) or FBM and poly IC for a month. In the seventh study, female C57BL/6 mice were treated with MCF for a week. MCF was ip injected daily at a dose of 200 mg/kg in 1.0% methylcellulose. Balb/c and C57BL/6 mice were chosen for our studies as they are considered to resemble Brown Norway and Lewis rats, respectively, when it comes to mounting an antigen specific T cell-mediated immune response (11). In our eighth study, female Brown Norway rats (one control and three treated) were administered a series of enzyme inhibitors, such as a P450 inhibitor, aminobenzotriazole (50 mg/kg/day, gavage), an agent to deplete glutathione, buthionine sulfoximine (30 mM in drinking water), an aldehyde dehydrogenase inhibitor, dicoumarol (75 mg/kg/day, ip), and a UDP-glucuronyltransferase inhibitor, valproic acid (100 mg/kg/day, ip), for 4 days in an attempt to increase the extent of 2-phenylpropenal formation and decrease FBM detoxification. All of the animals undergoing FBM or MCF treatment were followed for the changes in their total blood cell count as well as in the serum aspartate transaminase and alanine transaminase levels. Animals were subjected to treatments for up to a month, and if no evidence of bone marrow or liver toxicity was found, they were sacrificed and their tissues were analyzed for the presence of 2-phenylpropenal covalently modified proteins. SDS-PAGE and Immunoblotting. Liver tissues were homogenized in the ice-cold lysis buffer (10 mM Tris, 1 mM EDTA, 0,2% Triton X-100, pH 7.4, Sigma, Canada) containing EDTAfree protease inhibitors (Roche Pharmaceuticals, Germany). Liver tissues were mechanically homogenized for 10 s at 9500 rpm at 4 °C followed by a second homogenization for 20 s at the same speed. For bone marrow samples, both femurs and humeri of each rat and mouse were removed and aspirated with

2-Phenylpropenal-Triggered Immune Response in Mice fresh RPMI. Samples were centrifuged at 800 rpm for 10 min at 4 °C to remove dead cell debris and hair. The supernatant was discarded, and the pellet was resuspended via manual agitation in 5 mL of red blood cell (RBC) lysis buffer (10 × RBC lysis buffer: 1 mM EDTA, 100 mM KHCO3, 1.7 M NH4Cl, Sigma). Bone marrows were centrifuged again at 800 rpm for 10 min at 4 °C to remove the red cell debris. The supernatant was discarded, and the pellet was resuspended via manual agitation in 5 mL of fresh PBS (Sigma), pH 7.4. The suspension was strained using a 70 µm strainer to remove clumped white cells and the remaining RBC debris from the clean suspension of individual white cells. The samples were centrifuged once more at 800 rpm for 5 min at 4 °C to pellet white blood cells. The supernatant was discarded, and the pellet was resuspended via manual agitation in 0.3 mL of white blood cell lysis buffer [287 µL of phenyl methyl sulfonyl fluoride (Sigma) in 5 mL of lysis buffer]. The resulting protein samples were solubilized by boiling for 5 min in 6 × SDS/sample buffer (7 mL of 4 × TrisCl/ SDS, pH 6.8, 3.8 g of glycerol, 1 g of SDS, 1.2 mg of bromophenol blue, and (5% β-ME) before they were analyzed using SDS-gel electrophoresis. SDS-PAGE was performed using a minigel and the discontinuous buffer system described by Laemmli (12). Stacking and separating gels were 4 and 10% acrylamide, respectively. Proteins were run at 100 V for 90 min, and transfer to nitrocellulose membrane was carried out at 100 V in the mini trans-blot cell for 1 h. The subsequent steps were conducted at room temperature with constant shaking. The nitrocellulose was blocked overnight at 4 °C with 5% skim milk in TTBS (0.1% Tween 20, 100 mM TrisCl, pH 7.5, 0.9% NaCl) solution followed by incubation for 2 h at the room temperature with the anti2-phenylpropenal antiserum (1:1000) in TTBS. Unbound antibodies were removed by washing in TTBS (six times for 10 min). The nitrocellulose blot was incubated for 1 h with HRPconjugated goat anti-rabbit IgG antiserum (diluted 1:15,000 in TTBS), subsequently washed six times for 10 min with TTBS, and then incubated in Super Signal ECL reagent for 5 min. Binding was visualized using a FluorChem 8800 chemiluminescent detector (Ontario, CN). Preparation of CBMA. CBMA was prepared using the procedure outlined by Dess and Martin (13). To a solution of MCF (75 µmol in CH2Cl2) protected from light was added DessMartin periodinane (115.6 µmol in CH2Cl2). The solution was stirred for 1 h at room temperature at which time the reaction was complete as indicated by the disappearance of the original MCF peak on HPLC at 4 min and the appearance of a newly formed CBMA peak at 6.06 min (Phenomenex, C18 column 100 mm × 2 mm; 0.2 mL/min flow rate; H2O (0.1% CH3COONH4)/ CH3CN/CH3COOH, 79:20:1; UV detection, λ ) 254 nm). The crude material was purified by flash chromatography on silica gel (Aldrich, WI) using 1:1 diethyl ether/petroleum ether mixture. CBMA is unstable on silica because of conversion to 2-phenylpropenal; therefore, purification was performed rapidly (in less than 30 min) while being kept on ice. The ether was evaporated using a stream of nitrogen. The CBMA was immediately dissolved in 750 µL of 20% DMSO in PBS for use in the PLNA assay. The total amount of CBMA formed could not be quantified due to the lack of a CBMA standard; however, an increase in the presence of the CBMA peak and almost complete disappearance of the MCF peak led us to conclude that the reaction was almost complete. RA-PLNA. TNP-Ficoll was prepared as previously described (14). Naive Balb/c female mice, five per treatment group, received subcutaneous injections of a 50 µL freshly prepared mixture of a nonsensitizing dose of a reporter antigen, TNPFicoll (10 µg), and the chemical (5 µmol of FBM, MCF, or CBMA) into both hind foot pads. Five control animals were injected with a mixture of TNP-Ficoll and vehicle (20% DMSO/80% PBS). A week after injection, PLNs from both mice legs were extracted and used for analysis. PLN Cell Preparation. Both hind leg PLNs were excised, and the surrounding fatty tissue was removed. PLNs were isolated and put in a Petri dish containing 0.2-0.5 mL of ice-

Chem. Res. Toxicol., Vol. 17, No. 12, 2004 1571 cold RPMI with 2.5% FCS. Two microscopic slides were used to grind up and tear apart the lymph nodes into single-cell suspensions. The slides were rinsed off using RPMI medium, and the cells were placed in a 1 mL syringe and forced through a 25G needle to obtain single cell suspensions, transferred from the Petri dish into a Falcon tube, and spun down at 150g for 10 min at 4 °C. The pellet was resuspended in 1 mL of RPMI containing 10% FCS, and the final cell number was adjusted to 2.5 × 106 cells/mL using RPMI containing 10% FCS. One PLN from each treatment group was snap frozen in liquid nitrogen and stored at -70 °C for immunohistochemistry analysis. Immunohistochemistry. PLN cryostat sections (6 µm) were fixed in acetone for 10 min and air-dried for 10 min. The slides were incubated with rat anti-mouse B220 and hamster antimouse CD3 monoclonal antibody in PBS/1% BSA at room temperature for 1 h. After a single wash with PBS/BSA and two washes with PBS/0.05% Tween-20 (3 min total), the sections were incubated with a combination of polyclonal peroxidaseconjugated rabbit anti-rat antibody (1:40) and biotin-conjugated goat anti-hamster antibody (1:100) in PBS for 1 h. After three washes with PBS/0.05% Tween-20, the slides were incubated with streptavidin-AP antibody (1:600) for 1 h. The slides were rinsed with PBS/0.05% Tween-20 twice and AP buffer once (6.05 g of Tris, pH 9.5, 2.54 g of MgCl2, 1.46 g of NaCl, and distilled water up to 250 mL). T cells were stained for 7 min with paranitroblue tetrazolium (0.66 mL of 5 mg/mL) and 5-bromo-4chloro-3-indolyl phosphate toluidine salt (1.8 mg in 100 µL of dimethylformamide) in 9.34 mL of AP buffer. The slides were rinsed in 50 mM Tris buffer, pH 7.6, three times before they were stained for B cells for 5 min using 3,3′-diaminobenzidine (6 mg in 10 mL of Tris, 10 µL of H2O2). Afterward, the slides were washed with Tris buffer twice and acetate buffer (200 mM sodium acetate, pH 5.4) once. Subsequently, sections were incubated in Naphthol AS-BI phosphate (4 mg in 0.25 mL of dimethylformamide and 25 mL of acetate buffer). To this solution 35 mg of Fast red violet LB salt and 2 drops of MnCl2 were added and sections were incubated for 30 min at 37 °C. After they were washed in the running water, the sections were counterstained with Mayer’s hematoxylin. ELISPOT Assay for Detection of Anti-TNP Antibodies. This assay was essentially performed as described previously by Schielen et al. (15). Briefly, Immobilon-P membranes (Immobilon PVDF Transfer; Millipore, NL) were coated overnight at 4 °C with TNP-BSA (10-15 mL of 10 µg/mL) dissolved in PBS/0.05% Tween with constant shaking. The following morning, membranes were washed with PBS/0.05% Tween and then blocked for 1 h with PBS/Tween/1% BSA with shaking. The membranes were washed twice and then clamped into ELISPOT blocks, and 5 × 105 PLN cells were centrifuged onto the membranes (7 min at 150g) followed by incubation for 4 h at 37 °C. The membranes were then removed from the spot blocks, washed, and incubated overnight at 4 °C with 15 mL of either AP-conjugated IgG1, IgG2a, or IgM antibody (1:2000 dilution) in PBS/0.05% Tween. The following day, the membranes were washed four times with PBS/0.05% Tween, twice with PBS, and once with AP buffer (100 mM Tris, 100 mM NaCl, 5 mM MgCl2‚6H2O, pH to 9.5) in which they were left until staining. Incubation with a mixture of 15 mL of AP buffer, NBT, and BCIP in dimethylformamide followed until there was distinguishable color development of TNP specific antibody spots. Afterward, membranes were rinsed for 15 min with tap water and dried between filter paper for about 2 h. The number of specific antibody secreting cells (ASC) per 106 cells was determined based on the total spot number count performed by three independent observers using a stereomicroscope. The average results of the three independent ELISPOT analysis are presented as the number of ASC per million cells. Cell Culture and Cytokine Measurement. This procedure was performed following the protocol published by Nierkens et al. (16). Briefly, PLN cell suspensions (2.5 × 106 cells/mL in complete RPMI-1640 medium) were incubated with 50 µL of concavalin A (20 µg/mL) or lipopolysaccharide (8 µg/mL) in

1572


Costar 96 well plates (Cambridge, MA) overnight at 37 °C in 5% CO2. After a 10 min centrifugation at 150g, the supernatant was collected and stored at -70 °C until further analysis. Levels of IFN-γ and IL-4 were determined by sandwich ELISA; plates were coated with rat anti-mouse IFN-γ or rat anti-mouse IL-4 in carbonate buffer overnight at 4 °C, washed with PBS/Tween, and blocked with PBS/Tween/3% skim milk for 2 h at room temperature. Standards for IL-4 and IFN-γ, and sample supernatants were prepared in PBS/Tween/1% BSA and added to the plates to incubate overnight at 4 °C. After they were washed with PBS/Tween, the plates were incubated with biotinconjugated rat anti-mouse IFN-γ or IL-4 in PBS/Tween/1% BSA for 1 h at room temperature. The plates were then incubated with streptavidin-HRP (1:10,000 in PBS/Tween/1% BSA) for 45 min at room temperature. For TNF-R ELISA, 96 well plates were coated overnight with rat anti-mouse TNF-R in PBS at 4 °C. The plates were washed, blocked with PBS/0.5% BSA for 2 h at room temperature, and washed again, and then, TNF-R standards and sample supernatants were added to the wells. Immediately, biotinylated antiTNF-R antibody in PBS/Tween/0.5% BSA was added to the wells and incubated for 2 h at room temperature. The plates were washed with PBS/0.05% Tween and incubated with streptavidin-HRP in PBS/Tween/0.5% BSA for 45 min at room temperature. After the final washes, tetramethylbenzidine substrate (0.1 mg/mL) was added to the wells, and the color reaction was stopped after 10 min with 50 µL of 2 M sulfuric acid. The absorbance was measured at 450 nm. Flow Cytometry Analysis of PLN Cells. For flow cytometric analysis, 1 × 106 PLN cells were incubated with either FITC-, PE-, CY-, or biotin-conjugated monoclonal antibodies in 96 well plates in darkness for 30 min at 4 °C. The cells were first preincubated with anti-CD16/32 antibody for 20 min to block nonspecific Fc-receptor recognition. The samples were exposed to biotin-conjugated monoclonal antibodies and were then incubated with streptavidin-CY, washed, resuspended, stored in formalin (0.1%), and analyzed using FACScan with standard FACSflow and CellQuest software from BD Biosciences (Franklin Lakes, NJ). Statistics. Statistical significance between various treatment groups involved in the study was determined using one-way ANOVA, followed by Tukey-Kramer multiple comparisons tests; values of p e 0.05 were considered statistically significant.

Results Characterization of the 2-Phenylpropenal-KLH Antiserum. Sandwich ELISA analysis (Figure 3) demonstrated that the anti-2-phenylpropenal antiserum recognized 2-phenylpropenal-BSA but not BSA. Competitive ELISA showed antibody specificity for 2-phenylpropenal, and antibody did not appear cross-reactive with structurally related atropic acid or phenylalanine or the structurally unrelated drug clozapine (Figure 4A,B). However, a decrease in antibody binding to 2-phenylpropenal-BSA-coated ELISA plates was observed after the serum was preincubated with the increasing concentrations of 2-phenylpropenal-BSA. Preincubation of antibody with BSA did not decrease its recognition of 2-phenylpropenal. Attempts to Create an Animal Model of FBMInduced IDRs. Eight studies were completed using both sexes of several rat and mice strains and aiming to develop an animal model of FBM-induced IDRs. Throughout the studies, animals were monitored for changes in the complete blood cell counts or blood aspartate transaminase and alanine transaminase levels. None of these studies produced evidence of bone marrow or liver toxicity.

Popovic´ et al.

Figure 3. ELISA analysis showing binding of the anti-2phenylpropenal-KLH antiserum to wells of microtiter plates coated with 2-phenylpropenal-BSA conjugate (aa-BSA, 9) or BSA (2).

SDS-PAGE and Immunoblotting. Immunoblotting analysis of the liver and bone marrow tissues from the animals included in the eight studies did not reveal significant 2-phenylpropenal binding. High background binding was observed, which would make specific binding more difficult to detect. The aromatic ring of 2-phenylpropenal resembles the side chain of phenylalanine, and this might result in a high background, but preincubation of the serum with phenylalanine did not decrease the observed background binding. Assessment of Changes in the PLNs. A week after female Balb/c mice were subcutaneously injected with a combination of a reporter antigen and a chemical or a vehicle into the hind foot pads, the PLNs were inspected. There was no difference in the appearance of paws of animals injected with vehicle, FBM, or MCF. However, CBMA-injected animals developed inflamed, hard, crusty, and thickened foot pads. There was no difference in the PLN size between the control, the FBM-treated, and the MCF-treated mice; however, PLNs of CBMA-treated mice appeared larger. The total PLN cell number in the CBMA-treated mice was 8-fold greater than in the control mice; an increase was not observed in either FBM- or MCF-treated mice (Figure 5). Immunohistochemistry. No difference in the PLN morphology was observed between the control, the FBMtreated, and the MCF-treated mice; yet, development of germinal centers was observed in the CBMA-treated mice. Germinal centers contained clusters of larger B cells immediately surrounded by smaller B cells; no difference in the number and distribution of T cells or macrophages was observed in the CBMA-treated mice (Figure 6). TNP Specific Antibody Production. FBM and MCF treatments did not increase the number of IgG1, IgG2a, or IgM ASC cells above the baseline (Figure 7). In contrast, CBMA treatment induced significantly TNP specific IgG1 and IgM antibody production. Results show a 7-fold increase in the number of IgM ASC and a 64fold in the number of IgG1 antibody-secreting cells in CBMA-treated mice in comparison to the control group. There appeared to be a slight increase in the number of



Figure 5. Total PLN cell number 7 days after injection of 50 µL of freshly prepared mixture of a nonsensitizing dose of a reporter antigen TNP-Ficoll (10 µg) with FBM, MCF, CBMA, or a vehicle into both mice hind foot pads. Levels are expressed as means ( SEM of five mice. Statistical analyses were performed using one-way ANOVA, followed by Tukey-Kramer multiple comparisons tests: *** p < 0.001.

Figure 4. To check for serum cross-reactivity, competitive ELISA was performed using compounds structurally resembling 2-phenylpropenal such as atropic acid and phenylalanine and structurally distinct chemical clozapine (A). 2-Phenylpropenal antiserum was preincubated with increasing concentrations (0.38-380 µM) of 2-phenylpropenal-BSA (9), atropic acid (O), phenylalanine (*), or clozapine (•) for 2 h at 37 °C prior to being added to coated ELISA plates (B). Antiserum preincubation with atropic acid, phenylalanine, or clozapine did not decrease its recognition of 2-phenylpropenal (B); antiserum specificity toward BSA before (0) and after incubation with competitive inhibitor compounds did not change (2).

IgG2a ASC in the CBMA-treated mice; nevertheless, it did not reach statistical significance. It appears that B cells in CBMA-treated mice PLNs underwent activation, proliferation, and secretion of TNP specific IgG1 and IgM antibodies (Figure 7). Cytokine Production. We found that PLN cells of CBMA-treated mice had an increased production of IL-4 (Figure 8A) and IFN-γ (Figure 8B) after being restimulated with concanavalin A in vitro. The level of TNF-R did not significantly differ from control levels in any of the three treatment groups (data not shown). Flow Cytometry Analysis of PLN Cells. Results of PLN flow cytometry analyses were consistent with the immunohistochemistry findings: An increase in the B

Figure 6. B220 (B cell monoclonal) and CD3 (T cell) stained sections of PLN isolated on day 7 from female Balb/c mice injected with 50 µL/foot pad of a vehicle (A) or CBMA (B). CBMA treatment shows germinal centers (green GC label). B cells are stained gold-brown, T cells are stained purple-blue, and macrophages are stained red.

(CD19+) cell population from 26% in the controls to 43% in the CBMA-treated mice was observed (Figure 9). A corresponding decrease in the percent T (CD3+) cells in the CBMA-exposed PLNs was observed with CD4+ T cells dropping from 54% in controls to 35% in the CBMAtreated mice. It appeared that there might be a decrease in the percent CD8+ T cells; however, it was not statistically significant. The number and nature of antigen presenting cells were monitored as the percent stained F4/80+ (macrophage-restricted cell surface protein) or CD11c+ (transmembrane glycoprotein of B cell, monocyte, and macrophage) cells. An equivalent increase of 5% positively stained F4/80+ and CD11c+ cells was observed in the CBMA-treated mice in comparison to the controls. Some B cells expressed CD86+, while the others expressed the

1574


Popovic´ et al.

Figure 7. Number of TNP specific ASCs per 106 cells as determined by ELISPOT at day 7 after injection of compounds together with TNP-Ficoll. Levels are expressed as means ( SEM of five mice. Statistical analyses were performed using one-way ANOVA, followed by Tukey-Kramer multiple comparisons tests: * p < 0.05; ** p < 0.01.

Figure 8. PLN cell cytokine levels were determined by ELISA after in vitro restimulation with Con A. Levels are expressed as means ( SEM of five mice. Statistical analyses were performed using one-way ANOVA, followed by Tukey-Kramer multiple comparisons tests: * p < 0.05; ** p < 0.01.

intracellular adhesion molecule marker CD54+/ICAM-1 on their surface. The presence of both markers on the surface of B cells indicates their activation.

Discussion Despite the use of higher than clinical doses, various species, strains, sexes, inhibitors of detoxification pathways, and immune stimulation, we were unable to develop an animal model of FBM-induced bone marrow or liver toxicity. This is consistent with the low success rate of most attempts to develop animal models of IDRs. In addition, despite the production of an antibody against 2-phenylpropenal-modified protein, we were unable to detect in vivo covalent binding in FBM-treated animals. There are many possible reasons for this lack of success. One obvious possibility is that the antibody lacked sufficient sensitivity. It has been previously reported that humans can convert up to 6% of FBM to 2-phenylpropenal, while rats and mice produce less than 1% (7).

Figure 9. Expression of cell surface markers: CD19, CD3, CD4, and CD8 (A); F4/80 and CD11c (B); and CD 54, CD54/19, CD80, CD80/19, CD86, and CD86/19 (C) in the PLN cells on day 7 after coinjection of TNP-Ficoll together with vehicle, FBM, MCF, and CBMA in the hind foot pad of female Balb/c mice. Levels are expressed as means ( SEM of five mice. Statistical analyses were performed using one-way ANOVA, followed by Tukey-Kramer multiple comparisons tests: * p < 0.05; ** p < 0.01; *** p < 0.001.

Unfortunately, we had no tissue from FBM-treated patients to test. We tried to increase the amount of 2-phenylpropenal produced in rodents with various inhibitors, but covalent binding was still not detected. Another possibility is that the binding was lost during the experiments. We and Macdonald have previously found that 2-phenylpropenal binds to the cysteine, lysine, and histidine amino acids of protein (17). The reaction of Michael acceptors, such as 2-phenylpropenal, with sulfhydryl groups is reversible and so binding to cysteine could be lost, especially under reducing conditions of SDS-PAGE. Burcham et al. reported that even the Tris buffer used in most SDS-PAGE systems is sufficiently nucleophilic to interfere with the detection of acrolein covalent binding, and 2-phenylpropenal can be viewed as phenylacrolein (18). We were unable to find a satisfac-


tory buffer to replace Tris, but it is still possible that the use of a different buffer would make it possible to detect in vivo covalent binding. Although we were unable to confirm covalent binding with an antibody, covalent binding is presumed to occur in view of the previous finding of metabolites derived from glutathione conjugates of 2-phenylpropenal; the major goal of this study was to study the in vivo location and molecular targets of the presumed binding. Another important property of a reactive metabolite that may contribute to its ability to cause IDRs is its ability to induce an immune response. Initial studies using the direct PLN assay in which either 2.5 µmol of FBM or 2-phenylpropenal was directly injected into C57BL/6 female mice hind foot pad demonstrated an increase in the PLN weight and cellularity in the case of 2-phenylpropenal (results not shown). To analyze the events taking place in the PLNs of mice injected with either FBM or one of its metabolites, a RA-PLNA using TNP-Ficoll as a reporter antigen was performed (19). TNP-Ficoll is a sugar molecule, which is not recognized by T cells, but in the presence of an adjuvant, it can activate B cells to produce IgM antibodies. For switching to other antibody isotypes, B cells need direct help from T cells. Because T cells cannot recognize a polysaccharide, such as TNP-Ficoll, the help can only come from T cells that are activated by neoantigens. In our experiments, FBM, MCF, and CBMA were tested. Previous research has already demonstrated that CBMA spontaneously converts to 2-phenylpropenal in approximately 30 s under physiological conditions (20). The reason that we used CBMA instead of 2-phenylpropenal for the injections was that CBMA is more stable than 2-phenylpropenal. Of the three treatments, only the CBMA injection produced a positive response. CBMA increased the total number of cells, triggered germinal center formation, and increased the production of IFN-γ and IL-4 (labeled in green GC in Figure 6). In addition, approximately 7% of all B cells showed an upregulated expression of the costimulatory marker CD86 on their surface. An increase in the IgG1 and IgM TNP-Ficoll specific antibodies in the CBMA-injected mice was also detected. As there are no T cells specific for the TNPFicoll sugar present in the PLN, an antibody switch from IgM to IgG1 isotype can only take place in the presence of 2-phenylpropenal-activated T helper cells. Together with flow cytometry data showing an increase in costimulatory antigen-presenting cells and relative changes in lymphocyte numbers, the increase in IL-4 and IFN-γ production after treatment with 2-phenylpropenal indicates specific CD4+T cell sensitization, which can then provide help to B cells in the response to 2-phenylpropenal. Although an increase in both cytokines suggests a potential role of both Th2 and Th1 cells in mounting an immune response against 2-phenylpropenal, IFN-γ production was not as high as previously observed in the case of streptozotocin, a drug that generates a polarized Th1-mediated immune response and increases IFN-γ by a factor of approximately 40-fold (16). Hence, the involvement of the Th1 cells, albeit present, may not have been instrumental in mounting the immune response against 2-phenylpropenal. The increase in IL-4 production of up to 23 pg/mL is higher than previously reported for D-penicillamine, a classic Th2 type of response (16). The lack of an immune response in mice treated with FBM or MCF is most probably due to the fact that both


of these two chemicals require an enzymatic conversion to CBMA before they can trigger an immune response (refer to Figure 1) (21). As there is likely no adequate enzyme activity present in the mice PLNs, conversion to CBMA from FBM and MCF could not take place, although we have no direct data to confirm that. At the moment, our findings indicate that only 2-phenylpropenal, and not FBM or MCF, can trigger a cascade of immune events once formed in vivo. If continued, this immune cascade may represent an important step further in the induction of an IDR. As with other IDRs, it is unclear why only a very small fraction of treated patients develop an IDR. Presumably, the usual response to FBM in most animals and humans is immune tolerance. The detection by Macdonald et al. of an N-acetylcysteine conjugate is convincing evidence that 2-phenylpropenal is formed in rodents and humans (22). It remains to be determined how, where, and to what 2-phenylpropenal covalently binds in vivo and what other factors are required to induce an IDR.

Acknowledgment. We thank Dr. Timothy Macdonald and Dr. Webster Santos for 2-phenylpropenal, Dr. Lance Pohl and Dr. Cynthia Ju for allowing us to use their facility to make a rabbit polyclonal antiserum against 2-phenylpropenal, Shahla Yekta for her help in synthesizing CBMA, Marianne Bol for her expertise in ELISA and ELISPOT analysis, Rob Bleumink for his expertise in immunohistochemistry analysis, and Femke van Wijk for her help in performing flow cytometry analysis. We also thank Carter Wallace (Cranbury, NJ) for providing us with FBM. This work was supported by a grant from a Canadian Institute of Health Research (CIHR). M.P. is a recipient of a doctoral fellowship awarded by Rx&D HRF CIHR, and J.P.U. is a recipient of Canada Research Chair in Adverse Drug Reactions.

References (1) O’Neil, M. G., Perdun, C. S., Wilson, M. B., McGown, S. T., and Patel, S. (1996) Felbamate-associated fatal acute hepatic necrosis. Neurology 46, 1457-1459. (2) Kaufman, D. W., Kelly, J. P., Anderson, T., Harmon, D. C., and Shapiro, S. (1997) Evaluation of case reports of aplastic anemia among patients treated with felbamate. Epilepsia 38, 1265-1269. (3) Nightingale, S. L. (1994) From the Food and Drug Administration. J. Am. Med. Assoc. 272, 995. (4) Pirmohamed, M., Breckenridge, A. M., Kitteringham, N. R., and Park, B. K. (1998) Adverse drug reactions. Br. Med. J. 316, 12951298. (5) Park, B. K., Pirmohamed, M., and Kitteringham, N. R. (1998) Role of drug disposition in drug hypersensitivity: A chemical, molecular, and clinical perspective. Chem. Res. Toxicol. 11, 969-988. (6) Thompson, C. D., Kinter, M. T., and Macdonald, T. L. (1996) Synthesis and in vitro reactivity of 3-carbamoyl-2-phenylpropionaldehyde and 2-phenylpropenal: Putative reactive metabolites of felbamate. Chem. Res. Toxicol. 9, 1225-1229. (7) Dieckhaus, C. M., Miller, T. A., Sofia, R. D., and Macdonald, T. L. (2000) A mechanistic approach to understanding species differences in felbamate bioactivation: Relevance to drug-induced idiosyncratic reactions. Drug Metab. Dispos. 28, 814-822. (8) Alexopoulou, L., Holt, A. C., Medzhitov, R., and Flavell, R. A. (2001) Recognition of double-stranded RNA and activation of NFkappaB by Toll-like receptor 3. Nature 413, 732-738. (9) Knickle, L. C., Spencer, D. F., and Renton, K. W. (1992) The suppression of hepatic cytochrome P4504A mRNA mediated by the interferon inducer polyinosinic acid.polycytidylic acid. Biochem. Pharmacol. 44, 604-608. (10) Seguin, B., Teranishi, M., and Uetrecht, J. P. (2003) Modulation of D-penicillamine-induced autoimmunity in the Brown Norway rat using pharmacological agents that interfere with arachidonic acid metabolism or synthesis of inducible nitric oxide synthase. Toxicology 190, 267-278.

1576


(11) Heinzel, F. P., Sadick, M. D., Mutha, S. S., and Locksley, R. M. (1991) Production of interferon gamma, interleukin 2, interleukin 4, and interleukin 10 by CD4+ lymphocytes in vivo during healing and progressive murine leishmaniasis. Proc. Natl. Acad. Sci. U.S.A. 88, 7011-7015. (12) Laemmli, U. K., Beguin, F., and Gujer-Kellenberger, G. (1970) A factor preventing the major head protein of bacteriophage T4 from random aggregation. J. Mol. Biol. 47, 69-85. (13) Dess, D. B. M. J. C. (1983) Readily accessible 12-I-5 oxidant for the conversion of primary and secondary alcohols to aldehydes and ketones. J Org. Chem. 48, 4155-4156. (14) Albers, R., van der Pijl, A., Seinen, W., Pieters, R., and Bloksma, N. (1996) The autoimmunogenic chemicals HgCl2 and diphenylhydantoin stimulate IgG production to TNP-Ficoll and TNP-OVA, supporting and extending the graft-versus-host hypothesis for chemical induction of autoimmunity. Immunology 89, 468-473. (15) Schielen, P., van Rodijnen, W., Tekstra, J., Albers, R., and Seinen, W. (1995) Quantification of natural antibody producing B cells in rats by an improved ELISPOT technique using the poly(vinylidene difluoride) membrane as the solid support. J. Immunol. Methods 188, 33-41. (16) Nierkens, S., van Helden, P., Bol, M., Bleumink, R., van Kooten, P., Ramdien-Murli, S., Boon, L., and Pieters, R. (2002) Selective requirement for CD40-CD154 in drug-induced type 1 versus type 2 responses to trinitrophenyl-ovalbumin. J. Immunol. 168, 37473754.

Popovic´ et al. (17) Roller, S. G., Dieckhaus, C. M., Santos, W. L., Sofia, R. D., and Macdonald, T. L. (2002) Interaction between human serum albumin and the felbamate metabolites 4-Hydroxy-5-phenyl[1,3]oxazinan-2-one and 2-phenylpropenal. Chem. Res. Toxicol. 15, 815-824. (18) Burcham, P. C., Fontaine, F. R., Petersen, D. R., and Pyke, S. M. (2003) Reactivity with Tris(hydroxymethyl)aminomethane confounds immunodetection of acrolein-adducted proteins. Chem. Res. Toxicol. 16, 1196-1201. (19) Pieters, R. (2000) The popliteal lymph node assay in predictive testing for autoimmunity. Toxicol. Lett. 112-113, 453-459. (20) Dieckhaus, C. M., Roller, S. G., Santos, W. L., Sofia, R. D., and Macdonald, T. L. (2001) Role of glutathione S-transferases A1-1, M1-1, and P1-1 in the detoxification of 2-phenylpropenal, a reactive felbamate metabolite. Chem. Res. Toxicol. 14, 511-516. (21) Katsutani, N., and Shionoya, H. (1992) Popliteal lymph node enlargement induced by procainamide. Int. J. Immunopharmacol. 14, 681-686. (22) Thompson, C. D., Gulden, P. H., and Macdonald, T. L. (1997) Identification of modified atropaldehyde mercapturic acids in rat and human urine after felbamate administration. Chem. Res. Toxicol. 10, 457-462.

TX0498197

Investigating the Role of 2-Phenylpropenal in Felbamate-Induced

Recommend Documents